* [lustre-devel] [PATCH 000/124] missing patches from Lustre 2.7 release
@ 2016-09-18 20:36 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:36 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
This batch of patches includes various cleanups and fixes from
the Lustre 2.7 release. Most of these patches are removal of
cruft and unsigned to signed comparsion fixes. Also added in
the latest LNet fixes as well.
Alexander Boyko (3):
staging: lustre: at: net AT after connect
staging: lustre: mgc: add nid iteration
staging: lustre: ptlrpc: fix race between connect vs resend
Alexander Zarochentsev (1):
staging: lustre: obdclass: optimize busy loop wait
Alexey Lyashkov (1):
staging: lustre: ldlm: evict clients returning errors on ASTs
Amir Shehata (1):
staging: lustre: ptlrpc: Suppress error message when imp_sec is freed
Andreas Dilger (3):
staging: lustre: build: bump build version warnings to x.y.53
staging: lustre: ptlrpc: quiet errors on initial connection
staging: lustre: lprocfs: cleanup stats locking code
Artem Blagodarenko (1):
staging: lustre: client: Fix mkdir -i 1 from DNE2 client to DNE1 server
Bob Glossman (2):
staging: lustre: llite: fix dup flags names
staging: lustre: obdclass: eliminate NULL error return
Bobi Jam (7):
staging: lustre: llite: update ras stride offset
staging: lustre: clio: lu_ref_del() mismatch ref add scope
staging: lustre: fiemap: set FIEMAP_EXTENT_LAST correctly
staging: lustre: clio: add coo_getstripe interface
staging: lustre: osc: osc_object_ast_clear() LBUG
staging: lustre: clio: rename coo_attr_set to coo_attr_update
staging: lustre: clio: pass fid for OST setattr
Dmitry Eremin (18):
staging: lustre: fix comparison between signed and unsigned
staging: lustre: obdclass: fix comparison between signed and unsigned
staging: lustre: llite: Compare of unsigned value against 0 is always true
staging: lustre: obd: change type of lmv_tgt_desc->ltd_idx to u32
staging: lustre: lmv: change type of lmv_obd->tgts_size to u32
staging: lustre: lmv: remove dead code
staging: lustre: mdc: fix comparison between signed and unsigned
staging: lustre: ldlm: count of pools is unsigned long
staging: lustre: lu_dirent_calc_size() return type to size_t
staging: lustre: obdclass: change lu_site->ls_purge_start to unsigned
staging: lustre: changelog: fix comparison between signed and unsigned
staging: lustre: obdclass: lu_htable_order() return type to long
staging: lustre: obdclass: change loop indexes to unsigned
staging: lustre: obd: change brw_page->count to unsigned
staging: lustre: obdclass: change cl_fault_io->ft_nob to size_t
staging: lustre: ptlrpc: fix comparison between signed and unsigned
staging: lustre: osc: change cl_extent_tax and *grants to unsigned
staging: lustre: obd: change type of cl_conn_count to size_t
Doug Oucharek (3):
staging: lustre: lnet: Ensure routing is turned on first time
staging: lustre: lnet: Enable setting per NI peer_credits
staging: lustre: o2iblnd: Put back work queue check previously removed
Fan Yong (3):
staging: lustre: lmv: move some inline functions to lustre_lmv.h
staging: lustre: lmv: add testing for bad name hash
staging: lustre: llite: pack suppgid to MDS correctly
Henri Doreau (4):
staging: lustre: mdc: Report D_CHANGELOG messages as D_HSM
staging: lustre: Flexible changelog format.
staging: lustre: llite: Add ioctl to get parent fids from link EA.
staging: lustre: changelog: Proper record remapping
James Simmons (4):
staging: lustre: ldlm: restore some of the interval functionality
staging: lustre: obd: restore linkea support
staging: lustre: obd: use proper flags for call_usermodehelper
staging: lustre: update version to 2.5.99
Jian Yu (1):
staging: lustre: replace direct HZ access with kernel APIs
Jinshan Xiong (4):
staging: lustre: lov: adjust page bufsize after layout change
staging: lustre: llite: allow setting stripes to specify OSTs
staging: lustre: ldlm: revert the changes for lock canceling policy
staging: lustre: ldlm: move LDLM_GID_ANY to lustre_dlm.h
Johann Lombardi (1):
staging: lustre: grant: quiet message on grant waiting timeout
John L. Hammond (27):
staging: lustre: ldlm: fix a use after free in ldlm_resource_get()
staging: lustre: obd: add lnb_ prefix to members of struct niobuf_local
staging: lustre: obd: add rnb_ prefix to struct niobuf_remote members
staging: lustre: llite: add LL_LEASE_{RD,WR,UN}LCK
staging: lustre: lmv: fix some byte order issues
staging: lustre: lmv: release locks if lmv_intent_lock() fails
staging: lustre: lmv: release request in lmv_revalidate_slaves()
staging: lustre: remove RCU2HANDLE macro
staging: lustre: llite: remove lookup_flags from ll_lookup_it()
staging: lustre: llite: remove mode from ll_create_it()
staging: lustre: llite: turn mode to umode_t for ll_new_inode()
staging: lustre: llite: style cleanup for ll_mkdir
staging: lustre: llite: no need to check dentry is NULL
staging: lustre: cleanup lustre_lib.h
staging: lustre: remove lustre_lite.h
staging: lustre: obd: rename LUSTRE_STRIPE_MAXBYTES
staging: lustre: llite: don't call make_bad_inode() on an old inode
staging: lustre: obd: cleanup struct md_op_data and uses
staging: lustre: lov: remove LL_IOC_RECREATE_{FID,OBJ}
staging: lustre: lov: remove unused {get,set}_info handlers
staging: lustre: fix messages with missing newlines
staging: lustre: obd: remove unused obd methods
staging: lustre: echo: replace lov_stripe_md with lov_oinfo
staging: lustre: llite: remove ll_objects_destroy()
staging: lustre: lov: flatten struct lov_stripe_md
staging: lustre: llite: unlock inode size in ll_lov_setstripe_ea_info()
staging: lustre: remove lustre/include/linux/
Lai Siyao (5):
staging: lustre: statahead: statahead thread wait for RPCs to finish
staging: lustre: statahead: use dcache-like interface for sa entry
staging: lustre: statahead: ll_intent_drop_lock() called in spinlock
staging: lustre: statahead: race in start/stop statahead
staging: lustre: statahead: small fixes and cleanup
Li Xi (4):
staging: lustre: llite: fix ll_statahead_thread() problems on failure
staging: lustre: ptlrpc: enlarge OST_MAXREQSIZE for 4MB RPC
staging: lustre: llog: add newly opened llog at tail of handle list
staging: lustre: llite: enforce pool name length limit
Liang Zhen (1):
staging: lustre: libcfs: check mask returned by cpumask_of_node
Mikhail Pershin (1):
staging: lustre: ptlrpc: prevent request timeout grow due to recovery
Nathaniel Clark (1):
staging: lustre: ldlm: Recalculate interval in ldlm_pool_recalc()
Ned Bass (2):
staging: lustre: llite: make default_easize writeable in /sysfs
staging: lustre: mdc: cl_default_mds_easize not refreshed
Niu Yawei (3):
staging: lustre: obdclass: serialize lu_site purge
staging: lustre: osc: update kms in brw_interpret() properly
staging: lustre: recovery: don't replay closed open
Patrick Farrell (1):
staging: lustre: osc: debug to match extent to brw RPC
Prakash Surya (2):
staging: lustre: llite: Replace write mutex with range lock
staging: lustre: vvp: Use lockless __generic_file_aio_write
Quentin Bouget (1):
staging: lustre: lnet: potential deadlock in lnet
Sebastien Buisson (1):
staging: lustre: lnet: check if ni is in current net namespace
Stephen Champion (2):
staging: lustre: misc: Reduce exposure to overflow on page counters.
staging: lustre: llite: handle concurrent use of cob_transient_pages
Vitaly Fertman (3):
staging: lustre: ldlm: reconstruct proper flags on enqueue resend
staging: lustre: ldlm: resend AST callbacks
staging: lustre: ldlm: per-export lock callback timeout
Wang Shilong (1):
staging: lustre: ptlrpc: fix magic return value of ptlrpc_init_portals
Yang Sheng (1):
staging: lustre: llite: ensure all data flush out when umount
Yoshifumi Uemura (1):
staging: lustre: mdc: Proper accessing struct lov_user_md
frank zago (4):
staging: lustre: llog: prevent out-of-bound index
staging: lustre: misc: remove unnecessary EXPORT_SYMBOL
staging: lustre: obdclass: remove unnecessary EXPORT_SYMBOL
staging: lustre: ptlrpc: remove unnecessary EXPORT_SYMBOL
wang di (6):
staging: lustre: lmv: honor MDT index when creating volatile file
staging: lustre: lmv: Do not ignore ENOENT in lmv_unlink
staging: lustre: lmv: Do not revalidate stripes with master lock
staging: lustre: llite: lock the inode to be migrated
staging: lustre: mdt: add indexing option to default dir stripe
staging: lustre: lmv: fix parent FID for migration
.../lustre/include/linux/libcfs/libcfs_private.h | 12 +-
.../staging/lustre/include/linux/lnet/lib-types.h | 2 +
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c | 10 +
.../staging/lustre/lnet/libcfs/linux/linux-cpu.c | 17 +-
drivers/staging/lustre/lnet/lnet/api-ni.c | 46 +-
drivers/staging/lustre/lnet/lnet/config.c | 14 +
drivers/staging/lustre/lnet/lnet/lib-md.c | 24 +-
drivers/staging/lustre/lnet/lnet/lib-move.c | 11 +
drivers/staging/lustre/lnet/lnet/router.c | 10 +-
drivers/staging/lustre/lustre/fid/fid_lib.c | 2 -
drivers/staging/lustre/lustre/fld/fld_request.c | 2 -
drivers/staging/lustre/lustre/include/cl_object.h | 29 +-
.../staging/lustre/lustre/include/interval_tree.h | 26 +
.../lustre/lustre/include/linux/lustre_lite.h | 91 --
.../lustre/lustre/include/linux/lustre_user.h | 66 -
.../staging/lustre/lustre/include/lprocfs_status.h | 134 ++-
drivers/staging/lustre/lustre/include/lu_object.h | 16 +-
.../lustre/lustre/include/lustre/lustre_idl.h | 57 +-
.../lustre/lustre/include/lustre/lustre_user.h | 254 +++-
drivers/staging/lustre/lustre/include/lustre_cfg.h | 26 +-
.../{linux/lustre_compat25.h => lustre_compat.h} | 6 +-
drivers/staging/lustre/lustre/include/lustre_dlm.h | 5 +
.../staging/lustre/lustre/include/lustre_eacl.h | 1 +
.../staging/lustre/lustre/include/lustre_handles.h | 2 -
.../staging/lustre/lustre/include/lustre_import.h | 2 +
drivers/staging/lustre/lustre/include/lustre_lib.h | 36 -
.../staging/lustre/lustre/include/lustre_linkea.h | 79 ++
.../staging/lustre/lustre/include/lustre_lite.h | 85 --
drivers/staging/lustre/lustre/include/lustre_lmv.h | 69 +-
drivers/staging/lustre/lustre/include/lustre_mdc.h | 37 +-
drivers/staging/lustre/lustre/include/lustre_net.h | 56 +-
.../staging/lustre/lustre/include/lustre_param.h | 3 +
.../include/{linux => }/lustre_patchless_compat.h | 0
.../lustre/lustre/include/lustre_req_layout.h | 20 +-
drivers/staging/lustre/lustre/include/lustre_ver.h | 6 +-
drivers/staging/lustre/lustre/include/obd.h | 213 ++--
drivers/staging/lustre/lustre/include/obd_class.h | 65 +-
.../staging/lustre/lustre/include/obd_support.h | 9 +-
drivers/staging/lustre/lustre/ldlm/interval_tree.c | 100 ++-
drivers/staging/lustre/lustre/ldlm/ldlm_flock.c | 3 +-
drivers/staging/lustre/lustre/ldlm/ldlm_internal.h | 2 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lib.c | 8 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 5 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c | 5 +-
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 45 +-
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 70 +-
drivers/staging/lustre/lustre/ldlm/ldlm_resource.c | 9 +-
drivers/staging/lustre/lustre/llite/Makefile | 2 +-
drivers/staging/lustre/lustre/llite/dcache.c | 6 +-
drivers/staging/lustre/lustre/llite/dir.c | 39 +-
drivers/staging/lustre/lustre/llite/file.c | 323 ++---
drivers/staging/lustre/lustre/llite/glimpse.c | 1 -
drivers/staging/lustre/lustre/llite/lcommon_cl.c | 2 +-
drivers/staging/lustre/lustre/llite/lcommon_misc.c | 1 -
drivers/staging/lustre/lustre/llite/llite_close.c | 1 -
.../staging/lustre/lustre/llite/llite_internal.h | 299 +++--
drivers/staging/lustre/lustre/llite/llite_lib.c | 247 +++-
drivers/staging/lustre/lustre/llite/llite_mmap.c | 2 -
drivers/staging/lustre/lustre/llite/llite_nfs.c | 1 -
drivers/staging/lustre/lustre/llite/lproc_llite.c | 104 ++-
drivers/staging/lustre/lustre/llite/namei.c | 201 ++--
drivers/staging/lustre/lustre/llite/range_lock.c | 233 ++++
drivers/staging/lustre/lustre/llite/range_lock.h | 82 ++
drivers/staging/lustre/lustre/llite/rw.c | 20 +-
drivers/staging/lustre/lustre/llite/rw26.c | 19 +-
drivers/staging/lustre/lustre/llite/statahead.c | 1391 +++++++++-----------
drivers/staging/lustre/lustre/llite/super25.c | 1 -
drivers/staging/lustre/lustre/llite/symlink.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_dev.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_internal.h | 9 +-
drivers/staging/lustre/lustre/llite/vvp_io.c | 31 +-
drivers/staging/lustre/lustre/llite/vvp_lock.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_object.c | 13 +-
drivers/staging/lustre/lustre/llite/vvp_page.c | 22 +-
drivers/staging/lustre/lustre/llite/vvp_req.c | 3 +-
drivers/staging/lustre/lustre/llite/xattr.c | 20 +-
drivers/staging/lustre/lustre/llite/xattr_cache.c | 1 -
drivers/staging/lustre/lustre/lmv/lmv_fld.c | 16 +-
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 94 +-
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 17 +-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 310 +++--
drivers/staging/lustre/lustre/lmv/lproc_lmv.c | 2 +-
.../staging/lustre/lustre/lov/lov_cl_internal.h | 10 +-
drivers/staging/lustre/lustre/lov/lov_dev.c | 1 -
drivers/staging/lustre/lustre/lov/lov_ea.c | 14 +-
drivers/staging/lustre/lustre/lov/lov_internal.h | 9 -
drivers/staging/lustre/lustre/lov/lov_io.c | 3 +
drivers/staging/lustre/lustre/lov/lov_merge.c | 39 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 289 +----
drivers/staging/lustre/lustre/lov/lov_object.c | 49 +-
drivers/staging/lustre/lustre/lov/lov_pack.c | 60 +-
drivers/staging/lustre/lustre/lov/lov_request.c | 78 --
drivers/staging/lustre/lustre/lov/lovsub_object.c | 6 +-
drivers/staging/lustre/lustre/mdc/mdc_internal.h | 30 +-
drivers/staging/lustre/lustre/mdc/mdc_lib.c | 29 +-
drivers/staging/lustre/lustre/mdc/mdc_locks.c | 19 +-
drivers/staging/lustre/lustre/mdc/mdc_reint.c | 19 +-
drivers/staging/lustre/lustre/mdc/mdc_request.c | 74 +-
drivers/staging/lustre/lustre/mgc/mgc_request.c | 27 +-
drivers/staging/lustre/lustre/obdclass/Makefile | 2 +-
drivers/staging/lustre/lustre/obdclass/cl_io.c | 7 +-
drivers/staging/lustre/lustre/obdclass/cl_object.c | 48 +-
drivers/staging/lustre/lustre/obdclass/cl_page.c | 16 +-
drivers/staging/lustre/lustre/obdclass/class_obd.c | 15 +-
drivers/staging/lustre/lustre/obdclass/genops.c | 8 -
drivers/staging/lustre/lustre/obdclass/linkea.c | 201 +++
.../lustre/lustre/obdclass/linux/linux-module.c | 3 +-
.../lustre/lustre/obdclass/linux/linux-sysctl.c | 3 +-
drivers/staging/lustre/lustre/obdclass/llog.c | 8 +-
drivers/staging/lustre/lustre/obdclass/llog_cat.c | 6 +-
.../staging/lustre/lustre/obdclass/llog_internal.h | 5 +
drivers/staging/lustre/lustre/obdclass/llog_obd.c | 1 -
drivers/staging/lustre/lustre/obdclass/llog_swab.c | 25 +-
.../lustre/lustre/obdclass/lprocfs_status.c | 2 -
drivers/staging/lustre/lustre/obdclass/lu_object.c | 117 ++-
.../lustre/lustre/obdclass/lustre_handles.c | 7 +-
.../staging/lustre/lustre/obdclass/lustre_peer.c | 1 -
.../staging/lustre/lustre/obdclass/obd_config.c | 15 +-
drivers/staging/lustre/lustre/obdclass/obd_mount.c | 6 +-
.../staging/lustre/lustre/obdecho/echo_client.c | 168 +--
drivers/staging/lustre/lustre/osc/lproc_osc.c | 31 +-
drivers/staging/lustre/lustre/osc/osc_cache.c | 144 ++-
.../staging/lustre/lustre/osc/osc_cl_internal.h | 4 +-
drivers/staging/lustre/lustre/osc/osc_internal.h | 6 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 37 +-
drivers/staging/lustre/lustre/osc/osc_lock.c | 4 +-
drivers/staging/lustre/lustre/osc/osc_object.c | 7 +-
drivers/staging/lustre/lustre/osc/osc_page.c | 88 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 296 +----
drivers/staging/lustre/lustre/ptlrpc/client.c | 113 +-
drivers/staging/lustre/lustre/ptlrpc/connection.c | 5 -
drivers/staging/lustre/lustre/ptlrpc/events.c | 6 +-
drivers/staging/lustre/lustre/ptlrpc/import.c | 316 +++---
drivers/staging/lustre/lustre/ptlrpc/layout.c | 83 +-
drivers/staging/lustre/lustre/ptlrpc/niobuf.c | 33 +-
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 147 +--
drivers/staging/lustre/lustre/ptlrpc/pinger.c | 1 -
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 9 +-
drivers/staging/lustre/lustre/ptlrpc/recover.c | 2 -
drivers/staging/lustre/lustre/ptlrpc/sec.c | 26 +-
drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c | 13 +-
drivers/staging/lustre/lustre/ptlrpc/sec_config.c | 1 -
drivers/staging/lustre/lustre/ptlrpc/sec_gc.c | 5 +-
drivers/staging/lustre/lustre/ptlrpc/service.c | 13 +-
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 80 +-
145 files changed, 4253 insertions(+), 3944 deletions(-)
delete mode 100644 drivers/staging/lustre/lustre/include/linux/lustre_lite.h
delete mode 100644 drivers/staging/lustre/lustre/include/linux/lustre_user.h
rename drivers/staging/lustre/lustre/include/{linux/lustre_compat25.h => lustre_compat.h} (96%)
create mode 100644 drivers/staging/lustre/lustre/include/lustre_linkea.h
delete mode 100644 drivers/staging/lustre/lustre/include/lustre_lite.h
rename drivers/staging/lustre/lustre/include/{linux => }/lustre_patchless_compat.h (100%)
create mode 100644 drivers/staging/lustre/lustre/llite/range_lock.c
create mode 100644 drivers/staging/lustre/lustre/llite/range_lock.h
create mode 100644 drivers/staging/lustre/lustre/obdclass/linkea.c
^ permalink raw reply [flat|nested] 263+ messages in thread* [PATCH 001/124] staging: lustre: llite: fix ll_statahead_thread() problems on failure
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi, James Simmons
From: Li Xi <lixi@ddn.com>
When ll_prep_md_op_data() fails, ll_statahead_thread() does not
release reference of sai and parent dentry. It does not wake
up parent thread either. This patch fixes these problems.
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5274
Reviewed-on: http://review.whamcloud.com/10940
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/statahead.c | 11 +++++++----
1 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 4ac0d6a..016463b 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -1050,8 +1050,10 @@ static int ll_statahead_thread(void *arg)
op_data = ll_prep_md_op_data(NULL, dir, dir, NULL, 0, 0,
LUSTRE_OPC_ANY, dir);
- if (IS_ERR(op_data))
- return PTR_ERR(op_data);
+ if (IS_ERR(op_data)) {
+ rc = PTR_ERR(op_data);
+ goto out_put;
+ }
op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages;
@@ -1239,7 +1241,9 @@ do_it:
}
}
out:
+ ll_dir_chain_fini(&chain);
ll_finish_md_op_data(op_data);
+out_put:
if (sai->sai_agl_valid) {
spin_lock(&plli->lli_agl_lock);
thread_set_flags(agl_thread, SVC_STOPPING);
@@ -1255,7 +1259,6 @@ out:
/* Set agl_thread flags anyway. */
thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
}
- ll_dir_chain_fini(&chain);
spin_lock(&plli->lli_sa_lock);
if (!list_empty(&sai->sai_entries_received)) {
thread_set_flags(thread, SVC_STOPPING);
@@ -1272,9 +1275,9 @@ out:
wake_up(&sai->sai_waitq);
wake_up(&thread->t_ctl_waitq);
ll_sai_put(sai);
- dput(parent);
CDEBUG(D_READA, "statahead thread stopped: sai %p, parent %pd\n",
sai, parent);
+ dput(parent);
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 001/124] staging: lustre: llite: fix ll_statahead_thread() problems on failure
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi, James Simmons
From: Li Xi <lixi@ddn.com>
When ll_prep_md_op_data() fails, ll_statahead_thread() does not
release reference of sai and parent dentry. It does not wake
up parent thread either. This patch fixes these problems.
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5274
Reviewed-on: http://review.whamcloud.com/10940
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/statahead.c | 11 +++++++----
1 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 4ac0d6a..016463b 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -1050,8 +1050,10 @@ static int ll_statahead_thread(void *arg)
op_data = ll_prep_md_op_data(NULL, dir, dir, NULL, 0, 0,
LUSTRE_OPC_ANY, dir);
- if (IS_ERR(op_data))
- return PTR_ERR(op_data);
+ if (IS_ERR(op_data)) {
+ rc = PTR_ERR(op_data);
+ goto out_put;
+ }
op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages;
@@ -1239,7 +1241,9 @@ do_it:
}
}
out:
+ ll_dir_chain_fini(&chain);
ll_finish_md_op_data(op_data);
+out_put:
if (sai->sai_agl_valid) {
spin_lock(&plli->lli_agl_lock);
thread_set_flags(agl_thread, SVC_STOPPING);
@@ -1255,7 +1259,6 @@ out:
/* Set agl_thread flags anyway. */
thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
}
- ll_dir_chain_fini(&chain);
spin_lock(&plli->lli_sa_lock);
if (!list_empty(&sai->sai_entries_received)) {
thread_set_flags(thread, SVC_STOPPING);
@@ -1272,9 +1275,9 @@ out:
wake_up(&sai->sai_waitq);
wake_up(&thread->t_ctl_waitq);
ll_sai_put(sai);
- dput(parent);
CDEBUG(D_READA, "statahead thread stopped: sai %p, parent %pd\n",
sai, parent);
+ dput(parent);
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 002/124] staging: lustre: ptlrpc: enlarge OST_MAXREQSIZE for 4MB RPC
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi, James Simmons
From: Li Xi <lixi@ddn.com>
This patch enlarges OST_MAXREQSIZE so as to make the
request size large enough for 4MB RPC.
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4755
Reviewed-on: http://review.whamcloud.com/9599
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_net.h | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 24ddcca..6d82998 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -261,7 +261,10 @@
#define MDS_MAXREQSIZE (5 * 1024) /* >= 4736 */
-#define OST_MAXREQSIZE (5 * 1024)
+/**
+ * FIEMAP request can be 4K+ for now
+ */
+#define OST_MAXREQSIZE (16 * 1024)
/* Macro to hide a typecast. */
#define ptlrpc_req_async_args(req) ((void *)&req->rq_async_args)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 002/124] staging: lustre: ptlrpc: enlarge OST_MAXREQSIZE for 4MB RPC
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi, James Simmons
From: Li Xi <lixi@ddn.com>
This patch enlarges OST_MAXREQSIZE so as to make the
request size large enough for 4MB RPC.
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4755
Reviewed-on: http://review.whamcloud.com/9599
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_net.h | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 24ddcca..6d82998 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -261,7 +261,10 @@
#define MDS_MAXREQSIZE (5 * 1024) /* >= 4736 */
-#define OST_MAXREQSIZE (5 * 1024)
+/**
+ * FIEMAP request can be 4K+ for now
+ */
+#define OST_MAXREQSIZE (16 * 1024)
/* Macro to hide a typecast. */
#define ptlrpc_req_async_args(req) ((void *)&req->rq_async_args)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 003/124] staging: lustre: ldlm: fix a use after free in ldlm_resource_get()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
If lvbo initialization has failed then save the return status (from
lr_lvb_len) before putting the resource.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5305
Reviewed-on: http://review.whamcloud.com/11017
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_resource.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
index 62d9f6f..912cd68 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
@@ -1091,6 +1091,7 @@ ldlm_resource_get(struct ldlm_namespace *ns, struct ldlm_resource *parent,
struct cfs_hash_bd bd;
__u64 version;
int ns_refcount = 0;
+ int rc;
LASSERT(!parent);
LASSERT(ns->ns_rs_hash);
@@ -1140,8 +1141,9 @@ lvbo_init:
}
if (unlikely(res->lr_lvb_len < 0)) {
+ rc = res->lr_lvb_len;
ldlm_resource_putref(res);
- res = ERR_PTR(res->lr_lvb_len);
+ res = ERR_PTR(rc);
}
return res;
}
@@ -1152,8 +1154,6 @@ lvbo_init:
cfs_hash_bd_unlock(ns->ns_rs_hash, &bd, 1);
if (ns->ns_lvbo && ns->ns_lvbo->lvbo_init) {
- int rc;
-
OBD_FAIL_TIMEOUT(OBD_FAIL_LDLM_CREATE_RESOURCE, 2);
rc = ns->ns_lvbo->lvbo_init(res);
if (rc < 0) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 003/124] staging: lustre: ldlm: fix a use after free in ldlm_resource_get()
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
If lvbo initialization has failed then save the return status (from
lr_lvb_len) before putting the resource.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5305
Reviewed-on: http://review.whamcloud.com/11017
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_resource.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
index 62d9f6f..912cd68 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
@@ -1091,6 +1091,7 @@ ldlm_resource_get(struct ldlm_namespace *ns, struct ldlm_resource *parent,
struct cfs_hash_bd bd;
__u64 version;
int ns_refcount = 0;
+ int rc;
LASSERT(!parent);
LASSERT(ns->ns_rs_hash);
@@ -1140,8 +1141,9 @@ lvbo_init:
}
if (unlikely(res->lr_lvb_len < 0)) {
+ rc = res->lr_lvb_len;
ldlm_resource_putref(res);
- res = ERR_PTR(res->lr_lvb_len);
+ res = ERR_PTR(rc);
}
return res;
}
@@ -1152,8 +1154,6 @@ lvbo_init:
cfs_hash_bd_unlock(ns->ns_rs_hash, &bd, 1);
if (ns->ns_lvbo && ns->ns_lvbo->lvbo_init) {
- int rc;
-
OBD_FAIL_TIMEOUT(OBD_FAIL_LDLM_CREATE_RESOURCE, 2);
rc = ns->ns_lvbo->lvbo_init(res);
if (rc < 0) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 004/124] staging: lustre: lmv: honor MDT index when creating volatile file
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
LMV should honor MDT index embedded in the name of volatile
file, then during hsm restore, the file under striped dir can
be restored to the right MDT.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4691
Reviewed-on: http://review.whamcloud.com/10866
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 1 +
drivers/staging/lustre/lustre/llite/dir.c | 17 ++++++
drivers/staging/lustre/lustre/llite/llite_lib.c | 5 +-
drivers/staging/lustre/lustre/lmv/lmv_fld.c | 16 +++---
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 5 ++
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 52 ++++++++++++++++++-
6 files changed, 84 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 351fb4c..c69918b 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -244,6 +244,7 @@ struct ost_id {
#define LL_IOC_HSM_IMPORT _IOWR('f', 245, struct hsm_user_import)
#define LL_IOC_LMV_SET_DEFAULT_STRIPE _IOWR('f', 246, struct lmv_user_md)
#define LL_IOC_MIGRATE _IOR('f', 247, int)
+#define LL_IOC_FID2MDTIDX _IOWR('f', 248, struct lu_fid)
#define LL_STATFS_LMV 1
#define LL_STATFS_LOV 2
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index 532047b..c97a4a0 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -1568,6 +1568,23 @@ out_quotactl:
return rc;
case OBD_IOC_FID2PATH:
return ll_fid2path(inode, (void __user *)arg);
+ case LL_IOC_FID2MDTIDX: {
+ struct obd_export *exp = ll_i2mdexp(inode);
+ struct lu_fid fid;
+ __u32 index;
+
+ if (copy_from_user(&fid, (const struct lu_fid __user *)arg,
+ sizeof(fid)))
+ return -EFAULT;
+
+ /* Call mdc_iocontrol */
+ rc = obd_iocontrol(LL_IOC_FID2MDTIDX, exp, sizeof(fid), &fid,
+ &index);
+ if (rc)
+ return rc;
+
+ return index;
+ }
case LL_IOC_HSM_REQUEST: {
struct hsm_user_request *hur;
ssize_t totalsize;
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 1ff788e..99aba6b 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2373,9 +2373,10 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
op_data->op_bias = 0;
op_data->op_cli_flags = 0;
if ((opc == LUSTRE_OPC_CREATE) && name &&
- filename_is_volatile(name, namelen, NULL))
+ filename_is_volatile(name, namelen, &op_data->op_mds))
op_data->op_bias |= MDS_CREATE_VOLATILE;
- op_data->op_mds = 0;
+ else
+ op_data->op_mds = 0;
op_data->op_data = data;
/* When called by ll_setattr_raw, file is i1. */
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_fld.c b/drivers/staging/lustre/lustre/lmv/lmv_fld.c
index a3d170a..715179f 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_fld.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_fld.c
@@ -47,18 +47,20 @@
#include "../include/lprocfs_status.h"
#include "lmv_internal.h"
-int lmv_fld_lookup(struct lmv_obd *lmv,
- const struct lu_fid *fid,
- u32 *mds)
+int lmv_fld_lookup(struct lmv_obd *lmv, const struct lu_fid *fid, u32 *mds)
{
+ struct obd_device *obd = lmv2obd_dev(lmv);
int rc;
- /* FIXME: Currently ZFS still use local seq for ROOT unfortunately, and
+ /*
+ * FIXME: Currently ZFS still use local seq for ROOT unfortunately, and
* this fid_is_local check should be removed once LU-2240 is fixed
*/
- LASSERTF((fid_seq_in_fldb(fid_seq(fid)) ||
- fid_seq_is_local_file(fid_seq(fid))) &&
- fid_is_sane(fid), DFID" is insane!\n", PFID(fid));
+ if (!fid_is_sane(fid) || !(fid_seq_in_fldb(fid_seq(fid)) ||
+ fid_seq_is_local_file(fid_seq(fid)))) {
+ CERROR("%s: invalid FID " DFID "\n", obd->obd_name, PFID(fid));
+ return -EINVAL;
+ }
rc = fld_client_lookup(&lmv->lmv_fld, fid_seq(fid), mds,
LU_SEQ_RANGE_MDT, NULL);
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
index c4961d9..4a5e385 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h
+++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
@@ -62,6 +62,11 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags);
+static inline struct obd_device *lmv2obd_dev(struct lmv_obd *lmv)
+{
+ return container_of0(lmv, struct obd_device, u.lmv);
+}
+
static inline struct lmv_tgt_desc *
lmv_get_target(struct lmv_obd *lmv, u32 mdt_idx, int *index)
{
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 5783359..171c260 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1075,6 +1075,21 @@ static int lmv_iocontrol(unsigned int cmd, struct obd_export *exp,
rc = obd_iocontrol(cmd, tgt->ltd_exp, len, karg, uarg);
break;
}
+ case LL_IOC_FID2MDTIDX: {
+ struct lu_fid *fid = karg;
+ int mdt_index;
+
+ rc = lmv_fld_lookup(lmv, fid, &mdt_index);
+ if (rc)
+ return rc;
+
+ /*
+ * Note: this is from llite(see ll_dir_ioctl()), @uarg does not
+ * point to user space memory for FID2MDTIDX.
+ */
+ *(__u32 *)uarg = mdt_index;
+ break;
+ }
case OBD_IOC_FID2PATH: {
rc = lmv_fid2path(exp, len, karg, uarg);
break;
@@ -1673,13 +1688,44 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
* retval pointer to the lmv_tgt_desc if succeed.
* ERR_PTR(errno) if failed.
*/
-struct lmv_tgt_desc
-*lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data,
- struct lu_fid *fid)
+struct lmv_tgt_desc*
+lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data,
+ struct lu_fid *fid)
{
struct lmv_stripe_md *lsm = op_data->op_mea1;
struct lmv_tgt_desc *tgt;
+ /*
+ * During creating VOLATILE file, it should honor the mdt
+ * index if the file under striped dir is being restored, see
+ * ct_restore().
+ */
+ if (op_data->op_bias & MDS_CREATE_VOLATILE &&
+ (int)op_data->op_mds != -1 && lsm) {
+ int i;
+
+ tgt = lmv_get_target(lmv, op_data->op_mds, NULL);
+ if (IS_ERR(tgt))
+ return tgt;
+
+ /* refill the right parent fid */
+ for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
+ struct lmv_oinfo *oinfo;
+
+ oinfo = &lsm->lsm_md_oinfo[i];
+ if (oinfo->lmo_mds == op_data->op_mds) {
+ *fid = oinfo->lmo_fid;
+ break;
+ }
+ }
+
+ /* Hmm, can not find the stripe by mdt_index(op_mds) */
+ if (i == lsm->lsm_md_stripe_count)
+ tgt = ERR_PTR(-EINVAL);
+
+ return tgt;
+ }
+
if (!lsm || !op_data->op_namelen) {
tgt = lmv_find_target(lmv, fid);
if (IS_ERR(tgt))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 004/124] staging: lustre: lmv: honor MDT index when creating volatile file
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
LMV should honor MDT index embedded in the name of volatile
file, then during hsm restore, the file under striped dir can
be restored to the right MDT.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4691
Reviewed-on: http://review.whamcloud.com/10866
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 1 +
drivers/staging/lustre/lustre/llite/dir.c | 17 ++++++
drivers/staging/lustre/lustre/llite/llite_lib.c | 5 +-
drivers/staging/lustre/lustre/lmv/lmv_fld.c | 16 +++---
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 5 ++
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 52 ++++++++++++++++++-
6 files changed, 84 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 351fb4c..c69918b 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -244,6 +244,7 @@ struct ost_id {
#define LL_IOC_HSM_IMPORT _IOWR('f', 245, struct hsm_user_import)
#define LL_IOC_LMV_SET_DEFAULT_STRIPE _IOWR('f', 246, struct lmv_user_md)
#define LL_IOC_MIGRATE _IOR('f', 247, int)
+#define LL_IOC_FID2MDTIDX _IOWR('f', 248, struct lu_fid)
#define LL_STATFS_LMV 1
#define LL_STATFS_LOV 2
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index 532047b..c97a4a0 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -1568,6 +1568,23 @@ out_quotactl:
return rc;
case OBD_IOC_FID2PATH:
return ll_fid2path(inode, (void __user *)arg);
+ case LL_IOC_FID2MDTIDX: {
+ struct obd_export *exp = ll_i2mdexp(inode);
+ struct lu_fid fid;
+ __u32 index;
+
+ if (copy_from_user(&fid, (const struct lu_fid __user *)arg,
+ sizeof(fid)))
+ return -EFAULT;
+
+ /* Call mdc_iocontrol */
+ rc = obd_iocontrol(LL_IOC_FID2MDTIDX, exp, sizeof(fid), &fid,
+ &index);
+ if (rc)
+ return rc;
+
+ return index;
+ }
case LL_IOC_HSM_REQUEST: {
struct hsm_user_request *hur;
ssize_t totalsize;
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 1ff788e..99aba6b 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2373,9 +2373,10 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
op_data->op_bias = 0;
op_data->op_cli_flags = 0;
if ((opc == LUSTRE_OPC_CREATE) && name &&
- filename_is_volatile(name, namelen, NULL))
+ filename_is_volatile(name, namelen, &op_data->op_mds))
op_data->op_bias |= MDS_CREATE_VOLATILE;
- op_data->op_mds = 0;
+ else
+ op_data->op_mds = 0;
op_data->op_data = data;
/* When called by ll_setattr_raw, file is i1. */
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_fld.c b/drivers/staging/lustre/lustre/lmv/lmv_fld.c
index a3d170a..715179f 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_fld.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_fld.c
@@ -47,18 +47,20 @@
#include "../include/lprocfs_status.h"
#include "lmv_internal.h"
-int lmv_fld_lookup(struct lmv_obd *lmv,
- const struct lu_fid *fid,
- u32 *mds)
+int lmv_fld_lookup(struct lmv_obd *lmv, const struct lu_fid *fid, u32 *mds)
{
+ struct obd_device *obd = lmv2obd_dev(lmv);
int rc;
- /* FIXME: Currently ZFS still use local seq for ROOT unfortunately, and
+ /*
+ * FIXME: Currently ZFS still use local seq for ROOT unfortunately, and
* this fid_is_local check should be removed once LU-2240 is fixed
*/
- LASSERTF((fid_seq_in_fldb(fid_seq(fid)) ||
- fid_seq_is_local_file(fid_seq(fid))) &&
- fid_is_sane(fid), DFID" is insane!\n", PFID(fid));
+ if (!fid_is_sane(fid) || !(fid_seq_in_fldb(fid_seq(fid)) ||
+ fid_seq_is_local_file(fid_seq(fid)))) {
+ CERROR("%s: invalid FID " DFID "\n", obd->obd_name, PFID(fid));
+ return -EINVAL;
+ }
rc = fld_client_lookup(&lmv->lmv_fld, fid_seq(fid), mds,
LU_SEQ_RANGE_MDT, NULL);
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
index c4961d9..4a5e385 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h
+++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
@@ -62,6 +62,11 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags);
+static inline struct obd_device *lmv2obd_dev(struct lmv_obd *lmv)
+{
+ return container_of0(lmv, struct obd_device, u.lmv);
+}
+
static inline struct lmv_tgt_desc *
lmv_get_target(struct lmv_obd *lmv, u32 mdt_idx, int *index)
{
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 5783359..171c260 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1075,6 +1075,21 @@ static int lmv_iocontrol(unsigned int cmd, struct obd_export *exp,
rc = obd_iocontrol(cmd, tgt->ltd_exp, len, karg, uarg);
break;
}
+ case LL_IOC_FID2MDTIDX: {
+ struct lu_fid *fid = karg;
+ int mdt_index;
+
+ rc = lmv_fld_lookup(lmv, fid, &mdt_index);
+ if (rc)
+ return rc;
+
+ /*
+ * Note: this is from llite(see ll_dir_ioctl()), @uarg does not
+ * point to user space memory for FID2MDTIDX.
+ */
+ *(__u32 *)uarg = mdt_index;
+ break;
+ }
case OBD_IOC_FID2PATH: {
rc = lmv_fid2path(exp, len, karg, uarg);
break;
@@ -1673,13 +1688,44 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
* retval pointer to the lmv_tgt_desc if succeed.
* ERR_PTR(errno) if failed.
*/
-struct lmv_tgt_desc
-*lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data,
- struct lu_fid *fid)
+struct lmv_tgt_desc*
+lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data,
+ struct lu_fid *fid)
{
struct lmv_stripe_md *lsm = op_data->op_mea1;
struct lmv_tgt_desc *tgt;
+ /*
+ * During creating VOLATILE file, it should honor the mdt
+ * index if the file under striped dir is being restored, see
+ * ct_restore().
+ */
+ if (op_data->op_bias & MDS_CREATE_VOLATILE &&
+ (int)op_data->op_mds != -1 && lsm) {
+ int i;
+
+ tgt = lmv_get_target(lmv, op_data->op_mds, NULL);
+ if (IS_ERR(tgt))
+ return tgt;
+
+ /* refill the right parent fid */
+ for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
+ struct lmv_oinfo *oinfo;
+
+ oinfo = &lsm->lsm_md_oinfo[i];
+ if (oinfo->lmo_mds == op_data->op_mds) {
+ *fid = oinfo->lmo_fid;
+ break;
+ }
+ }
+
+ /* Hmm, can not find the stripe by mdt_index(op_mds) */
+ if (i == lsm->lsm_md_stripe_count)
+ tgt = ERR_PTR(-EINVAL);
+
+ return tgt;
+ }
+
if (!lsm || !op_data->op_namelen) {
tgt = lmv_find_target(lmv, fid);
if (IS_ERR(tgt))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 005/124] staging: lustre: obdclass: optimize busy loop wait
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Zarochentsev, James Simmons
From: Alexander Zarochentsev <alexander_zarochentsev@xyratex.com>
two optimizations in the busy loop wait in class_cleanup():
(1) remove unnecessary spinlocks protecting obd_conn_inprogress
access
(2) replace cond_resched by more suitable yield to unconditionally
yield the CPU if the condition check fails.
Signed-off-by: Alexander Zarochentsev <alexander_zarochentsev@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5397
Xyratex-bug-id: MRP-1991
Reviewed-on: http://review.whamcloud.com/11198
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/obdclass/obd_config.c | 13 ++++---------
1 files changed, 4 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_config.c b/drivers/staging/lustre/lustre/obdclass/obd_config.c
index 036f396..fddfc9c 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_config.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_config.c
@@ -414,17 +414,12 @@ static int class_cleanup(struct obd_device *obd, struct lustre_cfg *lcfg)
}
/* Leave this on forever */
obd->obd_stopping = 1;
-
- /* wait for already-arrived-connections to finish. */
- while (obd->obd_conn_inprogress > 0) {
- spin_unlock(&obd->obd_dev_lock);
-
- cond_resched();
-
- spin_lock(&obd->obd_dev_lock);
- }
spin_unlock(&obd->obd_dev_lock);
+ while (obd->obd_conn_inprogress > 0)
+ yield();
+ smp_rmb();
+
if (lcfg->lcfg_bufcount >= 2 && LUSTRE_CFG_BUFLEN(lcfg, 1) > 0) {
for (flag = lustre_cfg_string(lcfg, 1); *flag != 0; flag++)
switch (*flag) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 005/124] staging: lustre: obdclass: optimize busy loop wait
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Zarochentsev, James Simmons
From: Alexander Zarochentsev <alexander_zarochentsev@xyratex.com>
two optimizations in the busy loop wait in class_cleanup():
(1) remove unnecessary spinlocks protecting obd_conn_inprogress
access
(2) replace cond_resched by more suitable yield to unconditionally
yield the CPU if the condition check fails.
Signed-off-by: Alexander Zarochentsev <alexander_zarochentsev@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5397
Xyratex-bug-id: MRP-1991
Reviewed-on: http://review.whamcloud.com/11198
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/obdclass/obd_config.c | 13 ++++---------
1 files changed, 4 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_config.c b/drivers/staging/lustre/lustre/obdclass/obd_config.c
index 036f396..fddfc9c 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_config.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_config.c
@@ -414,17 +414,12 @@ static int class_cleanup(struct obd_device *obd, struct lustre_cfg *lcfg)
}
/* Leave this on forever */
obd->obd_stopping = 1;
-
- /* wait for already-arrived-connections to finish. */
- while (obd->obd_conn_inprogress > 0) {
- spin_unlock(&obd->obd_dev_lock);
-
- cond_resched();
-
- spin_lock(&obd->obd_dev_lock);
- }
spin_unlock(&obd->obd_dev_lock);
+ while (obd->obd_conn_inprogress > 0)
+ yield();
+ smp_rmb();
+
if (lcfg->lcfg_bufcount >= 2 && LUSTRE_CFG_BUFLEN(lcfg, 1) > 0) {
for (flag = lustre_cfg_string(lcfg, 1); *flag != 0; flag++)
switch (*flag) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 006/124] staging: lustre: lmv: Do not ignore ENOENT in lmv_unlink
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Return correct value (rc) in lmv_unlink. In lmv_unlink,
-ENOENT might be ingored for local directory unlink.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5371
Reviewed-on: http://review.whamcloud.com/11170
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 171c260..11dcbb6 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2643,7 +2643,7 @@ try_next_stripe:
/* Not cross-ref case, just get out of here. */
if (likely(!(body->mbo_valid & OBD_MD_MDS)))
- return 0;
+ return rc;
CDEBUG(D_INODE, "%s: try unlink to another MDT for "DFID"\n",
exp->exp_obd->obd_name, PFID(&body->mbo_fid1));
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 006/124] staging: lustre: lmv: Do not ignore ENOENT in lmv_unlink
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Return correct value (rc) in lmv_unlink. In lmv_unlink,
-ENOENT might be ingored for local directory unlink.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5371
Reviewed-on: http://review.whamcloud.com/11170
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 171c260..11dcbb6 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2643,7 +2643,7 @@ try_next_stripe:
/* Not cross-ref case, just get out of here. */
if (likely(!(body->mbo_valid & OBD_MD_MDS)))
- return 0;
+ return rc;
CDEBUG(D_INODE, "%s: try unlink to another MDT for "DFID"\n",
exp->exp_obd->obd_name, PFID(&body->mbo_fid1));
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 007/124] staging: lustre: obd: add lnb_ prefix to members of struct niobuf_local
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Add the prefix lnb_ to the members of struct niobuf_local that do not
already have it. Change the struct dentry *lnb_dentry member to void
*lnb_data as it is not used to hold a pointer to struct dentry.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5061
Reviewed-on: http://review.whamcloud.com/10451
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 10 +++++-----
.../staging/lustre/lustre/obdecho/echo_client.c | 6 +++---
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index f3d141b..f55e679 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -437,11 +437,11 @@ struct lmv_obd {
struct niobuf_local {
__u64 lnb_file_offset;
__u32 lnb_page_offset;
- __u32 len;
- __u32 flags;
- struct page *page;
- struct dentry *dentry;
- int rc;
+ __u32 lnb_len;
+ __u32 lnb_flags;
+ struct page *lnb_page;
+ void *lnb_data;
+ int lnb_rc;
};
#define LUSTRE_FLD_NAME "fld"
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index 7527112..e7c492e 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1452,14 +1452,14 @@ static int echo_client_prep_commit(const struct lu_env *env,
LASSERT(lpages == npages);
for (i = 0; i < lpages; i++) {
- struct page *page = lnb[i].page;
+ struct page *page = lnb[i].lnb_page;
/* read past eof? */
- if (!page && lnb[i].rc == 0)
+ if (!page && lnb[i].lnb_rc == 0)
continue;
if (async)
- lnb[i].flags |= OBD_BRW_ASYNC;
+ lnb[i].lnb_flags |= OBD_BRW_ASYNC;
if (ostid_id(&oa->o_oi) == ECHO_PERSISTENT_OBJID ||
(oa->o_valid & OBD_MD_FLFLAGS) == 0 ||
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 007/124] staging: lustre: obd: add lnb_ prefix to members of struct niobuf_local
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Add the prefix lnb_ to the members of struct niobuf_local that do not
already have it. Change the struct dentry *lnb_dentry member to void
*lnb_data as it is not used to hold a pointer to struct dentry.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5061
Reviewed-on: http://review.whamcloud.com/10451
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 10 +++++-----
.../staging/lustre/lustre/obdecho/echo_client.c | 6 +++---
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index f3d141b..f55e679 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -437,11 +437,11 @@ struct lmv_obd {
struct niobuf_local {
__u64 lnb_file_offset;
__u32 lnb_page_offset;
- __u32 len;
- __u32 flags;
- struct page *page;
- struct dentry *dentry;
- int rc;
+ __u32 lnb_len;
+ __u32 lnb_flags;
+ struct page *lnb_page;
+ void *lnb_data;
+ int lnb_rc;
};
#define LUSTRE_FLD_NAME "fld"
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index 7527112..e7c492e 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1452,14 +1452,14 @@ static int echo_client_prep_commit(const struct lu_env *env,
LASSERT(lpages == npages);
for (i = 0; i < lpages; i++) {
- struct page *page = lnb[i].page;
+ struct page *page = lnb[i].lnb_page;
/* read past eof? */
- if (!page && lnb[i].rc == 0)
+ if (!page && lnb[i].lnb_rc == 0)
continue;
if (async)
- lnb[i].flags |= OBD_BRW_ASYNC;
+ lnb[i].lnb_flags |= OBD_BRW_ASYNC;
if (ostid_id(&oa->o_oi) == ECHO_PERSISTENT_OBJID ||
(oa->o_valid & OBD_MD_FLFLAGS) == 0 ||
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 008/124] staging: lustre: obd: add rnb_ prefix to struct niobuf_remote members
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Add the prefix rnb_ to the members of struct niobuf_remote. Delete the
relevant compat macros from ofd_internal.h.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5061
Reviewed-on: http://review.whamcloud.com/10452
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 6 ++--
.../staging/lustre/lustre/obdecho/echo_client.c | 14 +++++-----
drivers/staging/lustre/lustre/osc/osc_request.c | 8 +++---
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 8 +++---
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 24 ++++++++++----------
5 files changed, 30 insertions(+), 30 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 3a0feac..6ff6615 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1803,9 +1803,9 @@ void lustre_swab_obd_ioobj(struct obd_ioobj *ioo);
/* multiple of 8 bytes => can array */
struct niobuf_remote {
- __u64 offset;
- __u32 len;
- __u32 flags;
+ __u64 rnb_offset;
+ __u32 rnb_len;
+ __u32 rnb_flags;
};
void lustre_swab_niobuf_remote(struct niobuf_remote *nbr);
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index e7c492e..6968953 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1437,9 +1437,9 @@ static int echo_client_prep_commit(const struct lu_env *env,
npages = tot_pages;
for (i = 0; i < npages; i++, off += PAGE_SIZE) {
- rnb[i].offset = off;
- rnb[i].len = PAGE_SIZE;
- rnb[i].flags = brw_flags;
+ rnb[i].rnb_offset = off;
+ rnb[i].rnb_len = PAGE_SIZE;
+ rnb[i].rnb_flags = brw_flags;
}
ioo.ioo_bufcnt = npages;
@@ -1469,13 +1469,13 @@ static int echo_client_prep_commit(const struct lu_env *env,
if (rw == OBD_BRW_WRITE)
echo_client_page_debug_setup(page, rw,
ostid_id(&oa->o_oi),
- rnb[i].offset,
- rnb[i].len);
+ rnb[i].rnb_offset,
+ rnb[i].rnb_len);
else
echo_client_page_debug_check(page,
ostid_id(&oa->o_oi),
- rnb[i].offset,
- rnb[i].len);
+ rnb[i].rnb_offset,
+ rnb[i].rnb_len);
}
ret = obd_commitrw(env, rw, exp, oa, 1, &ioo,
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index bdb329d..e861973 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1335,11 +1335,11 @@ static int osc_brw_prep_request(int cmd, struct client_obd *cli,
if (i > 0 && can_merge_pages(pg_prev, pg)) {
niobuf--;
- niobuf->len += pg->count;
+ niobuf->rnb_len += pg->count;
} else {
- niobuf->offset = pg->off;
- niobuf->len = pg->count;
- niobuf->flags = pg->flag;
+ niobuf->rnb_offset = pg->off;
+ niobuf->rnb_len = pg->count;
+ niobuf->rnb_flags = pg->flag;
}
pg_prev = pg;
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 2cf3a51..102f0aa 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1604,9 +1604,9 @@ EXPORT_SYMBOL(lustre_swab_obd_ioobj);
void lustre_swab_niobuf_remote(struct niobuf_remote *nbr)
{
- __swab64s(&nbr->offset);
- __swab32s(&nbr->len);
- __swab32s(&nbr->flags);
+ __swab64s(&nbr->rnb_offset);
+ __swab32s(&nbr->rnb_len);
+ __swab32s(&nbr->rnb_flags);
}
EXPORT_SYMBOL(lustre_swab_niobuf_remote);
@@ -2061,7 +2061,7 @@ EXPORT_SYMBOL(dump_ioo);
void dump_rniobuf(struct niobuf_remote *nb)
{
CDEBUG(D_RPCTRACE, "niobuf_remote: offset=%llu, len=%d, flags=%x\n",
- nb->offset, nb->len, nb->flags);
+ nb->rnb_offset, nb->rnb_len, nb->rnb_flags);
}
EXPORT_SYMBOL(dump_rniobuf);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index eb6d88e..2390255 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -1652,18 +1652,18 @@ void lustre_assert_wire_constants(void)
/* Checks for struct niobuf_remote */
LASSERTF((int)sizeof(struct niobuf_remote) == 16, "found %lld\n",
(long long)(int)sizeof(struct niobuf_remote));
- LASSERTF((int)offsetof(struct niobuf_remote, offset) == 0, "found %lld\n",
- (long long)(int)offsetof(struct niobuf_remote, offset));
- LASSERTF((int)sizeof(((struct niobuf_remote *)0)->offset) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct niobuf_remote *)0)->offset));
- LASSERTF((int)offsetof(struct niobuf_remote, len) == 8, "found %lld\n",
- (long long)(int)offsetof(struct niobuf_remote, len));
- LASSERTF((int)sizeof(((struct niobuf_remote *)0)->len) == 4, "found %lld\n",
- (long long)(int)sizeof(((struct niobuf_remote *)0)->len));
- LASSERTF((int)offsetof(struct niobuf_remote, flags) == 12, "found %lld\n",
- (long long)(int)offsetof(struct niobuf_remote, flags));
- LASSERTF((int)sizeof(((struct niobuf_remote *)0)->flags) == 4, "found %lld\n",
- (long long)(int)sizeof(((struct niobuf_remote *)0)->flags));
+ LASSERTF((int)offsetof(struct niobuf_remote, rnb_offset) == 0, "found %lld\n",
+ (long long)(int)offsetof(struct niobuf_remote, rnb_offset));
+ LASSERTF((int)sizeof(((struct niobuf_remote *)0)->rnb_offset) == 8, "found %lld\n",
+ (long long)(int)sizeof(((struct niobuf_remote *)0)->rnb_offset));
+ LASSERTF((int)offsetof(struct niobuf_remote, rnb_len) == 8, "found %lld\n",
+ (long long)(int)offsetof(struct niobuf_remote, rnb_len));
+ LASSERTF((int)sizeof(((struct niobuf_remote *)0)->rnb_len) == 4, "found %lld\n",
+ (long long)(int)sizeof(((struct niobuf_remote *)0)->rnb_len));
+ LASSERTF((int)offsetof(struct niobuf_remote, rnb_flags) == 12, "found %lld\n",
+ (long long)(int)offsetof(struct niobuf_remote, rnb_flags));
+ LASSERTF((int)sizeof(((struct niobuf_remote *)0)->rnb_flags) == 4, "found %lld\n",
+ (long long)(int)sizeof(((struct niobuf_remote *)0)->rnb_flags));
LASSERTF(OBD_BRW_READ == 0x01, "found 0x%.8x\n",
OBD_BRW_READ);
LASSERTF(OBD_BRW_WRITE == 0x02, "found 0x%.8x\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 008/124] staging: lustre: obd: add rnb_ prefix to struct niobuf_remote members
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Add the prefix rnb_ to the members of struct niobuf_remote. Delete the
relevant compat macros from ofd_internal.h.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5061
Reviewed-on: http://review.whamcloud.com/10452
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 6 ++--
.../staging/lustre/lustre/obdecho/echo_client.c | 14 +++++-----
drivers/staging/lustre/lustre/osc/osc_request.c | 8 +++---
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 8 +++---
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 24 ++++++++++----------
5 files changed, 30 insertions(+), 30 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 3a0feac..6ff6615 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1803,9 +1803,9 @@ void lustre_swab_obd_ioobj(struct obd_ioobj *ioo);
/* multiple of 8 bytes => can array */
struct niobuf_remote {
- __u64 offset;
- __u32 len;
- __u32 flags;
+ __u64 rnb_offset;
+ __u32 rnb_len;
+ __u32 rnb_flags;
};
void lustre_swab_niobuf_remote(struct niobuf_remote *nbr);
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index e7c492e..6968953 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1437,9 +1437,9 @@ static int echo_client_prep_commit(const struct lu_env *env,
npages = tot_pages;
for (i = 0; i < npages; i++, off += PAGE_SIZE) {
- rnb[i].offset = off;
- rnb[i].len = PAGE_SIZE;
- rnb[i].flags = brw_flags;
+ rnb[i].rnb_offset = off;
+ rnb[i].rnb_len = PAGE_SIZE;
+ rnb[i].rnb_flags = brw_flags;
}
ioo.ioo_bufcnt = npages;
@@ -1469,13 +1469,13 @@ static int echo_client_prep_commit(const struct lu_env *env,
if (rw == OBD_BRW_WRITE)
echo_client_page_debug_setup(page, rw,
ostid_id(&oa->o_oi),
- rnb[i].offset,
- rnb[i].len);
+ rnb[i].rnb_offset,
+ rnb[i].rnb_len);
else
echo_client_page_debug_check(page,
ostid_id(&oa->o_oi),
- rnb[i].offset,
- rnb[i].len);
+ rnb[i].rnb_offset,
+ rnb[i].rnb_len);
}
ret = obd_commitrw(env, rw, exp, oa, 1, &ioo,
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index bdb329d..e861973 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1335,11 +1335,11 @@ static int osc_brw_prep_request(int cmd, struct client_obd *cli,
if (i > 0 && can_merge_pages(pg_prev, pg)) {
niobuf--;
- niobuf->len += pg->count;
+ niobuf->rnb_len += pg->count;
} else {
- niobuf->offset = pg->off;
- niobuf->len = pg->count;
- niobuf->flags = pg->flag;
+ niobuf->rnb_offset = pg->off;
+ niobuf->rnb_len = pg->count;
+ niobuf->rnb_flags = pg->flag;
}
pg_prev = pg;
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 2cf3a51..102f0aa 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1604,9 +1604,9 @@ EXPORT_SYMBOL(lustre_swab_obd_ioobj);
void lustre_swab_niobuf_remote(struct niobuf_remote *nbr)
{
- __swab64s(&nbr->offset);
- __swab32s(&nbr->len);
- __swab32s(&nbr->flags);
+ __swab64s(&nbr->rnb_offset);
+ __swab32s(&nbr->rnb_len);
+ __swab32s(&nbr->rnb_flags);
}
EXPORT_SYMBOL(lustre_swab_niobuf_remote);
@@ -2061,7 +2061,7 @@ EXPORT_SYMBOL(dump_ioo);
void dump_rniobuf(struct niobuf_remote *nb)
{
CDEBUG(D_RPCTRACE, "niobuf_remote: offset=%llu, len=%d, flags=%x\n",
- nb->offset, nb->len, nb->flags);
+ nb->rnb_offset, nb->rnb_len, nb->rnb_flags);
}
EXPORT_SYMBOL(dump_rniobuf);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index eb6d88e..2390255 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -1652,18 +1652,18 @@ void lustre_assert_wire_constants(void)
/* Checks for struct niobuf_remote */
LASSERTF((int)sizeof(struct niobuf_remote) == 16, "found %lld\n",
(long long)(int)sizeof(struct niobuf_remote));
- LASSERTF((int)offsetof(struct niobuf_remote, offset) == 0, "found %lld\n",
- (long long)(int)offsetof(struct niobuf_remote, offset));
- LASSERTF((int)sizeof(((struct niobuf_remote *)0)->offset) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct niobuf_remote *)0)->offset));
- LASSERTF((int)offsetof(struct niobuf_remote, len) == 8, "found %lld\n",
- (long long)(int)offsetof(struct niobuf_remote, len));
- LASSERTF((int)sizeof(((struct niobuf_remote *)0)->len) == 4, "found %lld\n",
- (long long)(int)sizeof(((struct niobuf_remote *)0)->len));
- LASSERTF((int)offsetof(struct niobuf_remote, flags) == 12, "found %lld\n",
- (long long)(int)offsetof(struct niobuf_remote, flags));
- LASSERTF((int)sizeof(((struct niobuf_remote *)0)->flags) == 4, "found %lld\n",
- (long long)(int)sizeof(((struct niobuf_remote *)0)->flags));
+ LASSERTF((int)offsetof(struct niobuf_remote, rnb_offset) == 0, "found %lld\n",
+ (long long)(int)offsetof(struct niobuf_remote, rnb_offset));
+ LASSERTF((int)sizeof(((struct niobuf_remote *)0)->rnb_offset) == 8, "found %lld\n",
+ (long long)(int)sizeof(((struct niobuf_remote *)0)->rnb_offset));
+ LASSERTF((int)offsetof(struct niobuf_remote, rnb_len) == 8, "found %lld\n",
+ (long long)(int)offsetof(struct niobuf_remote, rnb_len));
+ LASSERTF((int)sizeof(((struct niobuf_remote *)0)->rnb_len) == 4, "found %lld\n",
+ (long long)(int)sizeof(((struct niobuf_remote *)0)->rnb_len));
+ LASSERTF((int)offsetof(struct niobuf_remote, rnb_flags) == 12, "found %lld\n",
+ (long long)(int)offsetof(struct niobuf_remote, rnb_flags));
+ LASSERTF((int)sizeof(((struct niobuf_remote *)0)->rnb_flags) == 4, "found %lld\n",
+ (long long)(int)sizeof(((struct niobuf_remote *)0)->rnb_flags));
LASSERTF(OBD_BRW_READ == 0x01, "found 0x%.8x\n",
OBD_BRW_READ);
LASSERTF(OBD_BRW_WRITE == 0x02, "found 0x%.8x\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 009/124] staging: lustre: obdclass: serialize lu_site purge
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Niu Yawei,
James Simmons
From: Niu Yawei <yawei.niu@intel.com>
Umount process relies on lu_site_purge(-1) to purge all
objects before umount, however, if there happen to have a
cache shrinker which calls lu_site_purge(nr) in parallel,
some objects may still being freed by cache shrinker even
after the lu_site_purge(-1) called by umount done.
This can be simply fixed by serializing purge threads,
since it doesn't make any sense to have them in parallel.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5331
Reviewed-on: http://review.whamcloud.com/11099
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 5 +++++
drivers/staging/lustre/lustre/obdclass/lu_object.c | 7 +++++++
2 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index 502bc41..fe40b42 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -623,6 +623,11 @@ struct lu_site {
spinlock_t ls_ld_lock;
/**
+ * Lock to serialize site purge.
+ */
+ struct mutex ls_purge_mutex;
+
+ /**
* lu_site stats
*/
struct lprocfs_stats *ls_stats;
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 9d1c96b..b6fd9af 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -354,6 +354,11 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
start = s->ls_purge_start;
bnr = (nr == ~0) ? -1 : nr / CFS_HASH_NBKT(s->ls_obj_hash) + 1;
again:
+ /*
+ * It doesn't make any sense to make purge threads parallel, that can
+ * only bring troubles to us. See LU-5331.
+ */
+ mutex_lock(&s->ls_purge_mutex);
did_sth = 0;
cfs_hash_for_each_bucket(s->ls_obj_hash, &bd, i) {
if (i < start)
@@ -399,6 +404,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
if (nr == 0)
break;
}
+ mutex_unlock(&s->ls_purge_mutex);
if (nr != 0 && did_sth && start != 0) {
start = 0; /* restart from the first bucket */
@@ -983,6 +989,7 @@ int lu_site_init(struct lu_site *s, struct lu_device *top)
char name[16];
memset(s, 0, sizeof(*s));
+ mutex_init(&s->ls_purge_mutex);
snprintf(name, 16, "lu_site_%s", top->ld_type->ldt_name);
for (bits = lu_htable_order(top); bits >= LU_SITE_BITS_MIN; bits--) {
s->ls_obj_hash = cfs_hash_create(name, bits, bits,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 009/124] staging: lustre: obdclass: serialize lu_site purge
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Niu Yawei,
James Simmons
From: Niu Yawei <yawei.niu@intel.com>
Umount process relies on lu_site_purge(-1) to purge all
objects before umount, however, if there happen to have a
cache shrinker which calls lu_site_purge(nr) in parallel,
some objects may still being freed by cache shrinker even
after the lu_site_purge(-1) called by umount done.
This can be simply fixed by serializing purge threads,
since it doesn't make any sense to have them in parallel.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5331
Reviewed-on: http://review.whamcloud.com/11099
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 5 +++++
drivers/staging/lustre/lustre/obdclass/lu_object.c | 7 +++++++
2 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index 502bc41..fe40b42 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -623,6 +623,11 @@ struct lu_site {
spinlock_t ls_ld_lock;
/**
+ * Lock to serialize site purge.
+ */
+ struct mutex ls_purge_mutex;
+
+ /**
* lu_site stats
*/
struct lprocfs_stats *ls_stats;
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 9d1c96b..b6fd9af 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -354,6 +354,11 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
start = s->ls_purge_start;
bnr = (nr == ~0) ? -1 : nr / CFS_HASH_NBKT(s->ls_obj_hash) + 1;
again:
+ /*
+ * It doesn't make any sense to make purge threads parallel, that can
+ * only bring troubles to us. See LU-5331.
+ */
+ mutex_lock(&s->ls_purge_mutex);
did_sth = 0;
cfs_hash_for_each_bucket(s->ls_obj_hash, &bd, i) {
if (i < start)
@@ -399,6 +404,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
if (nr == 0)
break;
}
+ mutex_unlock(&s->ls_purge_mutex);
if (nr != 0 && did_sth && start != 0) {
start = 0; /* restart from the first bucket */
@@ -983,6 +989,7 @@ int lu_site_init(struct lu_site *s, struct lu_device *top)
char name[16];
memset(s, 0, sizeof(*s));
+ mutex_init(&s->ls_purge_mutex);
snprintf(name, 16, "lu_site_%s", top->ld_type->ldt_name);
for (bits = lu_htable_order(top); bits >= LU_SITE_BITS_MIN; bits--) {
s->ls_obj_hash = cfs_hash_create(name, bits, bits,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 010/124] staging: lustre: llite: add LL_LEASE_{RD,WR,UN}LCK
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Define new constants LL_LEASE_{RD,WR,UN}LCK for use as the argument to
and return value from the LL_IOC_{GET,SET}_LEASE ioctls. As arguments,
these contants replace the use of F_{RD,WR,UN}LCK from fcntl.h. As
return values they replace the use of FMODE_{READ,WRITE} which are
internal to the Linux kernel source and not under the control of the
Lustre ioctl interface.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5013
Reviewed-on: http://review.whamcloud.com/10233
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 7 +++
drivers/staging/lustre/lustre/llite/file.c | 51 +++++++++++---------
2 files changed, 35 insertions(+), 23 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index c69918b..5e907af 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -246,6 +246,13 @@ struct ost_id {
#define LL_IOC_MIGRATE _IOR('f', 247, int)
#define LL_IOC_FID2MDTIDX _IOWR('f', 248, struct lu_fid)
+/* Lease types for use as arg and return of LL_IOC_{GET,SET}_LEASE ioctl. */
+enum ll_lease_type {
+ LL_LEASE_RDLCK = 0x1,
+ LL_LEASE_WRLCK = 0x2,
+ LL_LEASE_UNLCK = 0x4,
+};
+
#define LL_STATFS_LMV 1
#define LL_STATFS_LOV 2
#define LL_STATFS_NODELAY 4
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index bfac6dd..e9791e3 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -2245,6 +2245,12 @@ free_hss:
return rc;
}
+static inline long ll_lease_type_from_fmode(fmode_t fmode)
+{
+ return ((fmode & FMODE_READ) ? LL_LEASE_RDLCK : 0) |
+ ((fmode & FMODE_WRITE) ? LL_LEASE_WRLCK : 0);
+}
+
static long
ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
@@ -2449,20 +2455,20 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
struct ll_inode_info *lli = ll_i2info(inode);
struct obd_client_handle *och = NULL;
bool lease_broken;
- fmode_t mode = 0;
+ fmode_t fmode;
switch (arg) {
- case F_WRLCK:
+ case LL_LEASE_WRLCK:
if (!(file->f_mode & FMODE_WRITE))
return -EPERM;
- mode = FMODE_WRITE;
+ fmode = FMODE_WRITE;
break;
- case F_RDLCK:
+ case LL_LEASE_RDLCK:
if (!(file->f_mode & FMODE_READ))
return -EPERM;
- mode = FMODE_READ;
+ fmode = FMODE_READ;
break;
- case F_UNLCK:
+ case LL_LEASE_UNLCK:
mutex_lock(&lli->lli_och_mutex);
if (fd->fd_lease_och) {
och = fd->fd_lease_och;
@@ -2470,26 +2476,26 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
}
mutex_unlock(&lli->lli_och_mutex);
- if (och) {
- mode = och->och_flags &
- (FMODE_READ | FMODE_WRITE);
- rc = ll_lease_close(och, inode, &lease_broken);
- if (rc == 0 && lease_broken)
- mode = 0;
- } else {
- rc = -ENOLCK;
- }
+ if (!och)
+ return -ENOLCK;
- /* return the type of lease or error */
- return rc < 0 ? rc : (int)mode;
+ fmode = och->och_flags;
+ rc = ll_lease_close(och, inode, &lease_broken);
+ if (rc < 0)
+ return rc;
+
+ if (lease_broken)
+ fmode = 0;
+
+ return ll_lease_type_from_fmode(fmode);
default:
return -EINVAL;
}
- CDEBUG(D_INODE, "Set lease with mode %d\n", mode);
+ CDEBUG(D_INODE, "Set lease with mode %u\n", fmode);
/* apply for lease */
- och = ll_lease_open(inode, file, mode, 0);
+ och = ll_lease_open(inode, file, fmode, 0);
if (IS_ERR(och))
return PTR_ERR(och);
@@ -2510,8 +2516,8 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
case LL_IOC_GET_LEASE: {
struct ll_inode_info *lli = ll_i2info(inode);
struct ldlm_lock *lock = NULL;
+ fmode_t fmode = 0;
- rc = 0;
mutex_lock(&lli->lli_och_mutex);
if (fd->fd_lease_och) {
struct obd_client_handle *och = fd->fd_lease_och;
@@ -2520,14 +2526,13 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
if (lock) {
lock_res_and_lock(lock);
if (!ldlm_is_cancel(lock))
- rc = och->och_flags &
- (FMODE_READ | FMODE_WRITE);
+ fmode = och->och_flags;
unlock_res_and_lock(lock);
LDLM_LOCK_PUT(lock);
}
}
mutex_unlock(&lli->lli_och_mutex);
- return rc;
+ return ll_lease_type_from_fmode(fmode);
}
case LL_IOC_HSM_IMPORT: {
struct hsm_user_import *hui;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 010/124] staging: lustre: llite: add LL_LEASE_{RD, WR, UN}LCK
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Define new constants LL_LEASE_{RD,WR,UN}LCK for use as the argument to
and return value from the LL_IOC_{GET,SET}_LEASE ioctls. As arguments,
these contants replace the use of F_{RD,WR,UN}LCK from fcntl.h. As
return values they replace the use of FMODE_{READ,WRITE} which are
internal to the Linux kernel source and not under the control of the
Lustre ioctl interface.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5013
Reviewed-on: http://review.whamcloud.com/10233
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 7 +++
drivers/staging/lustre/lustre/llite/file.c | 51 +++++++++++---------
2 files changed, 35 insertions(+), 23 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index c69918b..5e907af 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -246,6 +246,13 @@ struct ost_id {
#define LL_IOC_MIGRATE _IOR('f', 247, int)
#define LL_IOC_FID2MDTIDX _IOWR('f', 248, struct lu_fid)
+/* Lease types for use as arg and return of LL_IOC_{GET,SET}_LEASE ioctl. */
+enum ll_lease_type {
+ LL_LEASE_RDLCK = 0x1,
+ LL_LEASE_WRLCK = 0x2,
+ LL_LEASE_UNLCK = 0x4,
+};
+
#define LL_STATFS_LMV 1
#define LL_STATFS_LOV 2
#define LL_STATFS_NODELAY 4
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index bfac6dd..e9791e3 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -2245,6 +2245,12 @@ free_hss:
return rc;
}
+static inline long ll_lease_type_from_fmode(fmode_t fmode)
+{
+ return ((fmode & FMODE_READ) ? LL_LEASE_RDLCK : 0) |
+ ((fmode & FMODE_WRITE) ? LL_LEASE_WRLCK : 0);
+}
+
static long
ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
@@ -2449,20 +2455,20 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
struct ll_inode_info *lli = ll_i2info(inode);
struct obd_client_handle *och = NULL;
bool lease_broken;
- fmode_t mode = 0;
+ fmode_t fmode;
switch (arg) {
- case F_WRLCK:
+ case LL_LEASE_WRLCK:
if (!(file->f_mode & FMODE_WRITE))
return -EPERM;
- mode = FMODE_WRITE;
+ fmode = FMODE_WRITE;
break;
- case F_RDLCK:
+ case LL_LEASE_RDLCK:
if (!(file->f_mode & FMODE_READ))
return -EPERM;
- mode = FMODE_READ;
+ fmode = FMODE_READ;
break;
- case F_UNLCK:
+ case LL_LEASE_UNLCK:
mutex_lock(&lli->lli_och_mutex);
if (fd->fd_lease_och) {
och = fd->fd_lease_och;
@@ -2470,26 +2476,26 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
}
mutex_unlock(&lli->lli_och_mutex);
- if (och) {
- mode = och->och_flags &
- (FMODE_READ | FMODE_WRITE);
- rc = ll_lease_close(och, inode, &lease_broken);
- if (rc == 0 && lease_broken)
- mode = 0;
- } else {
- rc = -ENOLCK;
- }
+ if (!och)
+ return -ENOLCK;
- /* return the type of lease or error */
- return rc < 0 ? rc : (int)mode;
+ fmode = och->och_flags;
+ rc = ll_lease_close(och, inode, &lease_broken);
+ if (rc < 0)
+ return rc;
+
+ if (lease_broken)
+ fmode = 0;
+
+ return ll_lease_type_from_fmode(fmode);
default:
return -EINVAL;
}
- CDEBUG(D_INODE, "Set lease with mode %d\n", mode);
+ CDEBUG(D_INODE, "Set lease with mode %u\n", fmode);
/* apply for lease */
- och = ll_lease_open(inode, file, mode, 0);
+ och = ll_lease_open(inode, file, fmode, 0);
if (IS_ERR(och))
return PTR_ERR(och);
@@ -2510,8 +2516,8 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
case LL_IOC_GET_LEASE: {
struct ll_inode_info *lli = ll_i2info(inode);
struct ldlm_lock *lock = NULL;
+ fmode_t fmode = 0;
- rc = 0;
mutex_lock(&lli->lli_och_mutex);
if (fd->fd_lease_och) {
struct obd_client_handle *och = fd->fd_lease_och;
@@ -2520,14 +2526,13 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
if (lock) {
lock_res_and_lock(lock);
if (!ldlm_is_cancel(lock))
- rc = och->och_flags &
- (FMODE_READ | FMODE_WRITE);
+ fmode = och->och_flags;
unlock_res_and_lock(lock);
LDLM_LOCK_PUT(lock);
}
}
mutex_unlock(&lli->lli_och_mutex);
- return rc;
+ return ll_lease_type_from_fmode(fmode);
}
case LL_IOC_HSM_IMPORT: {
struct hsm_user_import *hui;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 011/124] staging: lustre: llite: update ras stride offset
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
When a read ahead does not reach the end of the region reserved from
ras, we'd set ras::ras_next_readahead back to where we left off; For
stride read ahead, it needs to make sure that the offset is no less
than ras_stride_offset, so that the stride read ahead can work
correctly.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5263
Reviewed-on: http://review.whamcloud.com/11181
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/rw.c | 18 +++++++++++++++---
1 files changed, 15 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/rw.c b/drivers/staging/lustre/lustre/llite/rw.c
index 25de3bf..de20b7d 100644
--- a/drivers/staging/lustre/lustre/llite/rw.c
+++ b/drivers/staging/lustre/lustre/llite/rw.c
@@ -413,7 +413,7 @@ static int ll_read_ahead_pages(const struct lu_env *env,
* forward read-ahead, it will be fixed when backward
* read-ahead is implemented
*/
- LASSERTF(page_idx > ria->ria_stoff, "Invalid page_idx %lu rs %lu re %lu ro %lu rl %lu rp %lu\n",
+ LASSERTF(page_idx >= ria->ria_stoff, "Invalid page_idx %lu rs %lu re %lu ro %lu rl %lu rp %lu\n",
page_idx,
ria->ria_start, ria->ria_end, ria->ria_stoff,
ria->ria_length, ria->ria_pages);
@@ -474,10 +474,22 @@ int ll_readahead(const struct lu_env *env, struct cl_io *io,
}
/* Reserve a part of the read-ahead window that we'll be issuing */
- if (ras->ras_window_len) {
- start = ras->ras_next_readahead;
+ if (ras->ras_window_len > 0) {
+ /*
+ * Note: other thread might rollback the ras_next_readahead,
+ * if it can not get the full size of prepared pages, see the
+ * end of this function. For stride read ahead, it needs to
+ * make sure the offset is no less than ras_stride_offset,
+ * so that stride read ahead can work correctly.
+ */
+ if (stride_io_mode(ras))
+ start = max(ras->ras_next_readahead,
+ ras->ras_stride_offset);
+ else
+ start = ras->ras_next_readahead;
end = ras->ras_window_start + ras->ras_window_len - 1;
}
+
if (end != 0) {
unsigned long rpc_boundary;
/*
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 011/124] staging: lustre: llite: update ras stride offset
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
When a read ahead does not reach the end of the region reserved from
ras, we'd set ras::ras_next_readahead back to where we left off; For
stride read ahead, it needs to make sure that the offset is no less
than ras_stride_offset, so that the stride read ahead can work
correctly.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5263
Reviewed-on: http://review.whamcloud.com/11181
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/rw.c | 18 +++++++++++++++---
1 files changed, 15 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/rw.c b/drivers/staging/lustre/lustre/llite/rw.c
index 25de3bf..de20b7d 100644
--- a/drivers/staging/lustre/lustre/llite/rw.c
+++ b/drivers/staging/lustre/lustre/llite/rw.c
@@ -413,7 +413,7 @@ static int ll_read_ahead_pages(const struct lu_env *env,
* forward read-ahead, it will be fixed when backward
* read-ahead is implemented
*/
- LASSERTF(page_idx > ria->ria_stoff, "Invalid page_idx %lu rs %lu re %lu ro %lu rl %lu rp %lu\n",
+ LASSERTF(page_idx >= ria->ria_stoff, "Invalid page_idx %lu rs %lu re %lu ro %lu rl %lu rp %lu\n",
page_idx,
ria->ria_start, ria->ria_end, ria->ria_stoff,
ria->ria_length, ria->ria_pages);
@@ -474,10 +474,22 @@ int ll_readahead(const struct lu_env *env, struct cl_io *io,
}
/* Reserve a part of the read-ahead window that we'll be issuing */
- if (ras->ras_window_len) {
- start = ras->ras_next_readahead;
+ if (ras->ras_window_len > 0) {
+ /*
+ * Note: other thread might rollback the ras_next_readahead,
+ * if it can not get the full size of prepared pages, see the
+ * end of this function. For stride read ahead, it needs to
+ * make sure the offset is no less than ras_stride_offset,
+ * so that stride read ahead can work correctly.
+ */
+ if (stride_io_mode(ras))
+ start = max(ras->ras_next_readahead,
+ ras->ras_stride_offset);
+ else
+ start = ras->ras_next_readahead;
end = ras->ras_window_start + ras->ras_window_len - 1;
}
+
if (end != 0) {
unsigned long rpc_boundary;
/*
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 012/124] staging: lustre: lmv: fix some byte order issues
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In the handler for LL_IOC_LMV_GETSTRIPE convert stripe FIDs from
little to CPU endian when unpacking lmv_user_md. In lmv_unpack_md_v1()
fix a double conversion of the stripe count.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5342
Reviewed-on: http://review.whamcloud.com/11106
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/dir.c | 8 ++++----
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 2 +-
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index c97a4a0..97815cf 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -1252,16 +1252,16 @@ lmv_out_free:
tmp->lum_stripe_count = 0;
tmp->lum_stripe_offset = mdt_index;
for (i = 0; i < stripe_count; i++) {
- struct lu_fid *fid;
+ struct lu_fid fid;
- fid = &lmm->lmv_md_v1.lmv_stripe_fids[i];
- mdt_index = ll_get_mdt_idx_by_fid(sbi, fid);
+ fid_le_to_cpu(&fid, &lmm->lmv_md_v1.lmv_stripe_fids[i]);
+ mdt_index = ll_get_mdt_idx_by_fid(sbi, &fid);
if (mdt_index < 0) {
rc = mdt_index;
goto out_tmp;
}
tmp->lum_objects[i].lum_mds = mdt_index;
- tmp->lum_objects[i].lum_fid = *fid;
+ tmp->lum_objects[i].lum_fid = fid;
tmp->lum_stripe_count++;
}
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 11dcbb6..e1fac0b 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2909,7 +2909,7 @@ static int lmv_unpack_md_v1(struct obd_export *exp, struct lmv_stripe_md *lsm,
lsm->lsm_md_hash_type, lsm->lsm_md_layout_version);
stripe_count = le32_to_cpu(lmm1->lmv_stripe_count);
- for (i = 0; i < le32_to_cpu(stripe_count); i++) {
+ for (i = 0; i < stripe_count; i++) {
fid_le_to_cpu(&lsm->lsm_md_oinfo[i].lmo_fid,
&lmm1->lmv_stripe_fids[i]);
rc = lmv_fld_lookup(lmv, &lsm->lsm_md_oinfo[i].lmo_fid,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 012/124] staging: lustre: lmv: fix some byte order issues
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In the handler for LL_IOC_LMV_GETSTRIPE convert stripe FIDs from
little to CPU endian when unpacking lmv_user_md. In lmv_unpack_md_v1()
fix a double conversion of the stripe count.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5342
Reviewed-on: http://review.whamcloud.com/11106
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/dir.c | 8 ++++----
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 2 +-
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index c97a4a0..97815cf 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -1252,16 +1252,16 @@ lmv_out_free:
tmp->lum_stripe_count = 0;
tmp->lum_stripe_offset = mdt_index;
for (i = 0; i < stripe_count; i++) {
- struct lu_fid *fid;
+ struct lu_fid fid;
- fid = &lmm->lmv_md_v1.lmv_stripe_fids[i];
- mdt_index = ll_get_mdt_idx_by_fid(sbi, fid);
+ fid_le_to_cpu(&fid, &lmm->lmv_md_v1.lmv_stripe_fids[i]);
+ mdt_index = ll_get_mdt_idx_by_fid(sbi, &fid);
if (mdt_index < 0) {
rc = mdt_index;
goto out_tmp;
}
tmp->lum_objects[i].lum_mds = mdt_index;
- tmp->lum_objects[i].lum_fid = *fid;
+ tmp->lum_objects[i].lum_fid = fid;
tmp->lum_stripe_count++;
}
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 11dcbb6..e1fac0b 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2909,7 +2909,7 @@ static int lmv_unpack_md_v1(struct obd_export *exp, struct lmv_stripe_md *lsm,
lsm->lsm_md_hash_type, lsm->lsm_md_layout_version);
stripe_count = le32_to_cpu(lmm1->lmv_stripe_count);
- for (i = 0; i < le32_to_cpu(stripe_count); i++) {
+ for (i = 0; i < stripe_count; i++) {
fid_le_to_cpu(&lsm->lsm_md_oinfo[i].lmo_fid,
&lmm1->lmv_stripe_fids[i]);
rc = lmv_fld_lookup(lmv, &lsm->lsm_md_oinfo[i].lmo_fid,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 013/124] staging: lustre: osc: update kms in brw_interpret() properly
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Niu Yawei,
James Simmons
From: Niu Yawei <yawei.niu@intel.com>
In brw_interpret(), we forgot page offset when calculating
write offset, that leads to wrong kms for sync write.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5463
Reviewed-on: http://review.whamcloud.com/11374
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Li Dongyang <dongyang.li@anu.edu.au>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_request.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index e861973..e44b4fa 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1796,7 +1796,8 @@ static int brw_interpret(const struct lu_env *env,
if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
- loff_t last_off = last->oap_count + last->oap_obj_off;
+ loff_t last_off = last->oap_count + last->oap_obj_off +
+ last->oap_page_off;
/* Change file size if this is an out of quota or
* direct IO write and it extends the file size
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 013/124] staging: lustre: osc: update kms in brw_interpret() properly
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Niu Yawei,
James Simmons
From: Niu Yawei <yawei.niu@intel.com>
In brw_interpret(), we forgot page offset when calculating
write offset, that leads to wrong kms for sync write.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5463
Reviewed-on: http://review.whamcloud.com/11374
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Li Dongyang <dongyang.li@anu.edu.au>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_request.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index e861973..e44b4fa 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1796,7 +1796,8 @@ static int brw_interpret(const struct lu_env *env,
if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
- loff_t last_off = last->oap_count + last->oap_obj_off;
+ loff_t last_off = last->oap_count + last->oap_obj_off +
+ last->oap_page_off;
/* Change file size if this is an out of quota or
* direct IO write and it extends the file size
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 014/124] staging: lustre: lmv: release locks if lmv_intent_lock() fails
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In lmv_intent_lock() if we will return an error then first release any
locks referenced by the intent.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5431
Reviewed-on: http://review.whamcloud.com/11319
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 22 ++++++++++++++++++++++
1 files changed, 22 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 85cc5cb..fd3dc7c 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -533,5 +533,27 @@ int lmv_intent_lock(struct obd_export *exp, struct md_op_data *op_data,
extra_lock_flags);
else
LBUG();
+
+ if (rc < 0) {
+ struct lustre_handle lock_handle;
+
+ if (it->it_lock_mode) {
+ lock_handle.cookie = it->it_lock_handle;
+ ldlm_lock_decref(&lock_handle, it->it_lock_mode);
+ }
+
+ it->it_lock_handle = 0;
+ it->it_lock_mode = 0;
+
+ if (it->it_remote_lock_mode) {
+ lock_handle.cookie = it->it_remote_lock_handle;
+ ldlm_lock_decref(&lock_handle,
+ it->it_remote_lock_mode);
+ }
+
+ it->it_remote_lock_handle = 0;
+ it->it_remote_lock_mode = 0;
+ }
+
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 014/124] staging: lustre: lmv: release locks if lmv_intent_lock() fails
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In lmv_intent_lock() if we will return an error then first release any
locks referenced by the intent.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5431
Reviewed-on: http://review.whamcloud.com/11319
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 22 ++++++++++++++++++++++
1 files changed, 22 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 85cc5cb..fd3dc7c 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -533,5 +533,27 @@ int lmv_intent_lock(struct obd_export *exp, struct md_op_data *op_data,
extra_lock_flags);
else
LBUG();
+
+ if (rc < 0) {
+ struct lustre_handle lock_handle;
+
+ if (it->it_lock_mode) {
+ lock_handle.cookie = it->it_lock_handle;
+ ldlm_lock_decref(&lock_handle, it->it_lock_mode);
+ }
+
+ it->it_lock_handle = 0;
+ it->it_lock_mode = 0;
+
+ if (it->it_remote_lock_mode) {
+ lock_handle.cookie = it->it_remote_lock_handle;
+ ldlm_lock_decref(&lock_handle,
+ it->it_remote_lock_mode);
+ }
+
+ it->it_remote_lock_handle = 0;
+ it->it_remote_lock_mode = 0;
+ }
+
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 015/124] staging: lustre: clio: lu_ref_del() mismatch ref add scope
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
'commit 77605e41a26f ("staging/lustre/clio: add pages into writeback cache
in batches")' adds a page to a list aggregate issuing them to writeback
cache; A page add is referenced in llite/vvp io scope, while writeback
cache commit de-refers it under osc sub io scope, and enabling -lu_ref
will detect this scope mismatch.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4503
Reviewed-on: http://review.whamcloud.com/8970
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/vvp_io.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index 94916dc..09ccd1f 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -821,7 +821,7 @@ static void write_commit_callback(const struct lu_env *env, struct cl_io *io,
cl_page_disown(env, io, page);
/* held in ll_cl_init() */
- lu_ref_del(&page->cp_reference, "cl_io", io);
+ lu_ref_del(&page->cp_reference, "cl_io", cl_io_top(io));
cl_page_put(env, page);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 015/124] staging: lustre: clio: lu_ref_del() mismatch ref add scope
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
'commit 77605e41a26f ("staging/lustre/clio: add pages into writeback cache
in batches")' adds a page to a list aggregate issuing them to writeback
cache; A page add is referenced in llite/vvp io scope, while writeback
cache commit de-refers it under osc sub io scope, and enabling -lu_ref
will detect this scope mismatch.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4503
Reviewed-on: http://review.whamcloud.com/8970
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/vvp_io.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index 94916dc..09ccd1f 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -821,7 +821,7 @@ static void write_commit_callback(const struct lu_env *env, struct cl_io *io,
cl_page_disown(env, io, page);
/* held in ll_cl_init() */
- lu_ref_del(&page->cp_reference, "cl_io", io);
+ lu_ref_del(&page->cp_reference, "cl_io", cl_io_top(io));
cl_page_put(env, page);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 016/124] staging: lustre: fix comparison between signed and unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Cleanup in general headers.
* use size_t in cfs_size_round*()
* make unsigned index and len in lustre_cfg_*()
* make iteration variable the same type as comparing value
* make unsigned pages counters
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5417
Reviewed-on: http://review.whamcloud.com/11327
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/include/linux/libcfs/libcfs_private.h | 12 ++++----
.../staging/lustre/lustre/include/lprocfs_status.h | 2 +-
drivers/staging/lustre/lustre/include/lustre_cfg.h | 26 +++++++++----------
drivers/staging/lustre/lustre/include/lustre_lmv.h | 4 +-
drivers/staging/lustre/lustre/include/lustre_net.h | 10 ++++---
drivers/staging/lustre/lustre/include/obd.h | 20 +++++++-------
drivers/staging/lustre/lustre/osc/osc_request.c | 9 ++++---
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 4 +-
8 files changed, 44 insertions(+), 43 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
index d401ae1..e0e1a5d 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
@@ -310,13 +310,13 @@ do { \
#define MKSTR(ptr) ((ptr)) ? (ptr) : ""
-static inline int cfs_size_round4(int val)
+static inline size_t cfs_size_round4(int val)
{
return (val + 3) & (~0x3);
}
#ifndef HAVE_CFS_SIZE_ROUND
-static inline int cfs_size_round(int val)
+static inline size_t cfs_size_round(int val)
{
return (val + 7) & (~0x7);
}
@@ -324,17 +324,17 @@ static inline int cfs_size_round(int val)
#define HAVE_CFS_SIZE_ROUND
#endif
-static inline int cfs_size_round16(int val)
+static inline size_t cfs_size_round16(int val)
{
return (val + 0xf) & (~0xf);
}
-static inline int cfs_size_round32(int val)
+static inline size_t cfs_size_round32(int val)
{
return (val + 0x1f) & (~0x1f);
}
-static inline int cfs_size_round0(int val)
+static inline size_t cfs_size_round0(int val)
{
if (!val)
return 0;
@@ -343,7 +343,7 @@ static inline int cfs_size_round0(int val)
static inline size_t cfs_round_strlen(char *fset)
{
- return (size_t)cfs_size_round((int)strlen(fset) + 1);
+ return cfs_size_round((int)strlen(fset) + 1);
}
#define LOGL(var, len, ptr) \
diff --git a/drivers/staging/lustre/lustre/include/lprocfs_status.h b/drivers/staging/lustre/lustre/include/lprocfs_status.h
index 795d446..b1a8ca5 100644
--- a/drivers/staging/lustre/lustre/include/lprocfs_status.h
+++ b/drivers/staging/lustre/lustre/include/lprocfs_status.h
@@ -499,7 +499,7 @@ static inline __u64 lprocfs_stats_collector(struct lprocfs_stats *stats,
int idx,
enum lprocfs_fields_flags field)
{
- int i;
+ unsigned int i;
unsigned int num_cpu;
unsigned long flags = 0;
__u64 ret = 0;
diff --git a/drivers/staging/lustre/lustre/include/lustre_cfg.h b/drivers/staging/lustre/lustre/include/lustre_cfg.h
index 95a0be1..8eb394e 100644
--- a/drivers/staging/lustre/lustre/include/lustre_cfg.h
+++ b/drivers/staging/lustre/lustre/include/lustre_cfg.h
@@ -151,13 +151,11 @@ static inline void lustre_cfg_bufs_reset(struct lustre_cfg_bufs *bufs, char *nam
lustre_cfg_bufs_set_string(bufs, 0, name);
}
-static inline void *lustre_cfg_buf(struct lustre_cfg *lcfg, int index)
+static inline void *lustre_cfg_buf(struct lustre_cfg *lcfg, __u32 index)
{
- int i;
- int offset;
- int bufcount;
-
- LASSERT(index >= 0);
+ __u32 i;
+ size_t offset;
+ __u32 bufcount;
bufcount = lcfg->lcfg_bufcount;
if (index >= bufcount)
@@ -172,7 +170,7 @@ static inline void *lustre_cfg_buf(struct lustre_cfg *lcfg, int index)
static inline void lustre_cfg_bufs_init(struct lustre_cfg_bufs *bufs,
struct lustre_cfg *lcfg)
{
- int i;
+ __u32 i;
bufs->lcfg_bufcount = lcfg->lcfg_bufcount;
for (i = 0; i < bufs->lcfg_bufcount; i++) {
@@ -181,7 +179,7 @@ static inline void lustre_cfg_bufs_init(struct lustre_cfg_bufs *bufs,
}
}
-static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, int index)
+static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, __u32 index)
{
char *s;
@@ -197,8 +195,8 @@ static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, int index)
* of data. Try to use the padding first though.
*/
if (s[lcfg->lcfg_buflens[index] - 1] != '\0') {
- int last = min((int)lcfg->lcfg_buflens[index],
- cfs_size_round(lcfg->lcfg_buflens[index]) - 1);
+ size_t last = min((size_t)lcfg->lcfg_buflens[index],
+ cfs_size_round(lcfg->lcfg_buflens[index]) - 1);
char lost = s[last];
s[last] = '\0';
@@ -210,10 +208,10 @@ static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, int index)
return s;
}
-static inline int lustre_cfg_len(__u32 bufcount, __u32 *buflens)
+static inline __u32 lustre_cfg_len(__u32 bufcount, __u32 *buflens)
{
- int i;
- int len;
+ __u32 i;
+ __u32 len;
len = LCFG_HDR_SIZE(bufcount);
for (i = 0; i < bufcount; i++)
@@ -254,7 +252,7 @@ static inline void lustre_cfg_free(struct lustre_cfg *lcfg)
return;
}
-static inline int lustre_cfg_sanity_check(void *buf, int len)
+static inline int lustre_cfg_sanity_check(void *buf, size_t len)
{
struct lustre_cfg *lcfg = (struct lustre_cfg *)buf;
diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h
index 085e596..5a2fc30 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lmv.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h
@@ -55,7 +55,7 @@ struct lmv_stripe_md {
static inline bool
lsm_md_eq(const struct lmv_stripe_md *lsm1, const struct lmv_stripe_md *lsm2)
{
- int idx;
+ __u32 idx;
if (lsm1->lsm_md_magic != lsm2->lsm_md_magic ||
lsm1->lsm_md_stripe_count != lsm2->lsm_md_stripe_count ||
@@ -92,7 +92,7 @@ static inline void lmv_free_memmd(struct lmv_stripe_md *lsm)
static inline void lmv1_le_to_cpu(struct lmv_mds_md_v1 *lmv_dst,
const struct lmv_mds_md_v1 *lmv_src)
{
- int i;
+ __u32 i;
lmv_dst->lmv_magic = le32_to_cpu(lmv_src->lmv_magic);
lmv_dst->lmv_stripe_count = le32_to_cpu(lmv_src->lmv_stripe_count);
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 6d82998..531c628 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -1631,7 +1631,7 @@ static inline bool ptlrpc_nrs_req_can_move(struct ptlrpc_request *req)
/**
* Returns 1 if request buffer at offset \a index was already swabbed
*/
-static inline int lustre_req_swabbed(struct ptlrpc_request *req, int index)
+static inline int lustre_req_swabbed(struct ptlrpc_request *req, size_t index)
{
LASSERT(index < sizeof(req->rq_req_swab_mask) * 8);
return req->rq_req_swab_mask & (1 << index);
@@ -1640,7 +1640,7 @@ static inline int lustre_req_swabbed(struct ptlrpc_request *req, int index)
/**
* Returns 1 if request reply buffer at offset \a index was already swabbed
*/
-static inline int lustre_rep_swabbed(struct ptlrpc_request *req, int index)
+static inline int lustre_rep_swabbed(struct ptlrpc_request *req, size_t index)
{
LASSERT(index < sizeof(req->rq_rep_swab_mask) * 8);
return req->rq_rep_swab_mask & (1 << index);
@@ -1665,7 +1665,8 @@ static inline int ptlrpc_rep_need_swab(struct ptlrpc_request *req)
/**
* Mark request buffer at offset \a index that it was already swabbed
*/
-static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, int index)
+static inline void lustre_set_req_swabbed(struct ptlrpc_request *req,
+ size_t index)
{
LASSERT(index < sizeof(req->rq_req_swab_mask) * 8);
LASSERT((req->rq_req_swab_mask & (1 << index)) == 0);
@@ -1675,7 +1676,8 @@ static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, int index)
/**
* Mark request reply buffer at offset \a index that it was already swabbed
*/
-static inline void lustre_set_rep_swabbed(struct ptlrpc_request *req, int index)
+static inline void lustre_set_rep_swabbed(struct ptlrpc_request *req,
+ size_t index)
{
LASSERT(index < sizeof(req->rq_rep_swab_mask) * 8);
LASSERT((req->rq_rep_swab_mask & (1 << index)) == 0);
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index f55e679..37a3acb 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -222,11 +222,11 @@ struct client_obd {
struct sptlrpc_flavor cl_flvr_mgc; /* fixed flavor of mgc->mgs */
/* the grant values are protected by loi_list_lock below */
- long cl_dirty_pages; /* all _dirty_ in pahges */
- long cl_dirty_max_pages;/* allowed w/o rpc */
- long cl_dirty_transit; /* dirty synchronous */
- long cl_avail_grant; /* bytes of credit for ost */
- long cl_lost_grant; /* lost credits (trunc) */
+ unsigned long cl_dirty_pages; /* all _dirty_ in pahges */
+ unsigned long cl_dirty_max_pages; /* allowed w/o rpc */
+ unsigned long cl_dirty_transit; /* dirty synchronous */
+ unsigned long cl_avail_grant; /* bytes of credit for ost */
+ unsigned long cl_lost_grant; /* lost credits (trunc) */
/* since we allocate grant by blocks, we don't know how many grant will
* be used to add a page into cache. As a solution, we reserve maximum
@@ -268,13 +268,13 @@ struct client_obd {
struct list_head cl_loi_hp_ready_list;
struct list_head cl_loi_write_list;
struct list_head cl_loi_read_list;
- int cl_r_in_flight;
- int cl_w_in_flight;
+ __u32 cl_r_in_flight;
+ __u32 cl_w_in_flight;
/* just a sum of the loi/lop pending numbers to be exported by sysfs */
atomic_t cl_pending_w_pages;
atomic_t cl_pending_r_pages;
__u32 cl_max_pages_per_rpc;
- int cl_max_rpcs_in_flight;
+ __u32 cl_max_rpcs_in_flight;
struct obd_histogram cl_read_rpc_hist;
struct obd_histogram cl_write_rpc_hist;
struct obd_histogram cl_read_page_hist;
@@ -1183,8 +1183,8 @@ static inline void client_adjust_max_dirty(struct client_obd *cli)
cli->cl_dirty_max_pages =
(OSC_MAX_DIRTY_DEFAULT * 1024 * 1024) >> PAGE_SHIFT;
else {
- long dirty_max = cli->cl_max_rpcs_in_flight *
- cli->cl_max_pages_per_rpc;
+ unsigned long dirty_max = cli->cl_max_rpcs_in_flight *
+ cli->cl_max_pages_per_rpc;
if (dirty_max > cli->cl_dirty_max_pages)
cli->cl_dirty_max_pages = dirty_max;
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index e44b4fa..5cca2b6 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -823,9 +823,10 @@ static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
cli->cl_dirty_pages, cli->cl_dirty_max_pages);
oa->o_undirty = 0;
} else {
- long max_in_flight = (cli->cl_max_pages_per_rpc <<
- PAGE_SHIFT) *
- (cli->cl_max_rpcs_in_flight + 1);
+ unsigned long max_in_flight;
+
+ max_in_flight = (cli->cl_max_pages_per_rpc << PAGE_SHIFT) *
+ (cli->cl_max_rpcs_in_flight + 1);
oa->o_undirty = max(cli->cl_dirty_max_pages << PAGE_SHIFT,
max_in_flight);
}
@@ -2048,7 +2049,7 @@ int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
}
spin_unlock(&cli->cl_loi_list_lock);
- DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
+ DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %ur/%dw in flight",
page_count, aa, cli->cl_r_in_flight,
cli->cl_w_in_flight);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 102f0aa..03fd5d7 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1734,7 +1734,7 @@ EXPORT_SYMBOL(lustre_swab_mgs_target_info);
void lustre_swab_mgs_nidtbl_entry(struct mgs_nidtbl_entry *entry)
{
- int i;
+ __u8 i;
__swab64s(&entry->mne_version);
__swab32s(&entry->mne_instance);
@@ -1823,7 +1823,7 @@ static void lustre_swab_fiemap_extent(struct ll_fiemap_extent *fm_extent)
void lustre_swab_fiemap(struct ll_user_fiemap *fiemap)
{
- int i;
+ __u32 i;
__swab64s(&fiemap->fm_start);
__swab64s(&fiemap->fm_length);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 016/124] staging: lustre: fix comparison between signed and unsigned
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Cleanup in general headers.
* use size_t in cfs_size_round*()
* make unsigned index and len in lustre_cfg_*()
* make iteration variable the same type as comparing value
* make unsigned pages counters
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5417
Reviewed-on: http://review.whamcloud.com/11327
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/include/linux/libcfs/libcfs_private.h | 12 ++++----
.../staging/lustre/lustre/include/lprocfs_status.h | 2 +-
drivers/staging/lustre/lustre/include/lustre_cfg.h | 26 +++++++++----------
drivers/staging/lustre/lustre/include/lustre_lmv.h | 4 +-
drivers/staging/lustre/lustre/include/lustre_net.h | 10 ++++---
drivers/staging/lustre/lustre/include/obd.h | 20 +++++++-------
drivers/staging/lustre/lustre/osc/osc_request.c | 9 ++++---
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 4 +-
8 files changed, 44 insertions(+), 43 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
index d401ae1..e0e1a5d 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
@@ -310,13 +310,13 @@ do { \
#define MKSTR(ptr) ((ptr)) ? (ptr) : ""
-static inline int cfs_size_round4(int val)
+static inline size_t cfs_size_round4(int val)
{
return (val + 3) & (~0x3);
}
#ifndef HAVE_CFS_SIZE_ROUND
-static inline int cfs_size_round(int val)
+static inline size_t cfs_size_round(int val)
{
return (val + 7) & (~0x7);
}
@@ -324,17 +324,17 @@ static inline int cfs_size_round(int val)
#define HAVE_CFS_SIZE_ROUND
#endif
-static inline int cfs_size_round16(int val)
+static inline size_t cfs_size_round16(int val)
{
return (val + 0xf) & (~0xf);
}
-static inline int cfs_size_round32(int val)
+static inline size_t cfs_size_round32(int val)
{
return (val + 0x1f) & (~0x1f);
}
-static inline int cfs_size_round0(int val)
+static inline size_t cfs_size_round0(int val)
{
if (!val)
return 0;
@@ -343,7 +343,7 @@ static inline int cfs_size_round0(int val)
static inline size_t cfs_round_strlen(char *fset)
{
- return (size_t)cfs_size_round((int)strlen(fset) + 1);
+ return cfs_size_round((int)strlen(fset) + 1);
}
#define LOGL(var, len, ptr) \
diff --git a/drivers/staging/lustre/lustre/include/lprocfs_status.h b/drivers/staging/lustre/lustre/include/lprocfs_status.h
index 795d446..b1a8ca5 100644
--- a/drivers/staging/lustre/lustre/include/lprocfs_status.h
+++ b/drivers/staging/lustre/lustre/include/lprocfs_status.h
@@ -499,7 +499,7 @@ static inline __u64 lprocfs_stats_collector(struct lprocfs_stats *stats,
int idx,
enum lprocfs_fields_flags field)
{
- int i;
+ unsigned int i;
unsigned int num_cpu;
unsigned long flags = 0;
__u64 ret = 0;
diff --git a/drivers/staging/lustre/lustre/include/lustre_cfg.h b/drivers/staging/lustre/lustre/include/lustre_cfg.h
index 95a0be1..8eb394e 100644
--- a/drivers/staging/lustre/lustre/include/lustre_cfg.h
+++ b/drivers/staging/lustre/lustre/include/lustre_cfg.h
@@ -151,13 +151,11 @@ static inline void lustre_cfg_bufs_reset(struct lustre_cfg_bufs *bufs, char *nam
lustre_cfg_bufs_set_string(bufs, 0, name);
}
-static inline void *lustre_cfg_buf(struct lustre_cfg *lcfg, int index)
+static inline void *lustre_cfg_buf(struct lustre_cfg *lcfg, __u32 index)
{
- int i;
- int offset;
- int bufcount;
-
- LASSERT(index >= 0);
+ __u32 i;
+ size_t offset;
+ __u32 bufcount;
bufcount = lcfg->lcfg_bufcount;
if (index >= bufcount)
@@ -172,7 +170,7 @@ static inline void *lustre_cfg_buf(struct lustre_cfg *lcfg, int index)
static inline void lustre_cfg_bufs_init(struct lustre_cfg_bufs *bufs,
struct lustre_cfg *lcfg)
{
- int i;
+ __u32 i;
bufs->lcfg_bufcount = lcfg->lcfg_bufcount;
for (i = 0; i < bufs->lcfg_bufcount; i++) {
@@ -181,7 +179,7 @@ static inline void lustre_cfg_bufs_init(struct lustre_cfg_bufs *bufs,
}
}
-static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, int index)
+static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, __u32 index)
{
char *s;
@@ -197,8 +195,8 @@ static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, int index)
* of data. Try to use the padding first though.
*/
if (s[lcfg->lcfg_buflens[index] - 1] != '\0') {
- int last = min((int)lcfg->lcfg_buflens[index],
- cfs_size_round(lcfg->lcfg_buflens[index]) - 1);
+ size_t last = min((size_t)lcfg->lcfg_buflens[index],
+ cfs_size_round(lcfg->lcfg_buflens[index]) - 1);
char lost = s[last];
s[last] = '\0';
@@ -210,10 +208,10 @@ static inline char *lustre_cfg_string(struct lustre_cfg *lcfg, int index)
return s;
}
-static inline int lustre_cfg_len(__u32 bufcount, __u32 *buflens)
+static inline __u32 lustre_cfg_len(__u32 bufcount, __u32 *buflens)
{
- int i;
- int len;
+ __u32 i;
+ __u32 len;
len = LCFG_HDR_SIZE(bufcount);
for (i = 0; i < bufcount; i++)
@@ -254,7 +252,7 @@ static inline void lustre_cfg_free(struct lustre_cfg *lcfg)
return;
}
-static inline int lustre_cfg_sanity_check(void *buf, int len)
+static inline int lustre_cfg_sanity_check(void *buf, size_t len)
{
struct lustre_cfg *lcfg = (struct lustre_cfg *)buf;
diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h
index 085e596..5a2fc30 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lmv.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h
@@ -55,7 +55,7 @@ struct lmv_stripe_md {
static inline bool
lsm_md_eq(const struct lmv_stripe_md *lsm1, const struct lmv_stripe_md *lsm2)
{
- int idx;
+ __u32 idx;
if (lsm1->lsm_md_magic != lsm2->lsm_md_magic ||
lsm1->lsm_md_stripe_count != lsm2->lsm_md_stripe_count ||
@@ -92,7 +92,7 @@ static inline void lmv_free_memmd(struct lmv_stripe_md *lsm)
static inline void lmv1_le_to_cpu(struct lmv_mds_md_v1 *lmv_dst,
const struct lmv_mds_md_v1 *lmv_src)
{
- int i;
+ __u32 i;
lmv_dst->lmv_magic = le32_to_cpu(lmv_src->lmv_magic);
lmv_dst->lmv_stripe_count = le32_to_cpu(lmv_src->lmv_stripe_count);
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 6d82998..531c628 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -1631,7 +1631,7 @@ static inline bool ptlrpc_nrs_req_can_move(struct ptlrpc_request *req)
/**
* Returns 1 if request buffer at offset \a index was already swabbed
*/
-static inline int lustre_req_swabbed(struct ptlrpc_request *req, int index)
+static inline int lustre_req_swabbed(struct ptlrpc_request *req, size_t index)
{
LASSERT(index < sizeof(req->rq_req_swab_mask) * 8);
return req->rq_req_swab_mask & (1 << index);
@@ -1640,7 +1640,7 @@ static inline int lustre_req_swabbed(struct ptlrpc_request *req, int index)
/**
* Returns 1 if request reply buffer@offset \a index was already swabbed
*/
-static inline int lustre_rep_swabbed(struct ptlrpc_request *req, int index)
+static inline int lustre_rep_swabbed(struct ptlrpc_request *req, size_t index)
{
LASSERT(index < sizeof(req->rq_rep_swab_mask) * 8);
return req->rq_rep_swab_mask & (1 << index);
@@ -1665,7 +1665,8 @@ static inline int ptlrpc_rep_need_swab(struct ptlrpc_request *req)
/**
* Mark request buffer@offset \a index that it was already swabbed
*/
-static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, int index)
+static inline void lustre_set_req_swabbed(struct ptlrpc_request *req,
+ size_t index)
{
LASSERT(index < sizeof(req->rq_req_swab_mask) * 8);
LASSERT((req->rq_req_swab_mask & (1 << index)) == 0);
@@ -1675,7 +1676,8 @@ static inline void lustre_set_req_swabbed(struct ptlrpc_request *req, int index)
/**
* Mark request reply buffer@offset \a index that it was already swabbed
*/
-static inline void lustre_set_rep_swabbed(struct ptlrpc_request *req, int index)
+static inline void lustre_set_rep_swabbed(struct ptlrpc_request *req,
+ size_t index)
{
LASSERT(index < sizeof(req->rq_rep_swab_mask) * 8);
LASSERT((req->rq_rep_swab_mask & (1 << index)) == 0);
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index f55e679..37a3acb 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -222,11 +222,11 @@ struct client_obd {
struct sptlrpc_flavor cl_flvr_mgc; /* fixed flavor of mgc->mgs */
/* the grant values are protected by loi_list_lock below */
- long cl_dirty_pages; /* all _dirty_ in pahges */
- long cl_dirty_max_pages;/* allowed w/o rpc */
- long cl_dirty_transit; /* dirty synchronous */
- long cl_avail_grant; /* bytes of credit for ost */
- long cl_lost_grant; /* lost credits (trunc) */
+ unsigned long cl_dirty_pages; /* all _dirty_ in pahges */
+ unsigned long cl_dirty_max_pages; /* allowed w/o rpc */
+ unsigned long cl_dirty_transit; /* dirty synchronous */
+ unsigned long cl_avail_grant; /* bytes of credit for ost */
+ unsigned long cl_lost_grant; /* lost credits (trunc) */
/* since we allocate grant by blocks, we don't know how many grant will
* be used to add a page into cache. As a solution, we reserve maximum
@@ -268,13 +268,13 @@ struct client_obd {
struct list_head cl_loi_hp_ready_list;
struct list_head cl_loi_write_list;
struct list_head cl_loi_read_list;
- int cl_r_in_flight;
- int cl_w_in_flight;
+ __u32 cl_r_in_flight;
+ __u32 cl_w_in_flight;
/* just a sum of the loi/lop pending numbers to be exported by sysfs */
atomic_t cl_pending_w_pages;
atomic_t cl_pending_r_pages;
__u32 cl_max_pages_per_rpc;
- int cl_max_rpcs_in_flight;
+ __u32 cl_max_rpcs_in_flight;
struct obd_histogram cl_read_rpc_hist;
struct obd_histogram cl_write_rpc_hist;
struct obd_histogram cl_read_page_hist;
@@ -1183,8 +1183,8 @@ static inline void client_adjust_max_dirty(struct client_obd *cli)
cli->cl_dirty_max_pages =
(OSC_MAX_DIRTY_DEFAULT * 1024 * 1024) >> PAGE_SHIFT;
else {
- long dirty_max = cli->cl_max_rpcs_in_flight *
- cli->cl_max_pages_per_rpc;
+ unsigned long dirty_max = cli->cl_max_rpcs_in_flight *
+ cli->cl_max_pages_per_rpc;
if (dirty_max > cli->cl_dirty_max_pages)
cli->cl_dirty_max_pages = dirty_max;
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index e44b4fa..5cca2b6 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -823,9 +823,10 @@ static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
cli->cl_dirty_pages, cli->cl_dirty_max_pages);
oa->o_undirty = 0;
} else {
- long max_in_flight = (cli->cl_max_pages_per_rpc <<
- PAGE_SHIFT) *
- (cli->cl_max_rpcs_in_flight + 1);
+ unsigned long max_in_flight;
+
+ max_in_flight = (cli->cl_max_pages_per_rpc << PAGE_SHIFT) *
+ (cli->cl_max_rpcs_in_flight + 1);
oa->o_undirty = max(cli->cl_dirty_max_pages << PAGE_SHIFT,
max_in_flight);
}
@@ -2048,7 +2049,7 @@ int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
}
spin_unlock(&cli->cl_loi_list_lock);
- DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
+ DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %ur/%dw in flight",
page_count, aa, cli->cl_r_in_flight,
cli->cl_w_in_flight);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 102f0aa..03fd5d7 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1734,7 +1734,7 @@ EXPORT_SYMBOL(lustre_swab_mgs_target_info);
void lustre_swab_mgs_nidtbl_entry(struct mgs_nidtbl_entry *entry)
{
- int i;
+ __u8 i;
__swab64s(&entry->mne_version);
__swab32s(&entry->mne_instance);
@@ -1823,7 +1823,7 @@ static void lustre_swab_fiemap_extent(struct ll_fiemap_extent *fm_extent)
void lustre_swab_fiemap(struct ll_user_fiemap *fiemap)
{
- int i;
+ __u32 i;
__swab64s(&fiemap->fm_start);
__swab64s(&fiemap->fm_length);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 017/124] staging: lustre: lov: adjust page bufsize after layout change
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
Otherwise, the coh_page_bufsize keeps increasing when the file's
layout keeps changing in lov_init_raid0().
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5459
Reviewed-on: http://review.whamcloud.com/11394
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lov/lov_object.c | 8 ++++++++
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lov/lov_object.c b/drivers/staging/lustre/lustre/lov/lov_object.c
index 2a52d0c..4f766e0 100644
--- a/drivers/staging/lustre/lustre/lov/lov_object.c
+++ b/drivers/staging/lustre/lustre/lov/lov_object.c
@@ -195,6 +195,10 @@ static int lov_page_slice_fixup(struct lov_object *lov,
struct cl_object_header *hdr = cl_object_header(&lov->lo_cl);
struct cl_object *o;
+ if (!stripe)
+ return hdr->coh_page_bufsize - lov->lo_cl.co_slice_off -
+ cfs_size_round(sizeof(struct lov_page));
+
cl_object_for_each(o, stripe)
o->co_slice_off += hdr->coh_page_bufsize;
@@ -720,6 +724,10 @@ static int lov_layout_change(const struct lu_env *unused,
LASSERT(atomic_read(&lov->lo_active_ios) == 0);
lov->lo_type = LLT_EMPTY;
+ /* page bufsize fixup */
+ cl_object_header(&lov->lo_cl)->coh_page_bufsize -=
+ lov_page_slice_fixup(lov, NULL);
+
result = new_ops->llo_init(env,
lu2lov_dev(lov->lo_cl.co_lu.lo_dev),
lov, conf, state);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 017/124] staging: lustre: lov: adjust page bufsize after layout change
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
Otherwise, the coh_page_bufsize keeps increasing when the file's
layout keeps changing in lov_init_raid0().
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5459
Reviewed-on: http://review.whamcloud.com/11394
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lov/lov_object.c | 8 ++++++++
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lov/lov_object.c b/drivers/staging/lustre/lustre/lov/lov_object.c
index 2a52d0c..4f766e0 100644
--- a/drivers/staging/lustre/lustre/lov/lov_object.c
+++ b/drivers/staging/lustre/lustre/lov/lov_object.c
@@ -195,6 +195,10 @@ static int lov_page_slice_fixup(struct lov_object *lov,
struct cl_object_header *hdr = cl_object_header(&lov->lo_cl);
struct cl_object *o;
+ if (!stripe)
+ return hdr->coh_page_bufsize - lov->lo_cl.co_slice_off -
+ cfs_size_round(sizeof(struct lov_page));
+
cl_object_for_each(o, stripe)
o->co_slice_off += hdr->coh_page_bufsize;
@@ -720,6 +724,10 @@ static int lov_layout_change(const struct lu_env *unused,
LASSERT(atomic_read(&lov->lo_active_ios) == 0);
lov->lo_type = LLT_EMPTY;
+ /* page bufsize fixup */
+ cl_object_header(&lov->lo_cl)->coh_page_bufsize -=
+ lov_page_slice_fixup(lov, NULL);
+
result = new_ops->llo_init(env,
lu2lov_dev(lov->lo_cl.co_lu.lo_dev),
lov, conf, state);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 018/124] staging: lustre: obdclass: fix comparison between signed and unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Make lu_buf->lb_len unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5417
Reviewed-on: http://review.whamcloud.com/11281
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 2 +-
drivers/staging/lustre/lustre/obdclass/lu_object.c | 16 ++++++++--------
2 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index fe40b42..66ab98a 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -1290,7 +1290,7 @@ static inline bool lu_name_is_valid_2(const char *name, size_t name_len)
*/
struct lu_buf {
void *lb_buf;
- ssize_t lb_len;
+ size_t lb_len;
};
#define DLUBUF "(%p %zu)"
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index b6fd9af..44a854c 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -341,7 +341,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
int start;
int count;
int bnr;
- int i;
+ unsigned int i;
if (OBD_FAIL_CHECK(OBD_FAIL_OBD_NO_LRU))
return 0;
@@ -869,11 +869,11 @@ EXPORT_SYMBOL(lu_site_print);
/**
* Return desired hash table order.
*/
-static int lu_htable_order(struct lu_device *top)
+static unsigned int lu_htable_order(struct lu_device *top)
{
unsigned long bits_max = LU_SITE_BITS_MAX;
unsigned long cache_size;
- int bits;
+ unsigned int bits;
/*
* Calculate hash table size, assuming that we want reasonable
@@ -1325,7 +1325,7 @@ static unsigned key_set_version;
int lu_context_key_register(struct lu_context_key *key)
{
int result;
- int i;
+ unsigned int i;
LASSERT(key->lct_init);
LASSERT(key->lct_fini);
@@ -1531,7 +1531,7 @@ EXPORT_SYMBOL(lu_context_key_revive);
static void keys_fini(struct lu_context *ctx)
{
- int i;
+ unsigned int i;
if (!ctx->lc_value)
return;
@@ -1545,7 +1545,7 @@ static void keys_fini(struct lu_context *ctx)
static int keys_fill(struct lu_context *ctx)
{
- int i;
+ unsigned int i;
LINVRNT(ctx->lc_value);
for (i = 0; i < ARRAY_SIZE(lu_keys); ++i) {
@@ -1658,7 +1658,7 @@ EXPORT_SYMBOL(lu_context_enter);
*/
void lu_context_exit(struct lu_context *ctx)
{
- int i;
+ unsigned int i;
LINVRNT(ctx->lc_state == LCS_ENTERED);
ctx->lc_state = LCS_LEFT;
@@ -1740,7 +1740,7 @@ static void lu_site_stats_get(struct cfs_hash *hs,
struct lu_site_stats *stats, int populated)
{
struct cfs_hash_bd bd;
- int i;
+ unsigned int i;
cfs_hash_for_each_bucket(hs, &bd, i) {
struct lu_site_bkt_data *bkt = cfs_hash_bd_extra_get(hs, &bd);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 018/124] staging: lustre: obdclass: fix comparison between signed and unsigned
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Make lu_buf->lb_len unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5417
Reviewed-on: http://review.whamcloud.com/11281
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 2 +-
drivers/staging/lustre/lustre/obdclass/lu_object.c | 16 ++++++++--------
2 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index fe40b42..66ab98a 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -1290,7 +1290,7 @@ static inline bool lu_name_is_valid_2(const char *name, size_t name_len)
*/
struct lu_buf {
void *lb_buf;
- ssize_t lb_len;
+ size_t lb_len;
};
#define DLUBUF "(%p %zu)"
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index b6fd9af..44a854c 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -341,7 +341,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
int start;
int count;
int bnr;
- int i;
+ unsigned int i;
if (OBD_FAIL_CHECK(OBD_FAIL_OBD_NO_LRU))
return 0;
@@ -869,11 +869,11 @@ EXPORT_SYMBOL(lu_site_print);
/**
* Return desired hash table order.
*/
-static int lu_htable_order(struct lu_device *top)
+static unsigned int lu_htable_order(struct lu_device *top)
{
unsigned long bits_max = LU_SITE_BITS_MAX;
unsigned long cache_size;
- int bits;
+ unsigned int bits;
/*
* Calculate hash table size, assuming that we want reasonable
@@ -1325,7 +1325,7 @@ static unsigned key_set_version;
int lu_context_key_register(struct lu_context_key *key)
{
int result;
- int i;
+ unsigned int i;
LASSERT(key->lct_init);
LASSERT(key->lct_fini);
@@ -1531,7 +1531,7 @@ EXPORT_SYMBOL(lu_context_key_revive);
static void keys_fini(struct lu_context *ctx)
{
- int i;
+ unsigned int i;
if (!ctx->lc_value)
return;
@@ -1545,7 +1545,7 @@ static void keys_fini(struct lu_context *ctx)
static int keys_fill(struct lu_context *ctx)
{
- int i;
+ unsigned int i;
LINVRNT(ctx->lc_value);
for (i = 0; i < ARRAY_SIZE(lu_keys); ++i) {
@@ -1658,7 +1658,7 @@ EXPORT_SYMBOL(lu_context_enter);
*/
void lu_context_exit(struct lu_context *ctx)
{
- int i;
+ unsigned int i;
LINVRNT(ctx->lc_state == LCS_ENTERED);
ctx->lc_state = LCS_LEFT;
@@ -1740,7 +1740,7 @@ static void lu_site_stats_get(struct cfs_hash *hs,
struct lu_site_stats *stats, int populated)
{
struct cfs_hash_bd bd;
- int i;
+ unsigned int i;
cfs_hash_for_each_bucket(hs, &bd, i) {
struct lu_site_bkt_data *bkt = cfs_hash_bd_extra_get(hs, &bd);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 019/124] staging: lustre: ptlrpc: fix magic return value of ptlrpc_init_portals
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Wang Shilong,
James Simmons
From: Wang Shilong <wshilong@ddn.com>
Previously, when running 'modprobe lustre', it hit the following
error message which is becaue of network initialisation failure:
modprobe: ERROR: could not insert 'lustre': Input/output error
However, error code is there, just let it return to caller,
after this patch, error message will be something like:
modprobe: ERROR: could not insert 'lustre': Network is down
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5455
Reviewed-on: http://review.whamcloud.com/11337
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/events.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/events.c b/drivers/staging/lustre/lustre/ptlrpc/events.c
index b1ce725..283dfb2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/events.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/events.c
@@ -543,7 +543,7 @@ static int ptlrpc_ni_init(void)
rc = LNetNIInit(pid);
if (rc < 0) {
CDEBUG(D_NET, "Can't init network interface: %d\n", rc);
- return -ENOENT;
+ return rc;
}
/* CAVEAT EMPTOR: how we process portals events is _radically_
@@ -561,7 +561,7 @@ static int ptlrpc_ni_init(void)
CERROR("Failed to allocate event queue: %d\n", rc);
LNetNIFini();
- return -ENOMEM;
+ return rc;
}
int ptlrpc_init_portals(void)
@@ -570,7 +570,7 @@ int ptlrpc_init_portals(void)
if (rc != 0) {
CERROR("network initialisation failed\n");
- return -EIO;
+ return rc;
}
rc = ptlrpcd_addref();
if (rc == 0)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 019/124] staging: lustre: ptlrpc: fix magic return value of ptlrpc_init_portals
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Wang Shilong,
James Simmons
From: Wang Shilong <wshilong@ddn.com>
Previously, when running 'modprobe lustre', it hit the following
error message which is becaue of network initialisation failure:
modprobe: ERROR: could not insert 'lustre': Input/output error
However, error code is there, just let it return to caller,
after this patch, error message will be something like:
modprobe: ERROR: could not insert 'lustre': Network is down
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5455
Reviewed-on: http://review.whamcloud.com/11337
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/events.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/events.c b/drivers/staging/lustre/lustre/ptlrpc/events.c
index b1ce725..283dfb2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/events.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/events.c
@@ -543,7 +543,7 @@ static int ptlrpc_ni_init(void)
rc = LNetNIInit(pid);
if (rc < 0) {
CDEBUG(D_NET, "Can't init network interface: %d\n", rc);
- return -ENOENT;
+ return rc;
}
/* CAVEAT EMPTOR: how we process portals events is _radically_
@@ -561,7 +561,7 @@ static int ptlrpc_ni_init(void)
CERROR("Failed to allocate event queue: %d\n", rc);
LNetNIFini();
- return -ENOMEM;
+ return rc;
}
int ptlrpc_init_portals(void)
@@ -570,7 +570,7 @@ int ptlrpc_init_portals(void)
if (rc != 0) {
CERROR("network initialisation failed\n");
- return -EIO;
+ return rc;
}
rc = ptlrpcd_addref();
if (rc == 0)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 020/124] staging: lustre: lmv: release request in lmv_revalidate_slaves()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In lmv_revalidate_slaves() ensure that the request returned by
md_intent_lock() is properly released on all paths.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5452
Reviewed-on: http://review.whamcloud.com/11326
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 16 +++++++++-------
1 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index fd3dc7c..f24848f 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -155,6 +155,7 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
+ struct ptlrpc_request *req = NULL;
struct mdt_body *body;
struct md_op_data *op_data;
unsigned long size = 0;
@@ -178,7 +179,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
*/
for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
struct lookup_intent it = { .it_op = IT_GETATTR };
- struct ptlrpc_request *req = NULL;
struct lustre_handle *lockh = NULL;
struct lmv_tgt_desc *tgt = NULL;
struct inode *inode;
@@ -205,6 +205,11 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
CDEBUG(D_INODE, "Revalidate slave "DFID" -> mds #%d\n",
PFID(&fid), tgt->ltd_idx);
+ if (req) {
+ ptlrpc_req_finished(req);
+ req = NULL;
+ }
+
rc = md_intent_lock(tgt->ltd_exp, op_data, &it, &req,
cb_blocking, extra_lock_flags);
if (rc < 0)
@@ -228,9 +233,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
PFID(&lsm->lsm_md_oinfo[i].lmo_fid),
PFID(&lsm->lsm_md_oinfo[0].lmo_fid));
- if (req)
- ptlrpc_req_finished(req);
-
if (it.it_lock_mode && lockh) {
ldlm_lock_decref(lockh, it.it_lock_mode);
it.it_lock_mode = 0;
@@ -245,9 +247,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
LTIME_S(inode->i_atime) = body->mbo_atime;
LTIME_S(inode->i_ctime) = body->mbo_ctime;
LTIME_S(inode->i_mtime) = body->mbo_mtime;
-
- if (req)
- ptlrpc_req_finished(req);
}
md_set_lock_data(tgt->ltd_exp, lockh, inode, NULL);
@@ -288,6 +287,9 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
mbody->mbo_mtime = mtime;
}
cleanup:
+ if (req)
+ ptlrpc_req_finished(req);
+
kfree(op_data);
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 020/124] staging: lustre: lmv: release request in lmv_revalidate_slaves()
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In lmv_revalidate_slaves() ensure that the request returned by
md_intent_lock() is properly released on all paths.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5452
Reviewed-on: http://review.whamcloud.com/11326
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 16 +++++++++-------
1 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index fd3dc7c..f24848f 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -155,6 +155,7 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
+ struct ptlrpc_request *req = NULL;
struct mdt_body *body;
struct md_op_data *op_data;
unsigned long size = 0;
@@ -178,7 +179,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
*/
for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
struct lookup_intent it = { .it_op = IT_GETATTR };
- struct ptlrpc_request *req = NULL;
struct lustre_handle *lockh = NULL;
struct lmv_tgt_desc *tgt = NULL;
struct inode *inode;
@@ -205,6 +205,11 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
CDEBUG(D_INODE, "Revalidate slave "DFID" -> mds #%d\n",
PFID(&fid), tgt->ltd_idx);
+ if (req) {
+ ptlrpc_req_finished(req);
+ req = NULL;
+ }
+
rc = md_intent_lock(tgt->ltd_exp, op_data, &it, &req,
cb_blocking, extra_lock_flags);
if (rc < 0)
@@ -228,9 +233,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
PFID(&lsm->lsm_md_oinfo[i].lmo_fid),
PFID(&lsm->lsm_md_oinfo[0].lmo_fid));
- if (req)
- ptlrpc_req_finished(req);
-
if (it.it_lock_mode && lockh) {
ldlm_lock_decref(lockh, it.it_lock_mode);
it.it_lock_mode = 0;
@@ -245,9 +247,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
LTIME_S(inode->i_atime) = body->mbo_atime;
LTIME_S(inode->i_ctime) = body->mbo_ctime;
LTIME_S(inode->i_mtime) = body->mbo_mtime;
-
- if (req)
- ptlrpc_req_finished(req);
}
md_set_lock_data(tgt->ltd_exp, lockh, inode, NULL);
@@ -288,6 +287,9 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
mbody->mbo_mtime = mtime;
}
cleanup:
+ if (req)
+ ptlrpc_req_finished(req);
+
kfree(op_data);
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 021/124] staging: lustre: build: bump build version warnings to x.y.53
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Andreas Dilger, Emoly Liu, James Simmons
From: Andreas Dilger <andreas.dilger@intel.com>
Move the LUSTRE_VERSION_CODE checks to trigger on x.y.53 instead of
x.y.50, so that it is into the development cycle that they are hit
instead of right at the start.
In many cases, the #warning has been removed (to prevent build errors)
and instead the code is just disabled outright. The dead code can be
seen easily and removed in the future with less interruption to the
development process.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4217
Reviewed-on: http://review.whamcloud.com/8630
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Tested-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/lustre_import.h | 2 ++
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/mgc/mgc_request.c | 4 +---
drivers/staging/lustre/lustre/obdclass/obd_mount.c | 4 +---
drivers/staging/lustre/lustre/ptlrpc/import.c | 4 +---
5 files changed, 6 insertions(+), 10 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_import.h b/drivers/staging/lustre/lustre/include/lustre_import.h
index b7a7a74..5461ba3 100644
--- a/drivers/staging/lustre/lustre/include/lustre_import.h
+++ b/drivers/staging/lustre/lustre/include/lustre_import.h
@@ -285,8 +285,10 @@ struct obd_import {
imp_resend_replay:1,
/* disable normal recovery, for test only. */
imp_no_pinger_recover:1,
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
/* need IR MNE swab */
imp_need_mne_swab:1,
+#endif
/* import must be reconnected instead of
* chosing new connection
*/
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 37a3acb..838a428 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -317,7 +317,7 @@ struct client_obd {
/* used by quotacheck when the servers are older than 2.4 */
int cl_qchk_stat; /* quotacheck stat of the peer */
#define CL_NOT_QUOTACHECKED 1 /* client->cl_qchk_stat init value */
-#if LUSTRE_VERSION_CODE >= OBD_OCD_VERSION(2, 7, 50, 0)
+#if OBD_OCD_VERSION(2, 7, 53, 0) < LUSTRE_VERSION_CODE
#warning "please consider removing quotacheck compatibility code"
#endif
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index 9d0bd47..f3d4f7f 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -1428,14 +1428,12 @@ again:
}
mne_swab = !!ptlrpc_rep_need_swab(req);
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 2, 50, 0)
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
/* This import flag means the server did an extra swab of IR MNE
* records (fixed in LU-1252), reverse it here if needed. LU-1644
*/
if (unlikely(req->rq_import->imp_need_mne_swab))
mne_swab = !mne_swab;
-#else
-#warning "LU-1644: Remove old OBD_CONNECT_MNE_SWAB fixup and imp_need_mne_swab"
#endif
for (i = 0; i < nrpages && ealen > 0; i++) {
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_mount.c b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
index 0273768..0df1aa6 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_mount.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
@@ -384,10 +384,8 @@ int lustre_start_mgc(struct super_block *sb)
OBD_CONNECT_FULL20 | OBD_CONNECT_IMP_RECOV |
OBD_CONNECT_LVB_TYPE;
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 2, 50, 0)
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
data->ocd_connect_flags |= OBD_CONNECT_MNE_SWAB;
-#else
-#warning "LU-1644: Remove old OBD_CONNECT_MNE_SWAB fixup and imp_need_mne_swab"
#endif
if (lmd_is_client(lsi->lsi_lmd) &&
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 93b1e78..013a957 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -1042,7 +1042,7 @@ finish:
warned = true;
}
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 2, 50, 0)
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
/* Check if server has LU-1252 fix applied to not always swab
* the IR MNE entries. Do this only once per connection. This
* fixup is version-limited, because we don't want to carry the
@@ -1060,8 +1060,6 @@ finish:
imp->imp_need_mne_swab = 1;
else /* clear if server was upgraded since last connect */
imp->imp_need_mne_swab = 0;
-#else
-#warning "LU-1644: Remove old OBD_CONNECT_MNE_SWAB fixup and imp_need_mne_swab"
#endif
if (ocd->ocd_connect_flags & OBD_CONNECT_CKSUM) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 021/124] staging: lustre: build: bump build version warnings to x.y.53
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Emoly Liu,
James Simmons
From: Andreas Dilger <andreas.dilger@intel.com>
Move the LUSTRE_VERSION_CODE checks to trigger on x.y.53 instead of
x.y.50, so that it is into the development cycle that they are hit
instead of right at the start.
In many cases, the #warning has been removed (to prevent build errors)
and instead the code is just disabled outright. The dead code can be
seen easily and removed in the future with less interruption to the
development process.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Emoly Liu <emoly.liu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4217
Reviewed-on: http://review.whamcloud.com/8630
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Tested-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/lustre_import.h | 2 ++
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/mgc/mgc_request.c | 4 +---
drivers/staging/lustre/lustre/obdclass/obd_mount.c | 4 +---
drivers/staging/lustre/lustre/ptlrpc/import.c | 4 +---
5 files changed, 6 insertions(+), 10 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_import.h b/drivers/staging/lustre/lustre/include/lustre_import.h
index b7a7a74..5461ba3 100644
--- a/drivers/staging/lustre/lustre/include/lustre_import.h
+++ b/drivers/staging/lustre/lustre/include/lustre_import.h
@@ -285,8 +285,10 @@ struct obd_import {
imp_resend_replay:1,
/* disable normal recovery, for test only. */
imp_no_pinger_recover:1,
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
/* need IR MNE swab */
imp_need_mne_swab:1,
+#endif
/* import must be reconnected instead of
* chosing new connection
*/
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 37a3acb..838a428 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -317,7 +317,7 @@ struct client_obd {
/* used by quotacheck when the servers are older than 2.4 */
int cl_qchk_stat; /* quotacheck stat of the peer */
#define CL_NOT_QUOTACHECKED 1 /* client->cl_qchk_stat init value */
-#if LUSTRE_VERSION_CODE >= OBD_OCD_VERSION(2, 7, 50, 0)
+#if OBD_OCD_VERSION(2, 7, 53, 0) < LUSTRE_VERSION_CODE
#warning "please consider removing quotacheck compatibility code"
#endif
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index 9d0bd47..f3d4f7f 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -1428,14 +1428,12 @@ again:
}
mne_swab = !!ptlrpc_rep_need_swab(req);
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 2, 50, 0)
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
/* This import flag means the server did an extra swab of IR MNE
* records (fixed in LU-1252), reverse it here if needed. LU-1644
*/
if (unlikely(req->rq_import->imp_need_mne_swab))
mne_swab = !mne_swab;
-#else
-#warning "LU-1644: Remove old OBD_CONNECT_MNE_SWAB fixup and imp_need_mne_swab"
#endif
for (i = 0; i < nrpages && ealen > 0; i++) {
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_mount.c b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
index 0273768..0df1aa6 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_mount.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
@@ -384,10 +384,8 @@ int lustre_start_mgc(struct super_block *sb)
OBD_CONNECT_FULL20 | OBD_CONNECT_IMP_RECOV |
OBD_CONNECT_LVB_TYPE;
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 2, 50, 0)
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
data->ocd_connect_flags |= OBD_CONNECT_MNE_SWAB;
-#else
-#warning "LU-1644: Remove old OBD_CONNECT_MNE_SWAB fixup and imp_need_mne_swab"
#endif
if (lmd_is_client(lsi->lsi_lmd) &&
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 93b1e78..013a957 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -1042,7 +1042,7 @@ finish:
warned = true;
}
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 2, 50, 0)
+#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
/* Check if server has LU-1252 fix applied to not always swab
* the IR MNE entries. Do this only once per connection. This
* fixup is version-limited, because we don't want to carry the
@@ -1060,8 +1060,6 @@ finish:
imp->imp_need_mne_swab = 1;
else /* clear if server was upgraded since last connect */
imp->imp_need_mne_swab = 0;
-#else
-#warning "LU-1644: Remove old OBD_CONNECT_MNE_SWAB fixup and imp_need_mne_swab"
#endif
if (ocd->ocd_connect_flags & OBD_CONNECT_CKSUM) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 022/124] staging: lustre: llog: add newly opened llog at tail of handle list
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi,
Lai Siyao, James Simmons
From: Li Xi <lixi@ddn.com>
Add newly opened llog handle at the tail of handle list to increase
lookup speed, especially for cancel operation, because the canceled
log will be removed from the list, and lookup is from the beginning.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5405
Reviewed-on: http://review.whamcloud.com/11575
Reviewed-by: Ian Costello <costello.ian@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/llog_cat.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_cat.c b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
index d28751a..b27f08f 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_cat.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
@@ -107,7 +107,7 @@ static int llog_cat_id2handle(const struct lu_env *env,
}
down_write(&cathandle->lgh_lock);
- list_add(&loghandle->u.phd.phd_entry, &cathandle->u.chd.chd_head);
+ list_add_tail(&loghandle->u.phd.phd_entry, &cathandle->u.chd.chd_head);
up_write(&cathandle->lgh_lock);
loghandle->u.phd.phd_cat_handle = cathandle;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 022/124] staging: lustre: llog: add newly opened llog at tail of handle list
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi,
Lai Siyao, James Simmons
From: Li Xi <lixi@ddn.com>
Add newly opened llog handle at the tail of handle list to increase
lookup speed, especially for cancel operation, because the canceled
log will be removed from the list, and lookup is from the beginning.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5405
Reviewed-on: http://review.whamcloud.com/11575
Reviewed-by: Ian Costello <costello.ian@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/llog_cat.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_cat.c b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
index d28751a..b27f08f 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_cat.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
@@ -107,7 +107,7 @@ static int llog_cat_id2handle(const struct lu_env *env,
}
down_write(&cathandle->lgh_lock);
- list_add(&loghandle->u.phd.phd_entry, &cathandle->u.chd.chd_head);
+ list_add_tail(&loghandle->u.phd.phd_entry, &cathandle->u.chd.chd_head);
up_write(&cathandle->lgh_lock);
loghandle->u.phd.phd_cat_handle = cathandle;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 023/124] staging: lustre: mdc: Report D_CHANGELOG messages as D_HSM
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Removed the D_CHANGELOG pseudo-debug flag that wasn't actually defined
as a usable one. Report the D_CHANGELOG messages as D_HSM ones instead
since this is the primary user of these messages.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5538
Reviewed-on: http://review.whamcloud.com/11558
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mdc/mdc_request.c | 12 +++++-------
1 files changed, 5 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index cf94926..956b946 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1857,8 +1857,6 @@ static struct kuc_hdr *changelog_kuc_hdr(char *buf, int len, int flags)
return lh;
}
-#define D_CHANGELOG 0
-
struct changelog_show {
__u64 cs_startrec;
__u32 cs_flags;
@@ -1885,12 +1883,12 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
if (rec->cr.cr_index < cs->cs_startrec) {
/* Skip entries earlier than what we are interested in */
- CDEBUG(D_CHANGELOG, "rec=%llu start=%llu\n",
+ CDEBUG(D_HSM, "rec=%llu start=%llu\n",
rec->cr.cr_index, cs->cs_startrec);
return 0;
}
- CDEBUG(D_CHANGELOG, "%llu %02d%-5s %llu 0x%x t="DFID" p="DFID
+ CDEBUG(D_HSM, "%llu %02d%-5s %llu 0x%x t=" DFID " p=" DFID
" %.*s\n", rec->cr.cr_index, rec->cr.cr_type,
changelog_type2str(rec->cr.cr_type), rec->cr.cr_time,
rec->cr.cr_flags & CLF_FLAGMASK,
@@ -1904,7 +1902,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
memcpy(lh + 1, &rec->cr, len - sizeof(*lh));
rc = libcfs_kkuc_msg_put(cs->cs_fp, lh);
- CDEBUG(D_CHANGELOG, "kucmsg fp %p len %d rc %d\n", cs->cs_fp, len, rc);
+ CDEBUG(D_HSM, "kucmsg fp %p len %d rc %d\n", cs->cs_fp, len, rc);
return rc;
}
@@ -1917,7 +1915,7 @@ static int mdc_changelog_send_thread(void *csdata)
struct kuc_hdr *kuch;
int rc;
- CDEBUG(D_CHANGELOG, "changelog to fp=%p start %llu\n",
+ CDEBUG(D_HSM, "changelog to fp=%p start %llu\n",
cs->cs_fp, cs->cs_startrec);
cs->cs_buf = kzalloc(KUC_CHANGELOG_MSG_MAXSIZE, GFP_NOFS);
@@ -1996,7 +1994,7 @@ static int mdc_ioc_changelog_send(struct obd_device *obd,
kfree(cs);
} else {
rc = 0;
- CDEBUG(D_CHANGELOG, "%s: started changelog thread\n",
+ CDEBUG(D_HSM, "%s: started changelog thread\n",
obd->obd_name);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 023/124] staging: lustre: mdc: Report D_CHANGELOG messages as D_HSM
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Removed the D_CHANGELOG pseudo-debug flag that wasn't actually defined
as a usable one. Report the D_CHANGELOG messages as D_HSM ones instead
since this is the primary user of these messages.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5538
Reviewed-on: http://review.whamcloud.com/11558
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mdc/mdc_request.c | 12 +++++-------
1 files changed, 5 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index cf94926..956b946 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1857,8 +1857,6 @@ static struct kuc_hdr *changelog_kuc_hdr(char *buf, int len, int flags)
return lh;
}
-#define D_CHANGELOG 0
-
struct changelog_show {
__u64 cs_startrec;
__u32 cs_flags;
@@ -1885,12 +1883,12 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
if (rec->cr.cr_index < cs->cs_startrec) {
/* Skip entries earlier than what we are interested in */
- CDEBUG(D_CHANGELOG, "rec=%llu start=%llu\n",
+ CDEBUG(D_HSM, "rec=%llu start=%llu\n",
rec->cr.cr_index, cs->cs_startrec);
return 0;
}
- CDEBUG(D_CHANGELOG, "%llu %02d%-5s %llu 0x%x t="DFID" p="DFID
+ CDEBUG(D_HSM, "%llu %02d%-5s %llu 0x%x t=" DFID " p=" DFID
" %.*s\n", rec->cr.cr_index, rec->cr.cr_type,
changelog_type2str(rec->cr.cr_type), rec->cr.cr_time,
rec->cr.cr_flags & CLF_FLAGMASK,
@@ -1904,7 +1902,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
memcpy(lh + 1, &rec->cr, len - sizeof(*lh));
rc = libcfs_kkuc_msg_put(cs->cs_fp, lh);
- CDEBUG(D_CHANGELOG, "kucmsg fp %p len %d rc %d\n", cs->cs_fp, len, rc);
+ CDEBUG(D_HSM, "kucmsg fp %p len %d rc %d\n", cs->cs_fp, len, rc);
return rc;
}
@@ -1917,7 +1915,7 @@ static int mdc_changelog_send_thread(void *csdata)
struct kuc_hdr *kuch;
int rc;
- CDEBUG(D_CHANGELOG, "changelog to fp=%p start %llu\n",
+ CDEBUG(D_HSM, "changelog to fp=%p start %llu\n",
cs->cs_fp, cs->cs_startrec);
cs->cs_buf = kzalloc(KUC_CHANGELOG_MSG_MAXSIZE, GFP_NOFS);
@@ -1996,7 +1994,7 @@ static int mdc_ioc_changelog_send(struct obd_device *obd,
kfree(cs);
} else {
rc = 0;
- CDEBUG(D_CHANGELOG, "%s: started changelog thread\n",
+ CDEBUG(D_HSM, "%s: started changelog thread\n",
obd->obd_name);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 024/124] staging: lustre: remove RCU2HANDLE macro
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove RCU2HANDLE macro from lustre_handles.h.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11498
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/lustre_handles.h | 2 --
.../lustre/lustre/obdclass/lustre_handles.c | 7 +++++--
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_handles.h b/drivers/staging/lustre/lustre/include/lustre_handles.h
index bc1dd46..e071bac 100644
--- a/drivers/staging/lustre/lustre/include/lustre_handles.h
+++ b/drivers/staging/lustre/lustre/include/lustre_handles.h
@@ -76,8 +76,6 @@ struct portals_handle {
unsigned int h_in:1;
};
-#define RCU2HANDLE(rcu) container_of(rcu, struct portals_handle, h_rcu)
-
/* handles.c */
/* Add a handle to the hash table */
diff --git a/drivers/staging/lustre/lustre/obdclass/lustre_handles.c b/drivers/staging/lustre/lustre/obdclass/lustre_handles.c
index 8faa318..c9445e5 100644
--- a/drivers/staging/lustre/lustre/obdclass/lustre_handles.c
+++ b/drivers/staging/lustre/lustre/obdclass/lustre_handles.c
@@ -164,8 +164,11 @@ EXPORT_SYMBOL(class_handle2object);
void class_handle_free_cb(struct rcu_head *rcu)
{
- struct portals_handle *h = RCU2HANDLE(rcu);
- void *ptr = (void *)(unsigned long)h->h_cookie;
+ struct portals_handle *h;
+ void *ptr;
+
+ h = container_of(rcu, struct portals_handle, h_rcu);
+ ptr = (void *)(unsigned long)h->h_cookie;
if (h->h_ops->hop_free)
h->h_ops->hop_free(ptr, h->h_size);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 024/124] staging: lustre: remove RCU2HANDLE macro
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove RCU2HANDLE macro from lustre_handles.h.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11498
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/lustre_handles.h | 2 --
.../lustre/lustre/obdclass/lustre_handles.c | 7 +++++--
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_handles.h b/drivers/staging/lustre/lustre/include/lustre_handles.h
index bc1dd46..e071bac 100644
--- a/drivers/staging/lustre/lustre/include/lustre_handles.h
+++ b/drivers/staging/lustre/lustre/include/lustre_handles.h
@@ -76,8 +76,6 @@ struct portals_handle {
unsigned int h_in:1;
};
-#define RCU2HANDLE(rcu) container_of(rcu, struct portals_handle, h_rcu)
-
/* handles.c */
/* Add a handle to the hash table */
diff --git a/drivers/staging/lustre/lustre/obdclass/lustre_handles.c b/drivers/staging/lustre/lustre/obdclass/lustre_handles.c
index 8faa318..c9445e5 100644
--- a/drivers/staging/lustre/lustre/obdclass/lustre_handles.c
+++ b/drivers/staging/lustre/lustre/obdclass/lustre_handles.c
@@ -164,8 +164,11 @@ EXPORT_SYMBOL(class_handle2object);
void class_handle_free_cb(struct rcu_head *rcu)
{
- struct portals_handle *h = RCU2HANDLE(rcu);
- void *ptr = (void *)(unsigned long)h->h_cookie;
+ struct portals_handle *h;
+ void *ptr;
+
+ h = container_of(rcu, struct portals_handle, h_rcu);
+ ptr = (void *)(unsigned long)h->h_cookie;
if (h->h_ops->hop_free)
h->h_ops->hop_free(ptr, h->h_size);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 025/124] staging: lustre: llite: Compare of unsigned value against 0 is always true
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Comparison of unsigned value against 0 is always true.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5200
Reviewed-on: http://review.whamcloud.com/11217
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/lproc_llite.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index af233ea..fc2ba56 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -846,7 +846,7 @@ static ssize_t unstable_stats_store(struct kobject *kobj,
if (!count)
return 0;
- if (count < 0 || count >= sizeof(kernbuf))
+ if (count >= sizeof(kernbuf))
return -EINVAL;
if (copy_from_user(kernbuf, buffer, count))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 025/124] staging: lustre: llite: Compare of unsigned value against 0 is always true
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Comparison of unsigned value against 0 is always true.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5200
Reviewed-on: http://review.whamcloud.com/11217
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/lproc_llite.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index af233ea..fc2ba56 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -846,7 +846,7 @@ static ssize_t unstable_stats_store(struct kobject *kobj,
if (!count)
return 0;
- if (count < 0 || count >= sizeof(kernbuf))
+ if (count >= sizeof(kernbuf))
return -EINVAL;
if (copy_from_user(kernbuf, buffer, count))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 026/124] staging: lustre: statahead: statahead thread wait for RPCs to finish
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
Statahead thread should wait for inflight stat RPCs to finish in
case statahead RPC callback may access data allocated in statahead
thread context.
ll_sa_entry_fini() should keep old entry if stat RPC is not
finished yet.
Simplify sai refcounting:
* newly allocated sai will hold one refcount, and it will put it
after starting statahead thread.
* statahead thread holds one refcount.
* agl thread holds one refcount.
* stat process calls do_statahead_enter() which will try to get
sai, and if it's valid, it will revalidate from statahead cache,
and put refcount after use.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9663
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 1 -
drivers/staging/lustre/lustre/llite/dcache.c | 2 +-
drivers/staging/lustre/lustre/llite/file.c | 31 +-
.../staging/lustre/lustre/llite/llite_internal.h | 49 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 8 +
drivers/staging/lustre/lustre/llite/statahead.c | 849 +++++++++-----------
6 files changed, 434 insertions(+), 506 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 838a428..89633f7 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -806,7 +806,6 @@ struct md_enqueue_info {
int (*mi_cb)(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc);
__u64 mi_cbdata;
- unsigned int mi_generation;
};
struct obd_ops {
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index f4b6f38..8c00cc6 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -279,7 +279,7 @@ static int ll_revalidate_dentry(struct dentry *dentry,
if (lookup_flags & (LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE))
return 1;
- if (d_need_statahead(dir, dentry) <= 0)
+ if (!dentry_need_statahead(dir, dentry))
return 1;
if (lookup_flags & LOOKUP_RCU)
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index e9791e3..273b563 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -351,13 +351,11 @@ int ll_file_release(struct inode *inode, struct file *file)
fd = LUSTRE_FPRIVATE(file);
LASSERT(fd);
- /* The last ref on @file, maybe not be the owner pid of statahead.
- * Different processes can open the same dir, "ll_opendir_key" means:
- * it is me that should stop the statahead thread.
+ /* The last ref on @file, maybe not be the owner pid of statahead,
+ * because parent and child process can share the same file handle.
*/
- if (S_ISDIR(inode->i_mode) && lli->lli_opendir_key == fd &&
- lli->lli_opendir_pid != 0)
- ll_stop_statahead(inode, lli->lli_opendir_key);
+ if (S_ISDIR(inode->i_mode) && lli->lli_opendir_key == fd)
+ ll_deauthorize_statahead(inode, fd);
if (is_root_inode(inode)) {
LUSTRE_FPRIVATE(file) = NULL;
@@ -530,7 +528,7 @@ int ll_file_open(struct inode *inode, struct file *file)
struct obd_client_handle **och_p = NULL;
__u64 *och_usecount = NULL;
struct ll_file_data *fd;
- int rc = 0, opendir_set = 0;
+ int rc = 0;
CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p), flags %o\n",
PFID(ll_inode2fid(inode)), inode, file->f_flags);
@@ -545,16 +543,8 @@ int ll_file_open(struct inode *inode, struct file *file)
}
fd->fd_file = file;
- if (S_ISDIR(inode->i_mode)) {
- spin_lock(&lli->lli_sa_lock);
- if (!lli->lli_opendir_key && !lli->lli_sai &&
- lli->lli_opendir_pid == 0) {
- lli->lli_opendir_key = fd;
- lli->lli_opendir_pid = current_pid();
- opendir_set = 1;
- }
- spin_unlock(&lli->lli_sa_lock);
- }
+ if (S_ISDIR(inode->i_mode))
+ ll_authorize_statahead(inode, fd);
if (is_root_inode(inode)) {
LUSTRE_FPRIVATE(file) = fd;
@@ -713,9 +703,10 @@ out_och_free:
mutex_unlock(&lli->lli_och_mutex);
out_openerr:
- if (opendir_set != 0)
- ll_stop_statahead(inode, lli->lli_opendir_key);
- ll_file_data_put(fd);
+ if (lli->lli_opendir_key == fd)
+ ll_deauthorize_statahead(inode, fd);
+ if (fd)
+ ll_file_data_put(fd);
} else {
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_OPEN, 1);
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index cbd5bc5..f903f2a 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -172,6 +172,13 @@ struct ll_inode_info {
* -- I am the owner of dir statahead.
*/
pid_t d_opendir_pid;
+ /* stat will try to access statahead entries or start
+ * statahead if this flag is set, and this flag will be
+ * set upon dir open, and cleared when dir is closed,
+ * statahead hit ratio is too low, or start statahead
+ * thread failed.
+ */
+ unsigned int d_sa_enabled:1;
/* directory stripe information */
struct lmv_stripe_md *d_lsm_md;
/* striped directory size */
@@ -184,6 +191,7 @@ struct ll_inode_info {
#define lli_opendir_key u.d.d_opendir_key
#define lli_sai u.d.d_sai
#define lli_sa_lock u.d.d_sa_lock
+#define lli_sa_enabled u.d.d_sa_enabled
#define lli_opendir_pid u.d.d_opendir_pid
#define lli_lsm_md u.d.d_lsm_md
#define lli_stripe_dir_size u.d.d_stripe_size
@@ -495,6 +503,9 @@ struct ll_sb_info {
atomic_t ll_sa_wrong; /* statahead thread stopped for
* low hit ratio
*/
+ atomic_t ll_sa_running; /* running statahead thread
+ * count
+ */
atomic_t ll_agl_total; /* AGL thread started count */
dev_t ll_sdev_orig; /* save s_dev before assign for
@@ -1040,7 +1051,8 @@ struct ll_statahead_info {
int do_statahead_enter(struct inode *dir, struct dentry **dentry,
int only_unplug);
-void ll_stop_statahead(struct inode *dir, void *key);
+void ll_authorize_statahead(struct inode *dir, void *key);
+void ll_deauthorize_statahead(struct inode *dir, void *key);
blkcnt_t dirty_cnt(struct inode *inode);
@@ -1086,25 +1098,31 @@ ll_statahead_mark(struct inode *dir, struct dentry *dentry)
ldd->lld_sa_generation = sai->sai_generation;
}
-static inline int
-d_need_statahead(struct inode *dir, struct dentry *dentryp)
+static inline bool
+dentry_need_statahead(struct inode *dir, struct dentry *dentry)
{
struct ll_inode_info *lli;
struct ll_dentry_data *ldd;
if (ll_i2sbi(dir)->ll_sa_max == 0)
- return -EAGAIN;
+ return false;
lli = ll_i2info(dir);
+
+ /*
+ * statahead is not allowed for this dir, there may be three causes:
+ * 1. dir is not opened.
+ * 2. statahead hit ratio is too low.
+ * 3. previous stat started statahead thread failed.
+ */
+ if (!lli->lli_sa_enabled)
+ return false;
+
/* not the same process, don't statahead */
if (lli->lli_opendir_pid != current_pid())
- return -EAGAIN;
-
- /* statahead has been stopped */
- if (!lli->lli_opendir_key)
- return -EAGAIN;
+ return false;
- ldd = ll_d2d(dentryp);
+ ldd = ll_d2d(dentry);
/*
* When stats a dentry, the system trigger more than once "revalidate"
* or "lookup", for "getattr", for "getxattr", and maybe for others.
@@ -1122,19 +1140,16 @@ d_need_statahead(struct inode *dir, struct dentry *dentryp)
*/
if (ldd && lli->lli_sai &&
ldd->lld_sa_generation == lli->lli_sai->sai_generation)
- return -EAGAIN;
+ return false;
- return 1;
+ return true;
}
static inline int
ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug)
{
- int ret;
-
- ret = d_need_statahead(dir, *dentryp);
- if (ret <= 0)
- return ret;
+ if (!dentry_need_statahead(dir, *dentryp))
+ return -EAGAIN;
return do_statahead_enter(dir, dentryp, only_unplug);
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 99aba6b..93fd69b 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -116,6 +116,7 @@ static struct ll_sb_info *ll_init_sbi(struct super_block *sb)
sbi->ll_sa_max = LL_SA_RPC_DEF;
atomic_set(&sbi->ll_sa_total, 0);
atomic_set(&sbi->ll_sa_wrong, 0);
+ atomic_set(&sbi->ll_sa_running, 0);
atomic_set(&sbi->ll_agl_total, 0);
sbi->ll_flags |= LL_SBI_AGL_ENABLED;
@@ -630,6 +631,12 @@ void ll_kill_super(struct super_block *sb)
if (sbi) {
sb->s_dev = sbi->ll_sdev_orig;
sbi->ll_umounting = 1;
+
+ /* wait running statahead threads to quit */
+ while (atomic_read(&sbi->ll_sa_running) > 0) {
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ schedule_timeout(msecs_to_jiffies(MSEC_PER_SEC >> 3));
+ }
}
}
@@ -795,6 +802,7 @@ void ll_lli_init(struct ll_inode_info *lli)
lli->lli_sai = NULL;
spin_lock_init(&lli->lli_sa_lock);
lli->lli_opendir_pid = 0;
+ lli->lli_sa_enabled = 0;
} else {
mutex_init(&lli->lli_size_mutex);
lli->lli_symlink_name = NULL;
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 016463b..6577a66 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -281,25 +281,6 @@ ll_sa_entry_get_byindex(struct ll_statahead_info *sai, __u64 index)
return NULL;
}
-static void ll_sa_entry_cleanup(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry)
-{
- struct md_enqueue_info *minfo = entry->se_minfo;
- struct ptlrpc_request *req = entry->se_req;
-
- if (minfo) {
- entry->se_minfo = NULL;
- ll_intent_release(&minfo->mi_it);
- iput(minfo->mi_dir);
- kfree(minfo);
- }
-
- if (req) {
- entry->se_req = NULL;
- ptlrpc_req_finished(req);
- }
-}
-
static void ll_sa_entry_put(struct ll_statahead_info *sai,
struct ll_sa_entry *entry)
{
@@ -312,7 +293,6 @@ static void ll_sa_entry_put(struct ll_statahead_info *sai,
LASSERT(list_empty(&entry->se_list));
LASSERT(list_empty(&entry->se_hash));
- ll_sa_entry_cleanup(sai, entry);
iput(entry->se_inode);
kfree(entry);
@@ -355,7 +335,10 @@ ll_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
list_for_each_entry_safe(pos, next, &sai->sai_entries, se_link) {
if (!is_omitted_entry(sai, pos->se_index))
break;
- do_sa_entry_fini(sai, pos);
+ /* keep those whose statahead RPC not finished */
+ if (pos->se_stat == SA_ENTRY_SUCC ||
+ pos->se_stat == SA_ENTRY_INVA)
+ do_sa_entry_fini(sai, pos);
}
}
@@ -363,12 +346,14 @@ ll_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
* Inside lli_sa_lock.
*/
static void
-do_sa_entry_to_stated(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry, enum se_stat stat)
+__sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
+ enum se_stat stat)
{
struct ll_sa_entry *se;
struct list_head *pos = &sai->sai_entries_stated;
+ LASSERT(entry->se_stat == SA_ENTRY_INIT);
+
if (!list_empty(&entry->se_list))
list_del_init(&entry->se_list);
@@ -388,23 +373,30 @@ do_sa_entry_to_stated(struct ll_statahead_info *sai,
* \retval 1 -- entry to be destroyed.
* \retval 0 -- entry is inserted into stated list.
*/
-static int
-ll_sa_entry_to_stated(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry, enum se_stat stat)
+static void
+sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
+ enum se_stat stat)
{
struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
- int ret = 1;
+ struct md_enqueue_info *minfo = entry->se_minfo;
+ struct ptlrpc_request *req = entry->se_req;
+
+ /* release resources used in RPC */
+ if (minfo) {
+ entry->se_minfo = NULL;
+ ll_intent_release(&minfo->mi_it);
+ iput(minfo->mi_dir);
+ kfree(minfo);
+ }
- ll_sa_entry_cleanup(sai, entry);
+ if (req) {
+ entry->se_req = NULL;
+ ptlrpc_req_finished(req);
+ }
spin_lock(&lli->lli_sa_lock);
- if (likely(entry->se_stat != SA_ENTRY_DEST)) {
- do_sa_entry_to_stated(sai, entry, stat);
- ret = 0;
- }
+ __sa_entry_post_stat(sai, entry, stat);
spin_unlock(&lli->lli_sa_lock);
-
- return ret;
}
/*
@@ -475,56 +467,46 @@ static struct ll_statahead_info *ll_sai_alloc(void)
return sai;
}
-static inline struct ll_statahead_info *
-ll_sai_get(struct ll_statahead_info *sai)
+static inline struct ll_statahead_info *ll_sai_get(struct inode *dir)
{
- atomic_inc(&sai->sai_refcount);
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct ll_statahead_info *sai = NULL;
+
+ spin_lock(&lli->lli_sa_lock);
+ sai = lli->lli_sai;
+ if (sai)
+ atomic_inc(&sai->sai_refcount);
+ spin_unlock(&lli->lli_sa_lock);
+
return sai;
}
static void ll_sai_put(struct ll_statahead_info *sai)
{
- struct inode *inode = sai->sai_inode;
- struct ll_inode_info *lli = ll_i2info(inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
if (atomic_dec_and_lock(&sai->sai_refcount, &lli->lli_sa_lock)) {
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
struct ll_sa_entry *entry, *next;
- if (unlikely(atomic_read(&sai->sai_refcount) > 0)) {
- /* It is race case, the interpret callback just hold
- * a reference count
- */
- spin_unlock(&lli->lli_sa_lock);
- return;
- }
-
- LASSERT(!lli->lli_opendir_key);
- LASSERT(thread_is_stopped(&sai->sai_thread));
- LASSERT(thread_is_stopped(&sai->sai_agl_thread));
-
lli->lli_sai = NULL;
- lli->lli_opendir_pid = 0;
spin_unlock(&lli->lli_sa_lock);
- if (sai->sai_sent > sai->sai_replied)
- CDEBUG(D_READA, "statahead for dir "DFID
- " does not finish: [sent:%llu] [replied:%llu]\n",
- PFID(&lli->lli_fid),
- sai->sai_sent, sai->sai_replied);
+ LASSERT(thread_is_stopped(&sai->sai_thread));
+ LASSERT(thread_is_stopped(&sai->sai_agl_thread));
+ LASSERT(sai->sai_sent == sai->sai_replied);
list_for_each_entry_safe(entry, next, &sai->sai_entries,
se_link)
do_sa_entry_fini(sai, entry);
- LASSERT(list_empty(&sai->sai_entries));
- LASSERT(list_empty(&sai->sai_entries_received));
- LASSERT(list_empty(&sai->sai_entries_stated));
-
LASSERT(atomic_read(&sai->sai_cache_count) == 0);
LASSERT(list_empty(&sai->sai_entries_agl));
+ LASSERT(atomic_read(&sai->sai_refcount) == 0);
- iput(inode);
+ iput(sai->sai_inode);
kfree(sai);
+ atomic_dec(&sbi->ll_sa_running);
}
}
@@ -588,29 +570,18 @@ static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
iput(inode);
}
-static void ll_post_statahead(struct ll_statahead_info *sai)
+/* prepare inode for received statahead entry, and add it into agl list */
+static void sa_post_one(struct ll_statahead_info *sai,
+ struct ll_sa_entry *entry)
{
struct inode *dir = sai->sai_inode;
struct inode *child;
- struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_sa_entry *entry;
struct md_enqueue_info *minfo;
struct lookup_intent *it;
struct ptlrpc_request *req;
struct mdt_body *body;
int rc = 0;
- spin_lock(&lli->lli_sa_lock);
- if (unlikely(list_empty(&sai->sai_entries_received))) {
- spin_unlock(&lli->lli_sa_lock);
- return;
- }
- entry = list_entry(sai->sai_entries_received.next,
- struct ll_sa_entry, se_list);
- atomic_inc(&entry->se_refcount);
- list_del_init(&entry->se_list);
- spin_unlock(&lli->lli_sa_lock);
-
LASSERT(entry->se_handle != 0);
minfo = entry->se_minfo;
@@ -670,18 +641,56 @@ static void ll_post_statahead(struct ll_statahead_info *sai)
ll_agl_add(sai, child, entry->se_index);
out:
- /* The "ll_sa_entry_to_stated()" will drop related ldlm ibits lock
+ /* The "sa_entry_post_stat()" will drop related ldlm ibits lock
* reference count by calling "ll_intent_drop_lock()" in spite of the
* above operations failed or not. Do not worry about calling
* "ll_intent_drop_lock()" more than once.
*/
- rc = ll_sa_entry_to_stated(sai, entry,
- rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
- if (rc == 0 && entry->se_index == sai->sai_index_wait)
+ sa_entry_post_stat(sai, entry, rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
+ if (entry->se_index == sai->sai_index_wait)
wake_up(&sai->sai_waitq);
ll_sa_entry_put(sai, entry);
}
+static void ll_post_statahead(struct ll_statahead_info *sai)
+{
+ struct ll_inode_info *lli;
+
+ lli = ll_i2info(sai->sai_inode);
+
+ while (!sa_received_empty(sai)) {
+ struct ll_sa_entry *entry;
+
+ spin_lock(&lli->lli_sa_lock);
+ if (unlikely(sa_received_empty(sai))) {
+ spin_unlock(&lli->lli_sa_lock);
+ break;
+ }
+ entry = list_entry(sai->sai_entries_received.next,
+ struct ll_sa_entry, se_list);
+ atomic_inc(&entry->se_refcount);
+ list_del_init(&entry->se_list);
+ spin_unlock(&lli->lli_sa_lock);
+
+ sa_post_one(sai, entry);
+ }
+
+ spin_lock(&lli->lli_agl_lock);
+ while (!agl_list_empty(sai)) {
+ struct ll_inode_info *clli;
+
+ clli = list_entry(sai->sai_entries_agl.next,
+ struct ll_inode_info, lli_agl_list);
+ list_del_init(&clli->lli_agl_list);
+ spin_unlock(&lli->lli_agl_lock);
+
+ ll_agl_trigger(&clli->lli_vfs_inode, sai);
+
+ spin_lock(&lli->lli_agl_lock);
+ }
+ spin_unlock(&lli->lli_agl_lock);
+}
+
static int ll_statahead_interpret(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc)
{
@@ -690,72 +699,43 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_statahead_info *sai = NULL;
struct ll_sa_entry *entry;
- __u64 handle = 0;
int wakeup;
if (it_disposition(it, DISP_LOOKUP_NEG))
rc = -ENOENT;
- if (rc == 0) {
- /* release ibits lock ASAP to avoid deadlock when statahead
- * thread enqueues lock on parent in readdir and another
- * process enqueues lock on child with parent lock held, eg.
- * unlink.
- */
- handle = it->it_lock_handle;
- ll_intent_drop_lock(it);
- }
+ sai = ll_sai_get(dir);
+ LASSERT(sai);
+ LASSERT(!thread_is_stopped(&sai->sai_thread));
spin_lock(&lli->lli_sa_lock);
- /* stale entry */
- if (unlikely(!lli->lli_sai ||
- lli->lli_sai->sai_generation != minfo->mi_generation)) {
- spin_unlock(&lli->lli_sa_lock);
- rc = -ESTALE;
- goto out;
+ entry = ll_sa_entry_get_byindex(sai, minfo->mi_cbdata);
+ LASSERT(entry);
+ if (rc) {
+ __sa_entry_post_stat(sai, entry, SA_ENTRY_INVA);
+ wakeup = (entry->se_index == sai->sai_index_wait);
} else {
- sai = ll_sai_get(lli->lli_sai);
- if (unlikely(!thread_is_running(&sai->sai_thread))) {
- sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
- rc = -EBADFD;
- goto out;
- }
-
- entry = ll_sa_entry_get_byindex(sai, minfo->mi_cbdata);
- if (!entry) {
- sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
- rc = -EIDRM;
- goto out;
- }
-
- if (rc != 0) {
- do_sa_entry_to_stated(sai, entry, SA_ENTRY_INVA);
- wakeup = (entry->se_index == sai->sai_index_wait);
- } else {
- entry->se_minfo = minfo;
- entry->se_req = ptlrpc_request_addref(req);
- /* Release the async ibits lock ASAP to avoid deadlock
- * when statahead thread tries to enqueue lock on parent
- * for readpage and other tries to enqueue lock on child
- * with parent's lock held, for example: unlink.
- */
- entry->se_handle = handle;
- wakeup = list_empty(&sai->sai_entries_received);
- list_add_tail(&entry->se_list,
- &sai->sai_entries_received);
- }
- sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
-
- ll_sa_entry_put(sai, entry);
- if (wakeup)
- wake_up(&sai->sai_thread.t_ctl_waitq);
+ entry->se_minfo = minfo;
+ entry->se_req = ptlrpc_request_addref(req);
+ /*
+ * Release the async ibits lock ASAP to avoid deadlock
+ * when statahead thread tries to enqueue lock on parent
+ * for readpage and other tries to enqueue lock on child
+ * with parent's lock held, for example: unlink.
+ */
+ entry->se_handle = it->it_lock_handle;
+ ll_intent_drop_lock(it);
+ wakeup = sa_received_empty(sai);
+ list_add_tail(&entry->se_list, &sai->sai_entries_received);
}
+ sai->sai_replied++;
+ spin_unlock(&lli->lli_sa_lock);
-out:
- if (rc != 0) {
+ ll_sa_entry_put(sai, entry);
+ if (wakeup)
+ wake_up(&sai->sai_thread.t_ctl_waitq);
+
+ if (rc) {
ll_intent_release(it);
iput(dir);
kfree(minfo);
@@ -782,7 +762,6 @@ static int sa_args_init(struct inode *dir, struct inode *child,
struct ldlm_enqueue_info **pei)
{
const struct qstr *qstr = &entry->se_qstr;
- struct ll_inode_info *lli = ll_i2info(dir);
struct md_enqueue_info *minfo;
struct ldlm_enqueue_info *einfo;
struct md_op_data *op_data;
@@ -808,7 +787,6 @@ static int sa_args_init(struct inode *dir, struct inode *child,
minfo->mi_it.it_op = IT_GETATTR;
minfo->mi_dir = igrab(dir);
minfo->mi_cb = ll_statahead_interpret;
- minfo->mi_generation = lli->lli_sai->sai_generation;
minfo->mi_cbdata = entry->se_index;
einfo->ei_type = LDLM_IBITS;
@@ -889,8 +867,8 @@ static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
return rc;
}
-static void ll_statahead_one(struct dentry *parent, const char *entry_name,
- int entry_name_len)
+static void ll_statahead_one(struct dentry *parent, const char *name,
+ const int name_len)
{
struct inode *dir = d_inode(parent);
struct ll_inode_info *lli = ll_i2info(dir);
@@ -898,10 +876,9 @@ static void ll_statahead_one(struct dentry *parent, const char *entry_name,
struct dentry *dentry = NULL;
struct ll_sa_entry *entry;
int rc;
- int rc1;
- entry = ll_sa_entry_alloc(parent, sai, sai->sai_index, entry_name,
- entry_name_len);
+ entry = ll_sa_entry_alloc(parent, sai, sai->sai_index, name,
+ name_len);
if (IS_ERR(entry))
return;
@@ -912,15 +889,15 @@ static void ll_statahead_one(struct dentry *parent, const char *entry_name,
rc = do_sa_revalidate(dir, entry, dentry);
if (rc == 1 && agl_should_run(sai, d_inode(dentry)))
ll_agl_add(sai, d_inode(dentry), entry->se_index);
+ }
+ if (dentry)
dput(dentry);
- }
if (rc) {
- rc1 = ll_sa_entry_to_stated(sai, entry,
- rc < 0 ? SA_ENTRY_INVA :
- SA_ENTRY_SUCC);
- if (rc1 == 0 && entry->se_index == sai->sai_index_wait)
+ sa_entry_post_stat(sai, entry,
+ rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
+ if (entry->se_index == sai->sai_index_wait)
wake_up(&sai->sai_waitq);
} else {
sai->sai_sent++;
@@ -938,10 +915,12 @@ static int ll_agl_thread(void *arg)
struct ll_inode_info *plli = ll_i2info(dir);
struct ll_inode_info *clli;
struct ll_sb_info *sbi = ll_i2sbi(dir);
- struct ll_statahead_info *sai = ll_sai_get(plli->lli_sai);
- struct ptlrpc_thread *thread = &sai->sai_agl_thread;
+ struct ll_statahead_info *sai;
+ struct ptlrpc_thread *thread;
struct l_wait_info lwi = { 0 };
+ sai = ll_sai_get(dir);
+ thread = &sai->sai_agl_thread;
thread->t_pid = current_pid();
CDEBUG(D_READA, "agl thread started: sai %p, parent %pd\n",
sai, parent);
@@ -1030,12 +1009,11 @@ static int ll_statahead_thread(void *arg)
{
struct dentry *parent = arg;
struct inode *dir = d_inode(parent);
- struct ll_inode_info *plli = ll_i2info(dir);
- struct ll_inode_info *clli;
+ struct ll_inode_info *lli = ll_i2info(dir);
struct ll_sb_info *sbi = ll_i2sbi(dir);
- struct ll_statahead_info *sai = ll_sai_get(plli->lli_sai);
- struct ptlrpc_thread *thread = &sai->sai_thread;
- struct ptlrpc_thread *agl_thread = &sai->sai_agl_thread;
+ struct ll_statahead_info *sai;
+ struct ptlrpc_thread *thread;
+ struct ptlrpc_thread *agl_thread;
struct page *page = NULL;
__u64 pos = 0;
int first = 0;
@@ -1044,6 +1022,9 @@ static int ll_statahead_thread(void *arg)
struct ll_dir_chain chain;
struct l_wait_info lwi = { 0 };
+ sai = ll_sai_get(dir);
+ thread = &sai->sai_thread;
+ agl_thread = &sai->sai_agl_thread;
thread->t_pid = current_pid();
CDEBUG(D_READA, "statahead thread starting: sai %p, parent %pd\n",
sai, parent);
@@ -1052,7 +1033,7 @@ static int ll_statahead_thread(void *arg)
LUSTRE_OPC_ANY, dir);
if (IS_ERR(op_data)) {
rc = PTR_ERR(op_data);
- goto out_put;
+ goto out;
}
op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages;
@@ -1061,33 +1042,35 @@ static int ll_statahead_thread(void *arg)
ll_start_agl(parent, sai);
atomic_inc(&sbi->ll_sa_total);
- spin_lock(&plli->lli_sa_lock);
+ spin_lock(&lli->lli_sa_lock);
if (thread_is_init(thread))
/* If someone else has changed the thread state
* (e.g. already changed to SVC_STOPPING), we can't just
* blindly overwrite that setting.
*/
thread_set_flags(thread, SVC_RUNNING);
- spin_unlock(&plli->lli_sa_lock);
+ spin_unlock(&lli->lli_sa_lock);
wake_up(&thread->t_ctl_waitq);
ll_dir_chain_init(&chain);
- page = ll_get_dir_page(dir, op_data, pos, &chain);
-
- while (1) {
+ while (pos != MDS_DIR_END_OFF && thread_is_running(thread)) {
struct lu_dirpage *dp;
struct lu_dirent *ent;
+ sai->sai_in_readpage = 1;
+ page = ll_get_dir_page(dir, op_data, pos, &chain);
+ sai->sai_in_readpage = 0;
if (IS_ERR(page)) {
rc = PTR_ERR(page);
CDEBUG(D_READA, "error reading dir "DFID" at %llu/%llu: opendir_pid = %u: rc = %d\n",
PFID(ll_inode2fid(dir)), pos, sai->sai_index,
- plli->lli_opendir_pid, rc);
- goto out;
+ lli->lli_opendir_pid, rc);
+ break;
}
dp = page_address(page);
- for (ent = lu_dirent_start(dp); ent;
+ for (ent = lu_dirent_start(dp);
+ ent && thread_is_running(thread) && !sa_low_hit(sai);
ent = lu_dirent_next(ent)) {
__u64 hash;
int namelen;
@@ -1134,120 +1117,63 @@ static int ll_statahead_thread(void *arg)
if (unlikely(++first == 1))
continue;
-keep_it:
- l_wait_event(thread->t_ctl_waitq,
- !sa_sent_full(sai) ||
- !list_empty(&sai->sai_entries_received) ||
- !list_empty(&sai->sai_entries_agl) ||
- !thread_is_running(thread),
- &lwi);
-
-interpret_it:
- while (!list_empty(&sai->sai_entries_received))
+ /* wait for spare statahead window */
+ do {
+ l_wait_event(thread->t_ctl_waitq,
+ !sa_sent_full(sai) ||
+ !list_empty(&sai->sai_entries_received) ||
+ !list_empty(&sai->sai_entries_agl) ||
+ !thread_is_running(thread),
+ &lwi);
ll_post_statahead(sai);
+ } while (sa_sent_full(sai) &&
+ thread_is_running(thread));
- if (unlikely(!thread_is_running(thread))) {
- ll_release_page(dir, page, false);
- rc = 0;
- goto out;
- }
-
- /* If no window for metadata statahead, but there are
- * some AGL entries to be triggered, then try to help
- * to process the AGL entries.
- */
- if (sa_sent_full(sai)) {
- spin_lock(&plli->lli_agl_lock);
- while (!list_empty(&sai->sai_entries_agl)) {
- clli = list_entry(sai->sai_entries_agl.next,
- struct ll_inode_info, lli_agl_list);
- list_del_init(&clli->lli_agl_list);
- spin_unlock(&plli->lli_agl_lock);
- ll_agl_trigger(&clli->lli_vfs_inode,
- sai);
-
- if (!list_empty(&sai->sai_entries_received))
- goto interpret_it;
-
- if (unlikely(!thread_is_running(thread))) {
- ll_release_page(dir, page, false);
- rc = 0;
- goto out;
- }
-
- if (!sa_sent_full(sai))
- goto do_it;
-
- spin_lock(&plli->lli_agl_lock);
- }
- spin_unlock(&plli->lli_agl_lock);
-
- goto keep_it;
- }
-do_it:
ll_statahead_one(parent, name, namelen);
}
pos = le64_to_cpu(dp->ldp_hash_end);
- if (pos == MDS_DIR_END_OFF) {
- /*
- * End of directory reached.
- */
- ll_release_page(dir, page, false);
- while (1) {
- l_wait_event(thread->t_ctl_waitq,
- !list_empty(&sai->sai_entries_received) ||
- sai->sai_sent == sai->sai_replied ||
- !thread_is_running(thread),
- &lwi);
+ ll_release_page(dir, page,
+ le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE);
- while (!list_empty(&sai->sai_entries_received))
- ll_post_statahead(sai);
-
- if (unlikely(!thread_is_running(thread))) {
- rc = 0;
- goto out;
- }
+ if (sa_low_hit(sai)) {
+ rc = -EFAULT;
+ atomic_inc(&sbi->ll_sa_wrong);
+ CDEBUG(D_READA, "Statahead for dir "DFID" hit ratio too low: hit/miss %llu/%llu, sent/replied %llu/%llu, stopping statahead thread: pid %d\n",
+ PFID(&lli->lli_fid), sai->sai_hit,
+ sai->sai_miss, sai->sai_sent,
+ sai->sai_replied, current_pid());
+ break;
+ }
+ }
+ ll_dir_chain_fini(&chain);
+ ll_finish_md_op_data(op_data);
- if (sai->sai_sent == sai->sai_replied &&
- list_empty(&sai->sai_entries_received))
- break;
- }
+ if (rc < 0) {
+ spin_lock(&lli->lli_sa_lock);
+ thread_set_flags(thread, SVC_STOPPING);
+ lli->lli_sa_enabled = 0;
+ spin_unlock(&lli->lli_sa_lock);
+ }
- spin_lock(&plli->lli_agl_lock);
- while (!list_empty(&sai->sai_entries_agl) &&
- thread_is_running(thread)) {
- clli = list_entry(sai->sai_entries_agl.next,
- struct ll_inode_info, lli_agl_list);
- list_del_init(&clli->lli_agl_list);
- spin_unlock(&plli->lli_agl_lock);
- ll_agl_trigger(&clli->lli_vfs_inode, sai);
- spin_lock(&plli->lli_agl_lock);
- }
- spin_unlock(&plli->lli_agl_lock);
+ /*
+ * statahead is finished, but statahead entries need to be cached, wait
+ * for file release to stop me.
+ */
+ while (thread_is_running(thread)) {
+ l_wait_event(thread->t_ctl_waitq,
+ !sa_received_empty(sai) ||
+ !agl_list_empty(sai) ||
+ !thread_is_running(thread),
+ &lwi);
- rc = 0;
- goto out;
- } else {
- /*
- * chain is exhausted.
- * Normal case: continue to the next page.
- */
- ll_release_page(dir, page,
- le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE);
- sai->sai_in_readpage = 1;
- page = ll_get_dir_page(dir, op_data, pos, &chain);
- sai->sai_in_readpage = 0;
- }
+ ll_post_statahead(sai);
}
out:
- ll_dir_chain_fini(&chain);
- ll_finish_md_op_data(op_data);
-out_put:
if (sai->sai_agl_valid) {
- spin_lock(&plli->lli_agl_lock);
+ spin_lock(&lli->lli_agl_lock);
thread_set_flags(agl_thread, SVC_STOPPING);
- spin_unlock(&plli->lli_agl_lock);
+ spin_unlock(&lli->lli_agl_lock);
wake_up(&agl_thread->t_ctl_waitq);
CDEBUG(D_READA, "stop agl thread: sai %p pid %u\n",
@@ -1257,21 +1183,27 @@ out_put:
&lwi);
} else {
/* Set agl_thread flags anyway. */
- thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
+ thread_set_flags(agl_thread, SVC_STOPPED);
}
- spin_lock(&plli->lli_sa_lock);
- if (!list_empty(&sai->sai_entries_received)) {
- thread_set_flags(thread, SVC_STOPPING);
- spin_unlock(&plli->lli_sa_lock);
-
- /* To release the resources held by received entries. */
- while (!list_empty(&sai->sai_entries_received))
- ll_post_statahead(sai);
- spin_lock(&plli->lli_sa_lock);
+ /*
+ * wait for inflight statahead RPCs to finish, and then we can free sai
+ * safely because statahead RPC will access sai data
+ */
+ while (sai->sai_sent != sai->sai_replied) {
+ /* in case we're not woken up, timeout wait */
+ lwi = LWI_TIMEOUT(HZ >> 3, NULL, NULL);
+ l_wait_event(thread->t_ctl_waitq,
+ sai->sai_sent == sai->sai_replied, &lwi);
}
+
+ /* release resources held by received entries. */
+ ll_post_statahead(sai);
+
+ spin_lock(&lli->lli_sa_lock);
thread_set_flags(thread, SVC_STOPPED);
- spin_unlock(&plli->lli_sa_lock);
+ spin_unlock(&lli->lli_sa_lock);
+
wake_up(&sai->sai_waitq);
wake_up(&thread->t_ctl_waitq);
ll_sai_put(sai);
@@ -1281,52 +1213,54 @@ out_put:
return rc;
}
-/**
- * called in ll_file_release().
- */
-void ll_stop_statahead(struct inode *dir, void *key)
+/* authorize opened dir handle @key to statahead later */
+void ll_authorize_statahead(struct inode *dir, void *key)
{
struct ll_inode_info *lli = ll_i2info(dir);
- if (unlikely(!key))
- return;
-
spin_lock(&lli->lli_sa_lock);
- if (lli->lli_opendir_key != key || lli->lli_opendir_pid == 0) {
- spin_unlock(&lli->lli_sa_lock);
- return;
+ if (!lli->lli_opendir_key && !lli->lli_sai) {
+ /*
+ * if lli_sai is not NULL, it means previous statahead is not
+ * finished yet, we'd better not start a new statahead for now.
+ */
+ LASSERT(!lli->lli_opendir_pid);
+ lli->lli_opendir_key = key;
+ lli->lli_opendir_pid = current_pid();
+ lli->lli_sa_enabled = 1;
}
+ spin_unlock(&lli->lli_sa_lock);
+}
- lli->lli_opendir_key = NULL;
+/*
+ * deauthorize opened dir handle @key to statahead, but statahead thread may
+ * still be running, notify it to quit.
+ */
+void ll_deauthorize_statahead(struct inode *dir, void *key)
+{
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct ll_statahead_info *sai;
- if (lli->lli_sai) {
- struct l_wait_info lwi = { 0 };
- struct ptlrpc_thread *thread = &lli->lli_sai->sai_thread;
+ LASSERT(lli->lli_opendir_key == key);
+ LASSERT(lli->lli_opendir_pid);
- if (!thread_is_stopped(thread)) {
- thread_set_flags(thread, SVC_STOPPING);
- spin_unlock(&lli->lli_sa_lock);
- wake_up(&thread->t_ctl_waitq);
-
- CDEBUG(D_READA, "stop statahead thread: sai %p pid %u\n",
- lli->lli_sai, (unsigned int)thread->t_pid);
- l_wait_event(thread->t_ctl_waitq,
- thread_is_stopped(thread),
- &lwi);
- } else {
- spin_unlock(&lli->lli_sa_lock);
- }
+ CDEBUG(D_READA, "deauthorize statahead for "DFID"\n",
+ PFID(&lli->lli_fid));
+ spin_lock(&lli->lli_sa_lock);
+ lli->lli_opendir_key = NULL;
+ lli->lli_opendir_pid = 0;
+ lli->lli_sa_enabled = 0;
+ sai = lli->lli_sai;
+ if (sai && thread_is_running(&sai->sai_thread)) {
/*
- * Put the ref which was held when first statahead_enter.
- * It maybe not the last ref for some statahead requests
- * maybe inflight.
+ * statahead thread may not quit yet because it needs to cache
+ * stated entries, now it's time to tell it to quit.
*/
- ll_sai_put(lli->lli_sai);
- } else {
- lli->lli_opendir_pid = 0;
- spin_unlock(&lli->lli_sa_lock);
+ thread_set_flags(&sai->sai_thread, SVC_STOPPING);
+ wake_up(&sai->sai_thread.t_ctl_waitq);
}
+ spin_unlock(&lli->lli_sa_lock);
}
enum {
@@ -1465,175 +1399,137 @@ out:
static void
ll_sai_unplug(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
{
- struct ptlrpc_thread *thread = &sai->sai_thread;
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
- int hit;
+ if (entry && entry->se_stat == SA_ENTRY_SUCC) {
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
- if (entry && entry->se_stat == SA_ENTRY_SUCC)
- hit = 1;
- else
- hit = 0;
-
- ll_sa_entry_fini(sai, entry);
- if (hit) {
sai->sai_hit++;
sai->sai_consecutive_miss = 0;
sai->sai_max = min(2 * sai->sai_max, sbi->ll_sa_max);
} else {
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
-
sai->sai_miss++;
sai->sai_consecutive_miss++;
- if (sa_low_hit(sai) && thread_is_running(thread)) {
- atomic_inc(&sbi->ll_sa_wrong);
- CDEBUG(D_READA, "Statahead for dir " DFID " hit ratio too low: hit/miss %llu/%llu, sent/replied %llu/%llu, stopping statahead thread\n",
- PFID(&lli->lli_fid), sai->sai_hit,
- sai->sai_miss, sai->sai_sent,
- sai->sai_replied);
- spin_lock(&lli->lli_sa_lock);
- if (!thread_is_stopped(thread))
- thread_set_flags(thread, SVC_STOPPING);
- spin_unlock(&lli->lli_sa_lock);
- }
}
-
- if (!thread_is_stopped(thread))
- wake_up(&thread->t_ctl_waitq);
+ ll_sa_entry_fini(sai, entry);
+ wake_up(&sai->sai_thread.t_ctl_waitq);
}
-/**
- * Start statahead thread if this is the first dir entry.
- * Otherwise if a thread is started already, wait it until it is ahead of me.
- * \retval 1 -- find entry with lock in cache, the caller needs to do
- * nothing.
- * \retval 0 -- find entry in cache, but without lock, the caller needs
- * refresh from MDS.
- * \retval others -- the caller need to process as non-statahead.
- */
-int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
- int only_unplug)
+static int revalidate_statahead_dentry(struct inode *dir,
+ struct ll_statahead_info *sai,
+ struct dentry **dentryp,
+ int only_unplug)
{
- struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_statahead_info *sai = lli->lli_sai;
- struct dentry *parent;
- struct ll_sa_entry *entry;
- struct ptlrpc_thread *thread;
- struct l_wait_info lwi = { 0 };
- struct task_struct *task;
- int rc = 0;
- struct ll_inode_info *plli;
-
- LASSERT(lli->lli_opendir_pid == current_pid());
-
- if (sai) {
- thread = &sai->sai_thread;
- if (unlikely(thread_is_stopped(thread) &&
- list_empty(&sai->sai_entries_stated))) {
- /* to release resource */
- ll_stop_statahead(dir, lli->lli_opendir_key);
- return -EAGAIN;
- }
+ struct ll_sa_entry *entry = NULL;
+ struct l_wait_info lwi = { 0 };
+ int rc = 0;
- if ((*dentryp)->d_name.name[0] == '.') {
- if (sai->sai_ls_all ||
- sai->sai_miss_hidden >= sai->sai_skip_hidden) {
+ if ((*dentryp)->d_name.name[0] == '.') {
+ if (sai->sai_ls_all ||
+ sai->sai_miss_hidden >= sai->sai_skip_hidden) {
+ /*
+ * Hidden dentry is the first one, or statahead
+ * thread does not skip so many hidden dentries
+ * before "sai_ls_all" enabled as below.
+ */
+ } else {
+ if (!sai->sai_ls_all)
/*
- * Hidden dentry is the first one, or statahead
- * thread does not skip so many hidden dentries
- * before "sai_ls_all" enabled as below.
+ * It maybe because hidden dentry is not
+ * the first one, "sai_ls_all" was not
+ * set, then "ls -al" missed. Enable
+ * "sai_ls_all" for such case.
*/
- } else {
- if (!sai->sai_ls_all)
- /*
- * It maybe because hidden dentry is not
- * the first one, "sai_ls_all" was not
- * set, then "ls -al" missed. Enable
- * "sai_ls_all" for such case.
- */
- sai->sai_ls_all = 1;
+ sai->sai_ls_all = 1;
- /*
- * Such "getattr" has been skipped before
- * "sai_ls_all" enabled as above.
- */
- sai->sai_miss_hidden++;
- return -EAGAIN;
- }
+ /*
+ * Such "getattr" has been skipped before
+ * "sai_ls_all" enabled as above.
+ */
+ sai->sai_miss_hidden++;
+ return -EAGAIN;
}
+ }
- entry = ll_sa_entry_get_byname(sai, &(*dentryp)->d_name);
- if (!entry || only_unplug) {
+ entry = ll_sa_entry_get_byname(sai, &(*dentryp)->d_name);
+ if (!entry || only_unplug) {
+ ll_sai_unplug(sai, entry);
+ return entry ? 1 : -EAGAIN;
+ }
+
+ /* if statahead is busy in readdir, help it do post-work */
+ if (!ll_sa_entry_stated(entry) && sai->sai_in_readpage)
+ ll_post_statahead(sai);
+
+ if (!ll_sa_entry_stated(entry)) {
+ sai->sai_index_wait = entry->se_index;
+ lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
+ LWI_ON_SIGNAL_NOOP, NULL);
+ rc = l_wait_event(sai->sai_waitq,
+ ll_sa_entry_stated(entry) ||
+ thread_is_stopped(&sai->sai_thread),
+ &lwi);
+ if (rc < 0) {
ll_sai_unplug(sai, entry);
- return entry ? 1 : -EAGAIN;
+ return -EAGAIN;
}
+ }
- /* if statahead is busy in readdir, help it do post-work */
- while (!ll_sa_entry_stated(entry) && sai->sai_in_readpage &&
- !sa_received_empty(sai))
- ll_post_statahead(sai);
-
- if (!ll_sa_entry_stated(entry)) {
- sai->sai_index_wait = entry->se_index;
- lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
- LWI_ON_SIGNAL_NOOP, NULL);
- rc = l_wait_event(sai->sai_waitq,
- ll_sa_entry_stated(entry) ||
- thread_is_stopped(thread),
- &lwi);
- if (rc < 0) {
- ll_sai_unplug(sai, entry);
- return -EAGAIN;
- }
- }
+ if (entry->se_stat == SA_ENTRY_SUCC && entry->se_inode) {
+ struct inode *inode = entry->se_inode;
+ struct lookup_intent it = { .it_op = IT_GETATTR,
+ .it_lock_handle = entry->se_handle };
+ __u64 bits;
+
+ rc = md_revalidate_lock(ll_i2mdexp(dir), &it,
+ ll_inode2fid(inode), &bits);
+ if (rc == 1) {
+ if (!(*dentryp)->d_inode) {
+ struct dentry *alias;
- if (entry->se_stat == SA_ENTRY_SUCC && entry->se_inode) {
- struct inode *inode = entry->se_inode;
- struct lookup_intent it = { .it_op = IT_GETATTR,
- .it_lock_handle =
- entry->se_handle };
- __u64 bits;
-
- rc = md_revalidate_lock(ll_i2mdexp(dir), &it,
- ll_inode2fid(inode), &bits);
- if (rc == 1) {
- if (!d_inode(*dentryp)) {
- struct dentry *alias;
-
- alias = ll_splice_alias(inode,
- *dentryp);
- if (IS_ERR(alias)) {
- ll_sai_unplug(sai, entry);
- return PTR_ERR(alias);
- }
- *dentryp = alias;
- } else if (d_inode(*dentryp) != inode) {
- /* revalidate, but inode is recreated */
- CDEBUG(D_READA, "%s: stale dentry %pd inode "DFID", statahead inode "DFID"\n",
- ll_get_fsname(d_inode(*dentryp)->i_sb, NULL, 0),
- *dentryp,
- PFID(ll_inode2fid(d_inode(*dentryp))),
- PFID(ll_inode2fid(inode)));
- ll_intent_release(&it);
+ alias = ll_splice_alias(inode, *dentryp);
+ if (IS_ERR(alias)) {
ll_sai_unplug(sai, entry);
- return -ESTALE;
- } else {
- iput(inode);
+ return PTR_ERR(alias);
}
- entry->se_inode = NULL;
-
- if ((bits & MDS_INODELOCK_LOOKUP) &&
- d_lustre_invalid(*dentryp))
- d_lustre_revalidate(*dentryp);
- ll_intent_release(&it);
+ *dentryp = alias;
+ } else if ((*dentryp)->d_inode != inode) {
+ /* revalidate, but inode is recreated */
+ CDEBUG(D_READA,
+ "%s: stale dentry %pd inode "DFID", statahead inode "DFID"\n",
+ ll_get_fsname((*dentryp)->d_inode->i_sb,
+ NULL, 0),
+ *dentryp,
+ PFID(ll_inode2fid((*dentryp)->d_inode)),
+ PFID(ll_inode2fid(inode)));
+ rc = -ESTALE;
+ goto out_unplug;
+ } else {
+ iput(inode);
}
- }
+ entry->se_inode = NULL;
- ll_sai_unplug(sai, entry);
- return rc;
+ if ((bits & MDS_INODELOCK_LOOKUP) &&
+ d_lustre_invalid(*dentryp))
+ d_lustre_revalidate(*dentryp);
+ ll_intent_release(&it);
+ }
}
+out_unplug:
+ ll_sai_unplug(sai, entry);
+ return rc;
+}
+
+static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
+{
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct ll_statahead_info *sai = NULL;
+ struct l_wait_info lwi = { 0 };
+ struct ptlrpc_thread *thread;
+ struct task_struct *task;
+ struct dentry *parent;
+ int rc;
/* I am the "lli_opendir_pid" owner, only me can set "lli_sai". */
- rc = is_first_dirent(dir, *dentryp);
+ rc = is_first_dirent(dir, dentry);
if (rc == LS_NONE_FIRST_DE) {
/* It is not "ls -{a}l" operation, no need statahead for it. */
rc = -EAGAIN;
@@ -1656,13 +1552,12 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
}
/* get parent reference count here, and put it in ll_statahead_thread */
- parent = dget((*dentryp)->d_parent);
+ parent = dget(dentry->d_parent);
if (unlikely(sai->sai_inode != d_inode(parent))) {
struct ll_inode_info *nlli = ll_i2info(d_inode(parent));
CWARN("Race condition, someone changed %pd just now: old parent "DFID", new parent "DFID"\n",
- *dentryp,
- PFID(&lli->lli_fid), PFID(&nlli->lli_fid));
+ dentry, PFID(&lli->lli_fid), PFID(&nlli->lli_fid));
dput(parent);
iput(sai->sai_inode);
rc = -EAGAIN;
@@ -1672,30 +1567,18 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
CDEBUG(D_READA, "start statahead thread: sai %p, parent %pd\n",
sai, parent);
- /* The sai buffer already has one reference taken at allocation time,
- * but as soon as we expose the sai by attaching it to the lli that
- * default reference can be dropped by another thread calling
- * ll_stop_statahead. We need to take a local reference to protect
- * the sai buffer while we intend to access it.
- */
- ll_sai_get(sai);
lli->lli_sai = sai;
- plli = ll_i2info(d_inode(parent));
task = kthread_run(ll_statahead_thread, parent, "ll_sa_%u",
- plli->lli_opendir_pid);
+ lli->lli_opendir_pid);
thread = &sai->sai_thread;
if (IS_ERR(task)) {
rc = PTR_ERR(task);
- CERROR("can't start ll_sa thread, rc: %d\n", rc);
+ CERROR("cannot start ll_sa thread: rc = %d\n", rc);
dput(parent);
lli->lli_opendir_key = NULL;
thread_set_flags(thread, SVC_STOPPED);
thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
- /* Drop both our own local reference and the default
- * reference from allocation time.
- */
- ll_sai_put(sai);
ll_sai_put(sai);
LASSERT(!lli->lli_sai);
return -EAGAIN;
@@ -1704,6 +1587,7 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
l_wait_event(thread->t_ctl_waitq,
thread_is_running(thread) || thread_is_stopped(thread),
&lwi);
+ atomic_inc(&ll_i2sbi(d_inode(parent))->ll_sa_running);
ll_sai_put(sai);
/*
@@ -1717,6 +1601,37 @@ out:
spin_lock(&lli->lli_sa_lock);
lli->lli_opendir_key = NULL;
lli->lli_opendir_pid = 0;
+ lli->lli_sa_enabled = 0;
spin_unlock(&lli->lli_sa_lock);
+
return rc;
}
+
+/**
+ * Start statahead thread if this is the first dir entry.
+ * Otherwise if a thread is started already, wait it until it is ahead of me.
+ * \retval 1 -- find entry with lock in cache, the caller needs to do
+ * nothing.
+ * \retval 0 -- find entry in cache, but without lock, the caller needs
+ * refresh from MDS.
+ * \retval others -- the caller need to process as non-statahead.
+ */
+int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
+ int only_unplug)
+{
+ struct ll_statahead_info *sai;
+
+ sai = ll_sai_get(dir);
+ if (sai) {
+ int rc;
+
+ rc = revalidate_statahead_dentry(dir, sai, dentryp,
+ only_unplug);
+ CDEBUG(D_READA, "revalidate statahead %pd: %d.\n",
+ *dentryp, rc);
+ ll_sai_put(sai);
+ return rc;
+ }
+
+ return start_statahead_thread(dir, *dentryp);
+}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 026/124] staging: lustre: statahead: statahead thread wait for RPCs to finish
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
Statahead thread should wait for inflight stat RPCs to finish in
case statahead RPC callback may access data allocated in statahead
thread context.
ll_sa_entry_fini() should keep old entry if stat RPC is not
finished yet.
Simplify sai refcounting:
* newly allocated sai will hold one refcount, and it will put it
after starting statahead thread.
* statahead thread holds one refcount.
* agl thread holds one refcount.
* stat process calls do_statahead_enter() which will try to get
sai, and if it's valid, it will revalidate from statahead cache,
and put refcount after use.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9663
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 1 -
drivers/staging/lustre/lustre/llite/dcache.c | 2 +-
drivers/staging/lustre/lustre/llite/file.c | 31 +-
.../staging/lustre/lustre/llite/llite_internal.h | 49 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 8 +
drivers/staging/lustre/lustre/llite/statahead.c | 849 +++++++++-----------
6 files changed, 434 insertions(+), 506 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 838a428..89633f7 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -806,7 +806,6 @@ struct md_enqueue_info {
int (*mi_cb)(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc);
__u64 mi_cbdata;
- unsigned int mi_generation;
};
struct obd_ops {
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index f4b6f38..8c00cc6 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -279,7 +279,7 @@ static int ll_revalidate_dentry(struct dentry *dentry,
if (lookup_flags & (LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE))
return 1;
- if (d_need_statahead(dir, dentry) <= 0)
+ if (!dentry_need_statahead(dir, dentry))
return 1;
if (lookup_flags & LOOKUP_RCU)
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index e9791e3..273b563 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -351,13 +351,11 @@ int ll_file_release(struct inode *inode, struct file *file)
fd = LUSTRE_FPRIVATE(file);
LASSERT(fd);
- /* The last ref on @file, maybe not be the owner pid of statahead.
- * Different processes can open the same dir, "ll_opendir_key" means:
- * it is me that should stop the statahead thread.
+ /* The last ref on @file, maybe not be the owner pid of statahead,
+ * because parent and child process can share the same file handle.
*/
- if (S_ISDIR(inode->i_mode) && lli->lli_opendir_key == fd &&
- lli->lli_opendir_pid != 0)
- ll_stop_statahead(inode, lli->lli_opendir_key);
+ if (S_ISDIR(inode->i_mode) && lli->lli_opendir_key == fd)
+ ll_deauthorize_statahead(inode, fd);
if (is_root_inode(inode)) {
LUSTRE_FPRIVATE(file) = NULL;
@@ -530,7 +528,7 @@ int ll_file_open(struct inode *inode, struct file *file)
struct obd_client_handle **och_p = NULL;
__u64 *och_usecount = NULL;
struct ll_file_data *fd;
- int rc = 0, opendir_set = 0;
+ int rc = 0;
CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p), flags %o\n",
PFID(ll_inode2fid(inode)), inode, file->f_flags);
@@ -545,16 +543,8 @@ int ll_file_open(struct inode *inode, struct file *file)
}
fd->fd_file = file;
- if (S_ISDIR(inode->i_mode)) {
- spin_lock(&lli->lli_sa_lock);
- if (!lli->lli_opendir_key && !lli->lli_sai &&
- lli->lli_opendir_pid == 0) {
- lli->lli_opendir_key = fd;
- lli->lli_opendir_pid = current_pid();
- opendir_set = 1;
- }
- spin_unlock(&lli->lli_sa_lock);
- }
+ if (S_ISDIR(inode->i_mode))
+ ll_authorize_statahead(inode, fd);
if (is_root_inode(inode)) {
LUSTRE_FPRIVATE(file) = fd;
@@ -713,9 +703,10 @@ out_och_free:
mutex_unlock(&lli->lli_och_mutex);
out_openerr:
- if (opendir_set != 0)
- ll_stop_statahead(inode, lli->lli_opendir_key);
- ll_file_data_put(fd);
+ if (lli->lli_opendir_key == fd)
+ ll_deauthorize_statahead(inode, fd);
+ if (fd)
+ ll_file_data_put(fd);
} else {
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_OPEN, 1);
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index cbd5bc5..f903f2a 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -172,6 +172,13 @@ struct ll_inode_info {
* -- I am the owner of dir statahead.
*/
pid_t d_opendir_pid;
+ /* stat will try to access statahead entries or start
+ * statahead if this flag is set, and this flag will be
+ * set upon dir open, and cleared when dir is closed,
+ * statahead hit ratio is too low, or start statahead
+ * thread failed.
+ */
+ unsigned int d_sa_enabled:1;
/* directory stripe information */
struct lmv_stripe_md *d_lsm_md;
/* striped directory size */
@@ -184,6 +191,7 @@ struct ll_inode_info {
#define lli_opendir_key u.d.d_opendir_key
#define lli_sai u.d.d_sai
#define lli_sa_lock u.d.d_sa_lock
+#define lli_sa_enabled u.d.d_sa_enabled
#define lli_opendir_pid u.d.d_opendir_pid
#define lli_lsm_md u.d.d_lsm_md
#define lli_stripe_dir_size u.d.d_stripe_size
@@ -495,6 +503,9 @@ struct ll_sb_info {
atomic_t ll_sa_wrong; /* statahead thread stopped for
* low hit ratio
*/
+ atomic_t ll_sa_running; /* running statahead thread
+ * count
+ */
atomic_t ll_agl_total; /* AGL thread started count */
dev_t ll_sdev_orig; /* save s_dev before assign for
@@ -1040,7 +1051,8 @@ struct ll_statahead_info {
int do_statahead_enter(struct inode *dir, struct dentry **dentry,
int only_unplug);
-void ll_stop_statahead(struct inode *dir, void *key);
+void ll_authorize_statahead(struct inode *dir, void *key);
+void ll_deauthorize_statahead(struct inode *dir, void *key);
blkcnt_t dirty_cnt(struct inode *inode);
@@ -1086,25 +1098,31 @@ ll_statahead_mark(struct inode *dir, struct dentry *dentry)
ldd->lld_sa_generation = sai->sai_generation;
}
-static inline int
-d_need_statahead(struct inode *dir, struct dentry *dentryp)
+static inline bool
+dentry_need_statahead(struct inode *dir, struct dentry *dentry)
{
struct ll_inode_info *lli;
struct ll_dentry_data *ldd;
if (ll_i2sbi(dir)->ll_sa_max == 0)
- return -EAGAIN;
+ return false;
lli = ll_i2info(dir);
+
+ /*
+ * statahead is not allowed for this dir, there may be three causes:
+ * 1. dir is not opened.
+ * 2. statahead hit ratio is too low.
+ * 3. previous stat started statahead thread failed.
+ */
+ if (!lli->lli_sa_enabled)
+ return false;
+
/* not the same process, don't statahead */
if (lli->lli_opendir_pid != current_pid())
- return -EAGAIN;
-
- /* statahead has been stopped */
- if (!lli->lli_opendir_key)
- return -EAGAIN;
+ return false;
- ldd = ll_d2d(dentryp);
+ ldd = ll_d2d(dentry);
/*
* When stats a dentry, the system trigger more than once "revalidate"
* or "lookup", for "getattr", for "getxattr", and maybe for others.
@@ -1122,19 +1140,16 @@ d_need_statahead(struct inode *dir, struct dentry *dentryp)
*/
if (ldd && lli->lli_sai &&
ldd->lld_sa_generation == lli->lli_sai->sai_generation)
- return -EAGAIN;
+ return false;
- return 1;
+ return true;
}
static inline int
ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug)
{
- int ret;
-
- ret = d_need_statahead(dir, *dentryp);
- if (ret <= 0)
- return ret;
+ if (!dentry_need_statahead(dir, *dentryp))
+ return -EAGAIN;
return do_statahead_enter(dir, dentryp, only_unplug);
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 99aba6b..93fd69b 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -116,6 +116,7 @@ static struct ll_sb_info *ll_init_sbi(struct super_block *sb)
sbi->ll_sa_max = LL_SA_RPC_DEF;
atomic_set(&sbi->ll_sa_total, 0);
atomic_set(&sbi->ll_sa_wrong, 0);
+ atomic_set(&sbi->ll_sa_running, 0);
atomic_set(&sbi->ll_agl_total, 0);
sbi->ll_flags |= LL_SBI_AGL_ENABLED;
@@ -630,6 +631,12 @@ void ll_kill_super(struct super_block *sb)
if (sbi) {
sb->s_dev = sbi->ll_sdev_orig;
sbi->ll_umounting = 1;
+
+ /* wait running statahead threads to quit */
+ while (atomic_read(&sbi->ll_sa_running) > 0) {
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ schedule_timeout(msecs_to_jiffies(MSEC_PER_SEC >> 3));
+ }
}
}
@@ -795,6 +802,7 @@ void ll_lli_init(struct ll_inode_info *lli)
lli->lli_sai = NULL;
spin_lock_init(&lli->lli_sa_lock);
lli->lli_opendir_pid = 0;
+ lli->lli_sa_enabled = 0;
} else {
mutex_init(&lli->lli_size_mutex);
lli->lli_symlink_name = NULL;
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 016463b..6577a66 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -281,25 +281,6 @@ ll_sa_entry_get_byindex(struct ll_statahead_info *sai, __u64 index)
return NULL;
}
-static void ll_sa_entry_cleanup(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry)
-{
- struct md_enqueue_info *minfo = entry->se_minfo;
- struct ptlrpc_request *req = entry->se_req;
-
- if (minfo) {
- entry->se_minfo = NULL;
- ll_intent_release(&minfo->mi_it);
- iput(minfo->mi_dir);
- kfree(minfo);
- }
-
- if (req) {
- entry->se_req = NULL;
- ptlrpc_req_finished(req);
- }
-}
-
static void ll_sa_entry_put(struct ll_statahead_info *sai,
struct ll_sa_entry *entry)
{
@@ -312,7 +293,6 @@ static void ll_sa_entry_put(struct ll_statahead_info *sai,
LASSERT(list_empty(&entry->se_list));
LASSERT(list_empty(&entry->se_hash));
- ll_sa_entry_cleanup(sai, entry);
iput(entry->se_inode);
kfree(entry);
@@ -355,7 +335,10 @@ ll_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
list_for_each_entry_safe(pos, next, &sai->sai_entries, se_link) {
if (!is_omitted_entry(sai, pos->se_index))
break;
- do_sa_entry_fini(sai, pos);
+ /* keep those whose statahead RPC not finished */
+ if (pos->se_stat == SA_ENTRY_SUCC ||
+ pos->se_stat == SA_ENTRY_INVA)
+ do_sa_entry_fini(sai, pos);
}
}
@@ -363,12 +346,14 @@ ll_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
* Inside lli_sa_lock.
*/
static void
-do_sa_entry_to_stated(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry, enum se_stat stat)
+__sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
+ enum se_stat stat)
{
struct ll_sa_entry *se;
struct list_head *pos = &sai->sai_entries_stated;
+ LASSERT(entry->se_stat == SA_ENTRY_INIT);
+
if (!list_empty(&entry->se_list))
list_del_init(&entry->se_list);
@@ -388,23 +373,30 @@ do_sa_entry_to_stated(struct ll_statahead_info *sai,
* \retval 1 -- entry to be destroyed.
* \retval 0 -- entry is inserted into stated list.
*/
-static int
-ll_sa_entry_to_stated(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry, enum se_stat stat)
+static void
+sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
+ enum se_stat stat)
{
struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
- int ret = 1;
+ struct md_enqueue_info *minfo = entry->se_minfo;
+ struct ptlrpc_request *req = entry->se_req;
+
+ /* release resources used in RPC */
+ if (minfo) {
+ entry->se_minfo = NULL;
+ ll_intent_release(&minfo->mi_it);
+ iput(minfo->mi_dir);
+ kfree(minfo);
+ }
- ll_sa_entry_cleanup(sai, entry);
+ if (req) {
+ entry->se_req = NULL;
+ ptlrpc_req_finished(req);
+ }
spin_lock(&lli->lli_sa_lock);
- if (likely(entry->se_stat != SA_ENTRY_DEST)) {
- do_sa_entry_to_stated(sai, entry, stat);
- ret = 0;
- }
+ __sa_entry_post_stat(sai, entry, stat);
spin_unlock(&lli->lli_sa_lock);
-
- return ret;
}
/*
@@ -475,56 +467,46 @@ static struct ll_statahead_info *ll_sai_alloc(void)
return sai;
}
-static inline struct ll_statahead_info *
-ll_sai_get(struct ll_statahead_info *sai)
+static inline struct ll_statahead_info *ll_sai_get(struct inode *dir)
{
- atomic_inc(&sai->sai_refcount);
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct ll_statahead_info *sai = NULL;
+
+ spin_lock(&lli->lli_sa_lock);
+ sai = lli->lli_sai;
+ if (sai)
+ atomic_inc(&sai->sai_refcount);
+ spin_unlock(&lli->lli_sa_lock);
+
return sai;
}
static void ll_sai_put(struct ll_statahead_info *sai)
{
- struct inode *inode = sai->sai_inode;
- struct ll_inode_info *lli = ll_i2info(inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
if (atomic_dec_and_lock(&sai->sai_refcount, &lli->lli_sa_lock)) {
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
struct ll_sa_entry *entry, *next;
- if (unlikely(atomic_read(&sai->sai_refcount) > 0)) {
- /* It is race case, the interpret callback just hold
- * a reference count
- */
- spin_unlock(&lli->lli_sa_lock);
- return;
- }
-
- LASSERT(!lli->lli_opendir_key);
- LASSERT(thread_is_stopped(&sai->sai_thread));
- LASSERT(thread_is_stopped(&sai->sai_agl_thread));
-
lli->lli_sai = NULL;
- lli->lli_opendir_pid = 0;
spin_unlock(&lli->lli_sa_lock);
- if (sai->sai_sent > sai->sai_replied)
- CDEBUG(D_READA, "statahead for dir "DFID
- " does not finish: [sent:%llu] [replied:%llu]\n",
- PFID(&lli->lli_fid),
- sai->sai_sent, sai->sai_replied);
+ LASSERT(thread_is_stopped(&sai->sai_thread));
+ LASSERT(thread_is_stopped(&sai->sai_agl_thread));
+ LASSERT(sai->sai_sent == sai->sai_replied);
list_for_each_entry_safe(entry, next, &sai->sai_entries,
se_link)
do_sa_entry_fini(sai, entry);
- LASSERT(list_empty(&sai->sai_entries));
- LASSERT(list_empty(&sai->sai_entries_received));
- LASSERT(list_empty(&sai->sai_entries_stated));
-
LASSERT(atomic_read(&sai->sai_cache_count) == 0);
LASSERT(list_empty(&sai->sai_entries_agl));
+ LASSERT(atomic_read(&sai->sai_refcount) == 0);
- iput(inode);
+ iput(sai->sai_inode);
kfree(sai);
+ atomic_dec(&sbi->ll_sa_running);
}
}
@@ -588,29 +570,18 @@ static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
iput(inode);
}
-static void ll_post_statahead(struct ll_statahead_info *sai)
+/* prepare inode for received statahead entry, and add it into agl list */
+static void sa_post_one(struct ll_statahead_info *sai,
+ struct ll_sa_entry *entry)
{
struct inode *dir = sai->sai_inode;
struct inode *child;
- struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_sa_entry *entry;
struct md_enqueue_info *minfo;
struct lookup_intent *it;
struct ptlrpc_request *req;
struct mdt_body *body;
int rc = 0;
- spin_lock(&lli->lli_sa_lock);
- if (unlikely(list_empty(&sai->sai_entries_received))) {
- spin_unlock(&lli->lli_sa_lock);
- return;
- }
- entry = list_entry(sai->sai_entries_received.next,
- struct ll_sa_entry, se_list);
- atomic_inc(&entry->se_refcount);
- list_del_init(&entry->se_list);
- spin_unlock(&lli->lli_sa_lock);
-
LASSERT(entry->se_handle != 0);
minfo = entry->se_minfo;
@@ -670,18 +641,56 @@ static void ll_post_statahead(struct ll_statahead_info *sai)
ll_agl_add(sai, child, entry->se_index);
out:
- /* The "ll_sa_entry_to_stated()" will drop related ldlm ibits lock
+ /* The "sa_entry_post_stat()" will drop related ldlm ibits lock
* reference count by calling "ll_intent_drop_lock()" in spite of the
* above operations failed or not. Do not worry about calling
* "ll_intent_drop_lock()" more than once.
*/
- rc = ll_sa_entry_to_stated(sai, entry,
- rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
- if (rc == 0 && entry->se_index == sai->sai_index_wait)
+ sa_entry_post_stat(sai, entry, rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
+ if (entry->se_index == sai->sai_index_wait)
wake_up(&sai->sai_waitq);
ll_sa_entry_put(sai, entry);
}
+static void ll_post_statahead(struct ll_statahead_info *sai)
+{
+ struct ll_inode_info *lli;
+
+ lli = ll_i2info(sai->sai_inode);
+
+ while (!sa_received_empty(sai)) {
+ struct ll_sa_entry *entry;
+
+ spin_lock(&lli->lli_sa_lock);
+ if (unlikely(sa_received_empty(sai))) {
+ spin_unlock(&lli->lli_sa_lock);
+ break;
+ }
+ entry = list_entry(sai->sai_entries_received.next,
+ struct ll_sa_entry, se_list);
+ atomic_inc(&entry->se_refcount);
+ list_del_init(&entry->se_list);
+ spin_unlock(&lli->lli_sa_lock);
+
+ sa_post_one(sai, entry);
+ }
+
+ spin_lock(&lli->lli_agl_lock);
+ while (!agl_list_empty(sai)) {
+ struct ll_inode_info *clli;
+
+ clli = list_entry(sai->sai_entries_agl.next,
+ struct ll_inode_info, lli_agl_list);
+ list_del_init(&clli->lli_agl_list);
+ spin_unlock(&lli->lli_agl_lock);
+
+ ll_agl_trigger(&clli->lli_vfs_inode, sai);
+
+ spin_lock(&lli->lli_agl_lock);
+ }
+ spin_unlock(&lli->lli_agl_lock);
+}
+
static int ll_statahead_interpret(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc)
{
@@ -690,72 +699,43 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_statahead_info *sai = NULL;
struct ll_sa_entry *entry;
- __u64 handle = 0;
int wakeup;
if (it_disposition(it, DISP_LOOKUP_NEG))
rc = -ENOENT;
- if (rc == 0) {
- /* release ibits lock ASAP to avoid deadlock when statahead
- * thread enqueues lock on parent in readdir and another
- * process enqueues lock on child with parent lock held, eg.
- * unlink.
- */
- handle = it->it_lock_handle;
- ll_intent_drop_lock(it);
- }
+ sai = ll_sai_get(dir);
+ LASSERT(sai);
+ LASSERT(!thread_is_stopped(&sai->sai_thread));
spin_lock(&lli->lli_sa_lock);
- /* stale entry */
- if (unlikely(!lli->lli_sai ||
- lli->lli_sai->sai_generation != minfo->mi_generation)) {
- spin_unlock(&lli->lli_sa_lock);
- rc = -ESTALE;
- goto out;
+ entry = ll_sa_entry_get_byindex(sai, minfo->mi_cbdata);
+ LASSERT(entry);
+ if (rc) {
+ __sa_entry_post_stat(sai, entry, SA_ENTRY_INVA);
+ wakeup = (entry->se_index == sai->sai_index_wait);
} else {
- sai = ll_sai_get(lli->lli_sai);
- if (unlikely(!thread_is_running(&sai->sai_thread))) {
- sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
- rc = -EBADFD;
- goto out;
- }
-
- entry = ll_sa_entry_get_byindex(sai, minfo->mi_cbdata);
- if (!entry) {
- sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
- rc = -EIDRM;
- goto out;
- }
-
- if (rc != 0) {
- do_sa_entry_to_stated(sai, entry, SA_ENTRY_INVA);
- wakeup = (entry->se_index == sai->sai_index_wait);
- } else {
- entry->se_minfo = minfo;
- entry->se_req = ptlrpc_request_addref(req);
- /* Release the async ibits lock ASAP to avoid deadlock
- * when statahead thread tries to enqueue lock on parent
- * for readpage and other tries to enqueue lock on child
- * with parent's lock held, for example: unlink.
- */
- entry->se_handle = handle;
- wakeup = list_empty(&sai->sai_entries_received);
- list_add_tail(&entry->se_list,
- &sai->sai_entries_received);
- }
- sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
-
- ll_sa_entry_put(sai, entry);
- if (wakeup)
- wake_up(&sai->sai_thread.t_ctl_waitq);
+ entry->se_minfo = minfo;
+ entry->se_req = ptlrpc_request_addref(req);
+ /*
+ * Release the async ibits lock ASAP to avoid deadlock
+ * when statahead thread tries to enqueue lock on parent
+ * for readpage and other tries to enqueue lock on child
+ * with parent's lock held, for example: unlink.
+ */
+ entry->se_handle = it->it_lock_handle;
+ ll_intent_drop_lock(it);
+ wakeup = sa_received_empty(sai);
+ list_add_tail(&entry->se_list, &sai->sai_entries_received);
}
+ sai->sai_replied++;
+ spin_unlock(&lli->lli_sa_lock);
-out:
- if (rc != 0) {
+ ll_sa_entry_put(sai, entry);
+ if (wakeup)
+ wake_up(&sai->sai_thread.t_ctl_waitq);
+
+ if (rc) {
ll_intent_release(it);
iput(dir);
kfree(minfo);
@@ -782,7 +762,6 @@ static int sa_args_init(struct inode *dir, struct inode *child,
struct ldlm_enqueue_info **pei)
{
const struct qstr *qstr = &entry->se_qstr;
- struct ll_inode_info *lli = ll_i2info(dir);
struct md_enqueue_info *minfo;
struct ldlm_enqueue_info *einfo;
struct md_op_data *op_data;
@@ -808,7 +787,6 @@ static int sa_args_init(struct inode *dir, struct inode *child,
minfo->mi_it.it_op = IT_GETATTR;
minfo->mi_dir = igrab(dir);
minfo->mi_cb = ll_statahead_interpret;
- minfo->mi_generation = lli->lli_sai->sai_generation;
minfo->mi_cbdata = entry->se_index;
einfo->ei_type = LDLM_IBITS;
@@ -889,8 +867,8 @@ static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
return rc;
}
-static void ll_statahead_one(struct dentry *parent, const char *entry_name,
- int entry_name_len)
+static void ll_statahead_one(struct dentry *parent, const char *name,
+ const int name_len)
{
struct inode *dir = d_inode(parent);
struct ll_inode_info *lli = ll_i2info(dir);
@@ -898,10 +876,9 @@ static void ll_statahead_one(struct dentry *parent, const char *entry_name,
struct dentry *dentry = NULL;
struct ll_sa_entry *entry;
int rc;
- int rc1;
- entry = ll_sa_entry_alloc(parent, sai, sai->sai_index, entry_name,
- entry_name_len);
+ entry = ll_sa_entry_alloc(parent, sai, sai->sai_index, name,
+ name_len);
if (IS_ERR(entry))
return;
@@ -912,15 +889,15 @@ static void ll_statahead_one(struct dentry *parent, const char *entry_name,
rc = do_sa_revalidate(dir, entry, dentry);
if (rc == 1 && agl_should_run(sai, d_inode(dentry)))
ll_agl_add(sai, d_inode(dentry), entry->se_index);
+ }
+ if (dentry)
dput(dentry);
- }
if (rc) {
- rc1 = ll_sa_entry_to_stated(sai, entry,
- rc < 0 ? SA_ENTRY_INVA :
- SA_ENTRY_SUCC);
- if (rc1 == 0 && entry->se_index == sai->sai_index_wait)
+ sa_entry_post_stat(sai, entry,
+ rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
+ if (entry->se_index == sai->sai_index_wait)
wake_up(&sai->sai_waitq);
} else {
sai->sai_sent++;
@@ -938,10 +915,12 @@ static int ll_agl_thread(void *arg)
struct ll_inode_info *plli = ll_i2info(dir);
struct ll_inode_info *clli;
struct ll_sb_info *sbi = ll_i2sbi(dir);
- struct ll_statahead_info *sai = ll_sai_get(plli->lli_sai);
- struct ptlrpc_thread *thread = &sai->sai_agl_thread;
+ struct ll_statahead_info *sai;
+ struct ptlrpc_thread *thread;
struct l_wait_info lwi = { 0 };
+ sai = ll_sai_get(dir);
+ thread = &sai->sai_agl_thread;
thread->t_pid = current_pid();
CDEBUG(D_READA, "agl thread started: sai %p, parent %pd\n",
sai, parent);
@@ -1030,12 +1009,11 @@ static int ll_statahead_thread(void *arg)
{
struct dentry *parent = arg;
struct inode *dir = d_inode(parent);
- struct ll_inode_info *plli = ll_i2info(dir);
- struct ll_inode_info *clli;
+ struct ll_inode_info *lli = ll_i2info(dir);
struct ll_sb_info *sbi = ll_i2sbi(dir);
- struct ll_statahead_info *sai = ll_sai_get(plli->lli_sai);
- struct ptlrpc_thread *thread = &sai->sai_thread;
- struct ptlrpc_thread *agl_thread = &sai->sai_agl_thread;
+ struct ll_statahead_info *sai;
+ struct ptlrpc_thread *thread;
+ struct ptlrpc_thread *agl_thread;
struct page *page = NULL;
__u64 pos = 0;
int first = 0;
@@ -1044,6 +1022,9 @@ static int ll_statahead_thread(void *arg)
struct ll_dir_chain chain;
struct l_wait_info lwi = { 0 };
+ sai = ll_sai_get(dir);
+ thread = &sai->sai_thread;
+ agl_thread = &sai->sai_agl_thread;
thread->t_pid = current_pid();
CDEBUG(D_READA, "statahead thread starting: sai %p, parent %pd\n",
sai, parent);
@@ -1052,7 +1033,7 @@ static int ll_statahead_thread(void *arg)
LUSTRE_OPC_ANY, dir);
if (IS_ERR(op_data)) {
rc = PTR_ERR(op_data);
- goto out_put;
+ goto out;
}
op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages;
@@ -1061,33 +1042,35 @@ static int ll_statahead_thread(void *arg)
ll_start_agl(parent, sai);
atomic_inc(&sbi->ll_sa_total);
- spin_lock(&plli->lli_sa_lock);
+ spin_lock(&lli->lli_sa_lock);
if (thread_is_init(thread))
/* If someone else has changed the thread state
* (e.g. already changed to SVC_STOPPING), we can't just
* blindly overwrite that setting.
*/
thread_set_flags(thread, SVC_RUNNING);
- spin_unlock(&plli->lli_sa_lock);
+ spin_unlock(&lli->lli_sa_lock);
wake_up(&thread->t_ctl_waitq);
ll_dir_chain_init(&chain);
- page = ll_get_dir_page(dir, op_data, pos, &chain);
-
- while (1) {
+ while (pos != MDS_DIR_END_OFF && thread_is_running(thread)) {
struct lu_dirpage *dp;
struct lu_dirent *ent;
+ sai->sai_in_readpage = 1;
+ page = ll_get_dir_page(dir, op_data, pos, &chain);
+ sai->sai_in_readpage = 0;
if (IS_ERR(page)) {
rc = PTR_ERR(page);
CDEBUG(D_READA, "error reading dir "DFID" at %llu/%llu: opendir_pid = %u: rc = %d\n",
PFID(ll_inode2fid(dir)), pos, sai->sai_index,
- plli->lli_opendir_pid, rc);
- goto out;
+ lli->lli_opendir_pid, rc);
+ break;
}
dp = page_address(page);
- for (ent = lu_dirent_start(dp); ent;
+ for (ent = lu_dirent_start(dp);
+ ent && thread_is_running(thread) && !sa_low_hit(sai);
ent = lu_dirent_next(ent)) {
__u64 hash;
int namelen;
@@ -1134,120 +1117,63 @@ static int ll_statahead_thread(void *arg)
if (unlikely(++first == 1))
continue;
-keep_it:
- l_wait_event(thread->t_ctl_waitq,
- !sa_sent_full(sai) ||
- !list_empty(&sai->sai_entries_received) ||
- !list_empty(&sai->sai_entries_agl) ||
- !thread_is_running(thread),
- &lwi);
-
-interpret_it:
- while (!list_empty(&sai->sai_entries_received))
+ /* wait for spare statahead window */
+ do {
+ l_wait_event(thread->t_ctl_waitq,
+ !sa_sent_full(sai) ||
+ !list_empty(&sai->sai_entries_received) ||
+ !list_empty(&sai->sai_entries_agl) ||
+ !thread_is_running(thread),
+ &lwi);
ll_post_statahead(sai);
+ } while (sa_sent_full(sai) &&
+ thread_is_running(thread));
- if (unlikely(!thread_is_running(thread))) {
- ll_release_page(dir, page, false);
- rc = 0;
- goto out;
- }
-
- /* If no window for metadata statahead, but there are
- * some AGL entries to be triggered, then try to help
- * to process the AGL entries.
- */
- if (sa_sent_full(sai)) {
- spin_lock(&plli->lli_agl_lock);
- while (!list_empty(&sai->sai_entries_agl)) {
- clli = list_entry(sai->sai_entries_agl.next,
- struct ll_inode_info, lli_agl_list);
- list_del_init(&clli->lli_agl_list);
- spin_unlock(&plli->lli_agl_lock);
- ll_agl_trigger(&clli->lli_vfs_inode,
- sai);
-
- if (!list_empty(&sai->sai_entries_received))
- goto interpret_it;
-
- if (unlikely(!thread_is_running(thread))) {
- ll_release_page(dir, page, false);
- rc = 0;
- goto out;
- }
-
- if (!sa_sent_full(sai))
- goto do_it;
-
- spin_lock(&plli->lli_agl_lock);
- }
- spin_unlock(&plli->lli_agl_lock);
-
- goto keep_it;
- }
-do_it:
ll_statahead_one(parent, name, namelen);
}
pos = le64_to_cpu(dp->ldp_hash_end);
- if (pos == MDS_DIR_END_OFF) {
- /*
- * End of directory reached.
- */
- ll_release_page(dir, page, false);
- while (1) {
- l_wait_event(thread->t_ctl_waitq,
- !list_empty(&sai->sai_entries_received) ||
- sai->sai_sent == sai->sai_replied ||
- !thread_is_running(thread),
- &lwi);
+ ll_release_page(dir, page,
+ le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE);
- while (!list_empty(&sai->sai_entries_received))
- ll_post_statahead(sai);
-
- if (unlikely(!thread_is_running(thread))) {
- rc = 0;
- goto out;
- }
+ if (sa_low_hit(sai)) {
+ rc = -EFAULT;
+ atomic_inc(&sbi->ll_sa_wrong);
+ CDEBUG(D_READA, "Statahead for dir "DFID" hit ratio too low: hit/miss %llu/%llu, sent/replied %llu/%llu, stopping statahead thread: pid %d\n",
+ PFID(&lli->lli_fid), sai->sai_hit,
+ sai->sai_miss, sai->sai_sent,
+ sai->sai_replied, current_pid());
+ break;
+ }
+ }
+ ll_dir_chain_fini(&chain);
+ ll_finish_md_op_data(op_data);
- if (sai->sai_sent == sai->sai_replied &&
- list_empty(&sai->sai_entries_received))
- break;
- }
+ if (rc < 0) {
+ spin_lock(&lli->lli_sa_lock);
+ thread_set_flags(thread, SVC_STOPPING);
+ lli->lli_sa_enabled = 0;
+ spin_unlock(&lli->lli_sa_lock);
+ }
- spin_lock(&plli->lli_agl_lock);
- while (!list_empty(&sai->sai_entries_agl) &&
- thread_is_running(thread)) {
- clli = list_entry(sai->sai_entries_agl.next,
- struct ll_inode_info, lli_agl_list);
- list_del_init(&clli->lli_agl_list);
- spin_unlock(&plli->lli_agl_lock);
- ll_agl_trigger(&clli->lli_vfs_inode, sai);
- spin_lock(&plli->lli_agl_lock);
- }
- spin_unlock(&plli->lli_agl_lock);
+ /*
+ * statahead is finished, but statahead entries need to be cached, wait
+ * for file release to stop me.
+ */
+ while (thread_is_running(thread)) {
+ l_wait_event(thread->t_ctl_waitq,
+ !sa_received_empty(sai) ||
+ !agl_list_empty(sai) ||
+ !thread_is_running(thread),
+ &lwi);
- rc = 0;
- goto out;
- } else {
- /*
- * chain is exhausted.
- * Normal case: continue to the next page.
- */
- ll_release_page(dir, page,
- le32_to_cpu(dp->ldp_flags) & LDF_COLLIDE);
- sai->sai_in_readpage = 1;
- page = ll_get_dir_page(dir, op_data, pos, &chain);
- sai->sai_in_readpage = 0;
- }
+ ll_post_statahead(sai);
}
out:
- ll_dir_chain_fini(&chain);
- ll_finish_md_op_data(op_data);
-out_put:
if (sai->sai_agl_valid) {
- spin_lock(&plli->lli_agl_lock);
+ spin_lock(&lli->lli_agl_lock);
thread_set_flags(agl_thread, SVC_STOPPING);
- spin_unlock(&plli->lli_agl_lock);
+ spin_unlock(&lli->lli_agl_lock);
wake_up(&agl_thread->t_ctl_waitq);
CDEBUG(D_READA, "stop agl thread: sai %p pid %u\n",
@@ -1257,21 +1183,27 @@ out_put:
&lwi);
} else {
/* Set agl_thread flags anyway. */
- thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
+ thread_set_flags(agl_thread, SVC_STOPPED);
}
- spin_lock(&plli->lli_sa_lock);
- if (!list_empty(&sai->sai_entries_received)) {
- thread_set_flags(thread, SVC_STOPPING);
- spin_unlock(&plli->lli_sa_lock);
-
- /* To release the resources held by received entries. */
- while (!list_empty(&sai->sai_entries_received))
- ll_post_statahead(sai);
- spin_lock(&plli->lli_sa_lock);
+ /*
+ * wait for inflight statahead RPCs to finish, and then we can free sai
+ * safely because statahead RPC will access sai data
+ */
+ while (sai->sai_sent != sai->sai_replied) {
+ /* in case we're not woken up, timeout wait */
+ lwi = LWI_TIMEOUT(HZ >> 3, NULL, NULL);
+ l_wait_event(thread->t_ctl_waitq,
+ sai->sai_sent == sai->sai_replied, &lwi);
}
+
+ /* release resources held by received entries. */
+ ll_post_statahead(sai);
+
+ spin_lock(&lli->lli_sa_lock);
thread_set_flags(thread, SVC_STOPPED);
- spin_unlock(&plli->lli_sa_lock);
+ spin_unlock(&lli->lli_sa_lock);
+
wake_up(&sai->sai_waitq);
wake_up(&thread->t_ctl_waitq);
ll_sai_put(sai);
@@ -1281,52 +1213,54 @@ out_put:
return rc;
}
-/**
- * called in ll_file_release().
- */
-void ll_stop_statahead(struct inode *dir, void *key)
+/* authorize opened dir handle @key to statahead later */
+void ll_authorize_statahead(struct inode *dir, void *key)
{
struct ll_inode_info *lli = ll_i2info(dir);
- if (unlikely(!key))
- return;
-
spin_lock(&lli->lli_sa_lock);
- if (lli->lli_opendir_key != key || lli->lli_opendir_pid == 0) {
- spin_unlock(&lli->lli_sa_lock);
- return;
+ if (!lli->lli_opendir_key && !lli->lli_sai) {
+ /*
+ * if lli_sai is not NULL, it means previous statahead is not
+ * finished yet, we'd better not start a new statahead for now.
+ */
+ LASSERT(!lli->lli_opendir_pid);
+ lli->lli_opendir_key = key;
+ lli->lli_opendir_pid = current_pid();
+ lli->lli_sa_enabled = 1;
}
+ spin_unlock(&lli->lli_sa_lock);
+}
- lli->lli_opendir_key = NULL;
+/*
+ * deauthorize opened dir handle @key to statahead, but statahead thread may
+ * still be running, notify it to quit.
+ */
+void ll_deauthorize_statahead(struct inode *dir, void *key)
+{
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct ll_statahead_info *sai;
- if (lli->lli_sai) {
- struct l_wait_info lwi = { 0 };
- struct ptlrpc_thread *thread = &lli->lli_sai->sai_thread;
+ LASSERT(lli->lli_opendir_key == key);
+ LASSERT(lli->lli_opendir_pid);
- if (!thread_is_stopped(thread)) {
- thread_set_flags(thread, SVC_STOPPING);
- spin_unlock(&lli->lli_sa_lock);
- wake_up(&thread->t_ctl_waitq);
-
- CDEBUG(D_READA, "stop statahead thread: sai %p pid %u\n",
- lli->lli_sai, (unsigned int)thread->t_pid);
- l_wait_event(thread->t_ctl_waitq,
- thread_is_stopped(thread),
- &lwi);
- } else {
- spin_unlock(&lli->lli_sa_lock);
- }
+ CDEBUG(D_READA, "deauthorize statahead for "DFID"\n",
+ PFID(&lli->lli_fid));
+ spin_lock(&lli->lli_sa_lock);
+ lli->lli_opendir_key = NULL;
+ lli->lli_opendir_pid = 0;
+ lli->lli_sa_enabled = 0;
+ sai = lli->lli_sai;
+ if (sai && thread_is_running(&sai->sai_thread)) {
/*
- * Put the ref which was held when first statahead_enter.
- * It maybe not the last ref for some statahead requests
- * maybe inflight.
+ * statahead thread may not quit yet because it needs to cache
+ * stated entries, now it's time to tell it to quit.
*/
- ll_sai_put(lli->lli_sai);
- } else {
- lli->lli_opendir_pid = 0;
- spin_unlock(&lli->lli_sa_lock);
+ thread_set_flags(&sai->sai_thread, SVC_STOPPING);
+ wake_up(&sai->sai_thread.t_ctl_waitq);
}
+ spin_unlock(&lli->lli_sa_lock);
}
enum {
@@ -1465,175 +1399,137 @@ out:
static void
ll_sai_unplug(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
{
- struct ptlrpc_thread *thread = &sai->sai_thread;
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
- int hit;
+ if (entry && entry->se_stat == SA_ENTRY_SUCC) {
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
- if (entry && entry->se_stat == SA_ENTRY_SUCC)
- hit = 1;
- else
- hit = 0;
-
- ll_sa_entry_fini(sai, entry);
- if (hit) {
sai->sai_hit++;
sai->sai_consecutive_miss = 0;
sai->sai_max = min(2 * sai->sai_max, sbi->ll_sa_max);
} else {
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
-
sai->sai_miss++;
sai->sai_consecutive_miss++;
- if (sa_low_hit(sai) && thread_is_running(thread)) {
- atomic_inc(&sbi->ll_sa_wrong);
- CDEBUG(D_READA, "Statahead for dir " DFID " hit ratio too low: hit/miss %llu/%llu, sent/replied %llu/%llu, stopping statahead thread\n",
- PFID(&lli->lli_fid), sai->sai_hit,
- sai->sai_miss, sai->sai_sent,
- sai->sai_replied);
- spin_lock(&lli->lli_sa_lock);
- if (!thread_is_stopped(thread))
- thread_set_flags(thread, SVC_STOPPING);
- spin_unlock(&lli->lli_sa_lock);
- }
}
-
- if (!thread_is_stopped(thread))
- wake_up(&thread->t_ctl_waitq);
+ ll_sa_entry_fini(sai, entry);
+ wake_up(&sai->sai_thread.t_ctl_waitq);
}
-/**
- * Start statahead thread if this is the first dir entry.
- * Otherwise if a thread is started already, wait it until it is ahead of me.
- * \retval 1 -- find entry with lock in cache, the caller needs to do
- * nothing.
- * \retval 0 -- find entry in cache, but without lock, the caller needs
- * refresh from MDS.
- * \retval others -- the caller need to process as non-statahead.
- */
-int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
- int only_unplug)
+static int revalidate_statahead_dentry(struct inode *dir,
+ struct ll_statahead_info *sai,
+ struct dentry **dentryp,
+ int only_unplug)
{
- struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_statahead_info *sai = lli->lli_sai;
- struct dentry *parent;
- struct ll_sa_entry *entry;
- struct ptlrpc_thread *thread;
- struct l_wait_info lwi = { 0 };
- struct task_struct *task;
- int rc = 0;
- struct ll_inode_info *plli;
-
- LASSERT(lli->lli_opendir_pid == current_pid());
-
- if (sai) {
- thread = &sai->sai_thread;
- if (unlikely(thread_is_stopped(thread) &&
- list_empty(&sai->sai_entries_stated))) {
- /* to release resource */
- ll_stop_statahead(dir, lli->lli_opendir_key);
- return -EAGAIN;
- }
+ struct ll_sa_entry *entry = NULL;
+ struct l_wait_info lwi = { 0 };
+ int rc = 0;
- if ((*dentryp)->d_name.name[0] == '.') {
- if (sai->sai_ls_all ||
- sai->sai_miss_hidden >= sai->sai_skip_hidden) {
+ if ((*dentryp)->d_name.name[0] == '.') {
+ if (sai->sai_ls_all ||
+ sai->sai_miss_hidden >= sai->sai_skip_hidden) {
+ /*
+ * Hidden dentry is the first one, or statahead
+ * thread does not skip so many hidden dentries
+ * before "sai_ls_all" enabled as below.
+ */
+ } else {
+ if (!sai->sai_ls_all)
/*
- * Hidden dentry is the first one, or statahead
- * thread does not skip so many hidden dentries
- * before "sai_ls_all" enabled as below.
+ * It maybe because hidden dentry is not
+ * the first one, "sai_ls_all" was not
+ * set, then "ls -al" missed. Enable
+ * "sai_ls_all" for such case.
*/
- } else {
- if (!sai->sai_ls_all)
- /*
- * It maybe because hidden dentry is not
- * the first one, "sai_ls_all" was not
- * set, then "ls -al" missed. Enable
- * "sai_ls_all" for such case.
- */
- sai->sai_ls_all = 1;
+ sai->sai_ls_all = 1;
- /*
- * Such "getattr" has been skipped before
- * "sai_ls_all" enabled as above.
- */
- sai->sai_miss_hidden++;
- return -EAGAIN;
- }
+ /*
+ * Such "getattr" has been skipped before
+ * "sai_ls_all" enabled as above.
+ */
+ sai->sai_miss_hidden++;
+ return -EAGAIN;
}
+ }
- entry = ll_sa_entry_get_byname(sai, &(*dentryp)->d_name);
- if (!entry || only_unplug) {
+ entry = ll_sa_entry_get_byname(sai, &(*dentryp)->d_name);
+ if (!entry || only_unplug) {
+ ll_sai_unplug(sai, entry);
+ return entry ? 1 : -EAGAIN;
+ }
+
+ /* if statahead is busy in readdir, help it do post-work */
+ if (!ll_sa_entry_stated(entry) && sai->sai_in_readpage)
+ ll_post_statahead(sai);
+
+ if (!ll_sa_entry_stated(entry)) {
+ sai->sai_index_wait = entry->se_index;
+ lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
+ LWI_ON_SIGNAL_NOOP, NULL);
+ rc = l_wait_event(sai->sai_waitq,
+ ll_sa_entry_stated(entry) ||
+ thread_is_stopped(&sai->sai_thread),
+ &lwi);
+ if (rc < 0) {
ll_sai_unplug(sai, entry);
- return entry ? 1 : -EAGAIN;
+ return -EAGAIN;
}
+ }
- /* if statahead is busy in readdir, help it do post-work */
- while (!ll_sa_entry_stated(entry) && sai->sai_in_readpage &&
- !sa_received_empty(sai))
- ll_post_statahead(sai);
-
- if (!ll_sa_entry_stated(entry)) {
- sai->sai_index_wait = entry->se_index;
- lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
- LWI_ON_SIGNAL_NOOP, NULL);
- rc = l_wait_event(sai->sai_waitq,
- ll_sa_entry_stated(entry) ||
- thread_is_stopped(thread),
- &lwi);
- if (rc < 0) {
- ll_sai_unplug(sai, entry);
- return -EAGAIN;
- }
- }
+ if (entry->se_stat == SA_ENTRY_SUCC && entry->se_inode) {
+ struct inode *inode = entry->se_inode;
+ struct lookup_intent it = { .it_op = IT_GETATTR,
+ .it_lock_handle = entry->se_handle };
+ __u64 bits;
+
+ rc = md_revalidate_lock(ll_i2mdexp(dir), &it,
+ ll_inode2fid(inode), &bits);
+ if (rc == 1) {
+ if (!(*dentryp)->d_inode) {
+ struct dentry *alias;
- if (entry->se_stat == SA_ENTRY_SUCC && entry->se_inode) {
- struct inode *inode = entry->se_inode;
- struct lookup_intent it = { .it_op = IT_GETATTR,
- .it_lock_handle =
- entry->se_handle };
- __u64 bits;
-
- rc = md_revalidate_lock(ll_i2mdexp(dir), &it,
- ll_inode2fid(inode), &bits);
- if (rc == 1) {
- if (!d_inode(*dentryp)) {
- struct dentry *alias;
-
- alias = ll_splice_alias(inode,
- *dentryp);
- if (IS_ERR(alias)) {
- ll_sai_unplug(sai, entry);
- return PTR_ERR(alias);
- }
- *dentryp = alias;
- } else if (d_inode(*dentryp) != inode) {
- /* revalidate, but inode is recreated */
- CDEBUG(D_READA, "%s: stale dentry %pd inode "DFID", statahead inode "DFID"\n",
- ll_get_fsname(d_inode(*dentryp)->i_sb, NULL, 0),
- *dentryp,
- PFID(ll_inode2fid(d_inode(*dentryp))),
- PFID(ll_inode2fid(inode)));
- ll_intent_release(&it);
+ alias = ll_splice_alias(inode, *dentryp);
+ if (IS_ERR(alias)) {
ll_sai_unplug(sai, entry);
- return -ESTALE;
- } else {
- iput(inode);
+ return PTR_ERR(alias);
}
- entry->se_inode = NULL;
-
- if ((bits & MDS_INODELOCK_LOOKUP) &&
- d_lustre_invalid(*dentryp))
- d_lustre_revalidate(*dentryp);
- ll_intent_release(&it);
+ *dentryp = alias;
+ } else if ((*dentryp)->d_inode != inode) {
+ /* revalidate, but inode is recreated */
+ CDEBUG(D_READA,
+ "%s: stale dentry %pd inode "DFID", statahead inode "DFID"\n",
+ ll_get_fsname((*dentryp)->d_inode->i_sb,
+ NULL, 0),
+ *dentryp,
+ PFID(ll_inode2fid((*dentryp)->d_inode)),
+ PFID(ll_inode2fid(inode)));
+ rc = -ESTALE;
+ goto out_unplug;
+ } else {
+ iput(inode);
}
- }
+ entry->se_inode = NULL;
- ll_sai_unplug(sai, entry);
- return rc;
+ if ((bits & MDS_INODELOCK_LOOKUP) &&
+ d_lustre_invalid(*dentryp))
+ d_lustre_revalidate(*dentryp);
+ ll_intent_release(&it);
+ }
}
+out_unplug:
+ ll_sai_unplug(sai, entry);
+ return rc;
+}
+
+static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
+{
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct ll_statahead_info *sai = NULL;
+ struct l_wait_info lwi = { 0 };
+ struct ptlrpc_thread *thread;
+ struct task_struct *task;
+ struct dentry *parent;
+ int rc;
/* I am the "lli_opendir_pid" owner, only me can set "lli_sai". */
- rc = is_first_dirent(dir, *dentryp);
+ rc = is_first_dirent(dir, dentry);
if (rc == LS_NONE_FIRST_DE) {
/* It is not "ls -{a}l" operation, no need statahead for it. */
rc = -EAGAIN;
@@ -1656,13 +1552,12 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
}
/* get parent reference count here, and put it in ll_statahead_thread */
- parent = dget((*dentryp)->d_parent);
+ parent = dget(dentry->d_parent);
if (unlikely(sai->sai_inode != d_inode(parent))) {
struct ll_inode_info *nlli = ll_i2info(d_inode(parent));
CWARN("Race condition, someone changed %pd just now: old parent "DFID", new parent "DFID"\n",
- *dentryp,
- PFID(&lli->lli_fid), PFID(&nlli->lli_fid));
+ dentry, PFID(&lli->lli_fid), PFID(&nlli->lli_fid));
dput(parent);
iput(sai->sai_inode);
rc = -EAGAIN;
@@ -1672,30 +1567,18 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
CDEBUG(D_READA, "start statahead thread: sai %p, parent %pd\n",
sai, parent);
- /* The sai buffer already has one reference taken at allocation time,
- * but as soon as we expose the sai by attaching it to the lli that
- * default reference can be dropped by another thread calling
- * ll_stop_statahead. We need to take a local reference to protect
- * the sai buffer while we intend to access it.
- */
- ll_sai_get(sai);
lli->lli_sai = sai;
- plli = ll_i2info(d_inode(parent));
task = kthread_run(ll_statahead_thread, parent, "ll_sa_%u",
- plli->lli_opendir_pid);
+ lli->lli_opendir_pid);
thread = &sai->sai_thread;
if (IS_ERR(task)) {
rc = PTR_ERR(task);
- CERROR("can't start ll_sa thread, rc: %d\n", rc);
+ CERROR("cannot start ll_sa thread: rc = %d\n", rc);
dput(parent);
lli->lli_opendir_key = NULL;
thread_set_flags(thread, SVC_STOPPED);
thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
- /* Drop both our own local reference and the default
- * reference from allocation time.
- */
- ll_sai_put(sai);
ll_sai_put(sai);
LASSERT(!lli->lli_sai);
return -EAGAIN;
@@ -1704,6 +1587,7 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
l_wait_event(thread->t_ctl_waitq,
thread_is_running(thread) || thread_is_stopped(thread),
&lwi);
+ atomic_inc(&ll_i2sbi(d_inode(parent))->ll_sa_running);
ll_sai_put(sai);
/*
@@ -1717,6 +1601,37 @@ out:
spin_lock(&lli->lli_sa_lock);
lli->lli_opendir_key = NULL;
lli->lli_opendir_pid = 0;
+ lli->lli_sa_enabled = 0;
spin_unlock(&lli->lli_sa_lock);
+
return rc;
}
+
+/**
+ * Start statahead thread if this is the first dir entry.
+ * Otherwise if a thread is started already, wait it until it is ahead of me.
+ * \retval 1 -- find entry with lock in cache, the caller needs to do
+ * nothing.
+ * \retval 0 -- find entry in cache, but without lock, the caller needs
+ * refresh from MDS.
+ * \retval others -- the caller need to process as non-statahead.
+ */
+int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
+ int only_unplug)
+{
+ struct ll_statahead_info *sai;
+
+ sai = ll_sai_get(dir);
+ if (sai) {
+ int rc;
+
+ rc = revalidate_statahead_dentry(dir, sai, dentryp,
+ only_unplug);
+ CDEBUG(D_READA, "revalidate statahead %pd: %d.\n",
+ *dentryp, rc);
+ ll_sai_put(sai);
+ return rc;
+ }
+
+ return start_statahead_thread(dir, *dentryp);
+}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 027/124] staging: lustre: ldlm: reconstruct proper flags on enqueue resend
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Vitaly Fertman, James Simmons
From: Vitaly Fertman <vitaly_fertman@xyratex.com>
otherwise, waiting lock may get granted as no BLOCKED_GRANTED
flag is returned
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Xyratex-bug-id: MRP-1944
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5496
Reviewed-on: http://review.whamcloud.com/11644
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_flock.c | 3 +--
drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 3 +--
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 6 ++----
3 files changed, 4 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
index 78a8450..861f36f 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
@@ -470,8 +470,7 @@ ldlm_flock_completion_ast(struct ldlm_lock *lock, __u64 flags, void *data)
if (flags & LDLM_FL_FAILED)
goto granted;
- if (!(flags & (LDLM_FL_BLOCK_WAIT | LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV))) {
+ if (!(flags & LDLM_FL_BLOCKED_MASK)) {
if (!data)
/* mds granted the lock in the reply */
goto granted;
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
index 55b7460..f516106 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
@@ -1549,8 +1549,7 @@ enum ldlm_error ldlm_lock_enqueue(struct ldlm_namespace *ns,
* before we got a chance to actually enqueue it. We don't
* need to do anything else.
*/
- *flags &= ~(LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV | LDLM_FL_BLOCK_WAIT);
+ *flags &= ~LDLM_FL_BLOCKED_MASK;
goto out;
}
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index f4364f9..cc835cb 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -177,8 +177,7 @@ int ldlm_completion_ast_async(struct ldlm_lock *lock, __u64 flags, void *data)
return 0;
}
- if (!(flags & (LDLM_FL_BLOCK_WAIT | LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV))) {
+ if (!(flags & LDLM_FL_BLOCKED_MASK)) {
wake_up(&lock->l_waitq);
return ldlm_completion_tail(lock);
}
@@ -224,8 +223,7 @@ int ldlm_completion_ast(struct ldlm_lock *lock, __u64 flags, void *data)
goto noreproc;
}
- if (!(flags & (LDLM_FL_BLOCK_WAIT | LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV))) {
+ if (!(flags & LDLM_FL_BLOCKED_MASK)) {
wake_up(&lock->l_waitq);
return 0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 027/124] staging: lustre: ldlm: reconstruct proper flags on enqueue resend
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Vitaly Fertman, James Simmons
From: Vitaly Fertman <vitaly_fertman@xyratex.com>
otherwise, waiting lock may get granted as no BLOCKED_GRANTED
flag is returned
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Xyratex-bug-id: MRP-1944
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5496
Reviewed-on: http://review.whamcloud.com/11644
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_flock.c | 3 +--
drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 3 +--
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 6 ++----
3 files changed, 4 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
index 78a8450..861f36f 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
@@ -470,8 +470,7 @@ ldlm_flock_completion_ast(struct ldlm_lock *lock, __u64 flags, void *data)
if (flags & LDLM_FL_FAILED)
goto granted;
- if (!(flags & (LDLM_FL_BLOCK_WAIT | LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV))) {
+ if (!(flags & LDLM_FL_BLOCKED_MASK)) {
if (!data)
/* mds granted the lock in the reply */
goto granted;
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
index 55b7460..f516106 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
@@ -1549,8 +1549,7 @@ enum ldlm_error ldlm_lock_enqueue(struct ldlm_namespace *ns,
* before we got a chance to actually enqueue it. We don't
* need to do anything else.
*/
- *flags &= ~(LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV | LDLM_FL_BLOCK_WAIT);
+ *flags &= ~LDLM_FL_BLOCKED_MASK;
goto out;
}
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index f4364f9..cc835cb 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -177,8 +177,7 @@ int ldlm_completion_ast_async(struct ldlm_lock *lock, __u64 flags, void *data)
return 0;
}
- if (!(flags & (LDLM_FL_BLOCK_WAIT | LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV))) {
+ if (!(flags & LDLM_FL_BLOCKED_MASK)) {
wake_up(&lock->l_waitq);
return ldlm_completion_tail(lock);
}
@@ -224,8 +223,7 @@ int ldlm_completion_ast(struct ldlm_lock *lock, __u64 flags, void *data)
goto noreproc;
}
- if (!(flags & (LDLM_FL_BLOCK_WAIT | LDLM_FL_BLOCK_GRANTED |
- LDLM_FL_BLOCK_CONV))) {
+ if (!(flags & LDLM_FL_BLOCKED_MASK)) {
wake_up(&lock->l_waitq);
return 0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 028/124] staging: lustre: ldlm: resend AST callbacks
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Vitaly Fertman, James Simmons
From: Vitaly Fertman <vitaly_fertman@xyratex.com>
While clients will resend client->server RPCs, servers would not
resend server->client RPCs such as LDLM callbacks (blocking
or completion callbacks/ASTs). This could result in clients being
evicted from the server if blocking callbacks were dropped by the
network (a failed router or lossy network) and the client did not
cancel the requested lock in time.
In order to fix this problem, this patch adds the ability to resend
LDLM callbacks from the server and give the client a chance to
respond within the timeout period before it is evicted:
- resend BL AST within lock callback timeout period;
- still do not resend CANCEL_ON_BLOCK;
- regular resend for CP AST without BL AST embedded;
- prolong lock callback timeout on resend;
some fixes:
- recovery-small test_10 to actually evict the client
with dropped BL AST;
- ETIMEDOUT to be returned if send limit is expired;
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5520
Reviewed-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Reviewed-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Xyratex-bug-id: MRP-417
Reviewed-on: http://review.whamcloud.com/9335
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mdc/mdc_reint.c | 9 +++------
drivers/staging/lustre/lustre/ptlrpc/client.c | 4 ++--
2 files changed, 5 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_reint.c b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
index c018e3b..af5c92c 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_reint.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
@@ -113,8 +113,7 @@ int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
if (op_data->op_attr.ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID))
bits |= MDS_INODELOCK_LOOKUP;
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
- (fid_is_sane(&op_data->op_fid1)) &&
- !OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ (fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX, bits);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
@@ -305,14 +304,12 @@ int mdc_unlink(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(!req);
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
- (fid_is_sane(&op_data->op_fid1)) &&
- !OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ (fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX,
MDS_INODELOCK_UPDATE);
if ((op_data->op_flags & MF_MDC_CANCEL_FID3) &&
- (fid_is_sane(&op_data->op_fid3)) &&
- !OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ (fid_is_sane(&op_data->op_fid3)))
count += mdc_resource_get_unused(exp, &op_data->op_fid3,
&cancels, LCK_EX,
MDS_INODELOCK_FULL);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index bae91bd..bea1c16 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -1037,8 +1037,8 @@ static int ptlrpc_import_delay_req(struct obd_import *imp,
*status = -EIO;
} else if (ptlrpc_send_limit_expired(req)) {
/* probably doesn't need to be a D_ERROR after initial testing */
- DEBUG_REQ(D_ERROR, req, "send limit expired ");
- *status = -EIO;
+ DEBUG_REQ(D_HA, req, "send limit expired ");
+ *status = -ETIMEDOUT;
} else if (req->rq_send_state == LUSTRE_IMP_CONNECTING &&
imp->imp_state == LUSTRE_IMP_CONNECTING) {
/* allow CONNECT even if import is invalid */
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 028/124] staging: lustre: ldlm: resend AST callbacks
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Vitaly Fertman, James Simmons
From: Vitaly Fertman <vitaly_fertman@xyratex.com>
While clients will resend client->server RPCs, servers would not
resend server->client RPCs such as LDLM callbacks (blocking
or completion callbacks/ASTs). This could result in clients being
evicted from the server if blocking callbacks were dropped by the
network (a failed router or lossy network) and the client did not
cancel the requested lock in time.
In order to fix this problem, this patch adds the ability to resend
LDLM callbacks from the server and give the client a chance to
respond within the timeout period before it is evicted:
- resend BL AST within lock callback timeout period;
- still do not resend CANCEL_ON_BLOCK;
- regular resend for CP AST without BL AST embedded;
- prolong lock callback timeout on resend;
some fixes:
- recovery-small test_10 to actually evict the client
with dropped BL AST;
- ETIMEDOUT to be returned if send limit is expired;
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5520
Reviewed-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Reviewed-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Xyratex-bug-id: MRP-417
Reviewed-on: http://review.whamcloud.com/9335
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mdc/mdc_reint.c | 9 +++------
drivers/staging/lustre/lustre/ptlrpc/client.c | 4 ++--
2 files changed, 5 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_reint.c b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
index c018e3b..af5c92c 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_reint.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
@@ -113,8 +113,7 @@ int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
if (op_data->op_attr.ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID))
bits |= MDS_INODELOCK_LOOKUP;
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
- (fid_is_sane(&op_data->op_fid1)) &&
- !OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ (fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX, bits);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
@@ -305,14 +304,12 @@ int mdc_unlink(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(!req);
if ((op_data->op_flags & MF_MDC_CANCEL_FID1) &&
- (fid_is_sane(&op_data->op_fid1)) &&
- !OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ (fid_is_sane(&op_data->op_fid1)))
count = mdc_resource_get_unused(exp, &op_data->op_fid1,
&cancels, LCK_EX,
MDS_INODELOCK_UPDATE);
if ((op_data->op_flags & MF_MDC_CANCEL_FID3) &&
- (fid_is_sane(&op_data->op_fid3)) &&
- !OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ (fid_is_sane(&op_data->op_fid3)))
count += mdc_resource_get_unused(exp, &op_data->op_fid3,
&cancels, LCK_EX,
MDS_INODELOCK_FULL);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index bae91bd..bea1c16 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -1037,8 +1037,8 @@ static int ptlrpc_import_delay_req(struct obd_import *imp,
*status = -EIO;
} else if (ptlrpc_send_limit_expired(req)) {
/* probably doesn't need to be a D_ERROR after initial testing */
- DEBUG_REQ(D_ERROR, req, "send limit expired ");
- *status = -EIO;
+ DEBUG_REQ(D_HA, req, "send limit expired ");
+ *status = -ETIMEDOUT;
} else if (req->rq_send_state == LUSTRE_IMP_CONNECTING &&
imp->imp_state == LUSTRE_IMP_CONNECTING) {
/* allow CONNECT even if import is invalid */
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 029/124] staging: lustre: ldlm: restore some of the interval functionality
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
Earlier a bunch of interval handling got removed since it wasn't
used by the upstream client. Now some of it is needed again for
the client code so this patch restores what is needed.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/interval_tree.h | 26 +++++
drivers/staging/lustre/lustre/ldlm/interval_tree.c | 100 +++++++++++++++++++-
2 files changed, 125 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/interval_tree.h b/drivers/staging/lustre/lustre/include/interval_tree.h
index 4a15228..5d387d3 100644
--- a/drivers/staging/lustre/lustre/include/interval_tree.h
+++ b/drivers/staging/lustre/lustre/include/interval_tree.h
@@ -63,6 +63,11 @@ static inline int interval_is_intree(struct interval_node *node)
return node->in_intree == 1;
}
+static inline __u64 interval_low(struct interval_node *node)
+{
+ return node->in_extent.start;
+}
+
static inline __u64 interval_high(struct interval_node *node)
{
return node->in_extent.end;
@@ -77,8 +82,29 @@ static inline void interval_set(struct interval_node *node,
node->in_max_high = end;
}
+/*
+ * Rules to write an interval callback.
+ * - the callback returns INTERVAL_ITER_STOP when it thinks the iteration
+ * should be stopped. It will then cause the iteration function to return
+ * immediately with return value INTERVAL_ITER_STOP.
+ * - callbacks for interval_iterate and interval_iterate_reverse: Every
+ * nodes in the tree will be set to @node before the callback being called
+ * - callback for interval_search: Only overlapped node will be set to @node
+ * before the callback being called.
+ */
+typedef enum interval_iter (*interval_callback_t)(struct interval_node *node,
+ void *args);
+
struct interval_node *interval_insert(struct interval_node *node,
struct interval_node **root);
void interval_erase(struct interval_node *node, struct interval_node **root);
+/*
+ * Search the extents in the tree and call @func for each overlapped
+ * extents.
+ */
+enum interval_iter interval_search(struct interval_node *root,
+ struct interval_node_extent *ex,
+ interval_callback_t func, void *data);
+
#endif
diff --git a/drivers/staging/lustre/lustre/ldlm/interval_tree.c b/drivers/staging/lustre/lustre/ldlm/interval_tree.c
index f4a70eb..e134ecd 100644
--- a/drivers/staging/lustre/lustre/ldlm/interval_tree.c
+++ b/drivers/staging/lustre/lustre/ldlm/interval_tree.c
@@ -90,6 +90,17 @@ static inline int extent_equal(struct interval_node_extent *e1,
return (e1->start == e2->start) && (e1->end == e2->end);
}
+static inline int extent_overlapped(struct interval_node_extent *e1,
+ struct interval_node_extent *e2)
+{
+ return (e1->start <= e2->end) && (e2->start <= e1->end);
+}
+
+static inline int node_equal(struct interval_node *n1, struct interval_node *n2)
+{
+ return extent_equal(&n1->in_extent, &n2->in_extent);
+}
+
static inline __u64 max_u64(__u64 x, __u64 y)
{
return x > y ? x : y;
@@ -262,7 +273,7 @@ struct interval_node *interval_insert(struct interval_node *node,
p = root;
while (*p) {
parent = *p;
- if (extent_equal(&parent->in_extent, &node->in_extent))
+ if (node_equal(parent, node))
return parent;
/* max_high field must be updated after each iteration */
@@ -463,3 +474,90 @@ color:
interval_erase_color(child, parent, root);
}
EXPORT_SYMBOL(interval_erase);
+
+static inline int interval_may_overlap(struct interval_node *node,
+ struct interval_node_extent *ext)
+{
+ return (ext->start <= node->in_max_high &&
+ ext->end >= interval_low(node));
+}
+
+/*
+ * This function finds all intervals that overlap interval ext,
+ * and calls func to handle resulted intervals one by one.
+ * in lustre, this function will find all conflicting locks in
+ * the granted queue and add these locks to the ast work list.
+ *
+ * {
+ * if (!node)
+ * return 0;
+ * if (ext->end < interval_low(node)) {
+ * interval_search(node->in_left, ext, func, data);
+ * } else if (interval_may_overlap(node, ext)) {
+ * if (extent_overlapped(ext, &node->in_extent))
+ * func(node, data);
+ * interval_search(node->in_left, ext, func, data);
+ * interval_search(node->in_right, ext, func, data);
+ * }
+ * return 0;
+ * }
+ *
+ */
+enum interval_iter interval_search(struct interval_node *node,
+ struct interval_node_extent *ext,
+ interval_callback_t func,
+ void *data)
+{
+ enum interval_iter rc = INTERVAL_ITER_CONT;
+ struct interval_node *parent;
+
+ LASSERT(ext);
+ LASSERT(func);
+
+ while (node) {
+ if (ext->end < interval_low(node)) {
+ if (node->in_left) {
+ node = node->in_left;
+ continue;
+ }
+ } else if (interval_may_overlap(node, ext)) {
+ if (extent_overlapped(ext, &node->in_extent)) {
+ rc = func(node, data);
+ if (rc == INTERVAL_ITER_STOP)
+ break;
+ }
+
+ if (node->in_left) {
+ node = node->in_left;
+ continue;
+ }
+ if (node->in_right) {
+ node = node->in_right;
+ continue;
+ }
+ }
+
+ parent = node->in_parent;
+ while (parent) {
+ if (node_is_left_child(node) &&
+ parent->in_right) {
+ /*
+ * If we ever got the left, it means that the
+ * parent met ext->end<interval_low(parent), or
+ * may_overlap(parent). If the former is true,
+ * we needn't go back. So stop early and check
+ * may_overlap(parent) after this loop.
+ */
+ node = parent->in_right;
+ break;
+ }
+ node = parent;
+ parent = parent->in_parent;
+ }
+ if (!parent || !interval_may_overlap(parent, ext))
+ break;
+ }
+
+ return rc;
+}
+EXPORT_SYMBOL(interval_search);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 029/124] staging: lustre: ldlm: restore some of the interval functionality
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
Earlier a bunch of interval handling got removed since it wasn't
used by the upstream client. Now some of it is needed again for
the client code so this patch restores what is needed.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/interval_tree.h | 26 +++++
drivers/staging/lustre/lustre/ldlm/interval_tree.c | 100 +++++++++++++++++++-
2 files changed, 125 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/interval_tree.h b/drivers/staging/lustre/lustre/include/interval_tree.h
index 4a15228..5d387d3 100644
--- a/drivers/staging/lustre/lustre/include/interval_tree.h
+++ b/drivers/staging/lustre/lustre/include/interval_tree.h
@@ -63,6 +63,11 @@ static inline int interval_is_intree(struct interval_node *node)
return node->in_intree == 1;
}
+static inline __u64 interval_low(struct interval_node *node)
+{
+ return node->in_extent.start;
+}
+
static inline __u64 interval_high(struct interval_node *node)
{
return node->in_extent.end;
@@ -77,8 +82,29 @@ static inline void interval_set(struct interval_node *node,
node->in_max_high = end;
}
+/*
+ * Rules to write an interval callback.
+ * - the callback returns INTERVAL_ITER_STOP when it thinks the iteration
+ * should be stopped. It will then cause the iteration function to return
+ * immediately with return value INTERVAL_ITER_STOP.
+ * - callbacks for interval_iterate and interval_iterate_reverse: Every
+ * nodes in the tree will be set to @node before the callback being called
+ * - callback for interval_search: Only overlapped node will be set to @node
+ * before the callback being called.
+ */
+typedef enum interval_iter (*interval_callback_t)(struct interval_node *node,
+ void *args);
+
struct interval_node *interval_insert(struct interval_node *node,
struct interval_node **root);
void interval_erase(struct interval_node *node, struct interval_node **root);
+/*
+ * Search the extents in the tree and call @func for each overlapped
+ * extents.
+ */
+enum interval_iter interval_search(struct interval_node *root,
+ struct interval_node_extent *ex,
+ interval_callback_t func, void *data);
+
#endif
diff --git a/drivers/staging/lustre/lustre/ldlm/interval_tree.c b/drivers/staging/lustre/lustre/ldlm/interval_tree.c
index f4a70eb..e134ecd 100644
--- a/drivers/staging/lustre/lustre/ldlm/interval_tree.c
+++ b/drivers/staging/lustre/lustre/ldlm/interval_tree.c
@@ -90,6 +90,17 @@ static inline int extent_equal(struct interval_node_extent *e1,
return (e1->start == e2->start) && (e1->end == e2->end);
}
+static inline int extent_overlapped(struct interval_node_extent *e1,
+ struct interval_node_extent *e2)
+{
+ return (e1->start <= e2->end) && (e2->start <= e1->end);
+}
+
+static inline int node_equal(struct interval_node *n1, struct interval_node *n2)
+{
+ return extent_equal(&n1->in_extent, &n2->in_extent);
+}
+
static inline __u64 max_u64(__u64 x, __u64 y)
{
return x > y ? x : y;
@@ -262,7 +273,7 @@ struct interval_node *interval_insert(struct interval_node *node,
p = root;
while (*p) {
parent = *p;
- if (extent_equal(&parent->in_extent, &node->in_extent))
+ if (node_equal(parent, node))
return parent;
/* max_high field must be updated after each iteration */
@@ -463,3 +474,90 @@ color:
interval_erase_color(child, parent, root);
}
EXPORT_SYMBOL(interval_erase);
+
+static inline int interval_may_overlap(struct interval_node *node,
+ struct interval_node_extent *ext)
+{
+ return (ext->start <= node->in_max_high &&
+ ext->end >= interval_low(node));
+}
+
+/*
+ * This function finds all intervals that overlap interval ext,
+ * and calls func to handle resulted intervals one by one.
+ * in lustre, this function will find all conflicting locks in
+ * the granted queue and add these locks to the ast work list.
+ *
+ * {
+ * if (!node)
+ * return 0;
+ * if (ext->end < interval_low(node)) {
+ * interval_search(node->in_left, ext, func, data);
+ * } else if (interval_may_overlap(node, ext)) {
+ * if (extent_overlapped(ext, &node->in_extent))
+ * func(node, data);
+ * interval_search(node->in_left, ext, func, data);
+ * interval_search(node->in_right, ext, func, data);
+ * }
+ * return 0;
+ * }
+ *
+ */
+enum interval_iter interval_search(struct interval_node *node,
+ struct interval_node_extent *ext,
+ interval_callback_t func,
+ void *data)
+{
+ enum interval_iter rc = INTERVAL_ITER_CONT;
+ struct interval_node *parent;
+
+ LASSERT(ext);
+ LASSERT(func);
+
+ while (node) {
+ if (ext->end < interval_low(node)) {
+ if (node->in_left) {
+ node = node->in_left;
+ continue;
+ }
+ } else if (interval_may_overlap(node, ext)) {
+ if (extent_overlapped(ext, &node->in_extent)) {
+ rc = func(node, data);
+ if (rc == INTERVAL_ITER_STOP)
+ break;
+ }
+
+ if (node->in_left) {
+ node = node->in_left;
+ continue;
+ }
+ if (node->in_right) {
+ node = node->in_right;
+ continue;
+ }
+ }
+
+ parent = node->in_parent;
+ while (parent) {
+ if (node_is_left_child(node) &&
+ parent->in_right) {
+ /*
+ * If we ever got the left, it means that the
+ * parent met ext->end<interval_low(parent), or
+ * may_overlap(parent). If the former is true,
+ * we needn't go back. So stop early and check
+ * may_overlap(parent) after this loop.
+ */
+ node = parent->in_right;
+ break;
+ }
+ node = parent;
+ parent = parent->in_parent;
+ }
+ if (!parent || !interval_may_overlap(parent, ext))
+ break;
+ }
+
+ return rc;
+}
+EXPORT_SYMBOL(interval_search);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Prakash Surya, Patrick Farrell, Jinshan Xiong, James Simmons
From: Prakash Surya <surya1@llnl.gov>
Testing has shown the ll_inode_inode's lli_write_mutex to be a
limiting factor with single shared file write performance, when using
many writing threads on a single machine. Even if each thread is
writing to a unique portion of the file, the lli_write_mutex will
prevent no more than a single thread to ever write to the file
simultaneously.
This change attempts to remove this bottle neck, by replacing this
mutex with a range lock. This should allow multiple threads to write
to a single file simultaneously iff the threads are writing to unique
regions of the file.
Performance testing shows this change to garner a significant
performance boost to write bandwidth. Using FIO on a single machine
with 32 GB of RAM, write performance tests were run with and without
this change applied; the results are below:
+---------+-----------+---------+--------+--------+--------+
| fio v2.0.13 | Write Bandwidth (KB/s) |
+---------+-----------+---------+--------+--------+--------+
| # Tasks | GB / Task | Test 1 | Test 2 | Test 3 | Test 4 |
+---------+-----------+---------+--------+--------+--------+
| 1 | 64 | 452446 | 454623 | 457653 | 463737 |
| 2 | 32 | 850318 | 565373 | 602498 | 733027 |
| 4 | 16 | 1058900 | 463546 | 529107 | 976284 |
| 8 | 8 | 1026300 | 468190 | 576451 | 963404 |
| 16 | 4 | 1065500 | 503160 | 462902 | 830065 |
| 32 | 2 | 1068600 | 462228 | 466963 | 749733 |
| 64 | 1 | 991830 | 556618 | 557863 | 710912 |
+---------+-----------+---------+--------+--------+--------+
* Test 1: Lustre client running 04ec54f. File per process write
workload. This test was used as a baseline for what we
_could_ achieve in the single shared file tests if the
bottle necks were removed.
* Test 2: Lustre client running 04ec54f. Single shared file
workload, each task writing to a unique region.
* Test 3: Lustre client running 04ec54f + I0023132b. Single shared
file workload, each task writing to a unique region.
* Test 4: Lustre client running 04ec54f + this patch.
Single shared file workload, each task writing to a unique
region.
Direct IO does not use the page cache like normal IO, so
concurrent direct IO reads of the same pages are not safe.
As a result, direct IO reads must take the range lock
in ll_file_io_generic, otherwise they will attempt to work
on the same pages and hit assertions like:
(osc_request.c:1219:osc_brw_prep_request())
ASSERTION( i == 0 || pg->off > pg_prev->off ) failed:
i 3 p_c 10 pg ffffea00017a5208 [pri 0 ind 2771] off 16384
prev_pg ffffea00017a51d0 [pri 0 ind 2256] off 16384
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Patrick Farrell <paf@cray.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1669
Reviewed-on: http://review.whamcloud.com/6320
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6227
Reviewed-on: http://review.whamcloud.com/14385
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Reviewed-by: Hiroya Nozaki <nozaki.hiroya@jp.fujitsu.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/Makefile | 2 +-
drivers/staging/lustre/lustre/llite/file.c | 37 +++-
.../staging/lustre/lustre/llite/llite_internal.h | 5 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 2 +-
drivers/staging/lustre/lustre/llite/range_lock.c | 233 ++++++++++++++++++++
drivers/staging/lustre/lustre/llite/range_lock.h | 82 +++++++
6 files changed, 348 insertions(+), 13 deletions(-)
create mode 100644 drivers/staging/lustre/lustre/llite/range_lock.c
create mode 100644 drivers/staging/lustre/lustre/llite/range_lock.h
diff --git a/drivers/staging/lustre/lustre/llite/Makefile b/drivers/staging/lustre/lustre/llite/Makefile
index 2cbb1b8..1ac0940 100644
--- a/drivers/staging/lustre/lustre/llite/Makefile
+++ b/drivers/staging/lustre/lustre/llite/Makefile
@@ -1,6 +1,6 @@
obj-$(CONFIG_LUSTRE_FS) += lustre.o
lustre-y := dcache.o dir.o file.o llite_close.o llite_lib.o llite_nfs.o \
- rw.o namei.o symlink.o llite_mmap.o \
+ rw.o namei.o symlink.o llite_mmap.o range_lock.o \
xattr.o xattr_cache.o rw26.o super25.o statahead.o \
glimpse.o lcommon_cl.o lcommon_misc.o \
vvp_dev.o vvp_page.o vvp_lock.o vvp_io.o vvp_object.o vvp_req.o \
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 273b563..d9ee255 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1126,6 +1126,7 @@ ll_file_io_generic(const struct lu_env *env, struct vvp_io_args *args,
{
struct ll_inode_info *lli = ll_i2info(file_inode(file));
struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+ struct range_lock range;
struct cl_io *io;
ssize_t result;
@@ -1138,7 +1139,12 @@ restart:
if (cl_io_rw_init(env, io, iot, *ppos, count) == 0) {
struct vvp_io *vio = vvp_env_io(env);
- int write_mutex_locked = 0;
+ bool range_locked = false;
+
+ if (file->f_flags & O_APPEND)
+ range_lock_init(&range, 0, LUSTRE_EOF);
+ else
+ range_lock_init(&range, *ppos, *ppos + count - 1);
vio->vui_fd = LUSTRE_FPRIVATE(file);
vio->vui_io_subtype = args->via_io_subtype;
@@ -1147,14 +1153,23 @@ restart:
case IO_NORMAL:
vio->vui_iter = args->u.normal.via_iter;
vio->vui_iocb = args->u.normal.via_iocb;
- if ((iot == CIT_WRITE) &&
+ /*
+ * Direct IO reads must also take range lock,
+ * or multiple reads will try to work on the same pages
+ * See LU-6227 for details.
+ */
+ if (((iot == CIT_WRITE) ||
+ (iot == CIT_READ && (file->f_flags & O_DIRECT))) &&
!(vio->vui_fd->fd_flags & LL_FILE_GROUP_LOCKED)) {
- if (mutex_lock_interruptible(&lli->
- lli_write_mutex)) {
- result = -ERESTARTSYS;
+ CDEBUG(D_VFSTRACE, "Range lock [%llu, %llu]\n",
+ range.rl_node.in_extent.start,
+ range.rl_node.in_extent.end);
+ result = range_lock(&lli->lli_write_tree,
+ &range);
+ if (result < 0)
goto out;
- }
- write_mutex_locked = 1;
+
+ range_locked = true;
}
down_read(&lli->lli_trunc_sem);
break;
@@ -1171,8 +1186,12 @@ restart:
ll_cl_remove(file, env);
if (args->via_io_subtype == IO_NORMAL)
up_read(&lli->lli_trunc_sem);
- if (write_mutex_locked)
- mutex_unlock(&lli->lli_write_mutex);
+ if (range_locked) {
+ CDEBUG(D_VFSTRACE, "Range unlock [%llu, %llu]\n",
+ range.rl_node.in_extent.start,
+ range.rl_node.in_extent.end);
+ range_unlock(&lli->lli_write_tree, &range);
+ }
} else {
/* cl_io_rw_init() handled IO */
result = io->ci_result;
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index f903f2a..7f0a6c9 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -46,6 +46,7 @@
#include <linux/xattr.h>
#include <linux/posix_acl_xattr.h>
#include "vvp_internal.h"
+#include "range_lock.h"
#ifndef FMODE_EXEC
#define FMODE_EXEC 0
@@ -210,7 +211,7 @@ struct ll_inode_info {
* }
*/
struct rw_semaphore f_trunc_sem;
- struct mutex f_write_mutex;
+ struct range_lock_tree f_write_tree;
struct rw_semaphore f_glimpse_sem;
unsigned long f_glimpse_time;
@@ -235,7 +236,7 @@ struct ll_inode_info {
#define lli_symlink_name u.f.f_symlink_name
#define lli_maxbytes u.f.f_maxbytes
#define lli_trunc_sem u.f.f_trunc_sem
-#define lli_write_mutex u.f.f_write_mutex
+#define lli_write_tree u.f.f_write_tree
#define lli_glimpse_sem u.f.f_glimpse_sem
#define lli_glimpse_time u.f.f_glimpse_time
#define lli_agl_list u.f.f_agl_list
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 93fd69b..9112fa8 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -807,7 +807,7 @@ void ll_lli_init(struct ll_inode_info *lli)
mutex_init(&lli->lli_size_mutex);
lli->lli_symlink_name = NULL;
init_rwsem(&lli->lli_trunc_sem);
- mutex_init(&lli->lli_write_mutex);
+ range_lock_tree_init(&lli->lli_write_tree);
init_rwsem(&lli->lli_glimpse_sem);
lli->lli_glimpse_time = 0;
INIT_LIST_HEAD(&lli->lli_agl_list);
diff --git a/drivers/staging/lustre/lustre/llite/range_lock.c b/drivers/staging/lustre/lustre/llite/range_lock.c
new file mode 100644
index 0000000..94c818f
--- /dev/null
+++ b/drivers/staging/lustre/lustre/llite/range_lock.c
@@ -0,0 +1,233 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Range lock is used to allow multiple threads writing a single shared
+ * file given each thread is writing to a non-overlapping portion of the
+ * file.
+ *
+ * Refer to the possible upstream kernel version of range lock by
+ * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
+ *
+ * This file could later replaced by the upstream kernel version.
+ */
+/*
+ * Author: Prakash Surya <surya1@llnl.gov>
+ * Author: Bobi Jam <bobijam.xu@intel.com>
+ */
+#include "range_lock.h"
+#include "../include/lustre/lustre_user.h"
+
+/**
+ * Initialize a range lock tree
+ *
+ * \param tree [in] an empty range lock tree
+ *
+ * Pre: Caller should have allocated the range lock tree.
+ * Post: The range lock tree is ready to function.
+ */
+void range_lock_tree_init(struct range_lock_tree *tree)
+{
+ tree->rlt_root = NULL;
+ tree->rlt_sequence = 0;
+ spin_lock_init(&tree->rlt_lock);
+}
+
+/**
+ * Initialize a range lock node
+ *
+ * \param lock [in] an empty range lock node
+ * \param start [in] start of the covering region
+ * \param end [in] end of the covering region
+ *
+ * Pre: Caller should have allocated the range lock node.
+ * Post: The range lock node is meant to cover [start, end] region
+ */
+void range_lock_init(struct range_lock *lock, __u64 start, __u64 end)
+{
+ memset(&lock->rl_node, 0, sizeof(lock->rl_node));
+ if (end != LUSTRE_EOF)
+ end >>= PAGE_SHIFT;
+ interval_set(&lock->rl_node, start >> PAGE_SHIFT, end);
+ INIT_LIST_HEAD(&lock->rl_next_lock);
+ lock->rl_task = NULL;
+ lock->rl_lock_count = 0;
+ lock->rl_blocking_ranges = 0;
+ lock->rl_sequence = 0;
+}
+
+static inline struct range_lock *next_lock(struct range_lock *lock)
+{
+ return list_entry(lock->rl_next_lock.next, typeof(*lock), rl_next_lock);
+}
+
+/**
+ * Helper function of range_unlock()
+ *
+ * \param node [in] a range lock found overlapped during interval node
+ * search
+ * \param arg [in] the range lock to be tested
+ *
+ * \retval INTERVAL_ITER_CONT indicate to continue the search for next
+ * overlapping range node
+ * \retval INTERVAL_ITER_STOP indicate to stop the search
+ */
+static enum interval_iter range_unlock_cb(struct interval_node *node, void *arg)
+{
+ struct range_lock *lock = arg;
+ struct range_lock *overlap = node2rangelock(node);
+ struct range_lock *iter;
+
+ list_for_each_entry(iter, &overlap->rl_next_lock, rl_next_lock) {
+ if (iter->rl_sequence > lock->rl_sequence) {
+ --iter->rl_blocking_ranges;
+ LASSERT(iter->rl_blocking_ranges > 0);
+ }
+ }
+ if (overlap->rl_sequence > lock->rl_sequence) {
+ --overlap->rl_blocking_ranges;
+ if (overlap->rl_blocking_ranges == 0)
+ wake_up_process(overlap->rl_task);
+ }
+ return INTERVAL_ITER_CONT;
+}
+
+/**
+ * Unlock a range lock, wake up locks blocked by this lock.
+ *
+ * \param tree [in] range lock tree
+ * \param lock [in] range lock to be deleted
+ *
+ * If this lock has been granted, relase it; if not, just delete it from
+ * the tree or the same region lock list. Wake up those locks only blocked
+ * by this lock through range_unlock_cb().
+ */
+void range_unlock(struct range_lock_tree *tree, struct range_lock *lock)
+{
+ spin_lock(&tree->rlt_lock);
+ if (!list_empty(&lock->rl_next_lock)) {
+ struct range_lock *next;
+
+ if (interval_is_intree(&lock->rl_node)) { /* first lock */
+ /* Insert the next same range lock into the tree */
+ next = next_lock(lock);
+ next->rl_lock_count = lock->rl_lock_count - 1;
+ interval_erase(&lock->rl_node, &tree->rlt_root);
+ interval_insert(&next->rl_node, &tree->rlt_root);
+ } else {
+ /* find the first lock in tree */
+ list_for_each_entry(next, &lock->rl_next_lock,
+ rl_next_lock) {
+ if (!interval_is_intree(&next->rl_node))
+ continue;
+
+ LASSERT(next->rl_lock_count > 0);
+ next->rl_lock_count--;
+ break;
+ }
+ }
+ list_del_init(&lock->rl_next_lock);
+ } else {
+ LASSERT(interval_is_intree(&lock->rl_node));
+ interval_erase(&lock->rl_node, &tree->rlt_root);
+ }
+
+ interval_search(tree->rlt_root, &lock->rl_node.in_extent,
+ range_unlock_cb, lock);
+ spin_unlock(&tree->rlt_lock);
+}
+
+/**
+ * Helper function of range_lock()
+ *
+ * \param node [in] a range lock found overlapped during interval node
+ * search
+ * \param arg [in] the range lock to be tested
+ *
+ * \retval INTERVAL_ITER_CONT indicate to continue the search for next
+ * overlapping range node
+ * \retval INTERVAL_ITER_STOP indicate to stop the search
+ */
+static enum interval_iter range_lock_cb(struct interval_node *node, void *arg)
+{
+ struct range_lock *lock = (struct range_lock *)arg;
+ struct range_lock *overlap = node2rangelock(node);
+
+ lock->rl_blocking_ranges += overlap->rl_lock_count + 1;
+ return INTERVAL_ITER_CONT;
+}
+
+/**
+ * Lock a region
+ *
+ * \param tree [in] range lock tree
+ * \param lock [in] range lock node containing the region span
+ *
+ * \retval 0 get the range lock
+ * \retval <0 error code while not getting the range lock
+ *
+ * If there exists overlapping range lock, the new lock will wait and
+ * retry, if later it find that it is not the chosen one to wake up,
+ * it wait again.
+ */
+int range_lock(struct range_lock_tree *tree, struct range_lock *lock)
+{
+ struct interval_node *node;
+ int rc = 0;
+
+ spin_lock(&tree->rlt_lock);
+ /*
+ * We need to check for all conflicting intervals
+ * already in the tree.
+ */
+ interval_search(tree->rlt_root, &lock->rl_node.in_extent,
+ range_lock_cb, lock);
+ /*
+ * Insert to the tree if I am unique, otherwise I've been linked to
+ * the rl_next_lock of another lock which has the same range as mine
+ * in range_lock_cb().
+ */
+ node = interval_insert(&lock->rl_node, &tree->rlt_root);
+ if (node) {
+ struct range_lock *tmp = node2rangelock(node);
+
+ list_add_tail(&lock->rl_next_lock, &tmp->rl_next_lock);
+ tmp->rl_lock_count++;
+ }
+ lock->rl_sequence = ++tree->rlt_sequence;
+
+ while (lock->rl_blocking_ranges > 0) {
+ lock->rl_task = current;
+ __set_current_state(TASK_INTERRUPTIBLE);
+ spin_unlock(&tree->rlt_lock);
+ schedule();
+
+ if (signal_pending(current)) {
+ range_unlock(tree, lock);
+ rc = -EINTR;
+ goto out;
+ }
+ spin_lock(&tree->rlt_lock);
+ }
+ spin_unlock(&tree->rlt_lock);
+out:
+ return rc;
+}
diff --git a/drivers/staging/lustre/lustre/llite/range_lock.h b/drivers/staging/lustre/lustre/llite/range_lock.h
new file mode 100644
index 0000000..c6d04a6
--- /dev/null
+++ b/drivers/staging/lustre/lustre/llite/range_lock.h
@@ -0,0 +1,82 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Range lock is used to allow multiple threads writing a single shared
+ * file given each thread is writing to a non-overlapping portion of the
+ * file.
+ *
+ * Refer to the possible upstream kernel version of range lock by
+ * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
+ *
+ * This file could later replaced by the upstream kernel version.
+ */
+/*
+ * Author: Prakash Surya <surya1@llnl.gov>
+ * Author: Bobi Jam <bobijam.xu@intel.com>
+ */
+#ifndef _RANGE_LOCK_H
+#define _RANGE_LOCK_H
+
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../include/interval_tree.h"
+
+struct range_lock {
+ struct interval_node rl_node;
+ /**
+ * Process to enqueue this lock.
+ */
+ struct task_struct *rl_task;
+ /**
+ * List of locks with the same range.
+ */
+ struct list_head rl_next_lock;
+ /**
+ * Number of locks in the list rl_next_lock
+ */
+ unsigned int rl_lock_count;
+ /**
+ * Number of ranges which are blocking acquisition of the lock
+ */
+ unsigned int rl_blocking_ranges;
+ /**
+ * Sequence number of range lock. This number is used to get to know
+ * the order the locks are queued; this is required for range_cancel().
+ */
+ __u64 rl_sequence;
+};
+
+static inline struct range_lock *node2rangelock(const struct interval_node *n)
+{
+ return container_of(n, struct range_lock, rl_node);
+}
+
+struct range_lock_tree {
+ struct interval_node *rlt_root;
+ spinlock_t rlt_lock; /* protect range lock tree */
+ __u64 rlt_sequence;
+};
+
+void range_lock_tree_init(struct range_lock_tree *tree);
+void range_lock_init(struct range_lock *lock, __u64 start, __u64 end);
+int range_lock(struct range_lock_tree *tree, struct range_lock *lock);
+void range_unlock(struct range_lock_tree *tree, struct range_lock *lock);
+#endif
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Prakash Surya, Patrick Farrell, Jinshan Xiong, James Simmons
From: Prakash Surya <surya1@llnl.gov>
Testing has shown the ll_inode_inode's lli_write_mutex to be a
limiting factor with single shared file write performance, when using
many writing threads on a single machine. Even if each thread is
writing to a unique portion of the file, the lli_write_mutex will
prevent no more than a single thread to ever write to the file
simultaneously.
This change attempts to remove this bottle neck, by replacing this
mutex with a range lock. This should allow multiple threads to write
to a single file simultaneously iff the threads are writing to unique
regions of the file.
Performance testing shows this change to garner a significant
performance boost to write bandwidth. Using FIO on a single machine
with 32 GB of RAM, write performance tests were run with and without
this change applied; the results are below:
+---------+-----------+---------+--------+--------+--------+
| fio v2.0.13 | Write Bandwidth (KB/s) |
+---------+-----------+---------+--------+--------+--------+
| # Tasks | GB / Task | Test 1 | Test 2 | Test 3 | Test 4 |
+---------+-----------+---------+--------+--------+--------+
| 1 | 64 | 452446 | 454623 | 457653 | 463737 |
| 2 | 32 | 850318 | 565373 | 602498 | 733027 |
| 4 | 16 | 1058900 | 463546 | 529107 | 976284 |
| 8 | 8 | 1026300 | 468190 | 576451 | 963404 |
| 16 | 4 | 1065500 | 503160 | 462902 | 830065 |
| 32 | 2 | 1068600 | 462228 | 466963 | 749733 |
| 64 | 1 | 991830 | 556618 | 557863 | 710912 |
+---------+-----------+---------+--------+--------+--------+
* Test 1: Lustre client running 04ec54f. File per process write
workload. This test was used as a baseline for what we
_could_ achieve in the single shared file tests if the
bottle necks were removed.
* Test 2: Lustre client running 04ec54f. Single shared file
workload, each task writing to a unique region.
* Test 3: Lustre client running 04ec54f + I0023132b. Single shared
file workload, each task writing to a unique region.
* Test 4: Lustre client running 04ec54f + this patch.
Single shared file workload, each task writing to a unique
region.
Direct IO does not use the page cache like normal IO, so
concurrent direct IO reads of the same pages are not safe.
As a result, direct IO reads must take the range lock
in ll_file_io_generic, otherwise they will attempt to work
on the same pages and hit assertions like:
(osc_request.c:1219:osc_brw_prep_request())
ASSERTION( i == 0 || pg->off > pg_prev->off ) failed:
i 3 p_c 10 pg ffffea00017a5208 [pri 0 ind 2771] off 16384
prev_pg ffffea00017a51d0 [pri 0 ind 2256] off 16384
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Signed-off-by: Patrick Farrell <paf@cray.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1669
Reviewed-on: http://review.whamcloud.com/6320
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6227
Reviewed-on: http://review.whamcloud.com/14385
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Reviewed-by: Hiroya Nozaki <nozaki.hiroya@jp.fujitsu.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/Makefile | 2 +-
drivers/staging/lustre/lustre/llite/file.c | 37 +++-
.../staging/lustre/lustre/llite/llite_internal.h | 5 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 2 +-
drivers/staging/lustre/lustre/llite/range_lock.c | 233 ++++++++++++++++++++
drivers/staging/lustre/lustre/llite/range_lock.h | 82 +++++++
6 files changed, 348 insertions(+), 13 deletions(-)
create mode 100644 drivers/staging/lustre/lustre/llite/range_lock.c
create mode 100644 drivers/staging/lustre/lustre/llite/range_lock.h
diff --git a/drivers/staging/lustre/lustre/llite/Makefile b/drivers/staging/lustre/lustre/llite/Makefile
index 2cbb1b8..1ac0940 100644
--- a/drivers/staging/lustre/lustre/llite/Makefile
+++ b/drivers/staging/lustre/lustre/llite/Makefile
@@ -1,6 +1,6 @@
obj-$(CONFIG_LUSTRE_FS) += lustre.o
lustre-y := dcache.o dir.o file.o llite_close.o llite_lib.o llite_nfs.o \
- rw.o namei.o symlink.o llite_mmap.o \
+ rw.o namei.o symlink.o llite_mmap.o range_lock.o \
xattr.o xattr_cache.o rw26.o super25.o statahead.o \
glimpse.o lcommon_cl.o lcommon_misc.o \
vvp_dev.o vvp_page.o vvp_lock.o vvp_io.o vvp_object.o vvp_req.o \
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 273b563..d9ee255 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1126,6 +1126,7 @@ ll_file_io_generic(const struct lu_env *env, struct vvp_io_args *args,
{
struct ll_inode_info *lli = ll_i2info(file_inode(file));
struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+ struct range_lock range;
struct cl_io *io;
ssize_t result;
@@ -1138,7 +1139,12 @@ restart:
if (cl_io_rw_init(env, io, iot, *ppos, count) == 0) {
struct vvp_io *vio = vvp_env_io(env);
- int write_mutex_locked = 0;
+ bool range_locked = false;
+
+ if (file->f_flags & O_APPEND)
+ range_lock_init(&range, 0, LUSTRE_EOF);
+ else
+ range_lock_init(&range, *ppos, *ppos + count - 1);
vio->vui_fd = LUSTRE_FPRIVATE(file);
vio->vui_io_subtype = args->via_io_subtype;
@@ -1147,14 +1153,23 @@ restart:
case IO_NORMAL:
vio->vui_iter = args->u.normal.via_iter;
vio->vui_iocb = args->u.normal.via_iocb;
- if ((iot == CIT_WRITE) &&
+ /*
+ * Direct IO reads must also take range lock,
+ * or multiple reads will try to work on the same pages
+ * See LU-6227 for details.
+ */
+ if (((iot == CIT_WRITE) ||
+ (iot == CIT_READ && (file->f_flags & O_DIRECT))) &&
!(vio->vui_fd->fd_flags & LL_FILE_GROUP_LOCKED)) {
- if (mutex_lock_interruptible(&lli->
- lli_write_mutex)) {
- result = -ERESTARTSYS;
+ CDEBUG(D_VFSTRACE, "Range lock [%llu, %llu]\n",
+ range.rl_node.in_extent.start,
+ range.rl_node.in_extent.end);
+ result = range_lock(&lli->lli_write_tree,
+ &range);
+ if (result < 0)
goto out;
- }
- write_mutex_locked = 1;
+
+ range_locked = true;
}
down_read(&lli->lli_trunc_sem);
break;
@@ -1171,8 +1186,12 @@ restart:
ll_cl_remove(file, env);
if (args->via_io_subtype == IO_NORMAL)
up_read(&lli->lli_trunc_sem);
- if (write_mutex_locked)
- mutex_unlock(&lli->lli_write_mutex);
+ if (range_locked) {
+ CDEBUG(D_VFSTRACE, "Range unlock [%llu, %llu]\n",
+ range.rl_node.in_extent.start,
+ range.rl_node.in_extent.end);
+ range_unlock(&lli->lli_write_tree, &range);
+ }
} else {
/* cl_io_rw_init() handled IO */
result = io->ci_result;
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index f903f2a..7f0a6c9 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -46,6 +46,7 @@
#include <linux/xattr.h>
#include <linux/posix_acl_xattr.h>
#include "vvp_internal.h"
+#include "range_lock.h"
#ifndef FMODE_EXEC
#define FMODE_EXEC 0
@@ -210,7 +211,7 @@ struct ll_inode_info {
* }
*/
struct rw_semaphore f_trunc_sem;
- struct mutex f_write_mutex;
+ struct range_lock_tree f_write_tree;
struct rw_semaphore f_glimpse_sem;
unsigned long f_glimpse_time;
@@ -235,7 +236,7 @@ struct ll_inode_info {
#define lli_symlink_name u.f.f_symlink_name
#define lli_maxbytes u.f.f_maxbytes
#define lli_trunc_sem u.f.f_trunc_sem
-#define lli_write_mutex u.f.f_write_mutex
+#define lli_write_tree u.f.f_write_tree
#define lli_glimpse_sem u.f.f_glimpse_sem
#define lli_glimpse_time u.f.f_glimpse_time
#define lli_agl_list u.f.f_agl_list
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 93fd69b..9112fa8 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -807,7 +807,7 @@ void ll_lli_init(struct ll_inode_info *lli)
mutex_init(&lli->lli_size_mutex);
lli->lli_symlink_name = NULL;
init_rwsem(&lli->lli_trunc_sem);
- mutex_init(&lli->lli_write_mutex);
+ range_lock_tree_init(&lli->lli_write_tree);
init_rwsem(&lli->lli_glimpse_sem);
lli->lli_glimpse_time = 0;
INIT_LIST_HEAD(&lli->lli_agl_list);
diff --git a/drivers/staging/lustre/lustre/llite/range_lock.c b/drivers/staging/lustre/lustre/llite/range_lock.c
new file mode 100644
index 0000000..94c818f
--- /dev/null
+++ b/drivers/staging/lustre/lustre/llite/range_lock.c
@@ -0,0 +1,233 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Range lock is used to allow multiple threads writing a single shared
+ * file given each thread is writing to a non-overlapping portion of the
+ * file.
+ *
+ * Refer to the possible upstream kernel version of range lock by
+ * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
+ *
+ * This file could later replaced by the upstream kernel version.
+ */
+/*
+ * Author: Prakash Surya <surya1@llnl.gov>
+ * Author: Bobi Jam <bobijam.xu@intel.com>
+ */
+#include "range_lock.h"
+#include "../include/lustre/lustre_user.h"
+
+/**
+ * Initialize a range lock tree
+ *
+ * \param tree [in] an empty range lock tree
+ *
+ * Pre: Caller should have allocated the range lock tree.
+ * Post: The range lock tree is ready to function.
+ */
+void range_lock_tree_init(struct range_lock_tree *tree)
+{
+ tree->rlt_root = NULL;
+ tree->rlt_sequence = 0;
+ spin_lock_init(&tree->rlt_lock);
+}
+
+/**
+ * Initialize a range lock node
+ *
+ * \param lock [in] an empty range lock node
+ * \param start [in] start of the covering region
+ * \param end [in] end of the covering region
+ *
+ * Pre: Caller should have allocated the range lock node.
+ * Post: The range lock node is meant to cover [start, end] region
+ */
+void range_lock_init(struct range_lock *lock, __u64 start, __u64 end)
+{
+ memset(&lock->rl_node, 0, sizeof(lock->rl_node));
+ if (end != LUSTRE_EOF)
+ end >>= PAGE_SHIFT;
+ interval_set(&lock->rl_node, start >> PAGE_SHIFT, end);
+ INIT_LIST_HEAD(&lock->rl_next_lock);
+ lock->rl_task = NULL;
+ lock->rl_lock_count = 0;
+ lock->rl_blocking_ranges = 0;
+ lock->rl_sequence = 0;
+}
+
+static inline struct range_lock *next_lock(struct range_lock *lock)
+{
+ return list_entry(lock->rl_next_lock.next, typeof(*lock), rl_next_lock);
+}
+
+/**
+ * Helper function of range_unlock()
+ *
+ * \param node [in] a range lock found overlapped during interval node
+ * search
+ * \param arg [in] the range lock to be tested
+ *
+ * \retval INTERVAL_ITER_CONT indicate to continue the search for next
+ * overlapping range node
+ * \retval INTERVAL_ITER_STOP indicate to stop the search
+ */
+static enum interval_iter range_unlock_cb(struct interval_node *node, void *arg)
+{
+ struct range_lock *lock = arg;
+ struct range_lock *overlap = node2rangelock(node);
+ struct range_lock *iter;
+
+ list_for_each_entry(iter, &overlap->rl_next_lock, rl_next_lock) {
+ if (iter->rl_sequence > lock->rl_sequence) {
+ --iter->rl_blocking_ranges;
+ LASSERT(iter->rl_blocking_ranges > 0);
+ }
+ }
+ if (overlap->rl_sequence > lock->rl_sequence) {
+ --overlap->rl_blocking_ranges;
+ if (overlap->rl_blocking_ranges == 0)
+ wake_up_process(overlap->rl_task);
+ }
+ return INTERVAL_ITER_CONT;
+}
+
+/**
+ * Unlock a range lock, wake up locks blocked by this lock.
+ *
+ * \param tree [in] range lock tree
+ * \param lock [in] range lock to be deleted
+ *
+ * If this lock has been granted, relase it; if not, just delete it from
+ * the tree or the same region lock list. Wake up those locks only blocked
+ * by this lock through range_unlock_cb().
+ */
+void range_unlock(struct range_lock_tree *tree, struct range_lock *lock)
+{
+ spin_lock(&tree->rlt_lock);
+ if (!list_empty(&lock->rl_next_lock)) {
+ struct range_lock *next;
+
+ if (interval_is_intree(&lock->rl_node)) { /* first lock */
+ /* Insert the next same range lock into the tree */
+ next = next_lock(lock);
+ next->rl_lock_count = lock->rl_lock_count - 1;
+ interval_erase(&lock->rl_node, &tree->rlt_root);
+ interval_insert(&next->rl_node, &tree->rlt_root);
+ } else {
+ /* find the first lock in tree */
+ list_for_each_entry(next, &lock->rl_next_lock,
+ rl_next_lock) {
+ if (!interval_is_intree(&next->rl_node))
+ continue;
+
+ LASSERT(next->rl_lock_count > 0);
+ next->rl_lock_count--;
+ break;
+ }
+ }
+ list_del_init(&lock->rl_next_lock);
+ } else {
+ LASSERT(interval_is_intree(&lock->rl_node));
+ interval_erase(&lock->rl_node, &tree->rlt_root);
+ }
+
+ interval_search(tree->rlt_root, &lock->rl_node.in_extent,
+ range_unlock_cb, lock);
+ spin_unlock(&tree->rlt_lock);
+}
+
+/**
+ * Helper function of range_lock()
+ *
+ * \param node [in] a range lock found overlapped during interval node
+ * search
+ * \param arg [in] the range lock to be tested
+ *
+ * \retval INTERVAL_ITER_CONT indicate to continue the search for next
+ * overlapping range node
+ * \retval INTERVAL_ITER_STOP indicate to stop the search
+ */
+static enum interval_iter range_lock_cb(struct interval_node *node, void *arg)
+{
+ struct range_lock *lock = (struct range_lock *)arg;
+ struct range_lock *overlap = node2rangelock(node);
+
+ lock->rl_blocking_ranges += overlap->rl_lock_count + 1;
+ return INTERVAL_ITER_CONT;
+}
+
+/**
+ * Lock a region
+ *
+ * \param tree [in] range lock tree
+ * \param lock [in] range lock node containing the region span
+ *
+ * \retval 0 get the range lock
+ * \retval <0 error code while not getting the range lock
+ *
+ * If there exists overlapping range lock, the new lock will wait and
+ * retry, if later it find that it is not the chosen one to wake up,
+ * it wait again.
+ */
+int range_lock(struct range_lock_tree *tree, struct range_lock *lock)
+{
+ struct interval_node *node;
+ int rc = 0;
+
+ spin_lock(&tree->rlt_lock);
+ /*
+ * We need to check for all conflicting intervals
+ * already in the tree.
+ */
+ interval_search(tree->rlt_root, &lock->rl_node.in_extent,
+ range_lock_cb, lock);
+ /*
+ * Insert to the tree if I am unique, otherwise I've been linked to
+ * the rl_next_lock of another lock which has the same range as mine
+ * in range_lock_cb().
+ */
+ node = interval_insert(&lock->rl_node, &tree->rlt_root);
+ if (node) {
+ struct range_lock *tmp = node2rangelock(node);
+
+ list_add_tail(&lock->rl_next_lock, &tmp->rl_next_lock);
+ tmp->rl_lock_count++;
+ }
+ lock->rl_sequence = ++tree->rlt_sequence;
+
+ while (lock->rl_blocking_ranges > 0) {
+ lock->rl_task = current;
+ __set_current_state(TASK_INTERRUPTIBLE);
+ spin_unlock(&tree->rlt_lock);
+ schedule();
+
+ if (signal_pending(current)) {
+ range_unlock(tree, lock);
+ rc = -EINTR;
+ goto out;
+ }
+ spin_lock(&tree->rlt_lock);
+ }
+ spin_unlock(&tree->rlt_lock);
+out:
+ return rc;
+}
diff --git a/drivers/staging/lustre/lustre/llite/range_lock.h b/drivers/staging/lustre/lustre/llite/range_lock.h
new file mode 100644
index 0000000..c6d04a6
--- /dev/null
+++ b/drivers/staging/lustre/lustre/llite/range_lock.h
@@ -0,0 +1,82 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Range lock is used to allow multiple threads writing a single shared
+ * file given each thread is writing to a non-overlapping portion of the
+ * file.
+ *
+ * Refer to the possible upstream kernel version of range lock by
+ * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
+ *
+ * This file could later replaced by the upstream kernel version.
+ */
+/*
+ * Author: Prakash Surya <surya1@llnl.gov>
+ * Author: Bobi Jam <bobijam.xu@intel.com>
+ */
+#ifndef _RANGE_LOCK_H
+#define _RANGE_LOCK_H
+
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../include/interval_tree.h"
+
+struct range_lock {
+ struct interval_node rl_node;
+ /**
+ * Process to enqueue this lock.
+ */
+ struct task_struct *rl_task;
+ /**
+ * List of locks with the same range.
+ */
+ struct list_head rl_next_lock;
+ /**
+ * Number of locks in the list rl_next_lock
+ */
+ unsigned int rl_lock_count;
+ /**
+ * Number of ranges which are blocking acquisition of the lock
+ */
+ unsigned int rl_blocking_ranges;
+ /**
+ * Sequence number of range lock. This number is used to get to know
+ * the order the locks are queued; this is required for range_cancel().
+ */
+ __u64 rl_sequence;
+};
+
+static inline struct range_lock *node2rangelock(const struct interval_node *n)
+{
+ return container_of(n, struct range_lock, rl_node);
+}
+
+struct range_lock_tree {
+ struct interval_node *rlt_root;
+ spinlock_t rlt_lock; /* protect range lock tree */
+ __u64 rlt_sequence;
+};
+
+void range_lock_tree_init(struct range_lock_tree *tree);
+void range_lock_init(struct range_lock *lock, __u64 start, __u64 end);
+int range_lock(struct range_lock_tree *tree, struct range_lock *lock);
+void range_unlock(struct range_lock_tree *tree, struct range_lock *lock);
+#endif
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* Re: [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
2016-09-18 20:37 ` [lustre-devel] " James Simmons
@ 2016-09-19 7:28 ` Greg Kroah-Hartman
-1 siblings, 0 replies; 263+ messages in thread
From: Greg Kroah-Hartman @ 2016-09-19 7:28 UTC (permalink / raw)
To: James Simmons
Cc: devel, Andreas Dilger, Oleg Drokin, Patrick Farrell,
Linux Kernel Mailing List, Jinshan Xiong, Prakash Surya,
Lustre Development List
On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
> + * Range lock is used to allow multiple threads writing a single shared
> + * file given each thread is writing to a non-overlapping portion of the
> + * file.
> + *
> + * Refer to the possible upstream kernel version of range lock by
> + * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
> + *
> + * This file could later replaced by the upstream kernel version.
It doesn't look like range_lock ever got accepted in the kernel tree,
any idea what happened to it? Having a per-filesystem lock type seems
odd to me...
thanks,
greg k-h
^ permalink raw reply [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
@ 2016-09-19 7:28 ` Greg Kroah-Hartman
0 siblings, 0 replies; 263+ messages in thread
From: Greg Kroah-Hartman @ 2016-09-19 7:28 UTC (permalink / raw)
To: James Simmons
Cc: devel, Andreas Dilger, Oleg Drokin, Patrick Farrell,
Linux Kernel Mailing List, Jinshan Xiong, Prakash Surya,
Lustre Development List
On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
> + * Range lock is used to allow multiple threads writing a single shared
> + * file given each thread is writing to a non-overlapping portion of the
> + * file.
> + *
> + * Refer to the possible upstream kernel version of range lock by
> + * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
> + *
> + * This file could later replaced by the upstream kernel version.
It doesn't look like range_lock ever got accepted in the kernel tree,
any idea what happened to it? Having a per-filesystem lock type seems
odd to me...
thanks,
greg k-h
^ permalink raw reply [flat|nested] 263+ messages in thread
* Re: [lustre-devel] [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
@ 2016-09-19 9:25 ` Dilger, Andreas
0 siblings, 0 replies; 263+ messages in thread
From: Dilger, Andreas @ 2016-09-19 9:25 UTC (permalink / raw)
To: Greg Kroah-Hartman, Jan Kara
Cc: James Simmons, devel, Linux Kernel Mailing List, Drokin, Oleg,
Lustre Development List, linux-fsdevel, linux-ext4
On Sep 19, 2016, at 09:28, Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
>
> On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
>> + * Range lock is used to allow multiple threads writing a single shared
>> + * file given each thread is writing to a non-overlapping portion of the
>> + * file.
>> + *
>> + * Refer to the possible upstream kernel version of range lock by
>> + * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
>> + *
>> + * This file could later replaced by the upstream kernel version.
>
> It doesn't look like range_lock ever got accepted in the kernel tree,
> any idea what happened to it? Having a per-filesystem lock type seems
> odd to me...
I've added Jan and linux-fsdevel to the CC list to see what interest
there is in the range locking implementaion. At the time we added this
to Lustre it appeared that this was moving nicely torward landing, but
it seems to have stalled.
I think the range locking implementation is fairly generic, and if there
are other users in the kernel it could easily be pulled out of the staging
dir into vfs/. I'm not against it going into vfs/ directly either, but
not sure whether that is acceptable if the only user is in staging.
Cheers, Andreas
^ permalink raw reply [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
@ 2016-09-19 9:25 ` Dilger, Andreas
0 siblings, 0 replies; 263+ messages in thread
From: Dilger, Andreas @ 2016-09-19 9:25 UTC (permalink / raw)
To: Greg Kroah-Hartman, Jan Kara
Cc: devel-gWbeCf7V1WCQmaza687I9mD2FQJk+8+b,
Linux Kernel Mailing List, Drokin, Oleg,
linux-fsdevel-u79uwXL29TY76Z2rM5mHXA,
linux-ext4-u79uwXL29TY76Z2rM5mHXA, Lustre Development List
On Sep 19, 2016, at 09:28, Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
>
> On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
>> + * Range lock is used to allow multiple threads writing a single shared
>> + * file given each thread is writing to a non-overlapping portion of the
>> + * file.
>> + *
>> + * Refer to the possible upstream kernel version of range lock by
>> + * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
>> + *
>> + * This file could later replaced by the upstream kernel version.
>
> It doesn't look like range_lock ever got accepted in the kernel tree,
> any idea what happened to it? Having a per-filesystem lock type seems
> odd to me...
I've added Jan and linux-fsdevel to the CC list to see what interest
there is in the range locking implementaion. At the time we added this
to Lustre it appeared that this was moving nicely torward landing, but
it seems to have stalled.
I think the range locking implementation is fairly generic, and if there
are other users in the kernel it could easily be pulled out of the staging
dir into vfs/. I'm not against it going into vfs/ directly either, but
not sure whether that is acceptable if the only user is in staging.
Cheers, Andreas
^ permalink raw reply [flat|nested] 263+ messages in thread
* Re: [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
@ 2016-09-19 9:25 ` Dilger, Andreas
0 siblings, 0 replies; 263+ messages in thread
From: Dilger, Andreas @ 2016-09-19 9:25 UTC (permalink / raw)
To: Greg Kroah-Hartman, Jan Kara
Cc: devel-gWbeCf7V1WCQmaza687I9mD2FQJk+8+b,
Linux Kernel Mailing List, Drokin, Oleg,
linux-fsdevel-u79uwXL29TY76Z2rM5mHXA,
linux-ext4-u79uwXL29TY76Z2rM5mHXA, Lustre Development List
On Sep 19, 2016, at 09:28, Greg Kroah-Hartman <gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org> wrote:
>
> On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
>> + * Range lock is used to allow multiple threads writing a single shared
>> + * file given each thread is writing to a non-overlapping portion of the
>> + * file.
>> + *
>> + * Refer to the possible upstream kernel version of range lock by
>> + * Jan Kara <jack-AlSwsSmVLrQ@public.gmane.org>: https://lkml.org/lkml/2013/1/31/480
>> + *
>> + * This file could later replaced by the upstream kernel version.
>
> It doesn't look like range_lock ever got accepted in the kernel tree,
> any idea what happened to it? Having a per-filesystem lock type seems
> odd to me...
I've added Jan and linux-fsdevel to the CC list to see what interest
there is in the range locking implementaion. At the time we added this
to Lustre it appeared that this was moving nicely torward landing, but
it seems to have stalled.
I think the range locking implementation is fairly generic, and if there
are other users in the kernel it could easily be pulled out of the staging
dir into vfs/. I'm not against it going into vfs/ directly either, but
not sure whether that is acceptable if the only user is in staging.
Cheers, Andreas
^ permalink raw reply [flat|nested] 263+ messages in thread
* Re: [lustre-devel] [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
2016-09-19 9:25 ` Dilger, Andreas
@ 2016-09-19 9:59 ` Jan Kara
-1 siblings, 0 replies; 263+ messages in thread
From: Jan Kara @ 2016-09-19 9:59 UTC (permalink / raw)
To: Dilger, Andreas
Cc: Greg Kroah-Hartman, Jan Kara, James Simmons, devel,
Linux Kernel Mailing List, Drokin, Oleg, Lustre Development List,
linux-fsdevel, linux-ext4, dbueso
On Mon 19-09-16 09:25:48, Dilger, Andreas wrote:
> On Sep 19, 2016, at 09:28, Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
> >
> > On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
> >> + * Range lock is used to allow multiple threads writing a single shared
> >> + * file given each thread is writing to a non-overlapping portion of the
> >> + * file.
> >> + *
> >> + * Refer to the possible upstream kernel version of range lock by
> >> + * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
> >> + *
> >> + * This file could later replaced by the upstream kernel version.
> >
> > It doesn't look like range_lock ever got accepted in the kernel tree,
> > any idea what happened to it? Having a per-filesystem lock type seems
> > odd to me...
>
> I've added Jan and linux-fsdevel to the CC list to see what interest
> there is in the range locking implementaion. At the time we added this
> to Lustre it appeared that this was moving nicely torward landing, but
> it seems to have stalled.
>
> I think the range locking implementation is fairly generic, and if there
> are other users in the kernel it could easily be pulled out of the staging
> dir into vfs/. I'm not against it going into vfs/ directly either, but
> not sure whether that is acceptable if the only user is in staging.
Yeah, so the problem with my range_lock implementation (and your one looks
fairly similar from a quick look) is that it is fairly heavyweight. That
may be OK for an inode_lock replacement but it needs a careful
benchmarking. Davidlohr was looking into making the implementation more
efficient (he wanted to use it for mmap_sem) but I think it got preempted
by other work.
Honza
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
^ permalink raw reply [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 030/124] staging: lustre: llite: Replace write mutex with range lock
@ 2016-09-19 9:59 ` Jan Kara
0 siblings, 0 replies; 263+ messages in thread
From: Jan Kara @ 2016-09-19 9:59 UTC (permalink / raw)
To: Dilger, Andreas
Cc: Greg Kroah-Hartman, Jan Kara, James Simmons, devel,
Linux Kernel Mailing List, Drokin, Oleg, Lustre Development List,
linux-fsdevel, linux-ext4, dbueso
On Mon 19-09-16 09:25:48, Dilger, Andreas wrote:
> On Sep 19, 2016, at 09:28, Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
> >
> > On Sun, Sep 18, 2016 at 04:37:29PM -0400, James Simmons wrote:
> >> + * Range lock is used to allow multiple threads writing a single shared
> >> + * file given each thread is writing to a non-overlapping portion of the
> >> + * file.
> >> + *
> >> + * Refer to the possible upstream kernel version of range lock by
> >> + * Jan Kara <jack@suse.cz>: https://lkml.org/lkml/2013/1/31/480
> >> + *
> >> + * This file could later replaced by the upstream kernel version.
> >
> > It doesn't look like range_lock ever got accepted in the kernel tree,
> > any idea what happened to it? Having a per-filesystem lock type seems
> > odd to me...
>
> I've added Jan and linux-fsdevel to the CC list to see what interest
> there is in the range locking implementaion. At the time we added this
> to Lustre it appeared that this was moving nicely torward landing, but
> it seems to have stalled.
>
> I think the range locking implementation is fairly generic, and if there
> are other users in the kernel it could easily be pulled out of the staging
> dir into vfs/. I'm not against it going into vfs/ directly either, but
> not sure whether that is acceptable if the only user is in staging.
Yeah, so the problem with my range_lock implementation (and your one looks
fairly similar from a quick look) is that it is fairly heavyweight. That
may be OK for an inode_lock replacement but it needs a careful
benchmarking. Davidlohr was looking into making the implementation more
efficient (he wanted to use it for mmap_sem) but I think it got preempted
by other work.
Honza
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
^ permalink raw reply [flat|nested] 263+ messages in thread
* [PATCH 031/124] staging: lustre: vvp: Use lockless __generic_file_aio_write
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Prakash Surya, James Simmons
From: Prakash Surya <surya1@llnl.gov>
Testing multi-threaded single shard file write performance has shown
the inode mutex to be a limiting factor when using the
generic_file_write_iter function. To work around this bottle neck, this
change replaces the locked version of that call with the lock less
version, specifically, __generic_file_write_iter.
In order to maintain posix consistency, Lustre must now employ it's
own locking mechanism in the higher layers. Currently writes are
protected using the lli_write_mutex in the ll_inode_info structure.
To protect against simultaneous write and truncate operations, since
we no longer take the inode mutex during writes, we must down the
lli_trunc_sem semaphore.
Unfortunately, this change by itself does not garner any performance
benefits. Using FIO on a single machine with 32 GB of RAM, write
performance tests were ran with and without this change applied; the
results are below:
+---------+-----------+---------+--------+--------+
| fio v2.0.13 | Write Bandwidth (KB/s) |
+---------+-----------+---------+--------+--------+
| # Tasks | GB / Task | Test 1 | Test 2 | Test 3 |
+---------+-----------+---------+--------+--------+
| 1 | 64 | 452446 | 454623 | 457653 |
| 2 | 32 | 850318 | 565373 | 602498 |
| 4 | 16 | 1058900 | 463546 | 529107 |
| 8 | 8 | 1026300 | 468190 | 576451 |
| 16 | 4 | 1065500 | 503160 | 462902 |
| 32 | 2 | 1068600 | 462228 | 466963 |
| 64 | 1 | 991830 | 556618 | 557863 |
+---------+-----------+---------+--------+--------+
* Test 1: Lustre client running 04ec54f. File per process write
workload. This test was used as a baseline for what we
_could_ achieve in the single shared file tests if the
bottle necks were removed.
* Test 2: Lustre client running 04ec54f. Single shared file
workload, each task writing to a unique region.
* Test 3: Lustre client running 04ec54f + this patch. Single shared
file workload, each task writing to a unique region.
In order to garner any real performance benefits out of a single
shared file workload, the lli_write_mutex needs to be broken up into a
range lock. That would allow write operations to unique regions of a
file to be executed concurrently. This work is left to be done in a
follow up patch.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1669
Reviewed-on: http://review.whamcloud.com/6672
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/rw26.c | 9 -------
drivers/staging/lustre/lustre/llite/vvp_io.c | 26 +++++++++++++++++++--
drivers/staging/lustre/lustre/llite/vvp_page.c | 16 -------------
drivers/staging/lustre/lustre/obdclass/cl_page.c | 6 -----
4 files changed, 23 insertions(+), 34 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index 2f8ef54..66dba04 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -375,13 +375,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
io = vvp_env_io(env)->vui_cl.cis_io;
LASSERT(io);
- /* 0. Need locking between buffered and direct access. and race with
- * size changing by concurrent truncates and writes.
- * 1. Need inode mutex to operate transient pages.
- */
- if (iov_iter_rw(iter) == READ)
- inode_lock(inode);
-
LASSERT(obj->vob_transient_pages == 0);
while (iov_iter_count(iter)) {
struct page **pages;
@@ -431,8 +424,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
}
out:
LASSERT(obj->vob_transient_pages == 0);
- if (iov_iter_rw(iter) == READ)
- inode_unlock(inode);
if (tot_bytes > 0) {
struct vvp_io *vio = vvp_env_io(env);
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index 09ccd1f..f6dacee 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -959,10 +959,30 @@ static int vvp_io_write_start(const struct lu_env *env,
CDEBUG(D_VFSTRACE, "write: [%lli, %lli)\n", pos, pos + (long long)cnt);
- if (!vio->vui_iter) /* from a temp io in ll_cl_init(). */
+ if (!vio->vui_iter) {
+ /* from a temp io in ll_cl_init(). */
result = 0;
- else
- result = generic_file_write_iter(vio->vui_iocb, vio->vui_iter);
+ } else {
+ /*
+ * When using the locked AIO function (generic_file_aio_write())
+ * testing has shown the inode mutex to be a limiting factor
+ * with multi-threaded single shared file performance. To get
+ * around this, we now use the lockless version. To maintain
+ * consistency, proper locking to protect against writes,
+ * trucates, etc. is handled in the higher layers of lustre.
+ */
+ bool lock_node = !IS_NOSEC(inode);
+
+ if (lock_node)
+ inode_lock(inode);
+ result = __generic_file_write_iter(vio->vui_iocb,
+ vio->vui_iter);
+ if (lock_node)
+ inode_unlock(inode);
+
+ if (result > 0 || result == -EIOCBQUEUED)
+ result = generic_write_sync(vio->vui_iocb, result);
+ }
if (result > 0) {
result = vvp_io_write_commit(env, io);
diff --git a/drivers/staging/lustre/lustre/llite/vvp_page.c b/drivers/staging/lustre/lustre/llite/vvp_page.c
index 2818a68..bbbb0f1 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_page.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_page.c
@@ -444,18 +444,10 @@ static int vvp_transient_page_prep(const struct lu_env *env,
return 0;
}
-static void vvp_transient_page_verify(const struct cl_page *page)
-{
- struct inode *inode = vvp_object_inode(page->cp_obj);
-
- LASSERT(!inode_trylock(inode));
-}
-
static int vvp_transient_page_own(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused, int nonblock)
{
- vvp_transient_page_verify(slice->cpl_page);
return 0;
}
@@ -463,21 +455,18 @@ static void vvp_transient_page_assume(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_unassume(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_disown(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_discard(const struct lu_env *env,
@@ -486,8 +475,6 @@ static void vvp_transient_page_discard(const struct lu_env *env,
{
struct cl_page *page = slice->cpl_page;
- vvp_transient_page_verify(slice->cpl_page);
-
/*
* For transient pages, remove it from the radix tree.
*/
@@ -511,7 +498,6 @@ vvp_transient_page_completion(const struct lu_env *env,
const struct cl_page_slice *slice,
int ioret)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_fini(const struct lu_env *env,
@@ -522,7 +508,6 @@ static void vvp_transient_page_fini(const struct lu_env *env,
struct vvp_object *clobj = cl2vvp(clp->cp_obj);
vvp_page_fini_common(vpg);
- LASSERT(!inode_trylock(clobj->vob_inode));
clobj->vob_transient_pages--;
}
@@ -570,7 +555,6 @@ int vvp_page_init(const struct lu_env *env, struct cl_object *obj,
} else {
struct vvp_object *clobj = cl2vvp(obj);
- LASSERT(!inode_trylock(clobj->vob_inode));
cl_page_slice_add(page, &vpg->vpg_cl, obj, index,
&vvp_transient_page_ops);
clobj->vob_transient_pages++;
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index 80c6e0e..cae8d21 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -228,11 +228,6 @@ EXPORT_SYMBOL(cl_page_find);
static inline int cl_page_invariant(const struct cl_page *pg)
{
- /*
- * Page invariant is protected by a VM lock.
- */
- LINVRNT(cl_page_is_vmlocked(NULL, pg));
-
return cl_page_in_use_noref(pg);
}
@@ -864,7 +859,6 @@ void cl_page_completion(const struct lu_env *env,
(const struct lu_env *,
const struct cl_page_slice *, int), ioret);
if (anchor) {
- LASSERT(cl_page_is_vmlocked(env, pg));
LASSERT(pg->cp_sync_io == anchor);
pg->cp_sync_io = NULL;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 031/124] staging: lustre: vvp: Use lockless __generic_file_aio_write
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Prakash Surya, James Simmons
From: Prakash Surya <surya1@llnl.gov>
Testing multi-threaded single shard file write performance has shown
the inode mutex to be a limiting factor when using the
generic_file_write_iter function. To work around this bottle neck, this
change replaces the locked version of that call with the lock less
version, specifically, __generic_file_write_iter.
In order to maintain posix consistency, Lustre must now employ it's
own locking mechanism in the higher layers. Currently writes are
protected using the lli_write_mutex in the ll_inode_info structure.
To protect against simultaneous write and truncate operations, since
we no longer take the inode mutex during writes, we must down the
lli_trunc_sem semaphore.
Unfortunately, this change by itself does not garner any performance
benefits. Using FIO on a single machine with 32 GB of RAM, write
performance tests were ran with and without this change applied; the
results are below:
+---------+-----------+---------+--------+--------+
| fio v2.0.13 | Write Bandwidth (KB/s) |
+---------+-----------+---------+--------+--------+
| # Tasks | GB / Task | Test 1 | Test 2 | Test 3 |
+---------+-----------+---------+--------+--------+
| 1 | 64 | 452446 | 454623 | 457653 |
| 2 | 32 | 850318 | 565373 | 602498 |
| 4 | 16 | 1058900 | 463546 | 529107 |
| 8 | 8 | 1026300 | 468190 | 576451 |
| 16 | 4 | 1065500 | 503160 | 462902 |
| 32 | 2 | 1068600 | 462228 | 466963 |
| 64 | 1 | 991830 | 556618 | 557863 |
+---------+-----------+---------+--------+--------+
* Test 1: Lustre client running 04ec54f. File per process write
workload. This test was used as a baseline for what we
_could_ achieve in the single shared file tests if the
bottle necks were removed.
* Test 2: Lustre client running 04ec54f. Single shared file
workload, each task writing to a unique region.
* Test 3: Lustre client running 04ec54f + this patch. Single shared
file workload, each task writing to a unique region.
In order to garner any real performance benefits out of a single
shared file workload, the lli_write_mutex needs to be broken up into a
range lock. That would allow write operations to unique regions of a
file to be executed concurrently. This work is left to be done in a
follow up patch.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1669
Reviewed-on: http://review.whamcloud.com/6672
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/rw26.c | 9 -------
drivers/staging/lustre/lustre/llite/vvp_io.c | 26 +++++++++++++++++++--
drivers/staging/lustre/lustre/llite/vvp_page.c | 16 -------------
drivers/staging/lustre/lustre/obdclass/cl_page.c | 6 -----
4 files changed, 23 insertions(+), 34 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index 2f8ef54..66dba04 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -375,13 +375,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
io = vvp_env_io(env)->vui_cl.cis_io;
LASSERT(io);
- /* 0. Need locking between buffered and direct access. and race with
- * size changing by concurrent truncates and writes.
- * 1. Need inode mutex to operate transient pages.
- */
- if (iov_iter_rw(iter) == READ)
- inode_lock(inode);
-
LASSERT(obj->vob_transient_pages == 0);
while (iov_iter_count(iter)) {
struct page **pages;
@@ -431,8 +424,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
}
out:
LASSERT(obj->vob_transient_pages == 0);
- if (iov_iter_rw(iter) == READ)
- inode_unlock(inode);
if (tot_bytes > 0) {
struct vvp_io *vio = vvp_env_io(env);
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index 09ccd1f..f6dacee 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -959,10 +959,30 @@ static int vvp_io_write_start(const struct lu_env *env,
CDEBUG(D_VFSTRACE, "write: [%lli, %lli)\n", pos, pos + (long long)cnt);
- if (!vio->vui_iter) /* from a temp io in ll_cl_init(). */
+ if (!vio->vui_iter) {
+ /* from a temp io in ll_cl_init(). */
result = 0;
- else
- result = generic_file_write_iter(vio->vui_iocb, vio->vui_iter);
+ } else {
+ /*
+ * When using the locked AIO function (generic_file_aio_write())
+ * testing has shown the inode mutex to be a limiting factor
+ * with multi-threaded single shared file performance. To get
+ * around this, we now use the lockless version. To maintain
+ * consistency, proper locking to protect against writes,
+ * trucates, etc. is handled in the higher layers of lustre.
+ */
+ bool lock_node = !IS_NOSEC(inode);
+
+ if (lock_node)
+ inode_lock(inode);
+ result = __generic_file_write_iter(vio->vui_iocb,
+ vio->vui_iter);
+ if (lock_node)
+ inode_unlock(inode);
+
+ if (result > 0 || result == -EIOCBQUEUED)
+ result = generic_write_sync(vio->vui_iocb, result);
+ }
if (result > 0) {
result = vvp_io_write_commit(env, io);
diff --git a/drivers/staging/lustre/lustre/llite/vvp_page.c b/drivers/staging/lustre/lustre/llite/vvp_page.c
index 2818a68..bbbb0f1 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_page.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_page.c
@@ -444,18 +444,10 @@ static int vvp_transient_page_prep(const struct lu_env *env,
return 0;
}
-static void vvp_transient_page_verify(const struct cl_page *page)
-{
- struct inode *inode = vvp_object_inode(page->cp_obj);
-
- LASSERT(!inode_trylock(inode));
-}
-
static int vvp_transient_page_own(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused, int nonblock)
{
- vvp_transient_page_verify(slice->cpl_page);
return 0;
}
@@ -463,21 +455,18 @@ static void vvp_transient_page_assume(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_unassume(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_disown(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_discard(const struct lu_env *env,
@@ -486,8 +475,6 @@ static void vvp_transient_page_discard(const struct lu_env *env,
{
struct cl_page *page = slice->cpl_page;
- vvp_transient_page_verify(slice->cpl_page);
-
/*
* For transient pages, remove it from the radix tree.
*/
@@ -511,7 +498,6 @@ vvp_transient_page_completion(const struct lu_env *env,
const struct cl_page_slice *slice,
int ioret)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_fini(const struct lu_env *env,
@@ -522,7 +508,6 @@ static void vvp_transient_page_fini(const struct lu_env *env,
struct vvp_object *clobj = cl2vvp(clp->cp_obj);
vvp_page_fini_common(vpg);
- LASSERT(!inode_trylock(clobj->vob_inode));
clobj->vob_transient_pages--;
}
@@ -570,7 +555,6 @@ int vvp_page_init(const struct lu_env *env, struct cl_object *obj,
} else {
struct vvp_object *clobj = cl2vvp(obj);
- LASSERT(!inode_trylock(clobj->vob_inode));
cl_page_slice_add(page, &vpg->vpg_cl, obj, index,
&vvp_transient_page_ops);
clobj->vob_transient_pages++;
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index 80c6e0e..cae8d21 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -228,11 +228,6 @@ EXPORT_SYMBOL(cl_page_find);
static inline int cl_page_invariant(const struct cl_page *pg)
{
- /*
- * Page invariant is protected by a VM lock.
- */
- LINVRNT(cl_page_is_vmlocked(NULL, pg));
-
return cl_page_in_use_noref(pg);
}
@@ -864,7 +859,6 @@ void cl_page_completion(const struct lu_env *env,
(const struct lu_env *,
const struct cl_page_slice *, int), ioret);
if (anchor) {
- LASSERT(cl_page_is_vmlocked(env, pg));
LASSERT(pg->cp_sync_io == anchor);
pg->cp_sync_io = NULL;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 032/124] staging: lustre: llite: remove lookup_flags from ll_lookup_it()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove the effectively unused lookup_flags parameter from
ll_lookup_it().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 14 +++++---------
1 files changed, 5 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index b7d448f..2037c7c 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -502,7 +502,7 @@ out:
}
static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
- struct lookup_intent *it, int lookup_flags)
+ struct lookup_intent *it)
{
struct lookup_intent lookup_it = { .it_op = IT_LOOKUP };
struct dentry *save = dentry, *retval;
@@ -541,8 +541,7 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
opc = LUSTRE_OPC_ANY;
op_data = ll_prep_md_op_data(NULL, parent, NULL, dentry->d_name.name,
- dentry->d_name.len, lookup_flags, opc,
- NULL);
+ dentry->d_name.len, 0, opc, NULL);
if (IS_ERR(op_data))
return (void *)op_data;
@@ -606,7 +605,7 @@ static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry,
itp = NULL;
else
itp = ⁢
- de = ll_lookup_it(parent, dentry, itp, 0);
+ de = ll_lookup_it(parent, dentry, itp);
if (itp)
ll_intent_release(itp);
@@ -624,7 +623,6 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry,
{
struct lookup_intent *it;
struct dentry *de;
- long long lookup_flags = LOOKUP_OPEN;
int rc = 0;
CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p),file %p,open_flags %x,mode %x opened %d\n",
@@ -654,16 +652,14 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry,
return -ENOMEM;
it->it_op = IT_OPEN;
- if (open_flags & O_CREAT) {
+ if (open_flags & O_CREAT)
it->it_op |= IT_CREAT;
- lookup_flags |= LOOKUP_CREATE;
- }
it->it_create_mode = (mode & S_IALLUGO) | S_IFREG;
it->it_flags = (open_flags & ~O_ACCMODE) | OPEN_FMODE(open_flags);
it->it_flags &= ~MDS_OPEN_FL_INTERNAL;
/* Dentry added to dcache tree in ll_lookup_it */
- de = ll_lookup_it(dir, dentry, it, lookup_flags);
+ de = ll_lookup_it(dir, dentry, it);
if (IS_ERR(de))
rc = PTR_ERR(de);
else if (de)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 032/124] staging: lustre: llite: remove lookup_flags from ll_lookup_it()
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove the effectively unused lookup_flags parameter from
ll_lookup_it().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 14 +++++---------
1 files changed, 5 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index b7d448f..2037c7c 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -502,7 +502,7 @@ out:
}
static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
- struct lookup_intent *it, int lookup_flags)
+ struct lookup_intent *it)
{
struct lookup_intent lookup_it = { .it_op = IT_LOOKUP };
struct dentry *save = dentry, *retval;
@@ -541,8 +541,7 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
opc = LUSTRE_OPC_ANY;
op_data = ll_prep_md_op_data(NULL, parent, NULL, dentry->d_name.name,
- dentry->d_name.len, lookup_flags, opc,
- NULL);
+ dentry->d_name.len, 0, opc, NULL);
if (IS_ERR(op_data))
return (void *)op_data;
@@ -606,7 +605,7 @@ static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry,
itp = NULL;
else
itp = ⁢
- de = ll_lookup_it(parent, dentry, itp, 0);
+ de = ll_lookup_it(parent, dentry, itp);
if (itp)
ll_intent_release(itp);
@@ -624,7 +623,6 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry,
{
struct lookup_intent *it;
struct dentry *de;
- long long lookup_flags = LOOKUP_OPEN;
int rc = 0;
CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p),file %p,open_flags %x,mode %x opened %d\n",
@@ -654,16 +652,14 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry,
return -ENOMEM;
it->it_op = IT_OPEN;
- if (open_flags & O_CREAT) {
+ if (open_flags & O_CREAT)
it->it_op |= IT_CREAT;
- lookup_flags |= LOOKUP_CREATE;
- }
it->it_create_mode = (mode & S_IALLUGO) | S_IFREG;
it->it_flags = (open_flags & ~O_ACCMODE) | OPEN_FMODE(open_flags);
it->it_flags &= ~MDS_OPEN_FL_INTERNAL;
/* Dentry added to dcache tree in ll_lookup_it */
- de = ll_lookup_it(dir, dentry, it, lookup_flags);
+ de = ll_lookup_it(dir, dentry, it);
if (IS_ERR(de))
rc = PTR_ERR(de);
else if (de)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 033/124] staging: lustre: llite: remove mode from ll_create_it()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove the unused mode parameter from ll_create_it().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 2037c7c..8f92af4 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -47,8 +47,8 @@
#include "../include/lustre_ver.h"
#include "llite_internal.h"
-static int ll_create_it(struct inode *, struct dentry *,
- int, struct lookup_intent *);
+static int ll_create_it(struct inode *dir, struct dentry *dentry,
+ struct lookup_intent *it);
/* called from iget5_locked->find_inode() under inode_hash_lock spinlock */
static int ll_test_inode(struct inode *inode, void *opaque)
@@ -668,7 +668,7 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry,
if (!rc) {
if (it_disposition(it, DISP_OPEN_CREATE)) {
/* Dentry instantiated in ll_create_it. */
- rc = ll_create_it(dir, dentry, mode, it);
+ rc = ll_create_it(dir, dentry, it);
if (rc) {
/* We dget in ll_splice_alias. */
if (de)
@@ -754,7 +754,7 @@ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it)
* If the create succeeds, we fill in the inode information
* with d_instantiate().
*/
-static int ll_create_it(struct inode *dir, struct dentry *dentry, int mode,
+static int ll_create_it(struct inode *dir, struct dentry *dentry,
struct lookup_intent *it)
{
struct inode *inode;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 033/124] staging: lustre: llite: remove mode from ll_create_it()
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove the unused mode parameter from ll_create_it().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 2037c7c..8f92af4 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -47,8 +47,8 @@
#include "../include/lustre_ver.h"
#include "llite_internal.h"
-static int ll_create_it(struct inode *, struct dentry *,
- int, struct lookup_intent *);
+static int ll_create_it(struct inode *dir, struct dentry *dentry,
+ struct lookup_intent *it);
/* called from iget5_locked->find_inode() under inode_hash_lock spinlock */
static int ll_test_inode(struct inode *inode, void *opaque)
@@ -668,7 +668,7 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry,
if (!rc) {
if (it_disposition(it, DISP_OPEN_CREATE)) {
/* Dentry instantiated in ll_create_it. */
- rc = ll_create_it(dir, dentry, mode, it);
+ rc = ll_create_it(dir, dentry, it);
if (rc) {
/* We dget in ll_splice_alias. */
if (de)
@@ -754,7 +754,7 @@ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it)
* If the create succeeds, we fill in the inode information
* with d_instantiate().
*/
-static int ll_create_it(struct inode *dir, struct dentry *dentry, int mode,
+static int ll_create_it(struct inode *dir, struct dentry *dentry,
struct lookup_intent *it)
{
struct inode *inode;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 034/124] staging: lustre: llite: turn mode to umode_t for ll_new_inode()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Change int mode to umode_t.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 8f92af4..fadac44 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -794,7 +794,7 @@ void ll_update_times(struct ptlrpc_request *request, struct inode *inode)
}
static int ll_new_node(struct inode *dir, struct dentry *dentry,
- const char *tgt, int mode, int rdev,
+ const char *tgt, umode_t mode, int rdev,
__u32 opc)
{
struct ptlrpc_request *request = NULL;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 034/124] staging: lustre: llite: turn mode to umode_t for ll_new_inode()
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Change int mode to umode_t.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 8f92af4..fadac44 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -794,7 +794,7 @@ void ll_update_times(struct ptlrpc_request *request, struct inode *inode)
}
static int ll_new_node(struct inode *dir, struct dentry *dentry,
- const char *tgt, int mode, int rdev,
+ const char *tgt, umode_t mode, int rdev,
__u32 opc)
{
struct ptlrpc_request *request = NULL;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 035/124] staging: lustre: llite: style cleanup for ll_mkdir
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Style cleanup to make the code readable.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index fadac44..5e182fc 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -1015,8 +1015,8 @@ static int ll_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
if (!IS_POSIXACL(dir) || !exp_connect_umask(ll_i2mdexp(dir)))
mode &= ~current_umask();
mode = (mode & (S_IRWXUGO | S_ISVTX)) | S_IFDIR;
- err = ll_new_node(dir, dentry, NULL, mode, 0, LUSTRE_OPC_MKDIR);
+ err = ll_new_node(dir, dentry, NULL, mode, 0, LUSTRE_OPC_MKDIR);
if (!err)
ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_MKDIR, 1);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 035/124] staging: lustre: llite: style cleanup for ll_mkdir
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Style cleanup to make the code readable.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index fadac44..5e182fc 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -1015,8 +1015,8 @@ static int ll_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
if (!IS_POSIXACL(dir) || !exp_connect_umask(ll_i2mdexp(dir)))
mode &= ~current_umask();
mode = (mode & (S_IRWXUGO | S_ISVTX)) | S_IFDIR;
- err = ll_new_node(dir, dentry, NULL, mode, 0, LUSTRE_OPC_MKDIR);
+ err = ll_new_node(dir, dentry, NULL, mode, 0, LUSTRE_OPC_MKDIR);
if (!err)
ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_MKDIR, 1);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 036/124] staging: lustre: llite: no need to check dentry is NULL
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
We are already touching dentry in CDEBUG macros so it
will crash long before these checks. Since this is the
case no need to do an additional check.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 5e182fc..7a82ae4 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -987,7 +987,7 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild)
if (IS_ERR(op_data))
return PTR_ERR(op_data);
- if (dchild && dchild->d_inode)
+ if (dchild->d_inode)
op_data->op_fid3 = *ll_inode2fid(dchild->d_inode);
op_data->op_fid2 = op_data->op_fid3;
@@ -1039,7 +1039,7 @@ static int ll_rmdir(struct inode *dir, struct dentry *dchild)
if (IS_ERR(op_data))
return PTR_ERR(op_data);
- if (dchild && dchild->d_inode)
+ if (dchild->d_inode)
op_data->op_fid3 = *ll_inode2fid(dchild->d_inode);
op_data->op_fid2 = op_data->op_fid3;
@@ -1120,9 +1120,9 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild,
if (IS_ERR(op_data))
return PTR_ERR(op_data);
- if (src_dchild && src_dchild->d_inode)
+ if (src_dchild->d_inode)
op_data->op_fid3 = *ll_inode2fid(src_dchild->d_inode);
- if (tgt_dchild && tgt_dchild->d_inode)
+ if (tgt_dchild->d_inode)
op_data->op_fid4 = *ll_inode2fid(tgt_dchild->d_inode);
err = md_rename(sbi->ll_md_exp, op_data,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 036/124] staging: lustre: llite: no need to check dentry is NULL
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
We are already touching dentry in CDEBUG macros so it
will crash long before these checks. Since this is the
case no need to do an additional check.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/10769
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 5e182fc..7a82ae4 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -987,7 +987,7 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild)
if (IS_ERR(op_data))
return PTR_ERR(op_data);
- if (dchild && dchild->d_inode)
+ if (dchild->d_inode)
op_data->op_fid3 = *ll_inode2fid(dchild->d_inode);
op_data->op_fid2 = op_data->op_fid3;
@@ -1039,7 +1039,7 @@ static int ll_rmdir(struct inode *dir, struct dentry *dchild)
if (IS_ERR(op_data))
return PTR_ERR(op_data);
- if (dchild && dchild->d_inode)
+ if (dchild->d_inode)
op_data->op_fid3 = *ll_inode2fid(dchild->d_inode);
op_data->op_fid2 = op_data->op_fid3;
@@ -1120,9 +1120,9 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild,
if (IS_ERR(op_data))
return PTR_ERR(op_data);
- if (src_dchild && src_dchild->d_inode)
+ if (src_dchild->d_inode)
op_data->op_fid3 = *ll_inode2fid(src_dchild->d_inode);
- if (tgt_dchild && tgt_dchild->d_inode)
+ if (tgt_dchild->d_inode)
op_data->op_fid4 = *ll_inode2fid(tgt_dchild->d_inode);
err = md_rename(sbi->ll_md_exp, op_data,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 037/124] staging: lustre: cleanup lustre_lib.h
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove some unused declarations from lustre_lib.h and move
some others to more natural headers.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11500
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_lib.h | 36 --------------------
drivers/staging/lustre/lustre/include/obd.h | 17 +++++++++
drivers/staging/lustre/lustre/include/obd_class.h | 5 +++
3 files changed, 22 insertions(+), 36 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_lib.h b/drivers/staging/lustre/lustre/include/lustre_lib.h
index adb8c47..6b23191 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lib.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lib.h
@@ -51,7 +51,6 @@
#include "lustre_cfg.h"
/* target.c */
-struct kstatfs;
struct ptlrpc_request;
struct obd_export;
struct lu_target;
@@ -74,43 +73,8 @@ int do_set_info_async(struct obd_import *imp,
u32 vallen, void *val,
struct ptlrpc_request_set *set);
-#define OBD_RECOVERY_MAX_TIME (obd_timeout * 18) /* b13079 */
-
void target_send_reply(struct ptlrpc_request *req, int rc, int fail_id);
-/* client.c */
-
-int client_sanobd_setup(struct obd_device *obddev, struct lustre_cfg *lcfg);
-struct client_obd *client_conn2cli(struct lustre_handle *conn);
-
-struct md_open_data;
-struct obd_client_handle {
- struct lustre_handle och_fh;
- struct lu_fid och_fid;
- struct md_open_data *och_mod;
- struct lustre_handle och_lease_handle; /* open lock for lease */
- __u32 och_magic;
- fmode_t och_flags;
-};
-
-#define OBD_CLIENT_HANDLE_MAGIC 0xd15ea5ed
-
-/* statfs_pack.c */
-void statfs_unpack(struct kstatfs *sfs, struct obd_statfs *osfs);
-
-/* Until such time as we get_info the per-stripe maximum from the OST,
- * we define this to be 2T - 4k, which is the ext3 maxbytes.
- */
-#define LUSTRE_STRIPE_MAXBYTES 0x1fffffff000ULL
-
-/* Special values for remove LOV EA from disk */
-#define LOVEA_DELETE_VALUES(size, count, offset) (size == 0 && count == 0 && \
- offset == (typeof(offset))(-1))
-
-#define LMVEA_DELETE_VALUES(count, offset) ((count) == 0 && \
- (offset) == (typeof(offset))(-1))
-/* #define POISON_BULK 0 */
-
/*
* l_wait_event is a flexible sleeping function, permitting simple caller
* configuration of interrupt and timeout sensitivity along with actions to
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 89633f7..0917aaa 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -41,6 +41,7 @@
#include "lustre_export.h"
#include "lustre_fid.h"
#include "lustre_fld.h"
+#include "lustre_handles.h"
#include "lustre_intent.h"
#define MAX_OBD_DEVICES 8192
@@ -72,6 +73,11 @@ static inline void loi_init(struct lov_oinfo *loi)
{
}
+/* Until such time as we get_info the per-stripe maximum from the OST,
+ * we define this to be 2T - 4k, which is the ext3 maxbytes.
+ */
+#define LUSTRE_STRIPE_MAXBYTES 0x1fffffff000ULL
+
struct lov_stripe_md {
atomic_t lsm_refc;
spinlock_t lsm_lock;
@@ -949,6 +955,17 @@ struct md_open_data {
bool mod_is_create;
};
+struct obd_client_handle {
+ struct lustre_handle och_fh;
+ struct lu_fid och_fid;
+ struct md_open_data *och_mod;
+ struct lustre_handle och_lease_handle; /* open lock for lease */
+ __u32 och_magic;
+ int och_flags;
+};
+
+#define OBD_CLIENT_HANDLE_MAGIC 0xd15ea5ed
+
struct lookup_intent;
struct cl_attr;
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 8e8df08..6fc54bd 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1725,6 +1725,11 @@ int class_procfs_clean(void);
/* prng.c */
#define ll_generate_random_uuid(uuid_out) cfs_get_random_bytes(uuid_out, sizeof(class_uuid_t))
+/* statfs_pack.c */
+struct kstatfs;
+void statfs_pack(struct obd_statfs *osfs, struct kstatfs *sfs);
+void statfs_unpack(struct kstatfs *sfs, struct obd_statfs *osfs);
+
/* root squash info */
struct rw_semaphore;
struct root_squash_info {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 037/124] staging: lustre: cleanup lustre_lib.h
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove some unused declarations from lustre_lib.h and move
some others to more natural headers.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11500
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_lib.h | 36 --------------------
drivers/staging/lustre/lustre/include/obd.h | 17 +++++++++
drivers/staging/lustre/lustre/include/obd_class.h | 5 +++
3 files changed, 22 insertions(+), 36 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_lib.h b/drivers/staging/lustre/lustre/include/lustre_lib.h
index adb8c47..6b23191 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lib.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lib.h
@@ -51,7 +51,6 @@
#include "lustre_cfg.h"
/* target.c */
-struct kstatfs;
struct ptlrpc_request;
struct obd_export;
struct lu_target;
@@ -74,43 +73,8 @@ int do_set_info_async(struct obd_import *imp,
u32 vallen, void *val,
struct ptlrpc_request_set *set);
-#define OBD_RECOVERY_MAX_TIME (obd_timeout * 18) /* b13079 */
-
void target_send_reply(struct ptlrpc_request *req, int rc, int fail_id);
-/* client.c */
-
-int client_sanobd_setup(struct obd_device *obddev, struct lustre_cfg *lcfg);
-struct client_obd *client_conn2cli(struct lustre_handle *conn);
-
-struct md_open_data;
-struct obd_client_handle {
- struct lustre_handle och_fh;
- struct lu_fid och_fid;
- struct md_open_data *och_mod;
- struct lustre_handle och_lease_handle; /* open lock for lease */
- __u32 och_magic;
- fmode_t och_flags;
-};
-
-#define OBD_CLIENT_HANDLE_MAGIC 0xd15ea5ed
-
-/* statfs_pack.c */
-void statfs_unpack(struct kstatfs *sfs, struct obd_statfs *osfs);
-
-/* Until such time as we get_info the per-stripe maximum from the OST,
- * we define this to be 2T - 4k, which is the ext3 maxbytes.
- */
-#define LUSTRE_STRIPE_MAXBYTES 0x1fffffff000ULL
-
-/* Special values for remove LOV EA from disk */
-#define LOVEA_DELETE_VALUES(size, count, offset) (size == 0 && count == 0 && \
- offset == (typeof(offset))(-1))
-
-#define LMVEA_DELETE_VALUES(count, offset) ((count) == 0 && \
- (offset) == (typeof(offset))(-1))
-/* #define POISON_BULK 0 */
-
/*
* l_wait_event is a flexible sleeping function, permitting simple caller
* configuration of interrupt and timeout sensitivity along with actions to
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 89633f7..0917aaa 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -41,6 +41,7 @@
#include "lustre_export.h"
#include "lustre_fid.h"
#include "lustre_fld.h"
+#include "lustre_handles.h"
#include "lustre_intent.h"
#define MAX_OBD_DEVICES 8192
@@ -72,6 +73,11 @@ static inline void loi_init(struct lov_oinfo *loi)
{
}
+/* Until such time as we get_info the per-stripe maximum from the OST,
+ * we define this to be 2T - 4k, which is the ext3 maxbytes.
+ */
+#define LUSTRE_STRIPE_MAXBYTES 0x1fffffff000ULL
+
struct lov_stripe_md {
atomic_t lsm_refc;
spinlock_t lsm_lock;
@@ -949,6 +955,17 @@ struct md_open_data {
bool mod_is_create;
};
+struct obd_client_handle {
+ struct lustre_handle och_fh;
+ struct lu_fid och_fid;
+ struct md_open_data *och_mod;
+ struct lustre_handle och_lease_handle; /* open lock for lease */
+ __u32 och_magic;
+ int och_flags;
+};
+
+#define OBD_CLIENT_HANDLE_MAGIC 0xd15ea5ed
+
struct lookup_intent;
struct cl_attr;
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 8e8df08..6fc54bd 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1725,6 +1725,11 @@ int class_procfs_clean(void);
/* prng.c */
#define ll_generate_random_uuid(uuid_out) cfs_get_random_bytes(uuid_out, sizeof(class_uuid_t))
+/* statfs_pack.c */
+struct kstatfs;
+void statfs_pack(struct obd_statfs *osfs, struct kstatfs *sfs);
+void statfs_unpack(struct kstatfs *sfs, struct obd_statfs *osfs);
+
/* root squash info */
struct rw_semaphore;
struct root_squash_info {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 038/124] staging: lustre: osc: debug to match extent to brw RPC
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Patrick Farrell, James Simmons
From: Patrick Farrell <paf@cray.com>
Currently, it's difficult to match brw RPCs to objects and
extents from client logs. This patch adds a D_RPCTRACE
debug message giving the necessary information.
Signed-off-by: Patrick Farrell <paf@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5531
Reviewed-on: http://review.whamcloud.com/11548
Reviewed-by: Alexey Lyashkov <alexey.lyashkov@seagate.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Reviewed-by: Ryan Haasken <haasken@cray.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_request.c | 5 +++++
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 5cca2b6..d5d2963 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1419,6 +1419,11 @@ static int osc_brw_prep_request(int cmd, struct client_obd *cli,
INIT_LIST_HEAD(&aa->aa_oaps);
*reqp = req;
+ niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
+ CDEBUG(D_RPCTRACE, "brw rpc %p - object " DOSTID " offset %lld<>%lld\n",
+ req, POSTID(&oa->o_oi), niobuf[0].rnb_offset,
+ niobuf[niocount - 1].rnb_offset + niobuf[niocount - 1].rnb_len);
+
return 0;
out:
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 038/124] staging: lustre: osc: debug to match extent to brw RPC
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Patrick Farrell, James Simmons
From: Patrick Farrell <paf@cray.com>
Currently, it's difficult to match brw RPCs to objects and
extents from client logs. This patch adds a D_RPCTRACE
debug message giving the necessary information.
Signed-off-by: Patrick Farrell <paf@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5531
Reviewed-on: http://review.whamcloud.com/11548
Reviewed-by: Alexey Lyashkov <alexey.lyashkov@seagate.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Reviewed-by: Ryan Haasken <haasken@cray.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_request.c | 5 +++++
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 5cca2b6..d5d2963 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1419,6 +1419,11 @@ static int osc_brw_prep_request(int cmd, struct client_obd *cli,
INIT_LIST_HEAD(&aa->aa_oaps);
*reqp = req;
+ niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
+ CDEBUG(D_RPCTRACE, "brw rpc %p - object " DOSTID " offset %lld<>%lld\n",
+ req, POSTID(&oa->o_oi), niobuf[0].rnb_offset,
+ niobuf[niocount - 1].rnb_offset + niobuf[niocount - 1].rnb_len);
+
return 0;
out:
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 039/124] staging: lustre: remove lustre_lite.h
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Move several definition only used in lustre/llite/ to
lustre/llite/llite_internal.h.
Remove lustre/include/{,linux/}lustre_lite.h and fixup
the missing includes in other headers that this exposes.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11501
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/linux/lustre_lite.h | 91 --------------------
.../staging/lustre/lustre/include/lustre_eacl.h | 1 +
.../staging/lustre/lustre/include/lustre_lite.h | 85 ------------------
.../staging/lustre/lustre/include/lustre_param.h | 3 +
drivers/staging/lustre/lustre/llite/dcache.c | 1 -
drivers/staging/lustre/lustre/llite/dir.c | 12 +--
drivers/staging/lustre/lustre/llite/file.c | 1 -
drivers/staging/lustre/lustre/llite/glimpse.c | 1 -
drivers/staging/lustre/lustre/llite/lcommon_cl.c | 1 -
drivers/staging/lustre/lustre/llite/lcommon_misc.c | 1 -
drivers/staging/lustre/lustre/llite/llite_close.c | 1 -
.../staging/lustre/lustre/llite/llite_internal.h | 61 +++++++++++++-
drivers/staging/lustre/lustre/llite/llite_lib.c | 1 -
drivers/staging/lustre/lustre/llite/llite_mmap.c | 1 -
drivers/staging/lustre/lustre/llite/llite_nfs.c | 1 -
drivers/staging/lustre/lustre/llite/lproc_llite.c | 1 -
drivers/staging/lustre/lustre/llite/namei.c | 1 -
drivers/staging/lustre/lustre/llite/rw.c | 1 -
drivers/staging/lustre/lustre/llite/rw26.c | 1 -
drivers/staging/lustre/lustre/llite/statahead.c | 13 +--
drivers/staging/lustre/lustre/llite/super25.c | 1 -
drivers/staging/lustre/lustre/llite/symlink.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_dev.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_io.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_lock.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_object.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_page.c | 2 -
drivers/staging/lustre/lustre/llite/vvp_req.c | 1 -
drivers/staging/lustre/lustre/llite/xattr.c | 1 -
drivers/staging/lustre/lustre/llite/xattr_cache.c | 1 -
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 2 -
31 files changed, 70 insertions(+), 222 deletions(-)
delete mode 100644 drivers/staging/lustre/lustre/include/linux/lustre_lite.h
delete mode 100644 drivers/staging/lustre/lustre/include/lustre_lite.h
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_lite.h b/drivers/staging/lustre/lustre/include/linux/lustre_lite.h
deleted file mode 100644
index d18e8a7..0000000
--- a/drivers/staging/lustre/lustre/include/linux/lustre_lite.h
+++ /dev/null
@@ -1,91 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.gnu.org/licenses/gpl-2.0.html
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef _LINUX_LL_H
-#define _LINUX_LL_H
-
-#ifndef _LL_H
-#error Do not #include this file directly. #include <lustre_lite.h> instead
-#endif
-
-#include <linux/statfs.h>
-
-#include <linux/fs.h>
-#include <linux/dcache.h>
-
-#include "../obd_class.h"
-#include "../lustre_net.h"
-#include "../lustre_ha.h"
-
-#include <linux/rbtree.h>
-#include "../../include/linux/lustre_compat25.h"
-#include <linux/pagemap.h>
-
-/* lprocfs.c */
-enum {
- LPROC_LL_DIRTY_HITS = 0,
- LPROC_LL_DIRTY_MISSES,
- LPROC_LL_READ_BYTES,
- LPROC_LL_WRITE_BYTES,
- LPROC_LL_BRW_READ,
- LPROC_LL_BRW_WRITE,
- LPROC_LL_OSC_READ,
- LPROC_LL_OSC_WRITE,
- LPROC_LL_IOCTL,
- LPROC_LL_OPEN,
- LPROC_LL_RELEASE,
- LPROC_LL_MAP,
- LPROC_LL_LLSEEK,
- LPROC_LL_FSYNC,
- LPROC_LL_READDIR,
- LPROC_LL_SETATTR,
- LPROC_LL_TRUNC,
- LPROC_LL_FLOCK,
- LPROC_LL_GETATTR,
- LPROC_LL_CREATE,
- LPROC_LL_LINK,
- LPROC_LL_UNLINK,
- LPROC_LL_SYMLINK,
- LPROC_LL_MKDIR,
- LPROC_LL_RMDIR,
- LPROC_LL_MKNOD,
- LPROC_LL_RENAME,
- LPROC_LL_STAFS,
- LPROC_LL_ALLOC_INODE,
- LPROC_LL_SETXATTR,
- LPROC_LL_GETXATTR,
- LPROC_LL_GETXATTR_HITS,
- LPROC_LL_LISTXATTR,
- LPROC_LL_REMOVEXATTR,
- LPROC_LL_INODE_PERM,
- LPROC_LL_FILE_OPCODES
-};
-
-#endif
diff --git a/drivers/staging/lustre/lustre/include/lustre_eacl.h b/drivers/staging/lustre/lustre/include/lustre_eacl.h
index d1039e1..1e71a86 100644
--- a/drivers/staging/lustre/lustre/include/lustre_eacl.h
+++ b/drivers/staging/lustre/lustre/include/lustre_eacl.h
@@ -46,6 +46,7 @@
#ifdef CONFIG_FS_POSIX_ACL
+#include <linux/fs.h>
#include <linux/posix_acl_xattr.h>
typedef struct {
diff --git a/drivers/staging/lustre/lustre/include/lustre_lite.h b/drivers/staging/lustre/lustre/include/lustre_lite.h
deleted file mode 100644
index 8333d76..0000000
--- a/drivers/staging/lustre/lustre/include/lustre_lite.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.gnu.org/licenses/gpl-2.0.html
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2011, 2012, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef _LL_H
-#define _LL_H
-
-/** \defgroup lite lite
- *
- * @{
- */
-
-#include "linux/lustre_lite.h"
-
-#include "obd_class.h"
-#include "lustre_net.h"
-#include "lustre_ha.h"
-
-/* 4UL * 1024 * 1024 */
-#define LL_MAX_BLKSIZE_BITS (22)
-#define LL_MAX_BLKSIZE (1UL << LL_MAX_BLKSIZE_BITS)
-
-/*
- * This is embedded into llite super-blocks to keep track of
- * connect flags (capabilities) supported by all imports given mount is
- * connected to.
- */
-struct lustre_client_ocd {
- /*
- * This is conjunction of connect_flags across all imports (LOVs) this
- * mount is connected to. This field is updated by cl_ocd_update()
- * under ->lco_lock.
- */
- __u64 lco_flags;
- struct mutex lco_lock;
- struct obd_export *lco_md_exp;
- struct obd_export *lco_dt_exp;
-};
-
-/*
- * Chain of hash overflow pages.
- */
-struct ll_dir_chain {
- /* XXX something. Later */
-};
-
-static inline void ll_dir_chain_init(struct ll_dir_chain *chain)
-{
-}
-
-static inline void ll_dir_chain_fini(struct ll_dir_chain *chain)
-{
-}
-
-/** @} lite */
-
-#endif
diff --git a/drivers/staging/lustre/lustre/include/lustre_param.h b/drivers/staging/lustre/lustre/include/lustre_param.h
index 82aadd3..8061a04 100644
--- a/drivers/staging/lustre/lustre/include/lustre_param.h
+++ b/drivers/staging/lustre/lustre/include/lustre_param.h
@@ -39,6 +39,9 @@
#ifndef _LUSTRE_PARAM_H
#define _LUSTRE_PARAM_H
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../../include/linux/lnet/types.h"
+
/** \defgroup param param
*
* @{
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index 8c00cc6..8500080 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -37,7 +37,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre/lustre_idl.h"
#include "../include/lustre_dlm.h"
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index 97815cf..234095e 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -48,7 +48,6 @@
#include "../include/obd_class.h"
#include "../include/lustre/lustre_ioctl.h"
#include "../include/lustre_lib.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_fid.h"
#include "../include/lustre_kernelcomm.h"
@@ -135,7 +134,7 @@
*
*/
struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data,
- __u64 offset, struct ll_dir_chain *chain)
+ __u64 offset)
{
struct md_callback cb_op;
struct page *page;
@@ -202,13 +201,10 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data,
int is_api32 = ll_need_32bit_api(sbi);
int is_hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH;
struct page *page;
- struct ll_dir_chain chain;
bool done = false;
int rc = 0;
- ll_dir_chain_init(&chain);
-
- page = ll_get_dir_page(inode, op_data, pos, &chain);
+ page = ll_get_dir_page(inode, op_data, pos);
while (rc == 0 && !done) {
struct lu_dirpage *dp;
@@ -286,13 +282,11 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data,
le32_to_cpu(dp->ldp_flags) &
LDF_COLLIDE);
next = pos;
- page = ll_get_dir_page(inode, op_data, pos,
- &chain);
+ page = ll_get_dir_page(inode, op_data, pos);
}
}
ctx->pos = pos;
- ll_dir_chain_fini(&chain);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index d9ee255..127d0e0 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/lustre_dlm.h"
-#include "../include/lustre_lite.h"
#include <linux/pagemap.h>
#include <linux/file.h>
#include <linux/sched.h>
diff --git a/drivers/staging/lustre/lustre/llite/glimpse.c b/drivers/staging/lustre/lustre/llite/glimpse.c
index 92004a0..22507b9 100644
--- a/drivers/staging/lustre/lustre/llite/glimpse.c
+++ b/drivers/staging/lustre/lustre/llite/glimpse.c
@@ -42,7 +42,6 @@
#include "../include/obd.h"
#include "../include/lustre_dlm.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_mdc.h"
#include <linux/pagemap.h>
#include <linux/file.h>
diff --git a/drivers/staging/lustre/lustre/llite/lcommon_cl.c b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
index eed464b..2ded6b5 100644
--- a/drivers/staging/lustre/lustre/llite/lcommon_cl.c
+++ b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
@@ -49,7 +49,6 @@
#include "../include/obd.h"
#include "../include/obd_support.h"
#include "../include/lustre_fid.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "../include/lustre_mdc.h"
diff --git a/drivers/staging/lustre/lustre/llite/lcommon_misc.c b/drivers/staging/lustre/lustre/llite/lcommon_misc.c
index f6be105..fb346c1 100644
--- a/drivers/staging/lustre/lustre/llite/lcommon_misc.c
+++ b/drivers/staging/lustre/lustre/llite/lcommon_misc.c
@@ -38,7 +38,6 @@
#include "../include/obd.h"
#include "../include/cl_object.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
/* Initialize the default and maximum LOV EA and cookie sizes. This allows
diff --git a/drivers/staging/lustre/lustre/llite/llite_close.c b/drivers/staging/lustre/lustre/llite/llite_close.c
index 2326b40..8644631 100644
--- a/drivers/staging/lustre/lustre/llite/llite_close.c
+++ b/drivers/staging/lustre/lustre/llite/llite_close.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
/** records that a write is in flight */
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 7f0a6c9..f972ba3 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -60,6 +60,9 @@
#define LL_DIR_END_OFF 0x7fffffffffffffffULL
#define LL_DIR_END_OFF_32BIT 0x7fffffffUL
+/* 4UL * 1024 * 1024 */
+#define LL_MAX_BLKSIZE_BITS 22
+
#define LL_IT2STR(it) ((it) ? ldlm_it2str((it)->it_op) : "0")
#define LUSTRE_FPRIVATE(file) ((file)->private_data)
@@ -445,6 +448,23 @@ enum stats_track_type {
"norootsquash", \
}
+/*
+ * This is embedded into llite super-blocks to keep track of connect
+ * flags (capabilities) supported by all imports given mount is
+ * connected to.
+ */
+struct lustre_client_ocd {
+ /*
+ * This is conjunction of connect_flags across all imports
+ * (LOVs) this mount is connected to. This field is updated by
+ * cl_ocd_update() under ->lco_lock.
+ */
+ __u64 lco_flags;
+ struct mutex lco_lock;
+ struct obd_export *lco_md_exp;
+ struct obd_export *lco_dt_exp;
+};
+
struct ll_sb_info {
/* this protects pglist and ra_info. It isn't safe to
* grab from interrupt contexts
@@ -669,6 +689,45 @@ void ll_rw_stats_tally(struct ll_sb_info *sbi, pid_t pid,
struct ll_file_data *file, loff_t pos,
size_t count, int rw);
+enum {
+ LPROC_LL_DIRTY_HITS,
+ LPROC_LL_DIRTY_MISSES,
+ LPROC_LL_READ_BYTES,
+ LPROC_LL_WRITE_BYTES,
+ LPROC_LL_BRW_READ,
+ LPROC_LL_BRW_WRITE,
+ LPROC_LL_OSC_READ,
+ LPROC_LL_OSC_WRITE,
+ LPROC_LL_IOCTL,
+ LPROC_LL_OPEN,
+ LPROC_LL_RELEASE,
+ LPROC_LL_MAP,
+ LPROC_LL_LLSEEK,
+ LPROC_LL_FSYNC,
+ LPROC_LL_READDIR,
+ LPROC_LL_SETATTR,
+ LPROC_LL_TRUNC,
+ LPROC_LL_FLOCK,
+ LPROC_LL_GETATTR,
+ LPROC_LL_CREATE,
+ LPROC_LL_LINK,
+ LPROC_LL_UNLINK,
+ LPROC_LL_SYMLINK,
+ LPROC_LL_MKDIR,
+ LPROC_LL_RMDIR,
+ LPROC_LL_MKNOD,
+ LPROC_LL_RENAME,
+ LPROC_LL_STAFS,
+ LPROC_LL_ALLOC_INODE,
+ LPROC_LL_SETXATTR,
+ LPROC_LL_GETXATTR,
+ LPROC_LL_GETXATTR_HITS,
+ LPROC_LL_LISTXATTR,
+ LPROC_LL_REMOVEXATTR,
+ LPROC_LL_INODE_PERM,
+ LPROC_LL_FILE_OPCODES
+};
+
/* llite/dir.c */
extern const struct file_operations ll_dir_operations;
extern const struct inode_operations ll_dir_inode_operations;
@@ -677,7 +736,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data,
int ll_get_mdt_idx(struct inode *inode);
int ll_get_mdt_idx_by_fid(struct ll_sb_info *sbi, const struct lu_fid *fid);
struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data,
- __u64 offset, struct ll_dir_chain *chain);
+ __u64 offset);
void ll_release_page(struct inode *inode, struct page *page, bool remove);
/* llite/namei.c */
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 9112fa8..19d18e8 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -42,7 +42,6 @@
#include <linux/mm.h>
#include "../include/lustre/lustre_ioctl.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_ha.h"
#include "../include/lustre_dlm.h"
#include "../include/lprocfs_status.h"
diff --git a/drivers/staging/lustre/lustre/llite/llite_mmap.c b/drivers/staging/lustre/lustre/llite/llite_mmap.c
index 37f82ed..0cd25bf 100644
--- a/drivers/staging/lustre/lustre/llite/llite_mmap.c
+++ b/drivers/staging/lustre/lustre/llite/llite_mmap.c
@@ -43,7 +43,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "../include/linux/lustre_compat25.h"
diff --git a/drivers/staging/lustre/lustre/llite/llite_nfs.c b/drivers/staging/lustre/lustre/llite/llite_nfs.c
index 1e156dc..7092305 100644
--- a/drivers/staging/lustre/lustre/llite/llite_nfs.c
+++ b/drivers/staging/lustre/lustre/llite/llite_nfs.c
@@ -38,7 +38,6 @@
*/
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include <linux/exportfs.h>
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index fc2ba56..6123435 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -31,7 +31,6 @@
*/
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "../include/lprocfs_status.h"
#include <linux/seq_file.h>
#include "../include/obd_support.h"
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 7a82ae4..e2dc920 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -42,7 +42,6 @@
#include "../include/obd_support.h"
#include "../include/lustre_fid.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "llite_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/rw.c b/drivers/staging/lustre/lustre/llite/rw.c
index de20b7d..508cb46 100644
--- a/drivers/staging/lustre/lustre/llite/rw.c
+++ b/drivers/staging/lustre/lustre/llite/rw.c
@@ -50,7 +50,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "../include/obd_cksum.h"
#include "llite_internal.h"
#include "../include/linux/lustre_compat25.h"
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index 66dba04..6df89b0 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -51,7 +51,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "../include/linux/lustre_compat25.h"
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 6577a66..5760974 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -39,7 +39,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "llite_internal.h"
@@ -1019,7 +1018,6 @@ static int ll_statahead_thread(void *arg)
int first = 0;
int rc = 0;
struct md_op_data *op_data;
- struct ll_dir_chain chain;
struct l_wait_info lwi = { 0 };
sai = ll_sai_get(dir);
@@ -1052,13 +1050,12 @@ static int ll_statahead_thread(void *arg)
spin_unlock(&lli->lli_sa_lock);
wake_up(&thread->t_ctl_waitq);
- ll_dir_chain_init(&chain);
while (pos != MDS_DIR_END_OFF && thread_is_running(thread)) {
struct lu_dirpage *dp;
struct lu_dirent *ent;
sai->sai_in_readpage = 1;
- page = ll_get_dir_page(dir, op_data, pos, &chain);
+ page = ll_get_dir_page(dir, op_data, pos);
sai->sai_in_readpage = 0;
if (IS_ERR(page)) {
rc = PTR_ERR(page);
@@ -1146,7 +1143,6 @@ static int ll_statahead_thread(void *arg)
break;
}
}
- ll_dir_chain_fini(&chain);
ll_finish_md_op_data(op_data);
if (rc < 0) {
@@ -1280,7 +1276,6 @@ enum {
static int is_first_dirent(struct inode *dir, struct dentry *dentry)
{
- struct ll_dir_chain chain;
const struct qstr *target = &dentry->d_name;
struct md_op_data *op_data;
struct page *page;
@@ -1297,8 +1292,7 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
*/
op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages;
- ll_dir_chain_init(&chain);
- page = ll_get_dir_page(dir, op_data, pos, &chain);
+ page = ll_get_dir_page(dir, op_data, pos);
while (1) {
struct lu_dirpage *dp;
@@ -1387,11 +1381,10 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
ll_release_page(dir, page,
le32_to_cpu(dp->ldp_flags) &
LDF_COLLIDE);
- page = ll_get_dir_page(dir, op_data, pos, &chain);
+ page = ll_get_dir_page(dir, op_data, pos);
}
}
out:
- ll_dir_chain_fini(&chain);
ll_finish_md_op_data(op_data);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/super25.c b/drivers/staging/lustre/lustre/llite/super25.c
index 24c79c2..106cd00 100644
--- a/drivers/staging/lustre/lustre/llite/super25.c
+++ b/drivers/staging/lustre/lustre/llite/super25.c
@@ -34,7 +34,6 @@
#include <linux/module.h>
#include <linux/types.h>
-#include "../include/lustre_lite.h"
#include "../include/lustre_ha.h"
#include "../include/lustre_dlm.h"
#include <linux/init.h>
diff --git a/drivers/staging/lustre/lustre/llite/symlink.c b/drivers/staging/lustre/lustre/llite/symlink.c
index 47fb799..f8bc7ed 100644
--- a/drivers/staging/lustre/lustre/llite/symlink.c
+++ b/drivers/staging/lustre/lustre/llite/symlink.c
@@ -35,7 +35,6 @@
#include <linux/stat.h>
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
static int ll_readlink_internal(struct inode *inode,
diff --git a/drivers/staging/lustre/lustre/llite/vvp_dev.c b/drivers/staging/lustre/lustre/llite/vvp_dev.c
index 771c0bd..8aa8ecc 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_dev.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_dev.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index f6dacee..27fc0b0 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_lock.c b/drivers/staging/lustre/lustre/llite/vvp_lock.c
index 64be0c9..07eb26c 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_lock.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_lock.c
@@ -37,7 +37,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_object.c b/drivers/staging/lustre/lustre/llite/vvp_object.c
index e4080ba..01725fb 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_object.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_object.c
@@ -39,7 +39,6 @@
#include "../../include/linux/libcfs/libcfs.h"
#include "../include/obd.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_page.c b/drivers/staging/lustre/lustre/llite/vvp_page.c
index bbbb0f1..e763e7f 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_page.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_page.c
@@ -44,8 +44,6 @@
#include <linux/page-flags.h>
#include <linux/pagemap.h>
-#include "../include/lustre_lite.h"
-
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_req.c b/drivers/staging/lustre/lustre/llite/vvp_req.c
index 0567a15..a6e2672 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_req.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_req.c
@@ -32,7 +32,6 @@
#include "../include/cl_object.h"
#include "../include/obd.h"
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c
index 7b8d469..17ad04f 100644
--- a/drivers/staging/lustre/lustre/llite/xattr.c
+++ b/drivers/staging/lustre/lustre/llite/xattr.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "../include/lustre_eacl.h"
diff --git a/drivers/staging/lustre/lustre/llite/xattr_cache.c b/drivers/staging/lustre/lustre/llite/xattr_cache.c
index 0330d1a..50a19a4 100644
--- a/drivers/staging/lustre/lustre/llite/xattr_cache.c
+++ b/drivers/staging/lustre/lustre/llite/xattr_cache.c
@@ -13,7 +13,6 @@
#include <linux/sched.h>
#include <linux/mm.h>
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "llite_internal.h"
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index e1fac0b..208a44d 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -43,13 +43,11 @@
#include "../include/lustre/lustre_idl.h"
#include "../include/obd_support.h"
-#include "../include/lustre_lib.h"
#include "../include/lustre_net.h"
#include "../include/obd_class.h"
#include "../include/lustre_lmv.h"
#include "../include/lprocfs_status.h"
#include "../include/cl_object.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_fid.h"
#include "../include/lustre/lustre_ioctl.h"
#include "../include/lustre_kernelcomm.h"
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 039/124] staging: lustre: remove lustre_lite.h
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Move several definition only used in lustre/llite/ to
lustre/llite/llite_internal.h.
Remove lustre/include/{,linux/}lustre_lite.h and fixup
the missing includes in other headers that this exposes.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11501
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/linux/lustre_lite.h | 91 --------------------
.../staging/lustre/lustre/include/lustre_eacl.h | 1 +
.../staging/lustre/lustre/include/lustre_lite.h | 85 ------------------
.../staging/lustre/lustre/include/lustre_param.h | 3 +
drivers/staging/lustre/lustre/llite/dcache.c | 1 -
drivers/staging/lustre/lustre/llite/dir.c | 12 +--
drivers/staging/lustre/lustre/llite/file.c | 1 -
drivers/staging/lustre/lustre/llite/glimpse.c | 1 -
drivers/staging/lustre/lustre/llite/lcommon_cl.c | 1 -
drivers/staging/lustre/lustre/llite/lcommon_misc.c | 1 -
drivers/staging/lustre/lustre/llite/llite_close.c | 1 -
.../staging/lustre/lustre/llite/llite_internal.h | 61 +++++++++++++-
drivers/staging/lustre/lustre/llite/llite_lib.c | 1 -
drivers/staging/lustre/lustre/llite/llite_mmap.c | 1 -
drivers/staging/lustre/lustre/llite/llite_nfs.c | 1 -
drivers/staging/lustre/lustre/llite/lproc_llite.c | 1 -
drivers/staging/lustre/lustre/llite/namei.c | 1 -
drivers/staging/lustre/lustre/llite/rw.c | 1 -
drivers/staging/lustre/lustre/llite/rw26.c | 1 -
drivers/staging/lustre/lustre/llite/statahead.c | 13 +--
drivers/staging/lustre/lustre/llite/super25.c | 1 -
drivers/staging/lustre/lustre/llite/symlink.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_dev.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_io.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_lock.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_object.c | 1 -
drivers/staging/lustre/lustre/llite/vvp_page.c | 2 -
drivers/staging/lustre/lustre/llite/vvp_req.c | 1 -
drivers/staging/lustre/lustre/llite/xattr.c | 1 -
drivers/staging/lustre/lustre/llite/xattr_cache.c | 1 -
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 2 -
31 files changed, 70 insertions(+), 222 deletions(-)
delete mode 100644 drivers/staging/lustre/lustre/include/linux/lustre_lite.h
delete mode 100644 drivers/staging/lustre/lustre/include/lustre_lite.h
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_lite.h b/drivers/staging/lustre/lustre/include/linux/lustre_lite.h
deleted file mode 100644
index d18e8a7..0000000
--- a/drivers/staging/lustre/lustre/include/linux/lustre_lite.h
+++ /dev/null
@@ -1,91 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.gnu.org/licenses/gpl-2.0.html
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef _LINUX_LL_H
-#define _LINUX_LL_H
-
-#ifndef _LL_H
-#error Do not #include this file directly. #include <lustre_lite.h> instead
-#endif
-
-#include <linux/statfs.h>
-
-#include <linux/fs.h>
-#include <linux/dcache.h>
-
-#include "../obd_class.h"
-#include "../lustre_net.h"
-#include "../lustre_ha.h"
-
-#include <linux/rbtree.h>
-#include "../../include/linux/lustre_compat25.h"
-#include <linux/pagemap.h>
-
-/* lprocfs.c */
-enum {
- LPROC_LL_DIRTY_HITS = 0,
- LPROC_LL_DIRTY_MISSES,
- LPROC_LL_READ_BYTES,
- LPROC_LL_WRITE_BYTES,
- LPROC_LL_BRW_READ,
- LPROC_LL_BRW_WRITE,
- LPROC_LL_OSC_READ,
- LPROC_LL_OSC_WRITE,
- LPROC_LL_IOCTL,
- LPROC_LL_OPEN,
- LPROC_LL_RELEASE,
- LPROC_LL_MAP,
- LPROC_LL_LLSEEK,
- LPROC_LL_FSYNC,
- LPROC_LL_READDIR,
- LPROC_LL_SETATTR,
- LPROC_LL_TRUNC,
- LPROC_LL_FLOCK,
- LPROC_LL_GETATTR,
- LPROC_LL_CREATE,
- LPROC_LL_LINK,
- LPROC_LL_UNLINK,
- LPROC_LL_SYMLINK,
- LPROC_LL_MKDIR,
- LPROC_LL_RMDIR,
- LPROC_LL_MKNOD,
- LPROC_LL_RENAME,
- LPROC_LL_STAFS,
- LPROC_LL_ALLOC_INODE,
- LPROC_LL_SETXATTR,
- LPROC_LL_GETXATTR,
- LPROC_LL_GETXATTR_HITS,
- LPROC_LL_LISTXATTR,
- LPROC_LL_REMOVEXATTR,
- LPROC_LL_INODE_PERM,
- LPROC_LL_FILE_OPCODES
-};
-
-#endif
diff --git a/drivers/staging/lustre/lustre/include/lustre_eacl.h b/drivers/staging/lustre/lustre/include/lustre_eacl.h
index d1039e1..1e71a86 100644
--- a/drivers/staging/lustre/lustre/include/lustre_eacl.h
+++ b/drivers/staging/lustre/lustre/include/lustre_eacl.h
@@ -46,6 +46,7 @@
#ifdef CONFIG_FS_POSIX_ACL
+#include <linux/fs.h>
#include <linux/posix_acl_xattr.h>
typedef struct {
diff --git a/drivers/staging/lustre/lustre/include/lustre_lite.h b/drivers/staging/lustre/lustre/include/lustre_lite.h
deleted file mode 100644
index 8333d76..0000000
--- a/drivers/staging/lustre/lustre/include/lustre_lite.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.gnu.org/licenses/gpl-2.0.html
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2011, 2012, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef _LL_H
-#define _LL_H
-
-/** \defgroup lite lite
- *
- * @{
- */
-
-#include "linux/lustre_lite.h"
-
-#include "obd_class.h"
-#include "lustre_net.h"
-#include "lustre_ha.h"
-
-/* 4UL * 1024 * 1024 */
-#define LL_MAX_BLKSIZE_BITS (22)
-#define LL_MAX_BLKSIZE (1UL << LL_MAX_BLKSIZE_BITS)
-
-/*
- * This is embedded into llite super-blocks to keep track of
- * connect flags (capabilities) supported by all imports given mount is
- * connected to.
- */
-struct lustre_client_ocd {
- /*
- * This is conjunction of connect_flags across all imports (LOVs) this
- * mount is connected to. This field is updated by cl_ocd_update()
- * under ->lco_lock.
- */
- __u64 lco_flags;
- struct mutex lco_lock;
- struct obd_export *lco_md_exp;
- struct obd_export *lco_dt_exp;
-};
-
-/*
- * Chain of hash overflow pages.
- */
-struct ll_dir_chain {
- /* XXX something. Later */
-};
-
-static inline void ll_dir_chain_init(struct ll_dir_chain *chain)
-{
-}
-
-static inline void ll_dir_chain_fini(struct ll_dir_chain *chain)
-{
-}
-
-/** @} lite */
-
-#endif
diff --git a/drivers/staging/lustre/lustre/include/lustre_param.h b/drivers/staging/lustre/lustre/include/lustre_param.h
index 82aadd3..8061a04 100644
--- a/drivers/staging/lustre/lustre/include/lustre_param.h
+++ b/drivers/staging/lustre/lustre/include/lustre_param.h
@@ -39,6 +39,9 @@
#ifndef _LUSTRE_PARAM_H
#define _LUSTRE_PARAM_H
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../../include/linux/lnet/types.h"
+
/** \defgroup param param
*
* @{
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index 8c00cc6..8500080 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -37,7 +37,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre/lustre_idl.h"
#include "../include/lustre_dlm.h"
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index 97815cf..234095e 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -48,7 +48,6 @@
#include "../include/obd_class.h"
#include "../include/lustre/lustre_ioctl.h"
#include "../include/lustre_lib.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_fid.h"
#include "../include/lustre_kernelcomm.h"
@@ -135,7 +134,7 @@
*
*/
struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data,
- __u64 offset, struct ll_dir_chain *chain)
+ __u64 offset)
{
struct md_callback cb_op;
struct page *page;
@@ -202,13 +201,10 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data,
int is_api32 = ll_need_32bit_api(sbi);
int is_hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH;
struct page *page;
- struct ll_dir_chain chain;
bool done = false;
int rc = 0;
- ll_dir_chain_init(&chain);
-
- page = ll_get_dir_page(inode, op_data, pos, &chain);
+ page = ll_get_dir_page(inode, op_data, pos);
while (rc == 0 && !done) {
struct lu_dirpage *dp;
@@ -286,13 +282,11 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data,
le32_to_cpu(dp->ldp_flags) &
LDF_COLLIDE);
next = pos;
- page = ll_get_dir_page(inode, op_data, pos,
- &chain);
+ page = ll_get_dir_page(inode, op_data, pos);
}
}
ctx->pos = pos;
- ll_dir_chain_fini(&chain);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index d9ee255..127d0e0 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/lustre_dlm.h"
-#include "../include/lustre_lite.h"
#include <linux/pagemap.h>
#include <linux/file.h>
#include <linux/sched.h>
diff --git a/drivers/staging/lustre/lustre/llite/glimpse.c b/drivers/staging/lustre/lustre/llite/glimpse.c
index 92004a0..22507b9 100644
--- a/drivers/staging/lustre/lustre/llite/glimpse.c
+++ b/drivers/staging/lustre/lustre/llite/glimpse.c
@@ -42,7 +42,6 @@
#include "../include/obd.h"
#include "../include/lustre_dlm.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_mdc.h"
#include <linux/pagemap.h>
#include <linux/file.h>
diff --git a/drivers/staging/lustre/lustre/llite/lcommon_cl.c b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
index eed464b..2ded6b5 100644
--- a/drivers/staging/lustre/lustre/llite/lcommon_cl.c
+++ b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
@@ -49,7 +49,6 @@
#include "../include/obd.h"
#include "../include/obd_support.h"
#include "../include/lustre_fid.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "../include/lustre_mdc.h"
diff --git a/drivers/staging/lustre/lustre/llite/lcommon_misc.c b/drivers/staging/lustre/lustre/llite/lcommon_misc.c
index f6be105..fb346c1 100644
--- a/drivers/staging/lustre/lustre/llite/lcommon_misc.c
+++ b/drivers/staging/lustre/lustre/llite/lcommon_misc.c
@@ -38,7 +38,6 @@
#include "../include/obd.h"
#include "../include/cl_object.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
/* Initialize the default and maximum LOV EA and cookie sizes. This allows
diff --git a/drivers/staging/lustre/lustre/llite/llite_close.c b/drivers/staging/lustre/lustre/llite/llite_close.c
index 2326b40..8644631 100644
--- a/drivers/staging/lustre/lustre/llite/llite_close.c
+++ b/drivers/staging/lustre/lustre/llite/llite_close.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
/** records that a write is in flight */
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 7f0a6c9..f972ba3 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -60,6 +60,9 @@
#define LL_DIR_END_OFF 0x7fffffffffffffffULL
#define LL_DIR_END_OFF_32BIT 0x7fffffffUL
+/* 4UL * 1024 * 1024 */
+#define LL_MAX_BLKSIZE_BITS 22
+
#define LL_IT2STR(it) ((it) ? ldlm_it2str((it)->it_op) : "0")
#define LUSTRE_FPRIVATE(file) ((file)->private_data)
@@ -445,6 +448,23 @@ enum stats_track_type {
"norootsquash", \
}
+/*
+ * This is embedded into llite super-blocks to keep track of connect
+ * flags (capabilities) supported by all imports given mount is
+ * connected to.
+ */
+struct lustre_client_ocd {
+ /*
+ * This is conjunction of connect_flags across all imports
+ * (LOVs) this mount is connected to. This field is updated by
+ * cl_ocd_update() under ->lco_lock.
+ */
+ __u64 lco_flags;
+ struct mutex lco_lock;
+ struct obd_export *lco_md_exp;
+ struct obd_export *lco_dt_exp;
+};
+
struct ll_sb_info {
/* this protects pglist and ra_info. It isn't safe to
* grab from interrupt contexts
@@ -669,6 +689,45 @@ void ll_rw_stats_tally(struct ll_sb_info *sbi, pid_t pid,
struct ll_file_data *file, loff_t pos,
size_t count, int rw);
+enum {
+ LPROC_LL_DIRTY_HITS,
+ LPROC_LL_DIRTY_MISSES,
+ LPROC_LL_READ_BYTES,
+ LPROC_LL_WRITE_BYTES,
+ LPROC_LL_BRW_READ,
+ LPROC_LL_BRW_WRITE,
+ LPROC_LL_OSC_READ,
+ LPROC_LL_OSC_WRITE,
+ LPROC_LL_IOCTL,
+ LPROC_LL_OPEN,
+ LPROC_LL_RELEASE,
+ LPROC_LL_MAP,
+ LPROC_LL_LLSEEK,
+ LPROC_LL_FSYNC,
+ LPROC_LL_READDIR,
+ LPROC_LL_SETATTR,
+ LPROC_LL_TRUNC,
+ LPROC_LL_FLOCK,
+ LPROC_LL_GETATTR,
+ LPROC_LL_CREATE,
+ LPROC_LL_LINK,
+ LPROC_LL_UNLINK,
+ LPROC_LL_SYMLINK,
+ LPROC_LL_MKDIR,
+ LPROC_LL_RMDIR,
+ LPROC_LL_MKNOD,
+ LPROC_LL_RENAME,
+ LPROC_LL_STAFS,
+ LPROC_LL_ALLOC_INODE,
+ LPROC_LL_SETXATTR,
+ LPROC_LL_GETXATTR,
+ LPROC_LL_GETXATTR_HITS,
+ LPROC_LL_LISTXATTR,
+ LPROC_LL_REMOVEXATTR,
+ LPROC_LL_INODE_PERM,
+ LPROC_LL_FILE_OPCODES
+};
+
/* llite/dir.c */
extern const struct file_operations ll_dir_operations;
extern const struct inode_operations ll_dir_inode_operations;
@@ -677,7 +736,7 @@ int ll_dir_read(struct inode *inode, __u64 *ppos, struct md_op_data *op_data,
int ll_get_mdt_idx(struct inode *inode);
int ll_get_mdt_idx_by_fid(struct ll_sb_info *sbi, const struct lu_fid *fid);
struct page *ll_get_dir_page(struct inode *dir, struct md_op_data *op_data,
- __u64 offset, struct ll_dir_chain *chain);
+ __u64 offset);
void ll_release_page(struct inode *inode, struct page *page, bool remove);
/* llite/namei.c */
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 9112fa8..19d18e8 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -42,7 +42,6 @@
#include <linux/mm.h>
#include "../include/lustre/lustre_ioctl.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_ha.h"
#include "../include/lustre_dlm.h"
#include "../include/lprocfs_status.h"
diff --git a/drivers/staging/lustre/lustre/llite/llite_mmap.c b/drivers/staging/lustre/lustre/llite/llite_mmap.c
index 37f82ed..0cd25bf 100644
--- a/drivers/staging/lustre/lustre/llite/llite_mmap.c
+++ b/drivers/staging/lustre/lustre/llite/llite_mmap.c
@@ -43,7 +43,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "../include/linux/lustre_compat25.h"
diff --git a/drivers/staging/lustre/lustre/llite/llite_nfs.c b/drivers/staging/lustre/lustre/llite/llite_nfs.c
index 1e156dc..7092305 100644
--- a/drivers/staging/lustre/lustre/llite/llite_nfs.c
+++ b/drivers/staging/lustre/lustre/llite/llite_nfs.c
@@ -38,7 +38,6 @@
*/
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include <linux/exportfs.h>
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index fc2ba56..6123435 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -31,7 +31,6 @@
*/
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "../include/lprocfs_status.h"
#include <linux/seq_file.h>
#include "../include/obd_support.h"
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 7a82ae4..e2dc920 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -42,7 +42,6 @@
#include "../include/obd_support.h"
#include "../include/lustre_fid.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "llite_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/rw.c b/drivers/staging/lustre/lustre/llite/rw.c
index de20b7d..508cb46 100644
--- a/drivers/staging/lustre/lustre/llite/rw.c
+++ b/drivers/staging/lustre/lustre/llite/rw.c
@@ -50,7 +50,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "../include/obd_cksum.h"
#include "llite_internal.h"
#include "../include/linux/lustre_compat25.h"
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index 66dba04..6df89b0 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -51,7 +51,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "../include/linux/lustre_compat25.h"
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 6577a66..5760974 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -39,7 +39,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "llite_internal.h"
@@ -1019,7 +1018,6 @@ static int ll_statahead_thread(void *arg)
int first = 0;
int rc = 0;
struct md_op_data *op_data;
- struct ll_dir_chain chain;
struct l_wait_info lwi = { 0 };
sai = ll_sai_get(dir);
@@ -1052,13 +1050,12 @@ static int ll_statahead_thread(void *arg)
spin_unlock(&lli->lli_sa_lock);
wake_up(&thread->t_ctl_waitq);
- ll_dir_chain_init(&chain);
while (pos != MDS_DIR_END_OFF && thread_is_running(thread)) {
struct lu_dirpage *dp;
struct lu_dirent *ent;
sai->sai_in_readpage = 1;
- page = ll_get_dir_page(dir, op_data, pos, &chain);
+ page = ll_get_dir_page(dir, op_data, pos);
sai->sai_in_readpage = 0;
if (IS_ERR(page)) {
rc = PTR_ERR(page);
@@ -1146,7 +1143,6 @@ static int ll_statahead_thread(void *arg)
break;
}
}
- ll_dir_chain_fini(&chain);
ll_finish_md_op_data(op_data);
if (rc < 0) {
@@ -1280,7 +1276,6 @@ enum {
static int is_first_dirent(struct inode *dir, struct dentry *dentry)
{
- struct ll_dir_chain chain;
const struct qstr *target = &dentry->d_name;
struct md_op_data *op_data;
struct page *page;
@@ -1297,8 +1292,7 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
*/
op_data->op_max_pages = ll_i2sbi(dir)->ll_md_brw_pages;
- ll_dir_chain_init(&chain);
- page = ll_get_dir_page(dir, op_data, pos, &chain);
+ page = ll_get_dir_page(dir, op_data, pos);
while (1) {
struct lu_dirpage *dp;
@@ -1387,11 +1381,10 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
ll_release_page(dir, page,
le32_to_cpu(dp->ldp_flags) &
LDF_COLLIDE);
- page = ll_get_dir_page(dir, op_data, pos, &chain);
+ page = ll_get_dir_page(dir, op_data, pos);
}
}
out:
- ll_dir_chain_fini(&chain);
ll_finish_md_op_data(op_data);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/super25.c b/drivers/staging/lustre/lustre/llite/super25.c
index 24c79c2..106cd00 100644
--- a/drivers/staging/lustre/lustre/llite/super25.c
+++ b/drivers/staging/lustre/lustre/llite/super25.c
@@ -34,7 +34,6 @@
#include <linux/module.h>
#include <linux/types.h>
-#include "../include/lustre_lite.h"
#include "../include/lustre_ha.h"
#include "../include/lustre_dlm.h"
#include <linux/init.h>
diff --git a/drivers/staging/lustre/lustre/llite/symlink.c b/drivers/staging/lustre/lustre/llite/symlink.c
index 47fb799..f8bc7ed 100644
--- a/drivers/staging/lustre/lustre/llite/symlink.c
+++ b/drivers/staging/lustre/lustre/llite/symlink.c
@@ -35,7 +35,6 @@
#include <linux/stat.h>
#define DEBUG_SUBSYSTEM S_LLITE
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
static int ll_readlink_internal(struct inode *inode,
diff --git a/drivers/staging/lustre/lustre/llite/vvp_dev.c b/drivers/staging/lustre/lustre/llite/vvp_dev.c
index 771c0bd..8aa8ecc 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_dev.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_dev.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index f6dacee..27fc0b0 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_lock.c b/drivers/staging/lustre/lustre/llite/vvp_lock.c
index 64be0c9..07eb26c 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_lock.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_lock.c
@@ -37,7 +37,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_object.c b/drivers/staging/lustre/lustre/llite/vvp_object.c
index e4080ba..01725fb 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_object.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_object.c
@@ -39,7 +39,6 @@
#include "../../include/linux/libcfs/libcfs.h"
#include "../include/obd.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_page.c b/drivers/staging/lustre/lustre/llite/vvp_page.c
index bbbb0f1..e763e7f 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_page.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_page.c
@@ -44,8 +44,6 @@
#include <linux/page-flags.h>
#include <linux/pagemap.h>
-#include "../include/lustre_lite.h"
-
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/vvp_req.c b/drivers/staging/lustre/lustre/llite/vvp_req.c
index 0567a15..a6e2672 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_req.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_req.c
@@ -32,7 +32,6 @@
#include "../include/cl_object.h"
#include "../include/obd.h"
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "llite_internal.h"
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c
index 7b8d469..17ad04f 100644
--- a/drivers/staging/lustre/lustre/llite/xattr.c
+++ b/drivers/staging/lustre/lustre/llite/xattr.c
@@ -38,7 +38,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "../include/lustre_eacl.h"
diff --git a/drivers/staging/lustre/lustre/llite/xattr_cache.c b/drivers/staging/lustre/lustre/llite/xattr_cache.c
index 0330d1a..50a19a4 100644
--- a/drivers/staging/lustre/lustre/llite/xattr_cache.c
+++ b/drivers/staging/lustre/lustre/llite/xattr_cache.c
@@ -13,7 +13,6 @@
#include <linux/sched.h>
#include <linux/mm.h>
#include "../include/obd_support.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_dlm.h"
#include "../include/lustre_ver.h"
#include "llite_internal.h"
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index e1fac0b..208a44d 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -43,13 +43,11 @@
#include "../include/lustre/lustre_idl.h"
#include "../include/obd_support.h"
-#include "../include/lustre_lib.h"
#include "../include/lustre_net.h"
#include "../include/obd_class.h"
#include "../include/lustre_lmv.h"
#include "../include/lprocfs_status.h"
#include "../include/cl_object.h"
-#include "../include/lustre_lite.h"
#include "../include/lustre_fid.h"
#include "../include/lustre/lustre_ioctl.h"
#include "../include/lustre_kernelcomm.h"
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 040/124] staging: lustre: obd: rename LUSTRE_STRIPE_MAXBYTES
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Rename LUSTRE_STRIPE_MAXBYTES to LUSTRE_EXT3_STRIPE_MAXBYTES and
correct the comment describing its use.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11800
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 8 +++++---
drivers/staging/lustre/lustre/lov/lov_ea.c | 6 +++---
drivers/staging/lustre/lustre/lov/lov_pack.c | 2 +-
.../staging/lustre/lustre/obdecho/echo_client.c | 2 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 2 +-
5 files changed, 11 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 0917aaa..19230a9 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -73,10 +73,12 @@ static inline void loi_init(struct lov_oinfo *loi)
{
}
-/* Until such time as we get_info the per-stripe maximum from the OST,
- * we define this to be 2T - 4k, which is the ext3 maxbytes.
+/*
+ * If we are unable to get the maximum object size from the OST in
+ * ocd_maxbytes using OBD_CONNECT_MAXBYTES, then we fall back to using
+ * the old maximum object size from ext3.
*/
-#define LUSTRE_STRIPE_MAXBYTES 0x1fffffff000ULL
+#define LUSTRE_EXT3_STRIPE_MAXBYTES 0x1fffffff000ULL
struct lov_stripe_md {
atomic_t lsm_refc;
diff --git a/drivers/staging/lustre/lustre/lov/lov_ea.c b/drivers/staging/lustre/lustre/lov/lov_ea.c
index d88b81d..9f0b5b7 100644
--- a/drivers/staging/lustre/lustre/lov/lov_ea.c
+++ b/drivers/staging/lustre/lustre/lov/lov_ea.c
@@ -154,14 +154,14 @@ static int lsm_destroy_plain(struct lov_stripe_md *lsm, struct obdo *oa,
}
/* Find minimum stripe maxbytes value. For inactive or
- * reconnecting targets use LUSTRE_STRIPE_MAXBYTES.
+ * reconnecting targets use LUSTRE_EXT3_STRIPE_MAXBYTES.
*/
static void lov_tgt_maxbytes(struct lov_tgt_desc *tgt, __u64 *stripe_maxbytes)
{
struct obd_import *imp = tgt->ltd_obd->u.cli.cl_import;
if (!imp || !tgt->ltd_active) {
- *stripe_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ *stripe_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
return;
}
@@ -172,7 +172,7 @@ static void lov_tgt_maxbytes(struct lov_tgt_desc *tgt, __u64 *stripe_maxbytes)
if (*stripe_maxbytes > imp->imp_connect_data.ocd_maxbytes)
*stripe_maxbytes = imp->imp_connect_data.ocd_maxbytes;
} else {
- *stripe_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ *stripe_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
}
spin_unlock(&imp->imp_lock);
}
diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
index 869ef41..c654810 100644
--- a/drivers/staging/lustre/lustre/lov/lov_pack.c
+++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
@@ -284,7 +284,7 @@ int lov_alloc_memmd(struct lov_stripe_md **lsmp, __u16 stripe_count,
spin_lock_init(&(*lsmp)->lsm_lock);
(*lsmp)->lsm_magic = magic;
(*lsmp)->lsm_stripe_count = stripe_count;
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES * stripe_count;
+ (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES * stripe_count;
(*lsmp)->lsm_pattern = pattern;
(*lsmp)->lsm_pool_name[0] = '\0';
(*lsmp)->lsm_layout_gen = 0;
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index 6968953..cddc2d2 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -454,7 +454,7 @@ static int echo_alloc_memmd(struct echo_device *ed,
}
loi_init((*lsmp)->lsm_oinfo[0]);
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
ostid_set_seq_echo(&(*lsmp)->lsm_oi);
return lsm_size;
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index d5d2963..7f7c318 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -190,7 +190,7 @@ static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
(imp->imp_connect_data.ocd_connect_flags & OBD_CONNECT_MAXBYTES))
(*lsmp)->lsm_maxbytes = imp->imp_connect_data.ocd_maxbytes;
else
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
return lsm_size;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 040/124] staging: lustre: obd: rename LUSTRE_STRIPE_MAXBYTES
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Rename LUSTRE_STRIPE_MAXBYTES to LUSTRE_EXT3_STRIPE_MAXBYTES and
correct the comment describing its use.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11800
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 8 +++++---
drivers/staging/lustre/lustre/lov/lov_ea.c | 6 +++---
drivers/staging/lustre/lustre/lov/lov_pack.c | 2 +-
.../staging/lustre/lustre/obdecho/echo_client.c | 2 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 2 +-
5 files changed, 11 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 0917aaa..19230a9 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -73,10 +73,12 @@ static inline void loi_init(struct lov_oinfo *loi)
{
}
-/* Until such time as we get_info the per-stripe maximum from the OST,
- * we define this to be 2T - 4k, which is the ext3 maxbytes.
+/*
+ * If we are unable to get the maximum object size from the OST in
+ * ocd_maxbytes using OBD_CONNECT_MAXBYTES, then we fall back to using
+ * the old maximum object size from ext3.
*/
-#define LUSTRE_STRIPE_MAXBYTES 0x1fffffff000ULL
+#define LUSTRE_EXT3_STRIPE_MAXBYTES 0x1fffffff000ULL
struct lov_stripe_md {
atomic_t lsm_refc;
diff --git a/drivers/staging/lustre/lustre/lov/lov_ea.c b/drivers/staging/lustre/lustre/lov/lov_ea.c
index d88b81d..9f0b5b7 100644
--- a/drivers/staging/lustre/lustre/lov/lov_ea.c
+++ b/drivers/staging/lustre/lustre/lov/lov_ea.c
@@ -154,14 +154,14 @@ static int lsm_destroy_plain(struct lov_stripe_md *lsm, struct obdo *oa,
}
/* Find minimum stripe maxbytes value. For inactive or
- * reconnecting targets use LUSTRE_STRIPE_MAXBYTES.
+ * reconnecting targets use LUSTRE_EXT3_STRIPE_MAXBYTES.
*/
static void lov_tgt_maxbytes(struct lov_tgt_desc *tgt, __u64 *stripe_maxbytes)
{
struct obd_import *imp = tgt->ltd_obd->u.cli.cl_import;
if (!imp || !tgt->ltd_active) {
- *stripe_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ *stripe_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
return;
}
@@ -172,7 +172,7 @@ static void lov_tgt_maxbytes(struct lov_tgt_desc *tgt, __u64 *stripe_maxbytes)
if (*stripe_maxbytes > imp->imp_connect_data.ocd_maxbytes)
*stripe_maxbytes = imp->imp_connect_data.ocd_maxbytes;
} else {
- *stripe_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ *stripe_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
}
spin_unlock(&imp->imp_lock);
}
diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
index 869ef41..c654810 100644
--- a/drivers/staging/lustre/lustre/lov/lov_pack.c
+++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
@@ -284,7 +284,7 @@ int lov_alloc_memmd(struct lov_stripe_md **lsmp, __u16 stripe_count,
spin_lock_init(&(*lsmp)->lsm_lock);
(*lsmp)->lsm_magic = magic;
(*lsmp)->lsm_stripe_count = stripe_count;
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES * stripe_count;
+ (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES * stripe_count;
(*lsmp)->lsm_pattern = pattern;
(*lsmp)->lsm_pool_name[0] = '\0';
(*lsmp)->lsm_layout_gen = 0;
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index 6968953..cddc2d2 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -454,7 +454,7 @@ static int echo_alloc_memmd(struct echo_device *ed,
}
loi_init((*lsmp)->lsm_oinfo[0]);
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
ostid_set_seq_echo(&(*lsmp)->lsm_oi);
return lsm_size;
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index d5d2963..7f7c318 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -190,7 +190,7 @@ static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
(imp->imp_connect_data.ocd_connect_flags & OBD_CONNECT_MAXBYTES))
(*lsmp)->lsm_maxbytes = imp->imp_connect_data.ocd_maxbytes;
else
- (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
+ (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
return lsm_size;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 041/124] staging: lustre: llite: don't call make_bad_inode() on an old inode
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In ll_iget() if ll_update_inode() fails then do not call
make_bad_inode() on the inode since it may still be in use.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5468
Reviewed-on: http://review.whamcloud.com/11609
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index e2dc920..85f8ce7 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -132,7 +132,6 @@ struct inode *ll_iget(struct super_block *sb, ino_t hash,
CDEBUG(D_VFSTRACE, "got inode: "DFID"(%p): rc = %d\n",
PFID(&md->body->mbo_fid1), inode, rc);
if (rc) {
- make_bad_inode(inode);
iput(inode);
inode = ERR_PTR(rc);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 041/124] staging: lustre: llite: don't call make_bad_inode() on an old inode
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In ll_iget() if ll_update_inode() fails then do not call
make_bad_inode() on the inode since it may still be in use.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5468
Reviewed-on: http://review.whamcloud.com/11609
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/namei.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index e2dc920..85f8ce7 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -132,7 +132,6 @@ struct inode *ll_iget(struct super_block *sb, ino_t hash,
CDEBUG(D_VFSTRACE, "got inode: "DFID"(%p): rc = %d\n",
PFID(&md->body->mbo_fid1), inode, rc);
if (rc) {
- make_bad_inode(inode);
iput(inode);
inode = ERR_PTR(rc);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 042/124] staging: lustre: obd: change type of lmv_tgt_desc->ltd_idx to u32
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
ltd_idx is used as unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/11879
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 8 ++++----
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 6 +++---
drivers/staging/lustre/lustre/lmv/lproc_lmv.c | 2 +-
4 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 19230a9..4b72fb9 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -406,7 +406,7 @@ struct lov_obd {
struct lmv_tgt_desc {
struct obd_uuid ltd_uuid;
struct obd_export *ltd_exp;
- int ltd_idx;
+ u32 ltd_idx;
struct mutex ltd_fid_mutex;
unsigned long ltd_active:1; /* target up for requests */
};
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index f24848f..1952454 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -113,7 +113,7 @@ static int lmv_intent_remote(struct obd_export *exp, struct lookup_intent *it,
}
op_data->op_bias = MDS_CROSS_REF;
- CDEBUG(D_INODE, "REMOTE_INTENT with fid="DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "REMOTE_INTENT with fid=" DFID " -> mds #%u\n",
PFID(&body->mbo_fid1), tgt->ltd_idx);
rc = md_intent_lock(tgt->ltd_exp, op_data, it, &req, cb_blocking,
@@ -202,7 +202,7 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
goto cleanup;
}
- CDEBUG(D_INODE, "Revalidate slave "DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "Revalidate slave " DFID " -> mds #%u\n",
PFID(&fid), tgt->ltd_idx);
if (req) {
@@ -349,7 +349,7 @@ static int lmv_intent_open(struct obd_export *exp, struct md_op_data *op_data,
return rc;
}
- CDEBUG(D_INODE, "OPEN_INTENT with fid1=" DFID ", fid2=" DFID ", name='%s' -> mds #%d\n",
+ CDEBUG(D_INODE, "OPEN_INTENT with fid1=" DFID ", fid2=" DFID ", name='%s' -> mds #%u\n",
PFID(&op_data->op_fid1),
PFID(&op_data->op_fid2), op_data->op_name, tgt->ltd_idx);
@@ -427,7 +427,7 @@ static int lmv_intent_lookup(struct obd_export *exp,
if (!fid_is_sane(&op_data->op_fid2))
fid_zero(&op_data->op_fid2);
- CDEBUG(D_INODE, "LOOKUP_INTENT with fid1="DFID", fid2="DFID", name='%s' -> mds #%d lsm=%p lsm_magic=%x\n",
+ CDEBUG(D_INODE, "LOOKUP_INTENT with fid1=" DFID ", fid2=" DFID ", name='%s' -> mds #%u lsm=%p lsm_magic=%x\n",
PFID(&op_data->op_fid1), PFID(&op_data->op_fid2),
op_data->op_name ? op_data->op_name : "<NULL>",
tgt->ltd_idx, lsm, !lsm ? -1 : lsm->lsm_md_magic);
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 208a44d..b852a65 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1836,7 +1836,7 @@ lmv_enqueue(struct obd_export *exp, struct ldlm_enqueue_info *einfo,
if (IS_ERR(tgt))
return PTR_ERR(tgt);
- CDEBUG(D_INODE, "ENQUEUE '%s' on "DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "ENQUEUE '%s' on " DFID " -> mds #%u\n",
LL_IT2STR(it), PFID(&op_data->op_fid1), tgt->ltd_idx);
rc = md_enqueue(tgt->ltd_exp, einfo, policy, it, op_data, lockh,
@@ -1864,7 +1864,7 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data,
if (IS_ERR(tgt))
return PTR_ERR(tgt);
- CDEBUG(D_INODE, "GETATTR_NAME for %*s on "DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "GETATTR_NAME for %*s on " DFID " -> mds #%u\n",
op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
tgt->ltd_idx);
@@ -2609,7 +2609,7 @@ try_next_stripe:
if (rc != 0)
return rc;
- CDEBUG(D_INODE, "unlink with fid="DFID"/"DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "unlink with fid=" DFID "/" DFID " -> mds #%u\n",
PFID(&op_data->op_fid1), PFID(&op_data->op_fid2), tgt->ltd_idx);
rc = md_unlink(tgt->ltd_exp, op_data, request);
diff --git a/drivers/staging/lustre/lustre/lmv/lproc_lmv.c b/drivers/staging/lustre/lustre/lmv/lproc_lmv.c
index d2316c0..20bbdfc 100644
--- a/drivers/staging/lustre/lustre/lmv/lproc_lmv.c
+++ b/drivers/staging/lustre/lustre/lmv/lproc_lmv.c
@@ -169,7 +169,7 @@ static int lmv_tgt_seq_show(struct seq_file *p, void *v)
if (!tgt)
return 0;
- seq_printf(p, "%d: %s %sACTIVE\n",
+ seq_printf(p, "%u: %s %sACTIVE\n",
tgt->ltd_idx, tgt->ltd_uuid.uuid,
tgt->ltd_active ? "" : "IN");
return 0;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 042/124] staging: lustre: obd: change type of lmv_tgt_desc->ltd_idx to u32
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
ltd_idx is used as unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/11879
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 8 ++++----
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 6 +++---
drivers/staging/lustre/lustre/lmv/lproc_lmv.c | 2 +-
4 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 19230a9..4b72fb9 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -406,7 +406,7 @@ struct lov_obd {
struct lmv_tgt_desc {
struct obd_uuid ltd_uuid;
struct obd_export *ltd_exp;
- int ltd_idx;
+ u32 ltd_idx;
struct mutex ltd_fid_mutex;
unsigned long ltd_active:1; /* target up for requests */
};
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index f24848f..1952454 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -113,7 +113,7 @@ static int lmv_intent_remote(struct obd_export *exp, struct lookup_intent *it,
}
op_data->op_bias = MDS_CROSS_REF;
- CDEBUG(D_INODE, "REMOTE_INTENT with fid="DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "REMOTE_INTENT with fid=" DFID " -> mds #%u\n",
PFID(&body->mbo_fid1), tgt->ltd_idx);
rc = md_intent_lock(tgt->ltd_exp, op_data, it, &req, cb_blocking,
@@ -202,7 +202,7 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
goto cleanup;
}
- CDEBUG(D_INODE, "Revalidate slave "DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "Revalidate slave " DFID " -> mds #%u\n",
PFID(&fid), tgt->ltd_idx);
if (req) {
@@ -349,7 +349,7 @@ static int lmv_intent_open(struct obd_export *exp, struct md_op_data *op_data,
return rc;
}
- CDEBUG(D_INODE, "OPEN_INTENT with fid1=" DFID ", fid2=" DFID ", name='%s' -> mds #%d\n",
+ CDEBUG(D_INODE, "OPEN_INTENT with fid1=" DFID ", fid2=" DFID ", name='%s' -> mds #%u\n",
PFID(&op_data->op_fid1),
PFID(&op_data->op_fid2), op_data->op_name, tgt->ltd_idx);
@@ -427,7 +427,7 @@ static int lmv_intent_lookup(struct obd_export *exp,
if (!fid_is_sane(&op_data->op_fid2))
fid_zero(&op_data->op_fid2);
- CDEBUG(D_INODE, "LOOKUP_INTENT with fid1="DFID", fid2="DFID", name='%s' -> mds #%d lsm=%p lsm_magic=%x\n",
+ CDEBUG(D_INODE, "LOOKUP_INTENT with fid1=" DFID ", fid2=" DFID ", name='%s' -> mds #%u lsm=%p lsm_magic=%x\n",
PFID(&op_data->op_fid1), PFID(&op_data->op_fid2),
op_data->op_name ? op_data->op_name : "<NULL>",
tgt->ltd_idx, lsm, !lsm ? -1 : lsm->lsm_md_magic);
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 208a44d..b852a65 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1836,7 +1836,7 @@ lmv_enqueue(struct obd_export *exp, struct ldlm_enqueue_info *einfo,
if (IS_ERR(tgt))
return PTR_ERR(tgt);
- CDEBUG(D_INODE, "ENQUEUE '%s' on "DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "ENQUEUE '%s' on " DFID " -> mds #%u\n",
LL_IT2STR(it), PFID(&op_data->op_fid1), tgt->ltd_idx);
rc = md_enqueue(tgt->ltd_exp, einfo, policy, it, op_data, lockh,
@@ -1864,7 +1864,7 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data,
if (IS_ERR(tgt))
return PTR_ERR(tgt);
- CDEBUG(D_INODE, "GETATTR_NAME for %*s on "DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "GETATTR_NAME for %*s on " DFID " -> mds #%u\n",
op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
tgt->ltd_idx);
@@ -2609,7 +2609,7 @@ try_next_stripe:
if (rc != 0)
return rc;
- CDEBUG(D_INODE, "unlink with fid="DFID"/"DFID" -> mds #%d\n",
+ CDEBUG(D_INODE, "unlink with fid=" DFID "/" DFID " -> mds #%u\n",
PFID(&op_data->op_fid1), PFID(&op_data->op_fid2), tgt->ltd_idx);
rc = md_unlink(tgt->ltd_exp, op_data, request);
diff --git a/drivers/staging/lustre/lustre/lmv/lproc_lmv.c b/drivers/staging/lustre/lustre/lmv/lproc_lmv.c
index d2316c0..20bbdfc 100644
--- a/drivers/staging/lustre/lustre/lmv/lproc_lmv.c
+++ b/drivers/staging/lustre/lustre/lmv/lproc_lmv.c
@@ -169,7 +169,7 @@ static int lmv_tgt_seq_show(struct seq_file *p, void *v)
if (!tgt)
return 0;
- seq_printf(p, "%d: %s %sACTIVE\n",
+ seq_printf(p, "%u: %s %sACTIVE\n",
tgt->ltd_idx, tgt->ltd_uuid.uuid,
tgt->ltd_active ? "" : "IN");
return 0;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 043/124] staging: lustre: lmv: change type of lmv_obd->tgts_size to u32
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
tgts_size is used as unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/11881
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 4b72fb9..27fb4d7 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -435,7 +435,7 @@ struct lmv_obd {
int max_def_cookiesize;
int server_timeout;
- int tgts_size; /* size of tgts array */
+ u32 tgts_size; /* size of tgts array */
struct lmv_tgt_desc **tgts;
struct obd_connect_data conn_data;
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index b852a65..53de385 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1355,10 +1355,10 @@ static int lmv_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
return -EINVAL;
}
- lmv->tgts = kcalloc(32, sizeof(*lmv->tgts), GFP_NOFS);
+ lmv->tgts_size = 32U;
+ lmv->tgts = kcalloc(lmv->tgts_size, sizeof(*lmv->tgts), GFP_NOFS);
if (!lmv->tgts)
return -ENOMEM;
- lmv->tgts_size = 32;
obd_str2uuid(&lmv->desc.ld_uuid, desc->ld_uuid.uuid);
lmv->desc.ld_tgt_count = 0;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 043/124] staging: lustre: lmv: change type of lmv_obd->tgts_size to u32
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
tgts_size is used as unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/11881
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 4b72fb9..27fb4d7 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -435,7 +435,7 @@ struct lmv_obd {
int max_def_cookiesize;
int server_timeout;
- int tgts_size; /* size of tgts array */
+ u32 tgts_size; /* size of tgts array */
struct lmv_tgt_desc **tgts;
struct obd_connect_data conn_data;
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index b852a65..53de385 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1355,10 +1355,10 @@ static int lmv_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
return -EINVAL;
}
- lmv->tgts = kcalloc(32, sizeof(*lmv->tgts), GFP_NOFS);
+ lmv->tgts_size = 32U;
+ lmv->tgts = kcalloc(lmv->tgts_size, sizeof(*lmv->tgts), GFP_NOFS);
if (!lmv->tgts)
return -ENOMEM;
- lmv->tgts_size = 32;
obd_str2uuid(&lmv->desc.ld_uuid, desc->ld_uuid.uuid);
lmv->desc.ld_tgt_count = 0;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 044/124] staging: lustre: misc: Reduce exposure to overflow on page counters.
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Stephen Champion, James Simmons
From: Stephen Champion <schamp@sgi.com>
When the number of an object in use or circulation is tied to memory
size of the system, very large memory systems can overflow 32 bit
counters. This patch addresses overflow on page counters in the osc LRU
and obd accounting.
Signed-off-by: Stephen Champion <schamp@sgi.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4856
Reviewed-on: http://review.whamcloud.com/10537
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 4 +-
drivers/staging/lustre/lustre/include/obd.h | 8 +-
.../staging/lustre/lustre/include/obd_support.h | 6 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lib.c | 6 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 9 +-
drivers/staging/lustre/lustre/llite/lproc_llite.c | 50 +++++++-----
drivers/staging/lustre/lustre/obdclass/cl_page.c | 4 +-
drivers/staging/lustre/lustre/obdclass/class_obd.c | 6 +-
.../lustre/lustre/obdclass/linux/linux-sysctl.c | 3 +-
drivers/staging/lustre/lustre/osc/lproc_osc.c | 31 ++++---
drivers/staging/lustre/lustre/osc/osc_cache.c | 27 +++---
drivers/staging/lustre/lustre/osc/osc_internal.h | 6 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 14 ++--
drivers/staging/lustre/lustre/osc/osc_page.c | 88 +++++++++----------
drivers/staging/lustre/lustre/osc/osc_request.c | 18 ++--
15 files changed, 146 insertions(+), 134 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 41e0801..5039643 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -2326,7 +2326,7 @@ struct cl_client_cache {
/**
* # of LRU entries available
*/
- atomic_t ccc_lru_left;
+ atomic_long_t ccc_lru_left;
/**
* List of entities(OSCs) for this LRU cache
*/
@@ -2346,7 +2346,7 @@ struct cl_client_cache {
/**
* # of unstable pages for this mount point
*/
- atomic_t ccc_unstable_nr;
+ atomic_long_t ccc_unstable_nr;
/**
* Waitq for awaiting unstable pages to reach zero.
* Used at umounting time and signaled on BRW commit
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 27fb4d7..3fbb873 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -293,13 +293,13 @@ struct client_obd {
/* lru for osc caching pages */
struct cl_client_cache *cl_cache;
struct list_head cl_lru_osc; /* member of cl_cache->ccc_lru */
- atomic_t *cl_lru_left;
- atomic_t cl_lru_busy;
+ atomic_long_t *cl_lru_left;
+ atomic_long_t cl_lru_busy;
+ atomic_long_t cl_lru_in_list;
atomic_t cl_lru_shrinkers;
- atomic_t cl_lru_in_list;
struct list_head cl_lru_list; /* lru page list */
spinlock_t cl_lru_list_lock; /* page list protector */
- atomic_t cl_unstable_count;
+ atomic_long_t cl_unstable_count;
/* number of in flight destroy rpcs is limited to max_rpcs_in_flight */
atomic_t cl_destroy_in_flight;
diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h
index 4d7a5c8..d4c41d0 100644
--- a/drivers/staging/lustre/lustre/include/obd_support.h
+++ b/drivers/staging/lustre/lustre/include/obd_support.h
@@ -52,9 +52,9 @@ extern unsigned int at_max;
extern unsigned int at_history;
extern int at_early_margin;
extern int at_extra;
-extern unsigned int obd_max_dirty_pages;
-extern atomic_t obd_dirty_pages;
-extern atomic_t obd_dirty_transit_pages;
+extern unsigned long obd_max_dirty_pages;
+extern atomic_long_t obd_dirty_pages;
+extern atomic_long_t obd_dirty_transit_pages;
extern char obd_jobid_var[];
/* Some hash init argument constants */
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
index 0d466e2..9a5f56a 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
@@ -328,11 +328,11 @@ int client_obd_setup(struct obd_device *obddev, struct lustre_cfg *lcfg)
/* lru for osc. */
INIT_LIST_HEAD(&cli->cl_lru_osc);
atomic_set(&cli->cl_lru_shrinkers, 0);
- atomic_set(&cli->cl_lru_busy, 0);
- atomic_set(&cli->cl_lru_in_list, 0);
+ atomic_long_set(&cli->cl_lru_busy, 0);
+ atomic_long_set(&cli->cl_lru_in_list, 0);
INIT_LIST_HEAD(&cli->cl_lru_list);
spin_lock_init(&cli->cl_lru_list_lock);
- atomic_set(&cli->cl_unstable_count, 0);
+ atomic_long_set(&cli->cl_unstable_count, 0);
init_waitqueue_head(&cli->cl_destroy_waitq);
atomic_set(&cli->cl_destroy_in_flight, 0);
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 19d18e8..2aab396 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -926,7 +926,8 @@ void ll_put_super(struct super_block *sb)
struct lustre_sb_info *lsi = s2lsi(sb);
struct ll_sb_info *sbi = ll_s2sbi(sb);
char *profilenm = get_profile_name(sb);
- int ccc_count, next, force = 1, rc = 0;
+ int next, force = 1, rc = 0;
+ long ccc_count;
CDEBUG(D_VFSTRACE, "VFS Op: sb %p - %s\n", sb, profilenm);
@@ -947,13 +948,13 @@ void ll_put_super(struct super_block *sb)
struct l_wait_info lwi = LWI_INTR(LWI_ON_SIGNAL_NOOP, NULL);
rc = l_wait_event(sbi->ll_cache->ccc_unstable_waitq,
- !atomic_read(&sbi->ll_cache->ccc_unstable_nr),
+ !atomic_long_read(&sbi->ll_cache->ccc_unstable_nr),
&lwi);
}
- ccc_count = atomic_read(&sbi->ll_cache->ccc_unstable_nr);
+ ccc_count = atomic_long_read(&sbi->ll_cache->ccc_unstable_nr);
if (!force && rc != -EINTR)
- LASSERTF(!ccc_count, "count: %i\n", ccc_count);
+ LASSERTF(!ccc_count, "count: %li\n", ccc_count);
/* We need to set force before the lov_disconnect in
* lustre_common_put_super, since l_d cleans up osc's as well.
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index 6123435..188fd37 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -357,16 +357,16 @@ static int ll_max_cached_mb_seq_show(struct seq_file *m, void *v)
struct ll_sb_info *sbi = ll_s2sbi(sb);
struct cl_client_cache *cache = sbi->ll_cache;
int shift = 20 - PAGE_SHIFT;
- int max_cached_mb;
- int unused_mb;
+ long max_cached_mb;
+ long unused_mb;
max_cached_mb = cache->ccc_lru_max >> shift;
- unused_mb = atomic_read(&cache->ccc_lru_left) >> shift;
+ unused_mb = atomic_long_read(&cache->ccc_lru_left) >> shift;
seq_printf(m,
"users: %d\n"
- "max_cached_mb: %d\n"
- "used_mb: %d\n"
- "unused_mb: %d\n"
+ "max_cached_mb: %ld\n"
+ "used_mb: %ld\n"
+ "unused_mb: %ld\n"
"reclaim_count: %u\n",
atomic_read(&cache->ccc_users),
max_cached_mb,
@@ -384,10 +384,13 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
struct ll_sb_info *sbi = ll_s2sbi(sb);
struct cl_client_cache *cache = sbi->ll_cache;
struct lu_env *env;
+ long diff = 0;
+ long nrpages = 0;
int refcheck;
- int mult, rc, pages_number;
- int diff = 0;
- int nrpages = 0;
+ long pages_number;
+ int mult;
+ long rc;
+ u64 val;
char kernbuf[128];
if (count >= sizeof(kernbuf))
@@ -400,10 +403,14 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
mult = 1 << (20 - PAGE_SHIFT);
buffer += lprocfs_find_named_value(kernbuf, "max_cached_mb:", &count) -
kernbuf;
- rc = lprocfs_write_frac_helper(buffer, count, &pages_number, mult);
+ rc = lprocfs_write_frac_u64_helper(buffer, count, &val, mult);
if (rc)
return rc;
+ if (val > LONG_MAX)
+ return -ERANGE;
+ pages_number = (long)val;
+
if (pages_number < 0 || pages_number > totalram_pages) {
CERROR("%s: can't set max cache more than %lu MB\n",
ll_get_fsname(sb, NULL, 0),
@@ -417,7 +424,7 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
/* easy - add more LRU slots. */
if (diff >= 0) {
- atomic_add(diff, &cache->ccc_lru_left);
+ atomic_long_add(diff, &cache->ccc_lru_left);
rc = 0;
goto out;
}
@@ -428,18 +435,18 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
diff = -diff;
while (diff > 0) {
- int tmp;
+ long tmp;
/* reduce LRU budget from free slots. */
do {
- int ov, nv;
+ long ov, nv;
- ov = atomic_read(&cache->ccc_lru_left);
+ ov = atomic_long_read(&cache->ccc_lru_left);
if (ov == 0)
break;
nv = ov > diff ? ov - diff : 0;
- rc = atomic_cmpxchg(&cache->ccc_lru_left, ov, nv);
+ rc = atomic_long_cmpxchg(&cache->ccc_lru_left, ov, nv);
if (likely(ov == rc)) {
diff -= ov - nv;
nrpages += ov - nv;
@@ -473,7 +480,7 @@ out:
spin_unlock(&sbi->ll_lock);
rc = count;
} else {
- atomic_add(nrpages, &cache->ccc_lru_left);
+ atomic_long_add(nrpages, &cache->ccc_lru_left);
}
return rc;
}
@@ -822,14 +829,15 @@ static ssize_t unstable_stats_show(struct kobject *kobj,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kobj);
struct cl_client_cache *cache = sbi->ll_cache;
- int pages, mb;
+ long pages;
+ int mb;
- pages = atomic_read(&cache->ccc_unstable_nr);
+ pages = atomic_long_read(&cache->ccc_unstable_nr);
mb = (pages * PAGE_SIZE) >> 20;
- return sprintf(buf, "unstable_check: %8d\n"
- "unstable_pages: %8d\n"
- "unstable_mb: %8d\n",
+ return sprintf(buf, "unstable_check: %8d\n"
+ "unstable_pages: %12ld\n"
+ "unstable_mb: %8d\n",
cache->ccc_unstable_check, pages, mb);
}
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index cae8d21..5c89dbd 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -1073,11 +1073,11 @@ struct cl_client_cache *cl_cache_init(unsigned long lru_page_max)
/* Initialize cache data */
atomic_set(&cache->ccc_users, 1);
cache->ccc_lru_max = lru_page_max;
- atomic_set(&cache->ccc_lru_left, lru_page_max);
+ atomic_long_set(&cache->ccc_lru_left, lru_page_max);
spin_lock_init(&cache->ccc_lru_lock);
INIT_LIST_HEAD(&cache->ccc_lru);
- atomic_set(&cache->ccc_unstable_nr, 0);
+ atomic_long_set(&cache->ccc_unstable_nr, 0);
init_waitqueue_head(&cache->ccc_unstable_waitq);
return cache;
diff --git a/drivers/staging/lustre/lustre/obdclass/class_obd.c b/drivers/staging/lustre/lustre/obdclass/class_obd.c
index 2293f6a..629d8b5 100644
--- a/drivers/staging/lustre/lustre/obdclass/class_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/class_obd.c
@@ -55,9 +55,9 @@ unsigned int obd_dump_on_timeout;
EXPORT_SYMBOL(obd_dump_on_timeout);
unsigned int obd_dump_on_eviction;
EXPORT_SYMBOL(obd_dump_on_eviction);
-unsigned int obd_max_dirty_pages = 256;
+unsigned long obd_max_dirty_pages;
EXPORT_SYMBOL(obd_max_dirty_pages);
-atomic_t obd_dirty_pages;
+atomic_long_t obd_dirty_pages;
EXPORT_SYMBOL(obd_dirty_pages);
unsigned int obd_timeout = OBD_TIMEOUT_DEFAULT; /* seconds */
EXPORT_SYMBOL(obd_timeout);
@@ -75,7 +75,7 @@ EXPORT_SYMBOL(at_early_margin);
int at_extra = 30;
EXPORT_SYMBOL(at_extra);
-atomic_t obd_dirty_transit_pages;
+atomic_long_t obd_dirty_transit_pages;
EXPORT_SYMBOL(obd_dirty_transit_pages);
char obd_jobid_var[JOBSTATS_JOBID_VAR_MAX_LEN + 1] = JOBSTATS_DISABLE;
diff --git a/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c b/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c
index aea1abd..e6c785a 100644
--- a/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c
+++ b/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c
@@ -97,8 +97,7 @@ static ssize_t max_dirty_mb_show(struct kobject *kobj, struct attribute *attr,
char *buf)
{
return sprintf(buf, "%lu\n",
- (unsigned long)obd_max_dirty_pages /
- (1 << (20 - PAGE_SHIFT)));
+ obd_max_dirty_pages / (1 << (20 - PAGE_SHIFT)));
}
static ssize_t max_dirty_mb_store(struct kobject *kobj, struct attribute *attr,
diff --git a/drivers/staging/lustre/lustre/osc/lproc_osc.c b/drivers/staging/lustre/lustre/osc/lproc_osc.c
index 9172b78..f0062d4 100644
--- a/drivers/staging/lustre/lustre/osc/lproc_osc.c
+++ b/drivers/staging/lustre/lustre/osc/lproc_osc.c
@@ -182,11 +182,11 @@ static int osc_cached_mb_seq_show(struct seq_file *m, void *v)
int shift = 20 - PAGE_SHIFT;
seq_printf(m,
- "used_mb: %d\n"
- "busy_cnt: %d\n",
- (atomic_read(&cli->cl_lru_in_list) +
- atomic_read(&cli->cl_lru_busy)) >> shift,
- atomic_read(&cli->cl_lru_busy));
+ "used_mb: %ld\n"
+ "busy_cnt: %ld\n",
+ (atomic_long_read(&cli->cl_lru_in_list) +
+ atomic_long_read(&cli->cl_lru_busy)) >> shift,
+ atomic_long_read(&cli->cl_lru_busy));
return 0;
}
@@ -198,8 +198,10 @@ static ssize_t osc_cached_mb_seq_write(struct file *file,
{
struct obd_device *dev = ((struct seq_file *)file->private_data)->private;
struct client_obd *cli = &dev->u.cli;
- int pages_number, mult, rc;
+ long pages_number, rc;
char kernbuf[128];
+ int mult;
+ u64 val;
if (count >= sizeof(kernbuf))
return -EINVAL;
@@ -211,14 +213,18 @@ static ssize_t osc_cached_mb_seq_write(struct file *file,
mult = 1 << (20 - PAGE_SHIFT);
buffer += lprocfs_find_named_value(kernbuf, "used_mb:", &count) -
kernbuf;
- rc = lprocfs_write_frac_helper(buffer, count, &pages_number, mult);
+ rc = lprocfs_write_frac_u64_helper(buffer, count, &val, mult);
if (rc)
return rc;
+ if (val > LONG_MAX)
+ return -ERANGE;
+ pages_number = (long)val;
+
if (pages_number < 0)
return -ERANGE;
- rc = atomic_read(&cli->cl_lru_in_list) - pages_number;
+ rc = atomic_long_read(&cli->cl_lru_in_list) - pages_number;
if (rc > 0) {
struct lu_env *env;
int refcheck;
@@ -598,13 +604,14 @@ static ssize_t unstable_stats_show(struct kobject *kobj,
struct obd_device *dev = container_of(kobj, struct obd_device,
obd_kobj);
struct client_obd *cli = &dev->u.cli;
- int pages, mb;
+ long pages;
+ int mb;
- pages = atomic_read(&cli->cl_unstable_count);
+ pages = atomic_long_read(&cli->cl_unstable_count);
mb = (pages * PAGE_SIZE) >> 20;
- return sprintf(buf, "unstable_pages: %8d\n"
- "unstable_mb: %8d\n", pages, mb);
+ return sprintf(buf, "unstable_pages: %20ld\n"
+ "unstable_mb: %10d\n", pages, mb);
}
LUSTRE_RO_ATTR(unstable_stats);
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 97f936e..26ad360 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1383,16 +1383,16 @@ static int osc_completion(const struct lu_env *env, struct osc_async_page *oap,
#define OSC_DUMP_GRANT(lvl, cli, fmt, args...) do { \
struct client_obd *__tmp = (cli); \
- CDEBUG(lvl, "%s: grant { dirty: %ld/%ld dirty_pages: %d/%d " \
+ CDEBUG(lvl, "%s: grant { dirty: %ld/%ld dirty_pages: %ld/%lu " \
"dropped: %ld avail: %ld, reserved: %ld, flight: %d }" \
- "lru {in list: %d, left: %d, waiters: %d }" fmt, \
+ "lru {in list: %ld, left: %ld, waiters: %d }" fmt, \
__tmp->cl_import->imp_obd->obd_name, \
__tmp->cl_dirty_pages, __tmp->cl_dirty_max_pages, \
- atomic_read(&obd_dirty_pages), obd_max_dirty_pages, \
+ atomic_long_read(&obd_dirty_pages), obd_max_dirty_pages, \
__tmp->cl_lost_grant, __tmp->cl_avail_grant, \
__tmp->cl_reserved_grant, __tmp->cl_w_in_flight, \
- atomic_read(&__tmp->cl_lru_in_list), \
- atomic_read(&__tmp->cl_lru_busy), \
+ atomic_long_read(&__tmp->cl_lru_in_list), \
+ atomic_long_read(&__tmp->cl_lru_busy), \
atomic_read(&__tmp->cl_lru_shrinkers), ##args); \
} while (0)
@@ -1402,7 +1402,7 @@ static void osc_consume_write_grant(struct client_obd *cli,
{
assert_spin_locked(&cli->cl_loi_list_lock);
LASSERT(!(pga->flag & OBD_BRW_FROM_GRANT));
- atomic_inc(&obd_dirty_pages);
+ atomic_long_inc(&obd_dirty_pages);
cli->cl_dirty_pages++;
pga->flag |= OBD_BRW_FROM_GRANT;
CDEBUG(D_CACHE, "using %lu grant credits for brw %p page %p\n",
@@ -1422,11 +1422,11 @@ static void osc_release_write_grant(struct client_obd *cli,
}
pga->flag &= ~OBD_BRW_FROM_GRANT;
- atomic_dec(&obd_dirty_pages);
+ atomic_long_dec(&obd_dirty_pages);
cli->cl_dirty_pages--;
if (pga->flag & OBD_BRW_NOCACHE) {
pga->flag &= ~OBD_BRW_NOCACHE;
- atomic_dec(&obd_dirty_transit_pages);
+ atomic_long_dec(&obd_dirty_transit_pages);
cli->cl_dirty_transit--;
}
}
@@ -1495,7 +1495,7 @@ static void osc_free_grant(struct client_obd *cli, unsigned int nr_pages,
int grant = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
spin_lock(&cli->cl_loi_list_lock);
- atomic_sub(nr_pages, &obd_dirty_pages);
+ atomic_long_sub(nr_pages, &obd_dirty_pages);
cli->cl_dirty_pages -= nr_pages;
cli->cl_lost_grant += lost_grant;
if (cli->cl_avail_grant < grant && cli->cl_lost_grant >= grant) {
@@ -1540,11 +1540,11 @@ static int osc_enter_cache_try(struct client_obd *cli,
return 0;
if (cli->cl_dirty_pages <= cli->cl_dirty_max_pages &&
- atomic_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages) {
+ atomic_long_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages) {
osc_consume_write_grant(cli, &oap->oap_brw_page);
if (transient) {
cli->cl_dirty_transit++;
- atomic_inc(&obd_dirty_transit_pages);
+ atomic_long_inc(&obd_dirty_transit_pages);
oap->oap_brw_flags |= OBD_BRW_NOCACHE;
}
rc = 1;
@@ -1668,8 +1668,9 @@ void osc_wake_cache_waiters(struct client_obd *cli)
ocw->ocw_rc = -EDQUOT;
/* we can't dirty more */
if ((cli->cl_dirty_pages > cli->cl_dirty_max_pages) ||
- (atomic_read(&obd_dirty_pages) + 1 > obd_max_dirty_pages)) {
- CDEBUG(D_CACHE, "no dirty room: dirty: %ld osc max %ld, sys max %d\n",
+ (atomic_long_read(&obd_dirty_pages) + 1 >
+ obd_max_dirty_pages)) {
+ CDEBUG(D_CACHE, "no dirty room: dirty: %ld osc max %ld, sys max %ld\n",
cli->cl_dirty_pages, cli->cl_dirty_max_pages,
obd_max_dirty_pages);
goto wakeup;
diff --git a/drivers/staging/lustre/lustre/osc/osc_internal.h b/drivers/staging/lustre/lustre/osc/osc_internal.h
index eca5fef..67fe0a2 100644
--- a/drivers/staging/lustre/lustre/osc/osc_internal.h
+++ b/drivers/staging/lustre/lustre/osc/osc_internal.h
@@ -133,9 +133,9 @@ int osc_sync_base(struct obd_export *exp, struct obd_info *oinfo,
int osc_process_config_base(struct obd_device *obd, struct lustre_cfg *cfg);
int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
struct list_head *ext_list, int cmd);
-int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
- int target, bool force);
-int osc_lru_reclaim(struct client_obd *cli);
+long osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
+ long target, bool force);
+long osc_lru_reclaim(struct client_obd *cli);
unsigned long osc_ldlm_weigh_ast(struct ldlm_lock *dlmlock);
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index f6db60c..7698054 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -319,8 +319,8 @@ static int osc_io_rw_iter_init(const struct lu_env *env,
struct osc_object *osc = cl2osc(ios->cis_obj);
struct client_obd *cli = osc_cli(osc);
unsigned long c;
- unsigned int npages;
- unsigned int max_pages;
+ unsigned long npages;
+ unsigned long max_pages;
if (cl_io_is_append(io))
return 0;
@@ -333,15 +333,15 @@ static int osc_io_rw_iter_init(const struct lu_env *env,
if (npages > max_pages)
npages = max_pages;
- c = atomic_read(cli->cl_lru_left);
+ c = atomic_long_read(cli->cl_lru_left);
if (c < npages && osc_lru_reclaim(cli) > 0)
- c = atomic_read(cli->cl_lru_left);
+ c = atomic_long_read(cli->cl_lru_left);
while (c >= npages) {
- if (c == atomic_cmpxchg(cli->cl_lru_left, c, c - npages)) {
+ if (c == atomic_long_cmpxchg(cli->cl_lru_left, c, c - npages)) {
oio->oi_lru_reserved = npages;
break;
}
- c = atomic_read(cli->cl_lru_left);
+ c = atomic_long_read(cli->cl_lru_left);
}
return 0;
@@ -355,7 +355,7 @@ static void osc_io_rw_iter_fini(const struct lu_env *env,
struct client_obd *cli = osc_cli(osc);
if (oio->oi_lru_reserved > 0) {
- atomic_add(oio->oi_lru_reserved, cli->cl_lru_left);
+ atomic_long_add(oio->oi_lru_reserved, cli->cl_lru_left);
oio->oi_lru_reserved = 0;
}
oio->oi_write_osclock = NULL;
diff --git a/drivers/staging/lustre/lustre/osc/osc_page.c b/drivers/staging/lustre/lustre/osc/osc_page.c
index c8889ea..2a7a70a 100644
--- a/drivers/staging/lustre/lustre/osc/osc_page.c
+++ b/drivers/staging/lustre/lustre/osc/osc_page.c
@@ -380,7 +380,7 @@ static const int lru_shrink_max = 8 << (20 - PAGE_SHIFT); /* 8M */
static int osc_cache_too_much(struct client_obd *cli)
{
struct cl_client_cache *cache = cli->cl_cache;
- int pages = atomic_read(&cli->cl_lru_in_list);
+ long pages = atomic_long_read(&cli->cl_lru_in_list);
unsigned long budget;
budget = cache->ccc_lru_max / (atomic_read(&cache->ccc_users) - 2);
@@ -388,7 +388,7 @@ static int osc_cache_too_much(struct client_obd *cli)
/* if it's going to run out LRU slots, we should free some, but not
* too much to maintain fairness among OSCs.
*/
- if (atomic_read(cli->cl_lru_left) < cache->ccc_lru_max >> 4) {
+ if (atomic_long_read(cli->cl_lru_left) < cache->ccc_lru_max >> 4) {
if (pages >= budget)
return lru_shrink_max;
else if (pages >= budget / 2)
@@ -415,7 +415,7 @@ void osc_lru_add_batch(struct client_obd *cli, struct list_head *plist)
{
LIST_HEAD(lru);
struct osc_async_page *oap;
- int npages = 0;
+ long npages = 0;
list_for_each_entry(oap, plist, oap_pending_item) {
struct osc_page *opg = oap2osc_page(oap);
@@ -431,8 +431,8 @@ void osc_lru_add_batch(struct client_obd *cli, struct list_head *plist)
if (npages > 0) {
spin_lock(&cli->cl_lru_list_lock);
list_splice_tail(&lru, &cli->cl_lru_list);
- atomic_sub(npages, &cli->cl_lru_busy);
- atomic_add(npages, &cli->cl_lru_in_list);
+ atomic_long_sub(npages, &cli->cl_lru_busy);
+ atomic_long_add(npages, &cli->cl_lru_in_list);
spin_unlock(&cli->cl_lru_list_lock);
/* XXX: May set force to be true for better performance */
@@ -443,9 +443,9 @@ void osc_lru_add_batch(struct client_obd *cli, struct list_head *plist)
static void __osc_lru_del(struct client_obd *cli, struct osc_page *opg)
{
- LASSERT(atomic_read(&cli->cl_lru_in_list) > 0);
+ LASSERT(atomic_long_read(&cli->cl_lru_in_list) > 0);
list_del_init(&opg->ops_lru);
- atomic_dec(&cli->cl_lru_in_list);
+ atomic_long_dec(&cli->cl_lru_in_list);
}
/**
@@ -459,12 +459,12 @@ static void osc_lru_del(struct client_obd *cli, struct osc_page *opg)
if (!list_empty(&opg->ops_lru)) {
__osc_lru_del(cli, opg);
} else {
- LASSERT(atomic_read(&cli->cl_lru_busy) > 0);
- atomic_dec(&cli->cl_lru_busy);
+ LASSERT(atomic_long_read(&cli->cl_lru_busy) > 0);
+ atomic_long_dec(&cli->cl_lru_busy);
}
spin_unlock(&cli->cl_lru_list_lock);
- atomic_inc(cli->cl_lru_left);
+ atomic_long_inc(cli->cl_lru_left);
/* this is a great place to release more LRU pages if
* this osc occupies too many LRU pages and kernel is
* stealing one of them.
@@ -489,7 +489,7 @@ static void osc_lru_use(struct client_obd *cli, struct osc_page *opg)
spin_lock(&cli->cl_lru_list_lock);
__osc_lru_del(cli, opg);
spin_unlock(&cli->cl_lru_list_lock);
- atomic_inc(&cli->cl_lru_busy);
+ atomic_long_inc(&cli->cl_lru_busy);
}
}
@@ -535,8 +535,8 @@ static inline bool lru_page_busy(struct client_obd *cli, struct cl_page *page)
/**
* Drop @target of pages from LRU at most.
*/
-int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
- int target, bool force)
+long osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
+ long target, bool force)
{
struct cl_io *io;
struct cl_object *clobj = NULL;
@@ -544,12 +544,12 @@ int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
struct osc_page *opg;
struct osc_page *temp;
int maxscan = 0;
- int count = 0;
+ long count = 0;
int index = 0;
int rc = 0;
- LASSERT(atomic_read(&cli->cl_lru_in_list) >= 0);
- if (atomic_read(&cli->cl_lru_in_list) == 0 || target <= 0)
+ LASSERT(atomic_long_read(&cli->cl_lru_in_list) >= 0);
+ if (atomic_long_read(&cli->cl_lru_in_list) == 0 || target <= 0)
return 0;
if (!force) {
@@ -568,7 +568,7 @@ int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
io = &osc_env_info(env)->oti_io;
spin_lock(&cli->cl_lru_list_lock);
- maxscan = min(target << 1, atomic_read(&cli->cl_lru_in_list));
+ maxscan = min(target << 1, atomic_long_read(&cli->cl_lru_in_list));
list_for_each_entry_safe(opg, temp, &cli->cl_lru_list, ops_lru) {
struct cl_page *page;
bool will_free = false;
@@ -656,24 +656,19 @@ int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
atomic_dec(&cli->cl_lru_shrinkers);
if (count > 0) {
- atomic_add(count, cli->cl_lru_left);
+ atomic_long_add(count, cli->cl_lru_left);
wake_up_all(&osc_lru_waitq);
}
return count > 0 ? count : rc;
}
-static inline int max_to_shrink(struct client_obd *cli)
-{
- return min(atomic_read(&cli->cl_lru_in_list) >> 1, lru_shrink_max);
-}
-
-int osc_lru_reclaim(struct client_obd *cli)
+long osc_lru_reclaim(struct client_obd *cli)
{
struct cl_env_nest nest;
struct lu_env *env;
struct cl_client_cache *cache = cli->cl_cache;
int max_scans;
- int rc = 0;
+ long rc = 0;
LASSERT(cache);
@@ -686,15 +681,15 @@ int osc_lru_reclaim(struct client_obd *cli)
if (rc == -EBUSY)
rc = 0;
- CDEBUG(D_CACHE, "%s: Free %d pages from own LRU: %p.\n",
+ CDEBUG(D_CACHE, "%s: Free %ld pages from own LRU: %p.\n",
cli->cl_import->imp_obd->obd_name, rc, cli);
goto out;
}
- CDEBUG(D_CACHE, "%s: cli %p no free slots, pages: %d, busy: %d.\n",
+ CDEBUG(D_CACHE, "%s: cli %p no free slots, pages: %ld, busy: %ld.\n",
cli->cl_import->imp_obd->obd_name, cli,
- atomic_read(&cli->cl_lru_in_list),
- atomic_read(&cli->cl_lru_busy));
+ atomic_long_read(&cli->cl_lru_in_list),
+ atomic_long_read(&cli->cl_lru_busy));
/* Reclaim LRU slots from other client_obd as it can't free enough
* from its own. This should rarely happen.
@@ -710,10 +705,10 @@ int osc_lru_reclaim(struct client_obd *cli)
cli = list_entry(cache->ccc_lru.next, struct client_obd,
cl_lru_osc);
- CDEBUG(D_CACHE, "%s: cli %p LRU pages: %d, busy: %d.\n",
+ CDEBUG(D_CACHE, "%s: cli %p LRU pages: %ld, busy: %ld.\n",
cli->cl_import->imp_obd->obd_name, cli,
- atomic_read(&cli->cl_lru_in_list),
- atomic_read(&cli->cl_lru_busy));
+ atomic_long_read(&cli->cl_lru_in_list),
+ atomic_long_read(&cli->cl_lru_busy));
list_move_tail(&cli->cl_lru_osc, &cache->ccc_lru);
if (osc_cache_too_much(cli) > 0) {
@@ -730,7 +725,7 @@ int osc_lru_reclaim(struct client_obd *cli)
out:
cl_env_nested_put(&nest, env);
- CDEBUG(D_CACHE, "%s: cli %p freed %d pages.\n",
+ CDEBUG(D_CACHE, "%s: cli %p freed %ld pages.\n",
cli->cl_import->imp_obd->obd_name, cli, rc);
return rc;
}
@@ -758,8 +753,8 @@ static int osc_lru_reserve(const struct lu_env *env, struct osc_object *obj,
goto out;
}
- LASSERT(atomic_read(cli->cl_lru_left) >= 0);
- while (!atomic_add_unless(cli->cl_lru_left, -1, 0)) {
+ LASSERT(atomic_long_read(cli->cl_lru_left) >= 0);
+ while (!atomic_long_add_unless(cli->cl_lru_left, -1, 0)) {
/* run out of LRU spaces, try to drop some by itself */
rc = osc_lru_reclaim(cli);
if (rc < 0)
@@ -770,7 +765,7 @@ static int osc_lru_reserve(const struct lu_env *env, struct osc_object *obj,
cond_resched();
rc = l_wait_event(osc_lru_waitq,
- atomic_read(cli->cl_lru_left) > 0,
+ atomic_long_read(cli->cl_lru_left) > 0,
&lwi);
if (rc < 0)
@@ -779,7 +774,7 @@ static int osc_lru_reserve(const struct lu_env *env, struct osc_object *obj,
out:
if (rc >= 0) {
- atomic_inc(&cli->cl_lru_busy);
+ atomic_long_inc(&cli->cl_lru_busy);
opg->ops_in_lru = 1;
rc = 0;
}
@@ -847,16 +842,17 @@ void osc_dec_unstable_pages(struct ptlrpc_request *req)
struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
struct ptlrpc_bulk_desc *desc = req->rq_bulk;
int page_count = desc->bd_iov_count;
- int unstable_count;
+ long unstable_count;
LASSERT(page_count >= 0);
dec_unstable_page_accounting(desc);
- unstable_count = atomic_sub_return(page_count, &cli->cl_unstable_count);
+ unstable_count = atomic_long_sub_return(page_count,
+ &cli->cl_unstable_count);
LASSERT(unstable_count >= 0);
- unstable_count = atomic_sub_return(page_count,
- &cli->cl_cache->ccc_unstable_nr);
+ unstable_count = atomic_long_sub_return(page_count,
+ &cli->cl_cache->ccc_unstable_nr);
LASSERT(unstable_count >= 0);
if (!unstable_count)
wake_up_all(&cli->cl_cache->ccc_unstable_waitq);
@@ -872,15 +868,15 @@ void osc_inc_unstable_pages(struct ptlrpc_request *req)
{
struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
struct ptlrpc_bulk_desc *desc = req->rq_bulk;
- int page_count = desc->bd_iov_count;
+ long page_count = desc->bd_iov_count;
/* No unstable page tracking */
if (!cli->cl_cache || !cli->cl_cache->ccc_unstable_check)
return;
add_unstable_page_accounting(desc);
- atomic_add(page_count, &cli->cl_unstable_count);
- atomic_add(page_count, &cli->cl_cache->ccc_unstable_nr);
+ atomic_long_add(page_count, &cli->cl_unstable_count);
+ atomic_long_add(page_count, &cli->cl_cache->ccc_unstable_nr);
/*
* If the request has already been committed (i.e. brw_commit
@@ -912,8 +908,8 @@ bool osc_over_unstable_soft_limit(struct client_obd *cli)
if (!cli->cl_cache || !cli->cl_cache->ccc_unstable_check)
return false;
- osc_unstable_count = atomic_read(&cli->cl_unstable_count);
- unstable_nr = atomic_read(&cli->cl_cache->ccc_unstable_nr);
+ osc_unstable_count = atomic_long_read(&cli->cl_unstable_count);
+ unstable_nr = atomic_long_read(&cli->cl_cache->ccc_unstable_nr);
CDEBUG(D_CACHE,
"%s: cli: %p unstable pages: %lu, osc unstable pages: %lu\n",
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 7f7c318..7449b77 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -804,17 +804,17 @@ static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
cli->cl_dirty_pages, cli->cl_dirty_transit,
cli->cl_dirty_max_pages);
oa->o_undirty = 0;
- } else if (unlikely(atomic_read(&obd_dirty_pages) -
- atomic_read(&obd_dirty_transit_pages) >
- (long)(obd_max_dirty_pages + 1))) {
+ } else if (unlikely(atomic_long_read(&obd_dirty_pages) -
+ atomic_long_read(&obd_dirty_transit_pages) >
+ (obd_max_dirty_pages + 1))) {
/* The atomic_read() allowing the atomic_inc() are
* not covered by a lock thus they may safely race and trip
* this CERROR() unless we add in a small fudge factor (+1).
*/
- CERROR("%s: dirty %d + %d > system dirty_max %d\n",
+ CERROR("%s: dirty %ld + %ld > system dirty_max %lu\n",
cli->cl_import->imp_obd->obd_name,
- atomic_read(&obd_dirty_pages),
- atomic_read(&obd_dirty_transit_pages),
+ atomic_long_read(&obd_dirty_pages),
+ atomic_long_read(&obd_dirty_transit_pages),
obd_max_dirty_pages);
oa->o_undirty = 0;
} else if (unlikely(cli->cl_dirty_max_pages - cli->cl_dirty_pages >
@@ -2920,11 +2920,11 @@ static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
if (KEY_IS(KEY_CACHE_LRU_SHRINK)) {
struct client_obd *cli = &obd->u.cli;
- int nr = atomic_read(&cli->cl_lru_in_list) >> 1;
- int target = *(int *)val;
+ long nr = atomic_long_read(&cli->cl_lru_in_list) >> 1;
+ long target = *(long *)val;
nr = osc_lru_shrink(env, cli, min(nr, target), true);
- *(int *)val -= nr;
+ *(long *)val -= nr;
return 0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 044/124] staging: lustre: misc: Reduce exposure to overflow on page counters.
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Stephen Champion, James Simmons
From: Stephen Champion <schamp@sgi.com>
When the number of an object in use or circulation is tied to memory
size of the system, very large memory systems can overflow 32 bit
counters. This patch addresses overflow on page counters in the osc LRU
and obd accounting.
Signed-off-by: Stephen Champion <schamp@sgi.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4856
Reviewed-on: http://review.whamcloud.com/10537
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 4 +-
drivers/staging/lustre/lustre/include/obd.h | 8 +-
.../staging/lustre/lustre/include/obd_support.h | 6 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lib.c | 6 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 9 +-
drivers/staging/lustre/lustre/llite/lproc_llite.c | 50 +++++++-----
drivers/staging/lustre/lustre/obdclass/cl_page.c | 4 +-
drivers/staging/lustre/lustre/obdclass/class_obd.c | 6 +-
.../lustre/lustre/obdclass/linux/linux-sysctl.c | 3 +-
drivers/staging/lustre/lustre/osc/lproc_osc.c | 31 ++++---
drivers/staging/lustre/lustre/osc/osc_cache.c | 27 +++---
drivers/staging/lustre/lustre/osc/osc_internal.h | 6 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 14 ++--
drivers/staging/lustre/lustre/osc/osc_page.c | 88 +++++++++----------
drivers/staging/lustre/lustre/osc/osc_request.c | 18 ++--
15 files changed, 146 insertions(+), 134 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 41e0801..5039643 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -2326,7 +2326,7 @@ struct cl_client_cache {
/**
* # of LRU entries available
*/
- atomic_t ccc_lru_left;
+ atomic_long_t ccc_lru_left;
/**
* List of entities(OSCs) for this LRU cache
*/
@@ -2346,7 +2346,7 @@ struct cl_client_cache {
/**
* # of unstable pages for this mount point
*/
- atomic_t ccc_unstable_nr;
+ atomic_long_t ccc_unstable_nr;
/**
* Waitq for awaiting unstable pages to reach zero.
* Used at umounting time and signaled on BRW commit
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 27fb4d7..3fbb873 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -293,13 +293,13 @@ struct client_obd {
/* lru for osc caching pages */
struct cl_client_cache *cl_cache;
struct list_head cl_lru_osc; /* member of cl_cache->ccc_lru */
- atomic_t *cl_lru_left;
- atomic_t cl_lru_busy;
+ atomic_long_t *cl_lru_left;
+ atomic_long_t cl_lru_busy;
+ atomic_long_t cl_lru_in_list;
atomic_t cl_lru_shrinkers;
- atomic_t cl_lru_in_list;
struct list_head cl_lru_list; /* lru page list */
spinlock_t cl_lru_list_lock; /* page list protector */
- atomic_t cl_unstable_count;
+ atomic_long_t cl_unstable_count;
/* number of in flight destroy rpcs is limited to max_rpcs_in_flight */
atomic_t cl_destroy_in_flight;
diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h
index 4d7a5c8..d4c41d0 100644
--- a/drivers/staging/lustre/lustre/include/obd_support.h
+++ b/drivers/staging/lustre/lustre/include/obd_support.h
@@ -52,9 +52,9 @@ extern unsigned int at_max;
extern unsigned int at_history;
extern int at_early_margin;
extern int at_extra;
-extern unsigned int obd_max_dirty_pages;
-extern atomic_t obd_dirty_pages;
-extern atomic_t obd_dirty_transit_pages;
+extern unsigned long obd_max_dirty_pages;
+extern atomic_long_t obd_dirty_pages;
+extern atomic_long_t obd_dirty_transit_pages;
extern char obd_jobid_var[];
/* Some hash init argument constants */
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
index 0d466e2..9a5f56a 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
@@ -328,11 +328,11 @@ int client_obd_setup(struct obd_device *obddev, struct lustre_cfg *lcfg)
/* lru for osc. */
INIT_LIST_HEAD(&cli->cl_lru_osc);
atomic_set(&cli->cl_lru_shrinkers, 0);
- atomic_set(&cli->cl_lru_busy, 0);
- atomic_set(&cli->cl_lru_in_list, 0);
+ atomic_long_set(&cli->cl_lru_busy, 0);
+ atomic_long_set(&cli->cl_lru_in_list, 0);
INIT_LIST_HEAD(&cli->cl_lru_list);
spin_lock_init(&cli->cl_lru_list_lock);
- atomic_set(&cli->cl_unstable_count, 0);
+ atomic_long_set(&cli->cl_unstable_count, 0);
init_waitqueue_head(&cli->cl_destroy_waitq);
atomic_set(&cli->cl_destroy_in_flight, 0);
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 19d18e8..2aab396 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -926,7 +926,8 @@ void ll_put_super(struct super_block *sb)
struct lustre_sb_info *lsi = s2lsi(sb);
struct ll_sb_info *sbi = ll_s2sbi(sb);
char *profilenm = get_profile_name(sb);
- int ccc_count, next, force = 1, rc = 0;
+ int next, force = 1, rc = 0;
+ long ccc_count;
CDEBUG(D_VFSTRACE, "VFS Op: sb %p - %s\n", sb, profilenm);
@@ -947,13 +948,13 @@ void ll_put_super(struct super_block *sb)
struct l_wait_info lwi = LWI_INTR(LWI_ON_SIGNAL_NOOP, NULL);
rc = l_wait_event(sbi->ll_cache->ccc_unstable_waitq,
- !atomic_read(&sbi->ll_cache->ccc_unstable_nr),
+ !atomic_long_read(&sbi->ll_cache->ccc_unstable_nr),
&lwi);
}
- ccc_count = atomic_read(&sbi->ll_cache->ccc_unstable_nr);
+ ccc_count = atomic_long_read(&sbi->ll_cache->ccc_unstable_nr);
if (!force && rc != -EINTR)
- LASSERTF(!ccc_count, "count: %i\n", ccc_count);
+ LASSERTF(!ccc_count, "count: %li\n", ccc_count);
/* We need to set force before the lov_disconnect in
* lustre_common_put_super, since l_d cleans up osc's as well.
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index 6123435..188fd37 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -357,16 +357,16 @@ static int ll_max_cached_mb_seq_show(struct seq_file *m, void *v)
struct ll_sb_info *sbi = ll_s2sbi(sb);
struct cl_client_cache *cache = sbi->ll_cache;
int shift = 20 - PAGE_SHIFT;
- int max_cached_mb;
- int unused_mb;
+ long max_cached_mb;
+ long unused_mb;
max_cached_mb = cache->ccc_lru_max >> shift;
- unused_mb = atomic_read(&cache->ccc_lru_left) >> shift;
+ unused_mb = atomic_long_read(&cache->ccc_lru_left) >> shift;
seq_printf(m,
"users: %d\n"
- "max_cached_mb: %d\n"
- "used_mb: %d\n"
- "unused_mb: %d\n"
+ "max_cached_mb: %ld\n"
+ "used_mb: %ld\n"
+ "unused_mb: %ld\n"
"reclaim_count: %u\n",
atomic_read(&cache->ccc_users),
max_cached_mb,
@@ -384,10 +384,13 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
struct ll_sb_info *sbi = ll_s2sbi(sb);
struct cl_client_cache *cache = sbi->ll_cache;
struct lu_env *env;
+ long diff = 0;
+ long nrpages = 0;
int refcheck;
- int mult, rc, pages_number;
- int diff = 0;
- int nrpages = 0;
+ long pages_number;
+ int mult;
+ long rc;
+ u64 val;
char kernbuf[128];
if (count >= sizeof(kernbuf))
@@ -400,10 +403,14 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
mult = 1 << (20 - PAGE_SHIFT);
buffer += lprocfs_find_named_value(kernbuf, "max_cached_mb:", &count) -
kernbuf;
- rc = lprocfs_write_frac_helper(buffer, count, &pages_number, mult);
+ rc = lprocfs_write_frac_u64_helper(buffer, count, &val, mult);
if (rc)
return rc;
+ if (val > LONG_MAX)
+ return -ERANGE;
+ pages_number = (long)val;
+
if (pages_number < 0 || pages_number > totalram_pages) {
CERROR("%s: can't set max cache more than %lu MB\n",
ll_get_fsname(sb, NULL, 0),
@@ -417,7 +424,7 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
/* easy - add more LRU slots. */
if (diff >= 0) {
- atomic_add(diff, &cache->ccc_lru_left);
+ atomic_long_add(diff, &cache->ccc_lru_left);
rc = 0;
goto out;
}
@@ -428,18 +435,18 @@ static ssize_t ll_max_cached_mb_seq_write(struct file *file,
diff = -diff;
while (diff > 0) {
- int tmp;
+ long tmp;
/* reduce LRU budget from free slots. */
do {
- int ov, nv;
+ long ov, nv;
- ov = atomic_read(&cache->ccc_lru_left);
+ ov = atomic_long_read(&cache->ccc_lru_left);
if (ov == 0)
break;
nv = ov > diff ? ov - diff : 0;
- rc = atomic_cmpxchg(&cache->ccc_lru_left, ov, nv);
+ rc = atomic_long_cmpxchg(&cache->ccc_lru_left, ov, nv);
if (likely(ov == rc)) {
diff -= ov - nv;
nrpages += ov - nv;
@@ -473,7 +480,7 @@ out:
spin_unlock(&sbi->ll_lock);
rc = count;
} else {
- atomic_add(nrpages, &cache->ccc_lru_left);
+ atomic_long_add(nrpages, &cache->ccc_lru_left);
}
return rc;
}
@@ -822,14 +829,15 @@ static ssize_t unstable_stats_show(struct kobject *kobj,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kobj);
struct cl_client_cache *cache = sbi->ll_cache;
- int pages, mb;
+ long pages;
+ int mb;
- pages = atomic_read(&cache->ccc_unstable_nr);
+ pages = atomic_long_read(&cache->ccc_unstable_nr);
mb = (pages * PAGE_SIZE) >> 20;
- return sprintf(buf, "unstable_check: %8d\n"
- "unstable_pages: %8d\n"
- "unstable_mb: %8d\n",
+ return sprintf(buf, "unstable_check: %8d\n"
+ "unstable_pages: %12ld\n"
+ "unstable_mb: %8d\n",
cache->ccc_unstable_check, pages, mb);
}
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index cae8d21..5c89dbd 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -1073,11 +1073,11 @@ struct cl_client_cache *cl_cache_init(unsigned long lru_page_max)
/* Initialize cache data */
atomic_set(&cache->ccc_users, 1);
cache->ccc_lru_max = lru_page_max;
- atomic_set(&cache->ccc_lru_left, lru_page_max);
+ atomic_long_set(&cache->ccc_lru_left, lru_page_max);
spin_lock_init(&cache->ccc_lru_lock);
INIT_LIST_HEAD(&cache->ccc_lru);
- atomic_set(&cache->ccc_unstable_nr, 0);
+ atomic_long_set(&cache->ccc_unstable_nr, 0);
init_waitqueue_head(&cache->ccc_unstable_waitq);
return cache;
diff --git a/drivers/staging/lustre/lustre/obdclass/class_obd.c b/drivers/staging/lustre/lustre/obdclass/class_obd.c
index 2293f6a..629d8b5 100644
--- a/drivers/staging/lustre/lustre/obdclass/class_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/class_obd.c
@@ -55,9 +55,9 @@ unsigned int obd_dump_on_timeout;
EXPORT_SYMBOL(obd_dump_on_timeout);
unsigned int obd_dump_on_eviction;
EXPORT_SYMBOL(obd_dump_on_eviction);
-unsigned int obd_max_dirty_pages = 256;
+unsigned long obd_max_dirty_pages;
EXPORT_SYMBOL(obd_max_dirty_pages);
-atomic_t obd_dirty_pages;
+atomic_long_t obd_dirty_pages;
EXPORT_SYMBOL(obd_dirty_pages);
unsigned int obd_timeout = OBD_TIMEOUT_DEFAULT; /* seconds */
EXPORT_SYMBOL(obd_timeout);
@@ -75,7 +75,7 @@ EXPORT_SYMBOL(at_early_margin);
int at_extra = 30;
EXPORT_SYMBOL(at_extra);
-atomic_t obd_dirty_transit_pages;
+atomic_long_t obd_dirty_transit_pages;
EXPORT_SYMBOL(obd_dirty_transit_pages);
char obd_jobid_var[JOBSTATS_JOBID_VAR_MAX_LEN + 1] = JOBSTATS_DISABLE;
diff --git a/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c b/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c
index aea1abd..e6c785a 100644
--- a/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c
+++ b/drivers/staging/lustre/lustre/obdclass/linux/linux-sysctl.c
@@ -97,8 +97,7 @@ static ssize_t max_dirty_mb_show(struct kobject *kobj, struct attribute *attr,
char *buf)
{
return sprintf(buf, "%lu\n",
- (unsigned long)obd_max_dirty_pages /
- (1 << (20 - PAGE_SHIFT)));
+ obd_max_dirty_pages / (1 << (20 - PAGE_SHIFT)));
}
static ssize_t max_dirty_mb_store(struct kobject *kobj, struct attribute *attr,
diff --git a/drivers/staging/lustre/lustre/osc/lproc_osc.c b/drivers/staging/lustre/lustre/osc/lproc_osc.c
index 9172b78..f0062d4 100644
--- a/drivers/staging/lustre/lustre/osc/lproc_osc.c
+++ b/drivers/staging/lustre/lustre/osc/lproc_osc.c
@@ -182,11 +182,11 @@ static int osc_cached_mb_seq_show(struct seq_file *m, void *v)
int shift = 20 - PAGE_SHIFT;
seq_printf(m,
- "used_mb: %d\n"
- "busy_cnt: %d\n",
- (atomic_read(&cli->cl_lru_in_list) +
- atomic_read(&cli->cl_lru_busy)) >> shift,
- atomic_read(&cli->cl_lru_busy));
+ "used_mb: %ld\n"
+ "busy_cnt: %ld\n",
+ (atomic_long_read(&cli->cl_lru_in_list) +
+ atomic_long_read(&cli->cl_lru_busy)) >> shift,
+ atomic_long_read(&cli->cl_lru_busy));
return 0;
}
@@ -198,8 +198,10 @@ static ssize_t osc_cached_mb_seq_write(struct file *file,
{
struct obd_device *dev = ((struct seq_file *)file->private_data)->private;
struct client_obd *cli = &dev->u.cli;
- int pages_number, mult, rc;
+ long pages_number, rc;
char kernbuf[128];
+ int mult;
+ u64 val;
if (count >= sizeof(kernbuf))
return -EINVAL;
@@ -211,14 +213,18 @@ static ssize_t osc_cached_mb_seq_write(struct file *file,
mult = 1 << (20 - PAGE_SHIFT);
buffer += lprocfs_find_named_value(kernbuf, "used_mb:", &count) -
kernbuf;
- rc = lprocfs_write_frac_helper(buffer, count, &pages_number, mult);
+ rc = lprocfs_write_frac_u64_helper(buffer, count, &val, mult);
if (rc)
return rc;
+ if (val > LONG_MAX)
+ return -ERANGE;
+ pages_number = (long)val;
+
if (pages_number < 0)
return -ERANGE;
- rc = atomic_read(&cli->cl_lru_in_list) - pages_number;
+ rc = atomic_long_read(&cli->cl_lru_in_list) - pages_number;
if (rc > 0) {
struct lu_env *env;
int refcheck;
@@ -598,13 +604,14 @@ static ssize_t unstable_stats_show(struct kobject *kobj,
struct obd_device *dev = container_of(kobj, struct obd_device,
obd_kobj);
struct client_obd *cli = &dev->u.cli;
- int pages, mb;
+ long pages;
+ int mb;
- pages = atomic_read(&cli->cl_unstable_count);
+ pages = atomic_long_read(&cli->cl_unstable_count);
mb = (pages * PAGE_SIZE) >> 20;
- return sprintf(buf, "unstable_pages: %8d\n"
- "unstable_mb: %8d\n", pages, mb);
+ return sprintf(buf, "unstable_pages: %20ld\n"
+ "unstable_mb: %10d\n", pages, mb);
}
LUSTRE_RO_ATTR(unstable_stats);
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 97f936e..26ad360 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1383,16 +1383,16 @@ static int osc_completion(const struct lu_env *env, struct osc_async_page *oap,
#define OSC_DUMP_GRANT(lvl, cli, fmt, args...) do { \
struct client_obd *__tmp = (cli); \
- CDEBUG(lvl, "%s: grant { dirty: %ld/%ld dirty_pages: %d/%d " \
+ CDEBUG(lvl, "%s: grant { dirty: %ld/%ld dirty_pages: %ld/%lu " \
"dropped: %ld avail: %ld, reserved: %ld, flight: %d }" \
- "lru {in list: %d, left: %d, waiters: %d }" fmt, \
+ "lru {in list: %ld, left: %ld, waiters: %d }" fmt, \
__tmp->cl_import->imp_obd->obd_name, \
__tmp->cl_dirty_pages, __tmp->cl_dirty_max_pages, \
- atomic_read(&obd_dirty_pages), obd_max_dirty_pages, \
+ atomic_long_read(&obd_dirty_pages), obd_max_dirty_pages, \
__tmp->cl_lost_grant, __tmp->cl_avail_grant, \
__tmp->cl_reserved_grant, __tmp->cl_w_in_flight, \
- atomic_read(&__tmp->cl_lru_in_list), \
- atomic_read(&__tmp->cl_lru_busy), \
+ atomic_long_read(&__tmp->cl_lru_in_list), \
+ atomic_long_read(&__tmp->cl_lru_busy), \
atomic_read(&__tmp->cl_lru_shrinkers), ##args); \
} while (0)
@@ -1402,7 +1402,7 @@ static void osc_consume_write_grant(struct client_obd *cli,
{
assert_spin_locked(&cli->cl_loi_list_lock);
LASSERT(!(pga->flag & OBD_BRW_FROM_GRANT));
- atomic_inc(&obd_dirty_pages);
+ atomic_long_inc(&obd_dirty_pages);
cli->cl_dirty_pages++;
pga->flag |= OBD_BRW_FROM_GRANT;
CDEBUG(D_CACHE, "using %lu grant credits for brw %p page %p\n",
@@ -1422,11 +1422,11 @@ static void osc_release_write_grant(struct client_obd *cli,
}
pga->flag &= ~OBD_BRW_FROM_GRANT;
- atomic_dec(&obd_dirty_pages);
+ atomic_long_dec(&obd_dirty_pages);
cli->cl_dirty_pages--;
if (pga->flag & OBD_BRW_NOCACHE) {
pga->flag &= ~OBD_BRW_NOCACHE;
- atomic_dec(&obd_dirty_transit_pages);
+ atomic_long_dec(&obd_dirty_transit_pages);
cli->cl_dirty_transit--;
}
}
@@ -1495,7 +1495,7 @@ static void osc_free_grant(struct client_obd *cli, unsigned int nr_pages,
int grant = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
spin_lock(&cli->cl_loi_list_lock);
- atomic_sub(nr_pages, &obd_dirty_pages);
+ atomic_long_sub(nr_pages, &obd_dirty_pages);
cli->cl_dirty_pages -= nr_pages;
cli->cl_lost_grant += lost_grant;
if (cli->cl_avail_grant < grant && cli->cl_lost_grant >= grant) {
@@ -1540,11 +1540,11 @@ static int osc_enter_cache_try(struct client_obd *cli,
return 0;
if (cli->cl_dirty_pages <= cli->cl_dirty_max_pages &&
- atomic_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages) {
+ atomic_long_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages) {
osc_consume_write_grant(cli, &oap->oap_brw_page);
if (transient) {
cli->cl_dirty_transit++;
- atomic_inc(&obd_dirty_transit_pages);
+ atomic_long_inc(&obd_dirty_transit_pages);
oap->oap_brw_flags |= OBD_BRW_NOCACHE;
}
rc = 1;
@@ -1668,8 +1668,9 @@ void osc_wake_cache_waiters(struct client_obd *cli)
ocw->ocw_rc = -EDQUOT;
/* we can't dirty more */
if ((cli->cl_dirty_pages > cli->cl_dirty_max_pages) ||
- (atomic_read(&obd_dirty_pages) + 1 > obd_max_dirty_pages)) {
- CDEBUG(D_CACHE, "no dirty room: dirty: %ld osc max %ld, sys max %d\n",
+ (atomic_long_read(&obd_dirty_pages) + 1 >
+ obd_max_dirty_pages)) {
+ CDEBUG(D_CACHE, "no dirty room: dirty: %ld osc max %ld, sys max %ld\n",
cli->cl_dirty_pages, cli->cl_dirty_max_pages,
obd_max_dirty_pages);
goto wakeup;
diff --git a/drivers/staging/lustre/lustre/osc/osc_internal.h b/drivers/staging/lustre/lustre/osc/osc_internal.h
index eca5fef..67fe0a2 100644
--- a/drivers/staging/lustre/lustre/osc/osc_internal.h
+++ b/drivers/staging/lustre/lustre/osc/osc_internal.h
@@ -133,9 +133,9 @@ int osc_sync_base(struct obd_export *exp, struct obd_info *oinfo,
int osc_process_config_base(struct obd_device *obd, struct lustre_cfg *cfg);
int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
struct list_head *ext_list, int cmd);
-int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
- int target, bool force);
-int osc_lru_reclaim(struct client_obd *cli);
+long osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
+ long target, bool force);
+long osc_lru_reclaim(struct client_obd *cli);
unsigned long osc_ldlm_weigh_ast(struct ldlm_lock *dlmlock);
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index f6db60c..7698054 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -319,8 +319,8 @@ static int osc_io_rw_iter_init(const struct lu_env *env,
struct osc_object *osc = cl2osc(ios->cis_obj);
struct client_obd *cli = osc_cli(osc);
unsigned long c;
- unsigned int npages;
- unsigned int max_pages;
+ unsigned long npages;
+ unsigned long max_pages;
if (cl_io_is_append(io))
return 0;
@@ -333,15 +333,15 @@ static int osc_io_rw_iter_init(const struct lu_env *env,
if (npages > max_pages)
npages = max_pages;
- c = atomic_read(cli->cl_lru_left);
+ c = atomic_long_read(cli->cl_lru_left);
if (c < npages && osc_lru_reclaim(cli) > 0)
- c = atomic_read(cli->cl_lru_left);
+ c = atomic_long_read(cli->cl_lru_left);
while (c >= npages) {
- if (c == atomic_cmpxchg(cli->cl_lru_left, c, c - npages)) {
+ if (c == atomic_long_cmpxchg(cli->cl_lru_left, c, c - npages)) {
oio->oi_lru_reserved = npages;
break;
}
- c = atomic_read(cli->cl_lru_left);
+ c = atomic_long_read(cli->cl_lru_left);
}
return 0;
@@ -355,7 +355,7 @@ static void osc_io_rw_iter_fini(const struct lu_env *env,
struct client_obd *cli = osc_cli(osc);
if (oio->oi_lru_reserved > 0) {
- atomic_add(oio->oi_lru_reserved, cli->cl_lru_left);
+ atomic_long_add(oio->oi_lru_reserved, cli->cl_lru_left);
oio->oi_lru_reserved = 0;
}
oio->oi_write_osclock = NULL;
diff --git a/drivers/staging/lustre/lustre/osc/osc_page.c b/drivers/staging/lustre/lustre/osc/osc_page.c
index c8889ea..2a7a70a 100644
--- a/drivers/staging/lustre/lustre/osc/osc_page.c
+++ b/drivers/staging/lustre/lustre/osc/osc_page.c
@@ -380,7 +380,7 @@ static const int lru_shrink_max = 8 << (20 - PAGE_SHIFT); /* 8M */
static int osc_cache_too_much(struct client_obd *cli)
{
struct cl_client_cache *cache = cli->cl_cache;
- int pages = atomic_read(&cli->cl_lru_in_list);
+ long pages = atomic_long_read(&cli->cl_lru_in_list);
unsigned long budget;
budget = cache->ccc_lru_max / (atomic_read(&cache->ccc_users) - 2);
@@ -388,7 +388,7 @@ static int osc_cache_too_much(struct client_obd *cli)
/* if it's going to run out LRU slots, we should free some, but not
* too much to maintain fairness among OSCs.
*/
- if (atomic_read(cli->cl_lru_left) < cache->ccc_lru_max >> 4) {
+ if (atomic_long_read(cli->cl_lru_left) < cache->ccc_lru_max >> 4) {
if (pages >= budget)
return lru_shrink_max;
else if (pages >= budget / 2)
@@ -415,7 +415,7 @@ void osc_lru_add_batch(struct client_obd *cli, struct list_head *plist)
{
LIST_HEAD(lru);
struct osc_async_page *oap;
- int npages = 0;
+ long npages = 0;
list_for_each_entry(oap, plist, oap_pending_item) {
struct osc_page *opg = oap2osc_page(oap);
@@ -431,8 +431,8 @@ void osc_lru_add_batch(struct client_obd *cli, struct list_head *plist)
if (npages > 0) {
spin_lock(&cli->cl_lru_list_lock);
list_splice_tail(&lru, &cli->cl_lru_list);
- atomic_sub(npages, &cli->cl_lru_busy);
- atomic_add(npages, &cli->cl_lru_in_list);
+ atomic_long_sub(npages, &cli->cl_lru_busy);
+ atomic_long_add(npages, &cli->cl_lru_in_list);
spin_unlock(&cli->cl_lru_list_lock);
/* XXX: May set force to be true for better performance */
@@ -443,9 +443,9 @@ void osc_lru_add_batch(struct client_obd *cli, struct list_head *plist)
static void __osc_lru_del(struct client_obd *cli, struct osc_page *opg)
{
- LASSERT(atomic_read(&cli->cl_lru_in_list) > 0);
+ LASSERT(atomic_long_read(&cli->cl_lru_in_list) > 0);
list_del_init(&opg->ops_lru);
- atomic_dec(&cli->cl_lru_in_list);
+ atomic_long_dec(&cli->cl_lru_in_list);
}
/**
@@ -459,12 +459,12 @@ static void osc_lru_del(struct client_obd *cli, struct osc_page *opg)
if (!list_empty(&opg->ops_lru)) {
__osc_lru_del(cli, opg);
} else {
- LASSERT(atomic_read(&cli->cl_lru_busy) > 0);
- atomic_dec(&cli->cl_lru_busy);
+ LASSERT(atomic_long_read(&cli->cl_lru_busy) > 0);
+ atomic_long_dec(&cli->cl_lru_busy);
}
spin_unlock(&cli->cl_lru_list_lock);
- atomic_inc(cli->cl_lru_left);
+ atomic_long_inc(cli->cl_lru_left);
/* this is a great place to release more LRU pages if
* this osc occupies too many LRU pages and kernel is
* stealing one of them.
@@ -489,7 +489,7 @@ static void osc_lru_use(struct client_obd *cli, struct osc_page *opg)
spin_lock(&cli->cl_lru_list_lock);
__osc_lru_del(cli, opg);
spin_unlock(&cli->cl_lru_list_lock);
- atomic_inc(&cli->cl_lru_busy);
+ atomic_long_inc(&cli->cl_lru_busy);
}
}
@@ -535,8 +535,8 @@ static inline bool lru_page_busy(struct client_obd *cli, struct cl_page *page)
/**
* Drop @target of pages from LRU at most.
*/
-int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
- int target, bool force)
+long osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
+ long target, bool force)
{
struct cl_io *io;
struct cl_object *clobj = NULL;
@@ -544,12 +544,12 @@ int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
struct osc_page *opg;
struct osc_page *temp;
int maxscan = 0;
- int count = 0;
+ long count = 0;
int index = 0;
int rc = 0;
- LASSERT(atomic_read(&cli->cl_lru_in_list) >= 0);
- if (atomic_read(&cli->cl_lru_in_list) == 0 || target <= 0)
+ LASSERT(atomic_long_read(&cli->cl_lru_in_list) >= 0);
+ if (atomic_long_read(&cli->cl_lru_in_list) == 0 || target <= 0)
return 0;
if (!force) {
@@ -568,7 +568,7 @@ int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
io = &osc_env_info(env)->oti_io;
spin_lock(&cli->cl_lru_list_lock);
- maxscan = min(target << 1, atomic_read(&cli->cl_lru_in_list));
+ maxscan = min(target << 1, atomic_long_read(&cli->cl_lru_in_list));
list_for_each_entry_safe(opg, temp, &cli->cl_lru_list, ops_lru) {
struct cl_page *page;
bool will_free = false;
@@ -656,24 +656,19 @@ int osc_lru_shrink(const struct lu_env *env, struct client_obd *cli,
atomic_dec(&cli->cl_lru_shrinkers);
if (count > 0) {
- atomic_add(count, cli->cl_lru_left);
+ atomic_long_add(count, cli->cl_lru_left);
wake_up_all(&osc_lru_waitq);
}
return count > 0 ? count : rc;
}
-static inline int max_to_shrink(struct client_obd *cli)
-{
- return min(atomic_read(&cli->cl_lru_in_list) >> 1, lru_shrink_max);
-}
-
-int osc_lru_reclaim(struct client_obd *cli)
+long osc_lru_reclaim(struct client_obd *cli)
{
struct cl_env_nest nest;
struct lu_env *env;
struct cl_client_cache *cache = cli->cl_cache;
int max_scans;
- int rc = 0;
+ long rc = 0;
LASSERT(cache);
@@ -686,15 +681,15 @@ int osc_lru_reclaim(struct client_obd *cli)
if (rc == -EBUSY)
rc = 0;
- CDEBUG(D_CACHE, "%s: Free %d pages from own LRU: %p.\n",
+ CDEBUG(D_CACHE, "%s: Free %ld pages from own LRU: %p.\n",
cli->cl_import->imp_obd->obd_name, rc, cli);
goto out;
}
- CDEBUG(D_CACHE, "%s: cli %p no free slots, pages: %d, busy: %d.\n",
+ CDEBUG(D_CACHE, "%s: cli %p no free slots, pages: %ld, busy: %ld.\n",
cli->cl_import->imp_obd->obd_name, cli,
- atomic_read(&cli->cl_lru_in_list),
- atomic_read(&cli->cl_lru_busy));
+ atomic_long_read(&cli->cl_lru_in_list),
+ atomic_long_read(&cli->cl_lru_busy));
/* Reclaim LRU slots from other client_obd as it can't free enough
* from its own. This should rarely happen.
@@ -710,10 +705,10 @@ int osc_lru_reclaim(struct client_obd *cli)
cli = list_entry(cache->ccc_lru.next, struct client_obd,
cl_lru_osc);
- CDEBUG(D_CACHE, "%s: cli %p LRU pages: %d, busy: %d.\n",
+ CDEBUG(D_CACHE, "%s: cli %p LRU pages: %ld, busy: %ld.\n",
cli->cl_import->imp_obd->obd_name, cli,
- atomic_read(&cli->cl_lru_in_list),
- atomic_read(&cli->cl_lru_busy));
+ atomic_long_read(&cli->cl_lru_in_list),
+ atomic_long_read(&cli->cl_lru_busy));
list_move_tail(&cli->cl_lru_osc, &cache->ccc_lru);
if (osc_cache_too_much(cli) > 0) {
@@ -730,7 +725,7 @@ int osc_lru_reclaim(struct client_obd *cli)
out:
cl_env_nested_put(&nest, env);
- CDEBUG(D_CACHE, "%s: cli %p freed %d pages.\n",
+ CDEBUG(D_CACHE, "%s: cli %p freed %ld pages.\n",
cli->cl_import->imp_obd->obd_name, cli, rc);
return rc;
}
@@ -758,8 +753,8 @@ static int osc_lru_reserve(const struct lu_env *env, struct osc_object *obj,
goto out;
}
- LASSERT(atomic_read(cli->cl_lru_left) >= 0);
- while (!atomic_add_unless(cli->cl_lru_left, -1, 0)) {
+ LASSERT(atomic_long_read(cli->cl_lru_left) >= 0);
+ while (!atomic_long_add_unless(cli->cl_lru_left, -1, 0)) {
/* run out of LRU spaces, try to drop some by itself */
rc = osc_lru_reclaim(cli);
if (rc < 0)
@@ -770,7 +765,7 @@ static int osc_lru_reserve(const struct lu_env *env, struct osc_object *obj,
cond_resched();
rc = l_wait_event(osc_lru_waitq,
- atomic_read(cli->cl_lru_left) > 0,
+ atomic_long_read(cli->cl_lru_left) > 0,
&lwi);
if (rc < 0)
@@ -779,7 +774,7 @@ static int osc_lru_reserve(const struct lu_env *env, struct osc_object *obj,
out:
if (rc >= 0) {
- atomic_inc(&cli->cl_lru_busy);
+ atomic_long_inc(&cli->cl_lru_busy);
opg->ops_in_lru = 1;
rc = 0;
}
@@ -847,16 +842,17 @@ void osc_dec_unstable_pages(struct ptlrpc_request *req)
struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
struct ptlrpc_bulk_desc *desc = req->rq_bulk;
int page_count = desc->bd_iov_count;
- int unstable_count;
+ long unstable_count;
LASSERT(page_count >= 0);
dec_unstable_page_accounting(desc);
- unstable_count = atomic_sub_return(page_count, &cli->cl_unstable_count);
+ unstable_count = atomic_long_sub_return(page_count,
+ &cli->cl_unstable_count);
LASSERT(unstable_count >= 0);
- unstable_count = atomic_sub_return(page_count,
- &cli->cl_cache->ccc_unstable_nr);
+ unstable_count = atomic_long_sub_return(page_count,
+ &cli->cl_cache->ccc_unstable_nr);
LASSERT(unstable_count >= 0);
if (!unstable_count)
wake_up_all(&cli->cl_cache->ccc_unstable_waitq);
@@ -872,15 +868,15 @@ void osc_inc_unstable_pages(struct ptlrpc_request *req)
{
struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
struct ptlrpc_bulk_desc *desc = req->rq_bulk;
- int page_count = desc->bd_iov_count;
+ long page_count = desc->bd_iov_count;
/* No unstable page tracking */
if (!cli->cl_cache || !cli->cl_cache->ccc_unstable_check)
return;
add_unstable_page_accounting(desc);
- atomic_add(page_count, &cli->cl_unstable_count);
- atomic_add(page_count, &cli->cl_cache->ccc_unstable_nr);
+ atomic_long_add(page_count, &cli->cl_unstable_count);
+ atomic_long_add(page_count, &cli->cl_cache->ccc_unstable_nr);
/*
* If the request has already been committed (i.e. brw_commit
@@ -912,8 +908,8 @@ bool osc_over_unstable_soft_limit(struct client_obd *cli)
if (!cli->cl_cache || !cli->cl_cache->ccc_unstable_check)
return false;
- osc_unstable_count = atomic_read(&cli->cl_unstable_count);
- unstable_nr = atomic_read(&cli->cl_cache->ccc_unstable_nr);
+ osc_unstable_count = atomic_long_read(&cli->cl_unstable_count);
+ unstable_nr = atomic_long_read(&cli->cl_cache->ccc_unstable_nr);
CDEBUG(D_CACHE,
"%s: cli: %p unstable pages: %lu, osc unstable pages: %lu\n",
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 7f7c318..7449b77 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -804,17 +804,17 @@ static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
cli->cl_dirty_pages, cli->cl_dirty_transit,
cli->cl_dirty_max_pages);
oa->o_undirty = 0;
- } else if (unlikely(atomic_read(&obd_dirty_pages) -
- atomic_read(&obd_dirty_transit_pages) >
- (long)(obd_max_dirty_pages + 1))) {
+ } else if (unlikely(atomic_long_read(&obd_dirty_pages) -
+ atomic_long_read(&obd_dirty_transit_pages) >
+ (obd_max_dirty_pages + 1))) {
/* The atomic_read() allowing the atomic_inc() are
* not covered by a lock thus they may safely race and trip
* this CERROR() unless we add in a small fudge factor (+1).
*/
- CERROR("%s: dirty %d + %d > system dirty_max %d\n",
+ CERROR("%s: dirty %ld + %ld > system dirty_max %lu\n",
cli->cl_import->imp_obd->obd_name,
- atomic_read(&obd_dirty_pages),
- atomic_read(&obd_dirty_transit_pages),
+ atomic_long_read(&obd_dirty_pages),
+ atomic_long_read(&obd_dirty_transit_pages),
obd_max_dirty_pages);
oa->o_undirty = 0;
} else if (unlikely(cli->cl_dirty_max_pages - cli->cl_dirty_pages >
@@ -2920,11 +2920,11 @@ static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
if (KEY_IS(KEY_CACHE_LRU_SHRINK)) {
struct client_obd *cli = &obd->u.cli;
- int nr = atomic_read(&cli->cl_lru_in_list) >> 1;
- int target = *(int *)val;
+ long nr = atomic_long_read(&cli->cl_lru_in_list) >> 1;
+ long target = *(long *)val;
nr = osc_lru_shrink(env, cli, min(nr, target), true);
- *(int *)val -= nr;
+ *(long *)val -= nr;
return 0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 045/124] staging: lustre: lmv: remove dead code
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
The member lmv_obd->server_timeout and function lmv_set_timeouts()
are not used.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-991
Reviewed-on: http://review.whamcloud.com/11880
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 1 -
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 25 -------------------------
2 files changed, 0 insertions(+), 26 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 3fbb873..b84600a 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -433,7 +433,6 @@ struct lmv_obd {
int max_def_easize;
int max_cookiesize;
int max_def_cookiesize;
- int server_timeout;
u32 tgts_size; /* size of tgts array */
struct lmv_tgt_desc **tgts;
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 53de385..8033ae2 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -302,30 +302,6 @@ static int lmv_connect(const struct lu_env *env,
return rc;
}
-static void lmv_set_timeouts(struct obd_device *obd)
-{
- struct lmv_obd *lmv;
- u32 i;
-
- lmv = &obd->u.lmv;
- if (lmv->server_timeout == 0)
- return;
-
- if (lmv->connected == 0)
- return;
-
- for (i = 0; i < lmv->desc.ld_tgt_count; i++) {
- struct lmv_tgt_desc *tgt = lmv->tgts[i];
-
- tgt = lmv->tgts[i];
- if (!tgt || !tgt->ltd_exp || !tgt->ltd_active)
- continue;
-
- obd_set_info_async(NULL, tgt->ltd_exp, sizeof(KEY_INTERMDS),
- KEY_INTERMDS, 0, NULL, NULL);
- }
-}
-
static int lmv_init_ea_size(struct obd_export *exp, int easize,
int def_easize, int cookiesize, int def_cookiesize)
{
@@ -616,7 +592,6 @@ int lmv_check_connect(struct obd_device *obd)
goto out_disc;
}
- lmv_set_timeouts(obd);
class_export_put(lmv->exp);
lmv->connected = 1;
easize = lmv_mds_md_size(lmv->desc.ld_tgt_count, LMV_MAGIC);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 045/124] staging: lustre: lmv: remove dead code
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
The member lmv_obd->server_timeout and function lmv_set_timeouts()
are not used.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-991
Reviewed-on: http://review.whamcloud.com/11880
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 1 -
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 25 -------------------------
2 files changed, 0 insertions(+), 26 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 3fbb873..b84600a 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -433,7 +433,6 @@ struct lmv_obd {
int max_def_easize;
int max_cookiesize;
int max_def_cookiesize;
- int server_timeout;
u32 tgts_size; /* size of tgts array */
struct lmv_tgt_desc **tgts;
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 53de385..8033ae2 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -302,30 +302,6 @@ static int lmv_connect(const struct lu_env *env,
return rc;
}
-static void lmv_set_timeouts(struct obd_device *obd)
-{
- struct lmv_obd *lmv;
- u32 i;
-
- lmv = &obd->u.lmv;
- if (lmv->server_timeout == 0)
- return;
-
- if (lmv->connected == 0)
- return;
-
- for (i = 0; i < lmv->desc.ld_tgt_count; i++) {
- struct lmv_tgt_desc *tgt = lmv->tgts[i];
-
- tgt = lmv->tgts[i];
- if (!tgt || !tgt->ltd_exp || !tgt->ltd_active)
- continue;
-
- obd_set_info_async(NULL, tgt->ltd_exp, sizeof(KEY_INTERMDS),
- KEY_INTERMDS, 0, NULL, NULL);
- }
-}
-
static int lmv_init_ea_size(struct obd_export *exp, int easize,
int def_easize, int cookiesize, int def_cookiesize)
{
@@ -616,7 +592,6 @@ int lmv_check_connect(struct obd_device *obd)
goto out_disc;
}
- lmv_set_timeouts(obd);
class_export_put(lmv->exp);
lmv->connected = 1;
easize = lmv_mds_md_size(lmv->desc.ld_tgt_count, LMV_MAGIC);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 046/124] staging: lustre: llite: handle concurrent use of cob_transient_pages
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Stephen Champion, James Simmons
From: Stephen Champion <schamp@sgi.com>
With the lockless __generic_file_aio_write introduced in LU-1669,
ll_direct_IO_26 is no longer protected by the inode i_isem.
This renders obsoltete checks that all transient pages have been
handled before and after entry, and requires atomic access to their
counter.
Signed-off-by: Stephen Champion <schamp@sgi.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5700
Reviewed-on: http://review.whamcloud.com/12179
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/rw26.c | 4 ----
drivers/staging/lustre/lustre/llite/vvp_internal.h | 9 +++++----
drivers/staging/lustre/lustre/llite/vvp_object.c | 6 +++---
drivers/staging/lustre/lustre/llite/vvp_page.c | 4 ++--
4 files changed, 10 insertions(+), 13 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index 6df89b0..f35eb2e 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -345,7 +345,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
struct cl_io *io;
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
- struct vvp_object *obj = cl_inode2vvp(inode);
loff_t file_offset = iocb->ki_pos;
ssize_t count = iov_iter_count(iter);
ssize_t tot_bytes = 0, result = 0;
@@ -374,7 +373,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
io = vvp_env_io(env)->vui_cl.cis_io;
LASSERT(io);
- LASSERT(obj->vob_transient_pages == 0);
while (iov_iter_count(iter)) {
struct page **pages;
size_t offs;
@@ -422,8 +420,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
file_offset += result;
}
out:
- LASSERT(obj->vob_transient_pages == 0);
-
if (tot_bytes > 0) {
struct vvp_io *vio = vvp_env_io(env);
diff --git a/drivers/staging/lustre/lustre/llite/vvp_internal.h b/drivers/staging/lustre/lustre/llite/vvp_internal.h
index 99437b8..5802da8 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_internal.h
+++ b/drivers/staging/lustre/lustre/llite/vvp_internal.h
@@ -217,11 +217,12 @@ struct vvp_object {
struct list_head vob_pending_list;
/**
- * Access this counter is protected by inode->i_sem. Now that
- * the lifetime of transient pages must be covered by inode sem,
- * we don't need to hold any lock..
+ * Number of transient pages. This is no longer protected by i_sem,
+ * and needs to be atomic. This is not actually used for anything,
+ * and can probably be removed.
*/
- int vob_transient_pages;
+ atomic_t vob_transient_pages;
+
/**
* Number of outstanding mmaps on this file.
*
diff --git a/drivers/staging/lustre/lustre/llite/vvp_object.c b/drivers/staging/lustre/lustre/llite/vvp_object.c
index 01725fb..6b6a63d 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_object.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_object.c
@@ -67,8 +67,8 @@ static int vvp_object_print(const struct lu_env *env, void *cookie,
(*p)(env, cookie, "(%s %d %d) inode: %p ",
list_empty(&obj->vob_pending_list) ? "-" : "+",
- obj->vob_transient_pages, atomic_read(&obj->vob_mmap_cnt),
- inode);
+ atomic_read(&obj->vob_transient_pages),
+ atomic_read(&obj->vob_mmap_cnt), inode);
if (inode) {
lli = ll_i2info(inode);
(*p)(env, cookie, "%lu/%u %o %u %d %p "DFID,
@@ -220,7 +220,7 @@ static int vvp_object_init0(const struct lu_env *env,
const struct cl_object_conf *conf)
{
vob->vob_inode = conf->coc_inode;
- vob->vob_transient_pages = 0;
+ atomic_set(&vob->vob_transient_pages, 0);
cl_object_page_init(&vob->vob_cl, sizeof(struct vvp_page));
return 0;
}
diff --git a/drivers/staging/lustre/lustre/llite/vvp_page.c b/drivers/staging/lustre/lustre/llite/vvp_page.c
index e763e7f..5d79efc 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_page.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_page.c
@@ -506,7 +506,7 @@ static void vvp_transient_page_fini(const struct lu_env *env,
struct vvp_object *clobj = cl2vvp(clp->cp_obj);
vvp_page_fini_common(vpg);
- clobj->vob_transient_pages--;
+ atomic_dec(&clobj->vob_transient_pages);
}
static const struct cl_page_operations vvp_transient_page_ops = {
@@ -555,7 +555,7 @@ int vvp_page_init(const struct lu_env *env, struct cl_object *obj,
cl_page_slice_add(page, &vpg->vpg_cl, obj, index,
&vvp_transient_page_ops);
- clobj->vob_transient_pages++;
+ atomic_inc(&clobj->vob_transient_pages);
}
return 0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 046/124] staging: lustre: llite: handle concurrent use of cob_transient_pages
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Stephen Champion, James Simmons
From: Stephen Champion <schamp@sgi.com>
With the lockless __generic_file_aio_write introduced in LU-1669,
ll_direct_IO_26 is no longer protected by the inode i_isem.
This renders obsoltete checks that all transient pages have been
handled before and after entry, and requires atomic access to their
counter.
Signed-off-by: Stephen Champion <schamp@sgi.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5700
Reviewed-on: http://review.whamcloud.com/12179
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/rw26.c | 4 ----
drivers/staging/lustre/lustre/llite/vvp_internal.h | 9 +++++----
drivers/staging/lustre/lustre/llite/vvp_object.c | 6 +++---
drivers/staging/lustre/lustre/llite/vvp_page.c | 4 ++--
4 files changed, 10 insertions(+), 13 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index 6df89b0..f35eb2e 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -345,7 +345,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
struct cl_io *io;
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
- struct vvp_object *obj = cl_inode2vvp(inode);
loff_t file_offset = iocb->ki_pos;
ssize_t count = iov_iter_count(iter);
ssize_t tot_bytes = 0, result = 0;
@@ -374,7 +373,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
io = vvp_env_io(env)->vui_cl.cis_io;
LASSERT(io);
- LASSERT(obj->vob_transient_pages == 0);
while (iov_iter_count(iter)) {
struct page **pages;
size_t offs;
@@ -422,8 +420,6 @@ static ssize_t ll_direct_IO_26(struct kiocb *iocb, struct iov_iter *iter)
file_offset += result;
}
out:
- LASSERT(obj->vob_transient_pages == 0);
-
if (tot_bytes > 0) {
struct vvp_io *vio = vvp_env_io(env);
diff --git a/drivers/staging/lustre/lustre/llite/vvp_internal.h b/drivers/staging/lustre/lustre/llite/vvp_internal.h
index 99437b8..5802da8 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_internal.h
+++ b/drivers/staging/lustre/lustre/llite/vvp_internal.h
@@ -217,11 +217,12 @@ struct vvp_object {
struct list_head vob_pending_list;
/**
- * Access this counter is protected by inode->i_sem. Now that
- * the lifetime of transient pages must be covered by inode sem,
- * we don't need to hold any lock..
+ * Number of transient pages. This is no longer protected by i_sem,
+ * and needs to be atomic. This is not actually used for anything,
+ * and can probably be removed.
*/
- int vob_transient_pages;
+ atomic_t vob_transient_pages;
+
/**
* Number of outstanding mmaps on this file.
*
diff --git a/drivers/staging/lustre/lustre/llite/vvp_object.c b/drivers/staging/lustre/lustre/llite/vvp_object.c
index 01725fb..6b6a63d 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_object.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_object.c
@@ -67,8 +67,8 @@ static int vvp_object_print(const struct lu_env *env, void *cookie,
(*p)(env, cookie, "(%s %d %d) inode: %p ",
list_empty(&obj->vob_pending_list) ? "-" : "+",
- obj->vob_transient_pages, atomic_read(&obj->vob_mmap_cnt),
- inode);
+ atomic_read(&obj->vob_transient_pages),
+ atomic_read(&obj->vob_mmap_cnt), inode);
if (inode) {
lli = ll_i2info(inode);
(*p)(env, cookie, "%lu/%u %o %u %d %p "DFID,
@@ -220,7 +220,7 @@ static int vvp_object_init0(const struct lu_env *env,
const struct cl_object_conf *conf)
{
vob->vob_inode = conf->coc_inode;
- vob->vob_transient_pages = 0;
+ atomic_set(&vob->vob_transient_pages, 0);
cl_object_page_init(&vob->vob_cl, sizeof(struct vvp_page));
return 0;
}
diff --git a/drivers/staging/lustre/lustre/llite/vvp_page.c b/drivers/staging/lustre/lustre/llite/vvp_page.c
index e763e7f..5d79efc 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_page.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_page.c
@@ -506,7 +506,7 @@ static void vvp_transient_page_fini(const struct lu_env *env,
struct vvp_object *clobj = cl2vvp(clp->cp_obj);
vvp_page_fini_common(vpg);
- clobj->vob_transient_pages--;
+ atomic_dec(&clobj->vob_transient_pages);
}
static const struct cl_page_operations vvp_transient_page_ops = {
@@ -555,7 +555,7 @@ int vvp_page_init(const struct lu_env *env, struct cl_object *obj,
cl_page_slice_add(page, &vpg->vpg_cl, obj, index,
&vvp_transient_page_ops);
- clobj->vob_transient_pages++;
+ atomic_inc(&clobj->vob_transient_pages);
}
return 0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 047/124] staging: lustre: llite: enforce pool name length limit
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi, James Simmons
From: Li Xi <lixi@ddn.com>
The pool related codes have some inconsistency about the length
of pool name. Creating and setting a pool name of length 16
to a directory will succeed. However, creating a file under
that directory will fail.
This patch disables any pool name which is longer or equal to
16. And it changes LOV_MAXPOOLNAME from 16 to 15 which might
cause some invalid LLOG records of OST pools with 16 byte names.
It is not a problem since invalid LLOG records are just ignored.
And OST pools with 16 byte names won't work well anyway on the
old versions. There will be problem of inconsistency if part of
the servers have this patch and part of the servers don't. But
it would be safe to assume that this is not a normal
configuration.
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5054
Reviewed-on: http://review.whamcloud.com/10306
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 4 ++--
.../lustre/lustre/include/lustre/lustre_user.h | 8 ++++----
drivers/staging/lustre/lustre/include/lustre_lmv.h | 2 +-
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 2 +-
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 6ff6615..56b5129 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1620,7 +1620,7 @@ struct lov_mds_md_v3 { /* LOV EA mds/wire data (little-endian) */
/* lmm_stripe_count used to be __u32 */
__u16 lmm_stripe_count; /* num stripes in use for this object */
__u16 lmm_layout_gen; /* layout generation number */
- char lmm_pool_name[LOV_MAXPOOLNAME]; /* must be 32bit aligned */
+ char lmm_pool_name[LOV_MAXPOOLNAME + 1]; /* must be 32bit aligned */
struct lov_ost_data_v1 lmm_objects[0]; /* per-stripe data */
};
@@ -2498,7 +2498,7 @@ struct lmv_mds_md_v1 {
__u32 lmv_padding1;
__u64 lmv_padding2;
__u64 lmv_padding3;
- char lmv_pool_name[LOV_MAXPOOLNAME]; /* pool name */
+ char lmv_pool_name[LOV_MAXPOOLNAME + 1];/* pool name */
struct lu_fid lmv_stripe_fids[0]; /* FIDs for each stripe */
};
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 5e907af..79cd876 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -295,8 +295,8 @@ enum ll_lease_type {
#define LOV_PATTERN_F_HOLE 0x40000000 /* there is hole in LOV EA */
#define LOV_PATTERN_F_RELEASED 0x80000000 /* HSM released file */
-#define LOV_MAXPOOLNAME 16
-#define LOV_POOLNAMEF "%.16s"
+#define LOV_MAXPOOLNAME 15
+#define LOV_POOLNAMEF "%.15s"
#define LOV_MIN_STRIPE_BITS 16 /* maximum PAGE_SIZE (ia64), power of 2 */
#define LOV_MIN_STRIPE_SIZE (1 << LOV_MIN_STRIPE_BITS)
@@ -354,7 +354,7 @@ struct lov_user_md_v3 { /* LOV EA user data (host-endian) */
* used when reading
*/
};
- char lmm_pool_name[LOV_MAXPOOLNAME]; /* pool name */
+ char lmm_pool_name[LOV_MAXPOOLNAME + 1]; /* pool name */
struct lov_user_ost_data_v1 lmm_objects[0]; /* per-stripe data */
} __packed;
@@ -414,7 +414,7 @@ struct lmv_user_md_v1 {
__u32 lum_padding1;
__u32 lum_padding2;
__u32 lum_padding3;
- char lum_pool_name[LOV_MAXPOOLNAME];
+ char lum_pool_name[LOV_MAXPOOLNAME + 1];
struct lmv_user_mds_data lum_objects[0];
} __packed;
diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h
index 5a2fc30..388161e 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lmv.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h
@@ -48,7 +48,7 @@ struct lmv_stripe_md {
__u32 lsm_md_layout_version;
__u32 lsm_md_default_count;
__u32 lsm_md_default_index;
- char lsm_md_pool_name[LOV_MAXPOOLNAME];
+ char lsm_md_pool_name[LOV_MAXPOOLNAME + 1];
struct lmv_oinfo lsm_md_oinfo[0];
};
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index b84600a..1c884c0 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -99,7 +99,7 @@ struct lov_stripe_md {
__u32 lw_pattern; /* striping pattern (RAID0, RAID1) */
__u16 lw_stripe_count; /* number of objects being striped over */
__u16 lw_layout_gen; /* generation of the layout */
- char lw_pool_name[LOV_MAXPOOLNAME]; /* pool name */
+ char lw_pool_name[LOV_MAXPOOLNAME + 1]; /* pool name */
} lsm_wire;
struct lov_oinfo *lsm_oinfo[0];
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index 2390255..b33fd7b 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -1383,7 +1383,7 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct lov_mds_md_v3, lmm_layout_gen));
LASSERTF((int)sizeof(((struct lov_mds_md_v3 *)0)->lmm_layout_gen) == 2, "found %lld\n",
(long long)(int)sizeof(((struct lov_mds_md_v3 *)0)->lmm_layout_gen));
- CLASSERT(LOV_MAXPOOLNAME == 16);
+ CLASSERT(LOV_MAXPOOLNAME == 15);
LASSERTF((int)offsetof(struct lov_mds_md_v3, lmm_pool_name[16]) == 48, "found %lld\n",
(long long)(int)offsetof(struct lov_mds_md_v3, lmm_pool_name[16]));
LASSERTF((int)sizeof(((struct lov_mds_md_v3 *)0)->lmm_pool_name[16]) == 1, "found %lld\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 047/124] staging: lustre: llite: enforce pool name length limit
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Li Xi, James Simmons
From: Li Xi <lixi@ddn.com>
The pool related codes have some inconsistency about the length
of pool name. Creating and setting a pool name of length 16
to a directory will succeed. However, creating a file under
that directory will fail.
This patch disables any pool name which is longer or equal to
16. And it changes LOV_MAXPOOLNAME from 16 to 15 which might
cause some invalid LLOG records of OST pools with 16 byte names.
It is not a problem since invalid LLOG records are just ignored.
And OST pools with 16 byte names won't work well anyway on the
old versions. There will be problem of inconsistency if part of
the servers have this patch and part of the servers don't. But
it would be safe to assume that this is not a normal
configuration.
Signed-off-by: Li Xi <lixi@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5054
Reviewed-on: http://review.whamcloud.com/10306
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 4 ++--
.../lustre/lustre/include/lustre/lustre_user.h | 8 ++++----
drivers/staging/lustre/lustre/include/lustre_lmv.h | 2 +-
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 2 +-
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 6ff6615..56b5129 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1620,7 +1620,7 @@ struct lov_mds_md_v3 { /* LOV EA mds/wire data (little-endian) */
/* lmm_stripe_count used to be __u32 */
__u16 lmm_stripe_count; /* num stripes in use for this object */
__u16 lmm_layout_gen; /* layout generation number */
- char lmm_pool_name[LOV_MAXPOOLNAME]; /* must be 32bit aligned */
+ char lmm_pool_name[LOV_MAXPOOLNAME + 1]; /* must be 32bit aligned */
struct lov_ost_data_v1 lmm_objects[0]; /* per-stripe data */
};
@@ -2498,7 +2498,7 @@ struct lmv_mds_md_v1 {
__u32 lmv_padding1;
__u64 lmv_padding2;
__u64 lmv_padding3;
- char lmv_pool_name[LOV_MAXPOOLNAME]; /* pool name */
+ char lmv_pool_name[LOV_MAXPOOLNAME + 1];/* pool name */
struct lu_fid lmv_stripe_fids[0]; /* FIDs for each stripe */
};
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 5e907af..79cd876 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -295,8 +295,8 @@ enum ll_lease_type {
#define LOV_PATTERN_F_HOLE 0x40000000 /* there is hole in LOV EA */
#define LOV_PATTERN_F_RELEASED 0x80000000 /* HSM released file */
-#define LOV_MAXPOOLNAME 16
-#define LOV_POOLNAMEF "%.16s"
+#define LOV_MAXPOOLNAME 15
+#define LOV_POOLNAMEF "%.15s"
#define LOV_MIN_STRIPE_BITS 16 /* maximum PAGE_SIZE (ia64), power of 2 */
#define LOV_MIN_STRIPE_SIZE (1 << LOV_MIN_STRIPE_BITS)
@@ -354,7 +354,7 @@ struct lov_user_md_v3 { /* LOV EA user data (host-endian) */
* used when reading
*/
};
- char lmm_pool_name[LOV_MAXPOOLNAME]; /* pool name */
+ char lmm_pool_name[LOV_MAXPOOLNAME + 1]; /* pool name */
struct lov_user_ost_data_v1 lmm_objects[0]; /* per-stripe data */
} __packed;
@@ -414,7 +414,7 @@ struct lmv_user_md_v1 {
__u32 lum_padding1;
__u32 lum_padding2;
__u32 lum_padding3;
- char lum_pool_name[LOV_MAXPOOLNAME];
+ char lum_pool_name[LOV_MAXPOOLNAME + 1];
struct lmv_user_mds_data lum_objects[0];
} __packed;
diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h
index 5a2fc30..388161e 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lmv.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h
@@ -48,7 +48,7 @@ struct lmv_stripe_md {
__u32 lsm_md_layout_version;
__u32 lsm_md_default_count;
__u32 lsm_md_default_index;
- char lsm_md_pool_name[LOV_MAXPOOLNAME];
+ char lsm_md_pool_name[LOV_MAXPOOLNAME + 1];
struct lmv_oinfo lsm_md_oinfo[0];
};
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index b84600a..1c884c0 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -99,7 +99,7 @@ struct lov_stripe_md {
__u32 lw_pattern; /* striping pattern (RAID0, RAID1) */
__u16 lw_stripe_count; /* number of objects being striped over */
__u16 lw_layout_gen; /* generation of the layout */
- char lw_pool_name[LOV_MAXPOOLNAME]; /* pool name */
+ char lw_pool_name[LOV_MAXPOOLNAME + 1]; /* pool name */
} lsm_wire;
struct lov_oinfo *lsm_oinfo[0];
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index 2390255..b33fd7b 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -1383,7 +1383,7 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct lov_mds_md_v3, lmm_layout_gen));
LASSERTF((int)sizeof(((struct lov_mds_md_v3 *)0)->lmm_layout_gen) == 2, "found %lld\n",
(long long)(int)sizeof(((struct lov_mds_md_v3 *)0)->lmm_layout_gen));
- CLASSERT(LOV_MAXPOOLNAME == 16);
+ CLASSERT(LOV_MAXPOOLNAME == 15);
LASSERTF((int)offsetof(struct lov_mds_md_v3, lmm_pool_name[16]) == 48, "found %lld\n",
(long long)(int)offsetof(struct lov_mds_md_v3, lmm_pool_name[16]));
LASSERTF((int)sizeof(((struct lov_mds_md_v3 *)0)->lmm_pool_name[16]) == 1, "found %lld\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 048/124] staging: lustre: Flexible changelog format.
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Added jobid fields to Changelog records (and extended records). The
CLF_JOBID flags allows to check if the field is present or not (old
format) when reading an entry. Jobids are expressed as 32 chars long,
zero-terminated strings. Updated test_205 in sanity.sh.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1996
Reviewed-on: http://review.whamcloud.com/4060
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Aurelien Degremont <aurelien.degremont@cea.fr>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 2 +-
.../lustre/lustre/include/lustre/lustre_idl.h | 19 +--
.../lustre/lustre/include/lustre/lustre_user.h | 165 +++++++++++++++-----
.../staging/lustre/lustre/llite/llite_internal.h | 2 +-
drivers/staging/lustre/lustre/llite/vvp_req.c | 2 +-
drivers/staging/lustre/lustre/mdc/mdc_request.c | 22 ++-
drivers/staging/lustre/lustre/obdclass/class_obd.c | 6 +-
.../lustre/lustre/obdclass/linux/linux-module.c | 2 +-
drivers/staging/lustre/lustre/obdclass/llog.c | 4 +-
drivers/staging/lustre/lustre/obdclass/llog_cat.c | 4 +-
.../staging/lustre/lustre/obdclass/llog_internal.h | 5 +
drivers/staging/lustre/lustre/obdclass/llog_swab.c | 23 ++--
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 2 +-
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 54 +------
14 files changed, 188 insertions(+), 124 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 5039643..5af1153 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -1899,7 +1899,7 @@ struct cl_req_attr {
/** Generic attributes for the server consumption. */
struct obdo *cra_oa;
/** Jobid */
- char cra_jobid[JOBSTATS_JOBID_SIZE];
+ char cra_jobid[LUSTRE_JOBID_SIZE];
};
/**
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 56b5129..2dc550a 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1101,7 +1101,7 @@ struct lustre_msg_v2 {
/* without gss, ptlrpc_body is put at the first buffer. */
#define PTLRPC_NUM_VERSIONS 4
-#define JOBSTATS_JOBID_SIZE 32 /* 32 bytes string */
+
struct ptlrpc_body_v3 {
struct lustre_handle pb_handle;
__u32 pb_type;
@@ -1123,7 +1123,7 @@ struct ptlrpc_body_v3 {
__u64 pb_pre_versions[PTLRPC_NUM_VERSIONS];
/* padding for future needs */
__u64 pb_padding[4];
- char pb_jobid[JOBSTATS_JOBID_SIZE];
+ char pb_jobid[LUSTRE_JOBID_SIZE];
};
#define ptlrpc_body ptlrpc_body_v3
@@ -3088,15 +3088,9 @@ struct changelog_setinfo {
/** changelog record */
struct llog_changelog_rec {
- struct llog_rec_hdr cr_hdr;
- struct changelog_rec cr;
- struct llog_rec_tail cr_tail; /**< for_sizezof_only */
-} __packed;
-
-struct llog_changelog_ext_rec {
- struct llog_rec_hdr cr_hdr;
- struct changelog_ext_rec cr;
- struct llog_rec_tail cr_tail; /**< for_sizezof_only */
+ struct llog_rec_hdr cr_hdr;
+ struct changelog_rec cr; /**< Variable length field */
+ struct llog_rec_tail cr_do_not_use; /**< for_sizezof_only */
} __packed;
struct llog_changelog_user_rec {
@@ -3181,6 +3175,9 @@ enum llog_flag {
LLOG_F_ZAP_WHEN_EMPTY = 0x1,
LLOG_F_IS_CAT = 0x2,
LLOG_F_IS_PLAIN = 0x4,
+ LLOG_F_EXT_JOBID = BIT(3),
+
+ LLOG_F_EXT_MASK = LLOG_F_EXT_JOBID,
};
struct llog_log_hdr {
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 79cd876..cc0a786 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -673,11 +673,16 @@ static inline const char *changelog_type2str(int type)
}
/* per-record flags */
-#define CLF_VERSION 0x1000
-#define CLF_EXT_VERSION 0x2000
#define CLF_FLAGSHIFT 12
#define CLF_FLAGMASK ((1U << CLF_FLAGSHIFT) - 1)
#define CLF_VERMASK (~CLF_FLAGMASK)
+enum changelog_rec_flags {
+ CLF_VERSION = 0x1000,
+ CLF_RENAME = 0x2000,
+ CLF_JOBID = 0x4000,
+ CLF_SUPPORTED = CLF_VERSION | CLF_RENAME | CLF_JOBID
+};
+
/* Anything under the flagmask may be per-type (if desired) */
/* Flags for unlink */
#define CLF_UNLINK_LAST 0x0001 /* Unlink of last hardlink */
@@ -761,12 +766,35 @@ static inline void hsm_set_cl_error(int *flags, int error)
*flags |= (error << CLF_HSM_ERR_L);
}
-#define CR_MAXSIZE cfs_size_round(2 * NAME_MAX + 1 + \
- sizeof(struct changelog_ext_rec))
+enum changelog_send_flag {
+ /* Not yet implemented */
+ CHANGELOG_FLAG_FOLLOW = BIT(0),
+ /*
+ * Blocking IO makes sense in case of slow user parsing of the records,
+ * but it also prevents us from cleaning up if the records are not
+ * consumed.
+ */
+ CHANGELOG_FLAG_BLOCK = BIT(1),
+ /* Pack jobid into the changelog records if available. */
+ CHANGELOG_FLAG_JOBID = BIT(2),
+};
+
+#define CR_MAXSIZE cfs_size_round(2 * NAME_MAX + 2 + \
+ changelog_rec_offset(CLF_SUPPORTED))
+
+/* 31 usable bytes string + null terminator. */
+#define LUSTRE_JOBID_SIZE 32
+/*
+ * This is the minimal changelog record. It can contain extensions
+ * such as rename fields or process jobid. Its exact content is described
+ * by the cr_flags.
+ *
+ * Extensions are packed in the same order as their corresponding flags.
+ */
struct changelog_rec {
__u16 cr_namelen;
- __u16 cr_flags; /**< (flags&CLF_FLAGMASK)|CLF_VERSION */
+ __u16 cr_flags; /**< \a changelog_rec_flags */
__u32 cr_type; /**< \a changelog_rec_type */
__u64 cr_index; /**< changelog record number */
__u64 cr_prev; /**< last index for this target fid */
@@ -776,55 +804,116 @@ struct changelog_rec {
__u32 cr_markerflags; /**< CL_MARK flags */
};
struct lu_fid cr_pfid; /**< parent fid */
- char cr_name[0]; /**< last element */
} __packed;
-/* changelog_ext_rec is 2*sizeof(lu_fid) bigger than changelog_rec, to save
- * space, only rename uses changelog_ext_rec, while others use changelog_rec to
- * store records.
- */
-struct changelog_ext_rec {
- __u16 cr_namelen;
- __u16 cr_flags; /**< (flags & CLF_FLAGMASK) |
- * CLF_EXT_VERSION
- */
- __u32 cr_type; /**< \a changelog_rec_type */
- __u64 cr_index; /**< changelog record number */
- __u64 cr_prev; /**< last index for this target fid */
- __u64 cr_time;
- union {
- struct lu_fid cr_tfid; /**< target fid */
- __u32 cr_markerflags; /**< CL_MARK flags */
- };
- struct lu_fid cr_pfid; /**< target parent fid */
- struct lu_fid cr_sfid; /**< source fid, or zero */
- struct lu_fid cr_spfid; /**< source parent fid, or zero */
- char cr_name[0]; /**< last element */
-} __packed;
+/* Changelog extension for RENAME. */
+struct changelog_ext_rename {
+ struct lu_fid cr_sfid; /**< source fid, or zero */
+ struct lu_fid cr_spfid; /**< source parent fid, or zero */
+};
+
+/* Changelog extension to include JOBID. */
+struct changelog_ext_jobid {
+ char cr_jobid[LUSTRE_JOBID_SIZE]; /**< zero-terminated string. */
+};
+
+static inline unsigned int changelog_rec_offset(enum changelog_rec_flags crf)
+{
+ unsigned int size = sizeof(struct changelog_rec);
+
+ if (crf & CLF_RENAME)
+ size += sizeof(struct changelog_ext_rename);
+
+ if (crf & CLF_JOBID)
+ size += sizeof(struct changelog_ext_jobid);
-#define CHANGELOG_REC_EXTENDED(rec) \
- (((rec)->cr_flags & CLF_VERMASK) == CLF_EXT_VERSION)
+ return size;
+}
static inline int changelog_rec_size(struct changelog_rec *rec)
{
- return CHANGELOG_REC_EXTENDED(rec) ? sizeof(struct changelog_ext_rec) :
- sizeof(*rec);
+ return changelog_rec_offset(rec->cr_flags);
+}
+
+static inline unsigned int changelog_rec_varsize(struct changelog_rec *rec)
+{
+ return changelog_rec_size(rec) - sizeof(*rec) + rec->cr_namelen;
+}
+
+static inline
+struct changelog_ext_rename *changelog_rec_rename(struct changelog_rec *rec)
+{
+ enum changelog_rec_flags crf = rec->cr_flags & CLF_VERSION;
+
+ return (struct changelog_ext_rename *)((char *)rec +
+ changelog_rec_offset(crf));
+}
+
+/* The jobid follows the rename extension, if present */
+static inline
+struct changelog_ext_jobid *changelog_rec_jobid(struct changelog_rec *rec)
+{
+ enum changelog_rec_flags crf = rec->cr_flags &
+ (CLF_VERSION | CLF_RENAME);
+
+ return (struct changelog_ext_jobid *)((char *)rec +
+ changelog_rec_offset(crf));
}
+/* The name follows the rename and jobid extensions, if present */
static inline char *changelog_rec_name(struct changelog_rec *rec)
{
- return CHANGELOG_REC_EXTENDED(rec) ?
- ((struct changelog_ext_rec *)rec)->cr_name : rec->cr_name;
+ return (char *)rec + changelog_rec_offset(rec->cr_flags &
+ CLF_SUPPORTED);
+}
+
+static inline int changelog_rec_snamelen(struct changelog_rec *rec)
+{
+ return rec->cr_namelen - strlen(changelog_rec_name(rec)) - 1;
}
-static inline int changelog_rec_snamelen(struct changelog_ext_rec *rec)
+static inline char *changelog_rec_sname(struct changelog_rec *rec)
{
- return rec->cr_namelen - strlen(rec->cr_name) - 1;
+ char *cr_name = changelog_rec_name(rec);
+
+ return cr_name + strlen(cr_name) + 1;
}
-static inline char *changelog_rec_sname(struct changelog_ext_rec *rec)
+/*
+ * Remap a record to the desired format as specified by the crf flags.
+ * The record must be big enough to contain the final remapped version.
+ */
+static inline void changelog_remap_rec(struct changelog_rec *rec,
+ enum changelog_rec_flags crf)
{
- return rec->cr_name + strlen(rec->cr_name) + 1;
+ size_t var_size;
+ char *var_part;
+
+ crf &= CLF_SUPPORTED;
+
+ if ((rec->cr_flags & CLF_SUPPORTED) == crf)
+ return;
+
+ if ((crf & CLF_JOBID) && rec->cr_flags & CLF_JOBID) {
+ var_part = (char *)changelog_rec_jobid(rec);
+ var_size = rec->cr_namelen + sizeof(struct changelog_ext_jobid);
+ } else {
+ var_part = changelog_rec_name(rec);
+ var_size = rec->cr_namelen;
+ }
+
+ memmove((char *)rec + changelog_rec_offset(crf & ~CLF_JOBID), var_part,
+ var_size);
+
+ if ((crf & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
+ memset(changelog_rec_rename(rec), 0,
+ sizeof(struct changelog_ext_rename));
+
+ if ((crf & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
+ memset(changelog_rec_jobid(rec), 0,
+ sizeof(struct changelog_ext_jobid));
+
+ rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf;
}
struct ioc_changelog {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index f972ba3..fc707c2 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -232,7 +232,7 @@ struct ll_inode_info {
* so the read/write statistics for jobid will not be
* accurate if the file is shared by different jobs.
*/
- char f_jobid[JOBSTATS_JOBID_SIZE];
+ char f_jobid[LUSTRE_JOBID_SIZE];
} f;
#define lli_size_mutex u.f.f_size_mutex
diff --git a/drivers/staging/lustre/lustre/llite/vvp_req.c b/drivers/staging/lustre/lustre/llite/vvp_req.c
index a6e2672..e3f4c79 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_req.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_req.c
@@ -85,7 +85,7 @@ static void vvp_req_attr_set(const struct lu_env *env,
if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_INVALID_PFID))
oa->o_parent_oid++;
memcpy(attr->cra_jobid, ll_i2info(inode)->lli_jobid,
- JOBSTATS_JOBID_SIZE);
+ LUSTRE_JOBID_SIZE);
}
static void vvp_req_completion(const struct lu_env *env,
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 956b946..aeadcd6 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1859,12 +1859,17 @@ static struct kuc_hdr *changelog_kuc_hdr(char *buf, int len, int flags)
struct changelog_show {
__u64 cs_startrec;
- __u32 cs_flags;
+ enum changelog_send_flag cs_flags;
struct file *cs_fp;
char *cs_buf;
struct obd_device *cs_obd;
};
+static inline char *cs_obd_name(struct changelog_show *cs)
+{
+ return cs->cs_obd->obd_name;
+}
+
static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
struct llog_rec_hdr *hdr, void *data)
{
@@ -1876,7 +1881,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
if (rec->cr_hdr.lrh_type != CHANGELOG_REC) {
rc = -EINVAL;
CERROR("%s: not a changelog rec %x/%d: rc = %d\n",
- cs->cs_obd->obd_name, rec->cr_hdr.lrh_type,
+ cs_obd_name(cs), rec->cr_hdr.lrh_type,
rec->cr.cr_type, rc);
return rc;
}
@@ -1909,6 +1914,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
static int mdc_changelog_send_thread(void *csdata)
{
+ enum llog_flag flags = LLOG_F_IS_CAT;
struct changelog_show *cs = csdata;
struct llog_ctxt *ctxt = NULL;
struct llog_handle *llh = NULL;
@@ -1934,10 +1940,14 @@ static int mdc_changelog_send_thread(void *csdata)
LLOG_OPEN_EXISTS);
if (rc) {
CERROR("%s: fail to open changelog catalog: rc = %d\n",
- cs->cs_obd->obd_name, rc);
+ cs_obd_name(cs), rc);
goto out;
}
- rc = llog_init_handle(NULL, llh, LLOG_F_IS_CAT, NULL);
+
+ if (cs->cs_flags & CHANGELOG_FLAG_JOBID)
+ flags |= LLOG_F_EXT_JOBID;
+
+ rc = llog_init_handle(NULL, llh, flags, NULL);
if (rc) {
CERROR("llog_init_handle failed %d\n", rc);
goto out;
@@ -1990,12 +2000,12 @@ static int mdc_ioc_changelog_send(struct obd_device *obd,
if (IS_ERR(task)) {
rc = PTR_ERR(task);
CERROR("%s: can't start changelog thread: rc = %d\n",
- obd->obd_name, rc);
+ cs_obd_name(cs), rc);
kfree(cs);
} else {
rc = 0;
CDEBUG(D_HSM, "%s: started changelog thread\n",
- obd->obd_name);
+ cs_obd_name(cs));
}
CERROR("Failed to start changelog thread: %d\n", rc);
diff --git a/drivers/staging/lustre/lustre/obdclass/class_obd.c b/drivers/staging/lustre/lustre/obdclass/class_obd.c
index 629d8b5..4c30d6b 100644
--- a/drivers/staging/lustre/lustre/obdclass/class_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/class_obd.c
@@ -81,7 +81,7 @@ EXPORT_SYMBOL(obd_dirty_transit_pages);
char obd_jobid_var[JOBSTATS_JOBID_VAR_MAX_LEN + 1] = JOBSTATS_DISABLE;
EXPORT_SYMBOL(obd_jobid_var);
-char obd_jobid_node[JOBSTATS_JOBID_SIZE + 1];
+char obd_jobid_node[LUSTRE_JOBID_SIZE + 1];
/* Get jobid of current process from stored variable or calculate
* it from pid and user_id.
@@ -92,14 +92,14 @@ char obd_jobid_node[JOBSTATS_JOBID_SIZE + 1];
*/
int lustre_get_jobid(char *jobid)
{
- memset(jobid, 0, JOBSTATS_JOBID_SIZE);
+ memset(jobid, 0, LUSTRE_JOBID_SIZE);
/* Jobstats isn't enabled */
if (strcmp(obd_jobid_var, JOBSTATS_DISABLE) == 0)
return 0;
/* Use process name + fsuid as jobid */
if (strcmp(obd_jobid_var, JOBSTATS_PROCNAME_UID) == 0) {
- snprintf(jobid, JOBSTATS_JOBID_SIZE, "%s.%u",
+ snprintf(jobid, LUSTRE_JOBID_SIZE, "%s.%u",
current_comm(),
from_kuid(&init_user_ns, current_fsuid()));
return 0;
diff --git a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
index 2b691d8..ae172c4 100644
--- a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
+++ b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
@@ -292,7 +292,7 @@ static ssize_t jobid_name_store(struct kobject *kobj, struct attribute *attr,
const char *buffer,
size_t count)
{
- if (!count || count > JOBSTATS_JOBID_SIZE)
+ if (!count || count > LUSTRE_JOBID_SIZE)
return -EINVAL;
memcpy(obd_jobid_node, buffer, count);
diff --git a/drivers/staging/lustre/lustre/obdclass/llog.c b/drivers/staging/lustre/lustre/obdclass/llog.c
index 8f06141..119372c 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog.c
@@ -137,6 +137,7 @@ static int llog_read_header(const struct lu_env *env,
int llog_init_handle(const struct lu_env *env, struct llog_handle *handle,
int flags, struct obd_uuid *uuid)
{
+ enum llog_flag fmt = flags & LLOG_F_EXT_MASK;
struct llog_log_hdr *llh;
int rc;
@@ -194,6 +195,7 @@ int llog_init_handle(const struct lu_env *env, struct llog_handle *handle,
flags, LLOG_F_IS_CAT, LLOG_F_IS_PLAIN);
rc = -EINVAL;
}
+ llh->llh_flags |= fmt;
out:
if (rc) {
kfree(llh);
@@ -262,7 +264,7 @@ repeat:
*/
for (rec = (struct llog_rec_hdr *)buf;
(char *)rec < buf + LLOG_CHUNK_SIZE;
- rec = (struct llog_rec_hdr *)((char *)rec + rec->lrh_len)) {
+ rec = llog_rec_hdr_next(rec)) {
CDEBUG(D_OTHER, "processing rec 0x%p type %#x\n",
rec, rec->lrh_type);
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_cat.c b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
index b27f08f..ce8e2f6 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_cat.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
@@ -63,11 +63,13 @@ static int llog_cat_id2handle(const struct lu_env *env,
struct llog_logid *logid)
{
struct llog_handle *loghandle;
+ enum llog_flag fmt;
int rc = 0;
if (!cathandle)
return -EBADF;
+ fmt = cathandle->lgh_hdr->llh_flags & LLOG_F_EXT_MASK;
down_write(&cathandle->lgh_lock);
list_for_each_entry(loghandle, &cathandle->u.chd.chd_head,
u.phd.phd_entry) {
@@ -99,7 +101,7 @@ static int llog_cat_id2handle(const struct lu_env *env,
return rc;
}
- rc = llog_init_handle(env, loghandle, LLOG_F_IS_PLAIN, NULL);
+ rc = llog_init_handle(env, loghandle, fmt | LLOG_F_IS_PLAIN, NULL);
if (rc < 0) {
llog_close(env, loghandle);
loghandle = NULL;
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_internal.h b/drivers/staging/lustre/lustre/obdclass/llog_internal.h
index f794952..21a93c7 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_internal.h
+++ b/drivers/staging/lustre/lustre/obdclass/llog_internal.h
@@ -70,4 +70,9 @@ int llog_process_or_fork(const struct lu_env *env,
llog_cb_t cb, void *data, void *catdata, bool fork);
int llog_cat_cleanup(const struct lu_env *env, struct llog_handle *cathandle,
struct llog_handle *loghandle, int index);
+
+static inline struct llog_rec_hdr *llog_rec_hdr_next(struct llog_rec_hdr *rec)
+{
+ return (struct llog_rec_hdr *)((char *)rec + rec->lrh_len);
+}
#endif
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_swab.c b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
index 0ec6361..21ab90c 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_swab.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
@@ -172,20 +172,23 @@ void lustre_swab_llog_rec(struct llog_rec_hdr *rec)
__swab64s(&cr->cr.cr_time);
lustre_swab_lu_fid(&cr->cr.cr_tfid);
lustre_swab_lu_fid(&cr->cr.cr_pfid);
- if (CHANGELOG_REC_EXTENDED(&cr->cr)) {
- struct llog_changelog_ext_rec *ext =
- (struct llog_changelog_ext_rec *)rec;
-
- lustre_swab_lu_fid(&ext->cr.cr_sfid);
- lustre_swab_lu_fid(&ext->cr.cr_spfid);
- tail = &ext->cr_tail;
- } else {
- tail = &cr->cr_tail;
+ if (cr->cr.cr_flags & CLF_RENAME) {
+ struct changelog_ext_rename *rnm =
+ changelog_rec_rename(&cr->cr);
+
+ lustre_swab_lu_fid(&rnm->cr_sfid);
+ lustre_swab_lu_fid(&rnm->cr_spfid);
}
- tail = (struct llog_rec_tail *)((char *)tail +
+ /*
+ * Because the tail follows a variable-length structure we need
+ * to compute its location at runtime
+ */
+ tail = (struct llog_rec_tail *)((char *)&cr->cr +
+ changelog_rec_size(&cr->cr) +
cr->cr.cr_namelen);
break;
}
+
case CHANGELOG_USER_REC:
{
struct llog_changelog_user_rec *cur =
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 03fd5d7..e25596f 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1378,7 +1378,7 @@ void lustre_msg_set_jobid(struct lustre_msg *msg, char *jobid)
LASSERTF(pb, "invalid msg %p: no ptlrpc body!\n", msg);
if (jobid)
- memcpy(pb->pb_jobid, jobid, JOBSTATS_JOBID_SIZE);
+ memcpy(pb->pb_jobid, jobid, LUSTRE_JOBID_SIZE);
else if (pb->pb_jobid[0] == '\0')
lustre_get_jobid(pb->pb_jobid);
return;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index b33fd7b..e5945e2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -684,7 +684,7 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct ptlrpc_body_v3, pb_padding));
LASSERTF((int)sizeof(((struct ptlrpc_body_v3 *)0)->pb_padding) == 32, "found %lld\n",
(long long)(int)sizeof(((struct ptlrpc_body_v3 *)0)->pb_padding));
- CLASSERT(JOBSTATS_JOBID_SIZE == 32);
+ CLASSERT(LUSTRE_JOBID_SIZE == 32);
LASSERTF((int)offsetof(struct ptlrpc_body_v3, pb_jobid) == 152, "found %lld\n",
(long long)(int)offsetof(struct ptlrpc_body_v3, pb_jobid));
LASSERTF((int)sizeof(((struct ptlrpc_body_v3 *)0)->pb_jobid) == 32, "found %lld\n",
@@ -3300,50 +3300,6 @@ void lustre_assert_wire_constants(void)
LASSERTF((int)sizeof(((struct changelog_rec *)0)->cr_pfid) == 16, "found %lld\n",
(long long)(int)sizeof(((struct changelog_rec *)0)->cr_pfid));
- /* Checks for struct changelog_ext_rec */
- LASSERTF((int)sizeof(struct changelog_ext_rec) == 96, "found %lld\n",
- (long long)(int)sizeof(struct changelog_ext_rec));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_namelen) == 0, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_namelen));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_namelen) == 2, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_namelen));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_flags) == 2, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_flags));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_flags) == 2, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_flags));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_type) == 4, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_type));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_type) == 4, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_type));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_index) == 8, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_index));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_index) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_index));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_prev) == 16, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_prev));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_prev) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_prev));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_time) == 24, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_time));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_time) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_time));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_tfid) == 32, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_tfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_tfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_tfid));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_pfid) == 48, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_pfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_pfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_pfid));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_sfid) == 64, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_sfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_sfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_sfid));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_spfid) == 80, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_spfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_spfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_spfid));
-
/* Checks for struct changelog_setinfo */
LASSERTF((int)sizeof(struct changelog_setinfo) == 12, "found %lld\n",
(long long)(int)sizeof(struct changelog_setinfo));
@@ -3367,10 +3323,10 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct llog_changelog_rec, cr));
LASSERTF((int)sizeof(((struct llog_changelog_rec *)0)->cr) == 64, "found %lld\n",
(long long)(int)sizeof(((struct llog_changelog_rec *)0)->cr));
- LASSERTF((int)offsetof(struct llog_changelog_rec, cr_tail) == 80, "found %lld\n",
- (long long)(int)offsetof(struct llog_changelog_rec, cr_tail));
- LASSERTF((int)sizeof(((struct llog_changelog_rec *)0)->cr_tail) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct llog_changelog_rec *)0)->cr_tail));
+ LASSERTF((int)offsetof(struct llog_changelog_rec, cr_do_not_use) == 80, "found %lld\n",
+ (long long)(int)offsetof(struct llog_changelog_rec, cr_do_not_use));
+ LASSERTF((int)sizeof(((struct llog_changelog_rec *)0)->cr_do_not_use) == 8, "found %lld\n",
+ (long long)(int)sizeof(((struct llog_changelog_rec *)0)->cr_do_not_use));
/* Checks for struct llog_changelog_user_rec */
LASSERTF((int)sizeof(struct llog_changelog_user_rec) == 40, "found %lld\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 048/124] staging: lustre: Flexible changelog format.
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Added jobid fields to Changelog records (and extended records). The
CLF_JOBID flags allows to check if the field is present or not (old
format) when reading an entry. Jobids are expressed as 32 chars long,
zero-terminated strings. Updated test_205 in sanity.sh.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1996
Reviewed-on: http://review.whamcloud.com/4060
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Aurelien Degremont <aurelien.degremont@cea.fr>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 2 +-
.../lustre/lustre/include/lustre/lustre_idl.h | 19 +--
.../lustre/lustre/include/lustre/lustre_user.h | 165 +++++++++++++++-----
.../staging/lustre/lustre/llite/llite_internal.h | 2 +-
drivers/staging/lustre/lustre/llite/vvp_req.c | 2 +-
drivers/staging/lustre/lustre/mdc/mdc_request.c | 22 ++-
drivers/staging/lustre/lustre/obdclass/class_obd.c | 6 +-
.../lustre/lustre/obdclass/linux/linux-module.c | 2 +-
drivers/staging/lustre/lustre/obdclass/llog.c | 4 +-
drivers/staging/lustre/lustre/obdclass/llog_cat.c | 4 +-
.../staging/lustre/lustre/obdclass/llog_internal.h | 5 +
drivers/staging/lustre/lustre/obdclass/llog_swab.c | 23 ++--
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 2 +-
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 54 +------
14 files changed, 188 insertions(+), 124 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 5039643..5af1153 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -1899,7 +1899,7 @@ struct cl_req_attr {
/** Generic attributes for the server consumption. */
struct obdo *cra_oa;
/** Jobid */
- char cra_jobid[JOBSTATS_JOBID_SIZE];
+ char cra_jobid[LUSTRE_JOBID_SIZE];
};
/**
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 56b5129..2dc550a 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1101,7 +1101,7 @@ struct lustre_msg_v2 {
/* without gss, ptlrpc_body is put at the first buffer. */
#define PTLRPC_NUM_VERSIONS 4
-#define JOBSTATS_JOBID_SIZE 32 /* 32 bytes string */
+
struct ptlrpc_body_v3 {
struct lustre_handle pb_handle;
__u32 pb_type;
@@ -1123,7 +1123,7 @@ struct ptlrpc_body_v3 {
__u64 pb_pre_versions[PTLRPC_NUM_VERSIONS];
/* padding for future needs */
__u64 pb_padding[4];
- char pb_jobid[JOBSTATS_JOBID_SIZE];
+ char pb_jobid[LUSTRE_JOBID_SIZE];
};
#define ptlrpc_body ptlrpc_body_v3
@@ -3088,15 +3088,9 @@ struct changelog_setinfo {
/** changelog record */
struct llog_changelog_rec {
- struct llog_rec_hdr cr_hdr;
- struct changelog_rec cr;
- struct llog_rec_tail cr_tail; /**< for_sizezof_only */
-} __packed;
-
-struct llog_changelog_ext_rec {
- struct llog_rec_hdr cr_hdr;
- struct changelog_ext_rec cr;
- struct llog_rec_tail cr_tail; /**< for_sizezof_only */
+ struct llog_rec_hdr cr_hdr;
+ struct changelog_rec cr; /**< Variable length field */
+ struct llog_rec_tail cr_do_not_use; /**< for_sizezof_only */
} __packed;
struct llog_changelog_user_rec {
@@ -3181,6 +3175,9 @@ enum llog_flag {
LLOG_F_ZAP_WHEN_EMPTY = 0x1,
LLOG_F_IS_CAT = 0x2,
LLOG_F_IS_PLAIN = 0x4,
+ LLOG_F_EXT_JOBID = BIT(3),
+
+ LLOG_F_EXT_MASK = LLOG_F_EXT_JOBID,
};
struct llog_log_hdr {
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 79cd876..cc0a786 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -673,11 +673,16 @@ static inline const char *changelog_type2str(int type)
}
/* per-record flags */
-#define CLF_VERSION 0x1000
-#define CLF_EXT_VERSION 0x2000
#define CLF_FLAGSHIFT 12
#define CLF_FLAGMASK ((1U << CLF_FLAGSHIFT) - 1)
#define CLF_VERMASK (~CLF_FLAGMASK)
+enum changelog_rec_flags {
+ CLF_VERSION = 0x1000,
+ CLF_RENAME = 0x2000,
+ CLF_JOBID = 0x4000,
+ CLF_SUPPORTED = CLF_VERSION | CLF_RENAME | CLF_JOBID
+};
+
/* Anything under the flagmask may be per-type (if desired) */
/* Flags for unlink */
#define CLF_UNLINK_LAST 0x0001 /* Unlink of last hardlink */
@@ -761,12 +766,35 @@ static inline void hsm_set_cl_error(int *flags, int error)
*flags |= (error << CLF_HSM_ERR_L);
}
-#define CR_MAXSIZE cfs_size_round(2 * NAME_MAX + 1 + \
- sizeof(struct changelog_ext_rec))
+enum changelog_send_flag {
+ /* Not yet implemented */
+ CHANGELOG_FLAG_FOLLOW = BIT(0),
+ /*
+ * Blocking IO makes sense in case of slow user parsing of the records,
+ * but it also prevents us from cleaning up if the records are not
+ * consumed.
+ */
+ CHANGELOG_FLAG_BLOCK = BIT(1),
+ /* Pack jobid into the changelog records if available. */
+ CHANGELOG_FLAG_JOBID = BIT(2),
+};
+
+#define CR_MAXSIZE cfs_size_round(2 * NAME_MAX + 2 + \
+ changelog_rec_offset(CLF_SUPPORTED))
+
+/* 31 usable bytes string + null terminator. */
+#define LUSTRE_JOBID_SIZE 32
+/*
+ * This is the minimal changelog record. It can contain extensions
+ * such as rename fields or process jobid. Its exact content is described
+ * by the cr_flags.
+ *
+ * Extensions are packed in the same order as their corresponding flags.
+ */
struct changelog_rec {
__u16 cr_namelen;
- __u16 cr_flags; /**< (flags&CLF_FLAGMASK)|CLF_VERSION */
+ __u16 cr_flags; /**< \a changelog_rec_flags */
__u32 cr_type; /**< \a changelog_rec_type */
__u64 cr_index; /**< changelog record number */
__u64 cr_prev; /**< last index for this target fid */
@@ -776,55 +804,116 @@ struct changelog_rec {
__u32 cr_markerflags; /**< CL_MARK flags */
};
struct lu_fid cr_pfid; /**< parent fid */
- char cr_name[0]; /**< last element */
} __packed;
-/* changelog_ext_rec is 2*sizeof(lu_fid) bigger than changelog_rec, to save
- * space, only rename uses changelog_ext_rec, while others use changelog_rec to
- * store records.
- */
-struct changelog_ext_rec {
- __u16 cr_namelen;
- __u16 cr_flags; /**< (flags & CLF_FLAGMASK) |
- * CLF_EXT_VERSION
- */
- __u32 cr_type; /**< \a changelog_rec_type */
- __u64 cr_index; /**< changelog record number */
- __u64 cr_prev; /**< last index for this target fid */
- __u64 cr_time;
- union {
- struct lu_fid cr_tfid; /**< target fid */
- __u32 cr_markerflags; /**< CL_MARK flags */
- };
- struct lu_fid cr_pfid; /**< target parent fid */
- struct lu_fid cr_sfid; /**< source fid, or zero */
- struct lu_fid cr_spfid; /**< source parent fid, or zero */
- char cr_name[0]; /**< last element */
-} __packed;
+/* Changelog extension for RENAME. */
+struct changelog_ext_rename {
+ struct lu_fid cr_sfid; /**< source fid, or zero */
+ struct lu_fid cr_spfid; /**< source parent fid, or zero */
+};
+
+/* Changelog extension to include JOBID. */
+struct changelog_ext_jobid {
+ char cr_jobid[LUSTRE_JOBID_SIZE]; /**< zero-terminated string. */
+};
+
+static inline unsigned int changelog_rec_offset(enum changelog_rec_flags crf)
+{
+ unsigned int size = sizeof(struct changelog_rec);
+
+ if (crf & CLF_RENAME)
+ size += sizeof(struct changelog_ext_rename);
+
+ if (crf & CLF_JOBID)
+ size += sizeof(struct changelog_ext_jobid);
-#define CHANGELOG_REC_EXTENDED(rec) \
- (((rec)->cr_flags & CLF_VERMASK) == CLF_EXT_VERSION)
+ return size;
+}
static inline int changelog_rec_size(struct changelog_rec *rec)
{
- return CHANGELOG_REC_EXTENDED(rec) ? sizeof(struct changelog_ext_rec) :
- sizeof(*rec);
+ return changelog_rec_offset(rec->cr_flags);
+}
+
+static inline unsigned int changelog_rec_varsize(struct changelog_rec *rec)
+{
+ return changelog_rec_size(rec) - sizeof(*rec) + rec->cr_namelen;
+}
+
+static inline
+struct changelog_ext_rename *changelog_rec_rename(struct changelog_rec *rec)
+{
+ enum changelog_rec_flags crf = rec->cr_flags & CLF_VERSION;
+
+ return (struct changelog_ext_rename *)((char *)rec +
+ changelog_rec_offset(crf));
+}
+
+/* The jobid follows the rename extension, if present */
+static inline
+struct changelog_ext_jobid *changelog_rec_jobid(struct changelog_rec *rec)
+{
+ enum changelog_rec_flags crf = rec->cr_flags &
+ (CLF_VERSION | CLF_RENAME);
+
+ return (struct changelog_ext_jobid *)((char *)rec +
+ changelog_rec_offset(crf));
}
+/* The name follows the rename and jobid extensions, if present */
static inline char *changelog_rec_name(struct changelog_rec *rec)
{
- return CHANGELOG_REC_EXTENDED(rec) ?
- ((struct changelog_ext_rec *)rec)->cr_name : rec->cr_name;
+ return (char *)rec + changelog_rec_offset(rec->cr_flags &
+ CLF_SUPPORTED);
+}
+
+static inline int changelog_rec_snamelen(struct changelog_rec *rec)
+{
+ return rec->cr_namelen - strlen(changelog_rec_name(rec)) - 1;
}
-static inline int changelog_rec_snamelen(struct changelog_ext_rec *rec)
+static inline char *changelog_rec_sname(struct changelog_rec *rec)
{
- return rec->cr_namelen - strlen(rec->cr_name) - 1;
+ char *cr_name = changelog_rec_name(rec);
+
+ return cr_name + strlen(cr_name) + 1;
}
-static inline char *changelog_rec_sname(struct changelog_ext_rec *rec)
+/*
+ * Remap a record to the desired format as specified by the crf flags.
+ * The record must be big enough to contain the final remapped version.
+ */
+static inline void changelog_remap_rec(struct changelog_rec *rec,
+ enum changelog_rec_flags crf)
{
- return rec->cr_name + strlen(rec->cr_name) + 1;
+ size_t var_size;
+ char *var_part;
+
+ crf &= CLF_SUPPORTED;
+
+ if ((rec->cr_flags & CLF_SUPPORTED) == crf)
+ return;
+
+ if ((crf & CLF_JOBID) && rec->cr_flags & CLF_JOBID) {
+ var_part = (char *)changelog_rec_jobid(rec);
+ var_size = rec->cr_namelen + sizeof(struct changelog_ext_jobid);
+ } else {
+ var_part = changelog_rec_name(rec);
+ var_size = rec->cr_namelen;
+ }
+
+ memmove((char *)rec + changelog_rec_offset(crf & ~CLF_JOBID), var_part,
+ var_size);
+
+ if ((crf & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
+ memset(changelog_rec_rename(rec), 0,
+ sizeof(struct changelog_ext_rename));
+
+ if ((crf & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
+ memset(changelog_rec_jobid(rec), 0,
+ sizeof(struct changelog_ext_jobid));
+
+ rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf;
}
struct ioc_changelog {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index f972ba3..fc707c2 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -232,7 +232,7 @@ struct ll_inode_info {
* so the read/write statistics for jobid will not be
* accurate if the file is shared by different jobs.
*/
- char f_jobid[JOBSTATS_JOBID_SIZE];
+ char f_jobid[LUSTRE_JOBID_SIZE];
} f;
#define lli_size_mutex u.f.f_size_mutex
diff --git a/drivers/staging/lustre/lustre/llite/vvp_req.c b/drivers/staging/lustre/lustre/llite/vvp_req.c
index a6e2672..e3f4c79 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_req.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_req.c
@@ -85,7 +85,7 @@ static void vvp_req_attr_set(const struct lu_env *env,
if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_INVALID_PFID))
oa->o_parent_oid++;
memcpy(attr->cra_jobid, ll_i2info(inode)->lli_jobid,
- JOBSTATS_JOBID_SIZE);
+ LUSTRE_JOBID_SIZE);
}
static void vvp_req_completion(const struct lu_env *env,
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 956b946..aeadcd6 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1859,12 +1859,17 @@ static struct kuc_hdr *changelog_kuc_hdr(char *buf, int len, int flags)
struct changelog_show {
__u64 cs_startrec;
- __u32 cs_flags;
+ enum changelog_send_flag cs_flags;
struct file *cs_fp;
char *cs_buf;
struct obd_device *cs_obd;
};
+static inline char *cs_obd_name(struct changelog_show *cs)
+{
+ return cs->cs_obd->obd_name;
+}
+
static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
struct llog_rec_hdr *hdr, void *data)
{
@@ -1876,7 +1881,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
if (rec->cr_hdr.lrh_type != CHANGELOG_REC) {
rc = -EINVAL;
CERROR("%s: not a changelog rec %x/%d: rc = %d\n",
- cs->cs_obd->obd_name, rec->cr_hdr.lrh_type,
+ cs_obd_name(cs), rec->cr_hdr.lrh_type,
rec->cr.cr_type, rc);
return rc;
}
@@ -1909,6 +1914,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
static int mdc_changelog_send_thread(void *csdata)
{
+ enum llog_flag flags = LLOG_F_IS_CAT;
struct changelog_show *cs = csdata;
struct llog_ctxt *ctxt = NULL;
struct llog_handle *llh = NULL;
@@ -1934,10 +1940,14 @@ static int mdc_changelog_send_thread(void *csdata)
LLOG_OPEN_EXISTS);
if (rc) {
CERROR("%s: fail to open changelog catalog: rc = %d\n",
- cs->cs_obd->obd_name, rc);
+ cs_obd_name(cs), rc);
goto out;
}
- rc = llog_init_handle(NULL, llh, LLOG_F_IS_CAT, NULL);
+
+ if (cs->cs_flags & CHANGELOG_FLAG_JOBID)
+ flags |= LLOG_F_EXT_JOBID;
+
+ rc = llog_init_handle(NULL, llh, flags, NULL);
if (rc) {
CERROR("llog_init_handle failed %d\n", rc);
goto out;
@@ -1990,12 +2000,12 @@ static int mdc_ioc_changelog_send(struct obd_device *obd,
if (IS_ERR(task)) {
rc = PTR_ERR(task);
CERROR("%s: can't start changelog thread: rc = %d\n",
- obd->obd_name, rc);
+ cs_obd_name(cs), rc);
kfree(cs);
} else {
rc = 0;
CDEBUG(D_HSM, "%s: started changelog thread\n",
- obd->obd_name);
+ cs_obd_name(cs));
}
CERROR("Failed to start changelog thread: %d\n", rc);
diff --git a/drivers/staging/lustre/lustre/obdclass/class_obd.c b/drivers/staging/lustre/lustre/obdclass/class_obd.c
index 629d8b5..4c30d6b 100644
--- a/drivers/staging/lustre/lustre/obdclass/class_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/class_obd.c
@@ -81,7 +81,7 @@ EXPORT_SYMBOL(obd_dirty_transit_pages);
char obd_jobid_var[JOBSTATS_JOBID_VAR_MAX_LEN + 1] = JOBSTATS_DISABLE;
EXPORT_SYMBOL(obd_jobid_var);
-char obd_jobid_node[JOBSTATS_JOBID_SIZE + 1];
+char obd_jobid_node[LUSTRE_JOBID_SIZE + 1];
/* Get jobid of current process from stored variable or calculate
* it from pid and user_id.
@@ -92,14 +92,14 @@ char obd_jobid_node[JOBSTATS_JOBID_SIZE + 1];
*/
int lustre_get_jobid(char *jobid)
{
- memset(jobid, 0, JOBSTATS_JOBID_SIZE);
+ memset(jobid, 0, LUSTRE_JOBID_SIZE);
/* Jobstats isn't enabled */
if (strcmp(obd_jobid_var, JOBSTATS_DISABLE) == 0)
return 0;
/* Use process name + fsuid as jobid */
if (strcmp(obd_jobid_var, JOBSTATS_PROCNAME_UID) == 0) {
- snprintf(jobid, JOBSTATS_JOBID_SIZE, "%s.%u",
+ snprintf(jobid, LUSTRE_JOBID_SIZE, "%s.%u",
current_comm(),
from_kuid(&init_user_ns, current_fsuid()));
return 0;
diff --git a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
index 2b691d8..ae172c4 100644
--- a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
+++ b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
@@ -292,7 +292,7 @@ static ssize_t jobid_name_store(struct kobject *kobj, struct attribute *attr,
const char *buffer,
size_t count)
{
- if (!count || count > JOBSTATS_JOBID_SIZE)
+ if (!count || count > LUSTRE_JOBID_SIZE)
return -EINVAL;
memcpy(obd_jobid_node, buffer, count);
diff --git a/drivers/staging/lustre/lustre/obdclass/llog.c b/drivers/staging/lustre/lustre/obdclass/llog.c
index 8f06141..119372c 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog.c
@@ -137,6 +137,7 @@ static int llog_read_header(const struct lu_env *env,
int llog_init_handle(const struct lu_env *env, struct llog_handle *handle,
int flags, struct obd_uuid *uuid)
{
+ enum llog_flag fmt = flags & LLOG_F_EXT_MASK;
struct llog_log_hdr *llh;
int rc;
@@ -194,6 +195,7 @@ int llog_init_handle(const struct lu_env *env, struct llog_handle *handle,
flags, LLOG_F_IS_CAT, LLOG_F_IS_PLAIN);
rc = -EINVAL;
}
+ llh->llh_flags |= fmt;
out:
if (rc) {
kfree(llh);
@@ -262,7 +264,7 @@ repeat:
*/
for (rec = (struct llog_rec_hdr *)buf;
(char *)rec < buf + LLOG_CHUNK_SIZE;
- rec = (struct llog_rec_hdr *)((char *)rec + rec->lrh_len)) {
+ rec = llog_rec_hdr_next(rec)) {
CDEBUG(D_OTHER, "processing rec 0x%p type %#x\n",
rec, rec->lrh_type);
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_cat.c b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
index b27f08f..ce8e2f6 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_cat.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_cat.c
@@ -63,11 +63,13 @@ static int llog_cat_id2handle(const struct lu_env *env,
struct llog_logid *logid)
{
struct llog_handle *loghandle;
+ enum llog_flag fmt;
int rc = 0;
if (!cathandle)
return -EBADF;
+ fmt = cathandle->lgh_hdr->llh_flags & LLOG_F_EXT_MASK;
down_write(&cathandle->lgh_lock);
list_for_each_entry(loghandle, &cathandle->u.chd.chd_head,
u.phd.phd_entry) {
@@ -99,7 +101,7 @@ static int llog_cat_id2handle(const struct lu_env *env,
return rc;
}
- rc = llog_init_handle(env, loghandle, LLOG_F_IS_PLAIN, NULL);
+ rc = llog_init_handle(env, loghandle, fmt | LLOG_F_IS_PLAIN, NULL);
if (rc < 0) {
llog_close(env, loghandle);
loghandle = NULL;
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_internal.h b/drivers/staging/lustre/lustre/obdclass/llog_internal.h
index f794952..21a93c7 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_internal.h
+++ b/drivers/staging/lustre/lustre/obdclass/llog_internal.h
@@ -70,4 +70,9 @@ int llog_process_or_fork(const struct lu_env *env,
llog_cb_t cb, void *data, void *catdata, bool fork);
int llog_cat_cleanup(const struct lu_env *env, struct llog_handle *cathandle,
struct llog_handle *loghandle, int index);
+
+static inline struct llog_rec_hdr *llog_rec_hdr_next(struct llog_rec_hdr *rec)
+{
+ return (struct llog_rec_hdr *)((char *)rec + rec->lrh_len);
+}
#endif
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_swab.c b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
index 0ec6361..21ab90c 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_swab.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
@@ -172,20 +172,23 @@ void lustre_swab_llog_rec(struct llog_rec_hdr *rec)
__swab64s(&cr->cr.cr_time);
lustre_swab_lu_fid(&cr->cr.cr_tfid);
lustre_swab_lu_fid(&cr->cr.cr_pfid);
- if (CHANGELOG_REC_EXTENDED(&cr->cr)) {
- struct llog_changelog_ext_rec *ext =
- (struct llog_changelog_ext_rec *)rec;
-
- lustre_swab_lu_fid(&ext->cr.cr_sfid);
- lustre_swab_lu_fid(&ext->cr.cr_spfid);
- tail = &ext->cr_tail;
- } else {
- tail = &cr->cr_tail;
+ if (cr->cr.cr_flags & CLF_RENAME) {
+ struct changelog_ext_rename *rnm =
+ changelog_rec_rename(&cr->cr);
+
+ lustre_swab_lu_fid(&rnm->cr_sfid);
+ lustre_swab_lu_fid(&rnm->cr_spfid);
}
- tail = (struct llog_rec_tail *)((char *)tail +
+ /*
+ * Because the tail follows a variable-length structure we need
+ * to compute its location at runtime
+ */
+ tail = (struct llog_rec_tail *)((char *)&cr->cr +
+ changelog_rec_size(&cr->cr) +
cr->cr.cr_namelen);
break;
}
+
case CHANGELOG_USER_REC:
{
struct llog_changelog_user_rec *cur =
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 03fd5d7..e25596f 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1378,7 +1378,7 @@ void lustre_msg_set_jobid(struct lustre_msg *msg, char *jobid)
LASSERTF(pb, "invalid msg %p: no ptlrpc body!\n", msg);
if (jobid)
- memcpy(pb->pb_jobid, jobid, JOBSTATS_JOBID_SIZE);
+ memcpy(pb->pb_jobid, jobid, LUSTRE_JOBID_SIZE);
else if (pb->pb_jobid[0] == '\0')
lustre_get_jobid(pb->pb_jobid);
return;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index b33fd7b..e5945e2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -684,7 +684,7 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct ptlrpc_body_v3, pb_padding));
LASSERTF((int)sizeof(((struct ptlrpc_body_v3 *)0)->pb_padding) == 32, "found %lld\n",
(long long)(int)sizeof(((struct ptlrpc_body_v3 *)0)->pb_padding));
- CLASSERT(JOBSTATS_JOBID_SIZE == 32);
+ CLASSERT(LUSTRE_JOBID_SIZE == 32);
LASSERTF((int)offsetof(struct ptlrpc_body_v3, pb_jobid) == 152, "found %lld\n",
(long long)(int)offsetof(struct ptlrpc_body_v3, pb_jobid));
LASSERTF((int)sizeof(((struct ptlrpc_body_v3 *)0)->pb_jobid) == 32, "found %lld\n",
@@ -3300,50 +3300,6 @@ void lustre_assert_wire_constants(void)
LASSERTF((int)sizeof(((struct changelog_rec *)0)->cr_pfid) == 16, "found %lld\n",
(long long)(int)sizeof(((struct changelog_rec *)0)->cr_pfid));
- /* Checks for struct changelog_ext_rec */
- LASSERTF((int)sizeof(struct changelog_ext_rec) == 96, "found %lld\n",
- (long long)(int)sizeof(struct changelog_ext_rec));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_namelen) == 0, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_namelen));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_namelen) == 2, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_namelen));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_flags) == 2, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_flags));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_flags) == 2, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_flags));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_type) == 4, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_type));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_type) == 4, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_type));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_index) == 8, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_index));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_index) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_index));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_prev) == 16, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_prev));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_prev) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_prev));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_time) == 24, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_time));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_time) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_time));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_tfid) == 32, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_tfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_tfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_tfid));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_pfid) == 48, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_pfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_pfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_pfid));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_sfid) == 64, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_sfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_sfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_sfid));
- LASSERTF((int)offsetof(struct changelog_ext_rec, cr_spfid) == 80, "found %lld\n",
- (long long)(int)offsetof(struct changelog_ext_rec, cr_spfid));
- LASSERTF((int)sizeof(((struct changelog_ext_rec *)0)->cr_spfid) == 16, "found %lld\n",
- (long long)(int)sizeof(((struct changelog_ext_rec *)0)->cr_spfid));
-
/* Checks for struct changelog_setinfo */
LASSERTF((int)sizeof(struct changelog_setinfo) == 12, "found %lld\n",
(long long)(int)sizeof(struct changelog_setinfo));
@@ -3367,10 +3323,10 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct llog_changelog_rec, cr));
LASSERTF((int)sizeof(((struct llog_changelog_rec *)0)->cr) == 64, "found %lld\n",
(long long)(int)sizeof(((struct llog_changelog_rec *)0)->cr));
- LASSERTF((int)offsetof(struct llog_changelog_rec, cr_tail) == 80, "found %lld\n",
- (long long)(int)offsetof(struct llog_changelog_rec, cr_tail));
- LASSERTF((int)sizeof(((struct llog_changelog_rec *)0)->cr_tail) == 8, "found %lld\n",
- (long long)(int)sizeof(((struct llog_changelog_rec *)0)->cr_tail));
+ LASSERTF((int)offsetof(struct llog_changelog_rec, cr_do_not_use) == 80, "found %lld\n",
+ (long long)(int)offsetof(struct llog_changelog_rec, cr_do_not_use));
+ LASSERTF((int)sizeof(((struct llog_changelog_rec *)0)->cr_do_not_use) == 8, "found %lld\n",
+ (long long)(int)sizeof(((struct llog_changelog_rec *)0)->cr_do_not_use));
/* Checks for struct llog_changelog_user_rec */
LASSERTF((int)sizeof(struct llog_changelog_user_rec) == 40, "found %lld\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 049/124] staging: lustre: lmv: move some inline functions to lustre_lmv.h
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Fan Yong,
James Simmons
From: Fan Yong <fan.yong@intel.com>
Move some inline code out of lmv core into lustre_lmv.h.
This is to prepare for use outside of the lmv layer in
the future of these functions. Change from passing in
struct lmv_stripe_md to just int for lmv_is_known_hash_type.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5519
Reviewed-on: http://review.whamcloud.com/11845
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_lmv.h | 63 ++++++++++++++++++++
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 2 +-
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 8 +--
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 59 +------------------
4 files changed, 66 insertions(+), 66 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h
index 388161e..d7f7afa 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lmv.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h
@@ -118,4 +118,67 @@ static inline void lmv_le_to_cpu(union lmv_mds_md *lmv_dst,
}
}
+/* This hash is only for testing purpose */
+static inline unsigned int
+lmv_hash_all_chars(unsigned int count, const char *name, int namelen)
+{
+ const unsigned char *p = (const unsigned char *)name;
+ unsigned int c = 0;
+
+ while (--namelen >= 0)
+ c += p[namelen];
+
+ c = c % count;
+
+ return c;
+}
+
+static inline unsigned int
+lmv_hash_fnv1a(unsigned int count, const char *name, int namelen)
+{
+ __u64 hash;
+
+ hash = lustre_hash_fnv_1a_64(name, namelen);
+
+ return do_div(hash, count);
+}
+
+static inline int lmv_name_to_stripe_index(__u32 lmv_hash_type,
+ unsigned int stripe_count,
+ const char *name, int namelen)
+{
+ __u32 hash_type = lmv_hash_type & LMV_HASH_TYPE_MASK;
+ int idx;
+
+ LASSERT(namelen > 0);
+ if (stripe_count <= 1)
+ return 0;
+
+ /* for migrating object, always start from 0 stripe */
+ if (lmv_hash_type & LMV_HASH_FLAG_MIGRATION)
+ return 0;
+
+ switch (hash_type) {
+ case LMV_HASH_TYPE_ALL_CHARS:
+ idx = lmv_hash_all_chars(stripe_count, name, namelen);
+ break;
+ case LMV_HASH_TYPE_FNV_1A_64:
+ idx = lmv_hash_fnv1a(stripe_count, name, namelen);
+ break;
+ default:
+ idx = -EBADFD;
+ break;
+ }
+ CDEBUG(D_INFO, "name %.*s hash_type %d idx %d\n", namelen, name,
+ hash_type, idx);
+
+ return idx;
+}
+
+static inline bool lmv_is_known_hash_type(__u32 type)
+{
+ return (type & LMV_HASH_TYPE_MASK) == LMV_HASH_TYPE_FNV_1A_64 ||
+ (type & LMV_HASH_TYPE_MASK) == LMV_HASH_TYPE_ALL_CHARS;
+}
+
#endif
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 1952454..94577f3 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -414,7 +414,7 @@ static int lmv_intent_lookup(struct obd_export *exp,
* Both migrating dir and unknown hash dir need to try
* all of sub-stripes
*/
- if (lsm && !lmv_is_known_hash_type(lsm)) {
+ if (lsm && !lmv_is_known_hash_type(lsm->lsm_md_hash_type)) {
struct lmv_oinfo *oinfo = &lsm->lsm_md_oinfo[0];
op_data->op_fid1 = oinfo->lmo_fid;
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
index 4a5e385..8f703ea 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h
+++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
@@ -151,15 +151,9 @@ lsm_name_to_stripe_info(const struct lmv_stripe_md *lsm, const char *name,
return &lsm->lsm_md_oinfo[stripe_index];
}
-static inline bool lmv_is_known_hash_type(const struct lmv_stripe_md *lsm)
-{
- return lsm->lsm_md_hash_type == LMV_HASH_TYPE_FNV_1A_64 ||
- lsm->lsm_md_hash_type == LMV_HASH_TYPE_ALL_CHARS;
-}
-
static inline bool lmv_need_try_all_stripes(const struct lmv_stripe_md *lsm)
{
- return !lmv_is_known_hash_type(lsm) ||
+ return !lmv_is_known_hash_type(lsm->lsm_md_hash_type) ||
lsm->lsm_md_hash_type & LMV_HASH_FLAG_MIGRATION;
}
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 8033ae2..d9f8003 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -53,63 +53,6 @@
#include "../include/lustre_kernelcomm.h"
#include "lmv_internal.h"
-/* This hash is only for testing purpose */
-static inline unsigned int
-lmv_hash_all_chars(unsigned int count, const char *name, int namelen)
-{
- const unsigned char *p = (const unsigned char *)name;
- unsigned int c = 0;
-
- while (--namelen >= 0)
- c += p[namelen];
-
- c = c % count;
-
- return c;
-}
-
-static inline unsigned int
-lmv_hash_fnv1a(unsigned int count, const char *name, int namelen)
-{
- __u64 hash;
-
- hash = lustre_hash_fnv_1a_64(name, namelen);
-
- return do_div(hash, count);
-}
-
-int lmv_name_to_stripe_index(__u32 lmv_hash_type, unsigned int stripe_count,
- const char *name, int namelen)
-{
- __u32 hash_type = lmv_hash_type & LMV_HASH_TYPE_MASK;
- int idx;
-
- LASSERT(namelen > 0);
- if (stripe_count <= 1)
- return 0;
-
- /* for migrating object, always start from 0 stripe */
- if (lmv_hash_type & LMV_HASH_FLAG_MIGRATION)
- return 0;
-
- switch (hash_type) {
- case LMV_HASH_TYPE_ALL_CHARS:
- idx = lmv_hash_all_chars(stripe_count, name, namelen);
- break;
- case LMV_HASH_TYPE_FNV_1A_64:
- idx = lmv_hash_fnv1a(stripe_count, name, namelen);
- break;
- default:
- idx = -EBADFD;
- break;
- }
-
- CDEBUG(D_INFO, "name %.*s hash_type %d idx %d\n", namelen, name,
- hash_type, idx);
-
- return idx;
-}
-
static void lmv_activate_target(struct lmv_obd *lmv,
struct lmv_tgt_desc *tgt,
int activate)
@@ -2531,7 +2474,7 @@ retry_unlink:
* inside lmv_locate_target_for_name(), so we only check
* unknown hash type directory here
*/
- if (!lmv_is_known_hash_type(lsm)) {
+ if (!lmv_is_known_hash_type(lsm->lsm_md_hash_type)) {
struct lmv_oinfo *oinfo;
oinfo = &lsm->lsm_md_oinfo[stripe_index];
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 049/124] staging: lustre: lmv: move some inline functions to lustre_lmv.h
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Fan Yong,
James Simmons
From: Fan Yong <fan.yong@intel.com>
Move some inline code out of lmv core into lustre_lmv.h.
This is to prepare for use outside of the lmv layer in
the future of these functions. Change from passing in
struct lmv_stripe_md to just int for lmv_is_known_hash_type.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5519
Reviewed-on: http://review.whamcloud.com/11845
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_lmv.h | 63 ++++++++++++++++++++
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 2 +-
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 8 +--
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 59 +------------------
4 files changed, 66 insertions(+), 66 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_lmv.h b/drivers/staging/lustre/lustre/include/lustre_lmv.h
index 388161e..d7f7afa 100644
--- a/drivers/staging/lustre/lustre/include/lustre_lmv.h
+++ b/drivers/staging/lustre/lustre/include/lustre_lmv.h
@@ -118,4 +118,67 @@ static inline void lmv_le_to_cpu(union lmv_mds_md *lmv_dst,
}
}
+/* This hash is only for testing purpose */
+static inline unsigned int
+lmv_hash_all_chars(unsigned int count, const char *name, int namelen)
+{
+ const unsigned char *p = (const unsigned char *)name;
+ unsigned int c = 0;
+
+ while (--namelen >= 0)
+ c += p[namelen];
+
+ c = c % count;
+
+ return c;
+}
+
+static inline unsigned int
+lmv_hash_fnv1a(unsigned int count, const char *name, int namelen)
+{
+ __u64 hash;
+
+ hash = lustre_hash_fnv_1a_64(name, namelen);
+
+ return do_div(hash, count);
+}
+
+static inline int lmv_name_to_stripe_index(__u32 lmv_hash_type,
+ unsigned int stripe_count,
+ const char *name, int namelen)
+{
+ __u32 hash_type = lmv_hash_type & LMV_HASH_TYPE_MASK;
+ int idx;
+
+ LASSERT(namelen > 0);
+ if (stripe_count <= 1)
+ return 0;
+
+ /* for migrating object, always start from 0 stripe */
+ if (lmv_hash_type & LMV_HASH_FLAG_MIGRATION)
+ return 0;
+
+ switch (hash_type) {
+ case LMV_HASH_TYPE_ALL_CHARS:
+ idx = lmv_hash_all_chars(stripe_count, name, namelen);
+ break;
+ case LMV_HASH_TYPE_FNV_1A_64:
+ idx = lmv_hash_fnv1a(stripe_count, name, namelen);
+ break;
+ default:
+ idx = -EBADFD;
+ break;
+ }
+ CDEBUG(D_INFO, "name %.*s hash_type %d idx %d\n", namelen, name,
+ hash_type, idx);
+
+ return idx;
+}
+
+static inline bool lmv_is_known_hash_type(__u32 type)
+{
+ return (type & LMV_HASH_TYPE_MASK) == LMV_HASH_TYPE_FNV_1A_64 ||
+ (type & LMV_HASH_TYPE_MASK) == LMV_HASH_TYPE_ALL_CHARS;
+}
+
#endif
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 1952454..94577f3 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -414,7 +414,7 @@ static int lmv_intent_lookup(struct obd_export *exp,
* Both migrating dir and unknown hash dir need to try
* all of sub-stripes
*/
- if (lsm && !lmv_is_known_hash_type(lsm)) {
+ if (lsm && !lmv_is_known_hash_type(lsm->lsm_md_hash_type)) {
struct lmv_oinfo *oinfo = &lsm->lsm_md_oinfo[0];
op_data->op_fid1 = oinfo->lmo_fid;
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
index 4a5e385..8f703ea 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h
+++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
@@ -151,15 +151,9 @@ lsm_name_to_stripe_info(const struct lmv_stripe_md *lsm, const char *name,
return &lsm->lsm_md_oinfo[stripe_index];
}
-static inline bool lmv_is_known_hash_type(const struct lmv_stripe_md *lsm)
-{
- return lsm->lsm_md_hash_type == LMV_HASH_TYPE_FNV_1A_64 ||
- lsm->lsm_md_hash_type == LMV_HASH_TYPE_ALL_CHARS;
-}
-
static inline bool lmv_need_try_all_stripes(const struct lmv_stripe_md *lsm)
{
- return !lmv_is_known_hash_type(lsm) ||
+ return !lmv_is_known_hash_type(lsm->lsm_md_hash_type) ||
lsm->lsm_md_hash_type & LMV_HASH_FLAG_MIGRATION;
}
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 8033ae2..d9f8003 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -53,63 +53,6 @@
#include "../include/lustre_kernelcomm.h"
#include "lmv_internal.h"
-/* This hash is only for testing purpose */
-static inline unsigned int
-lmv_hash_all_chars(unsigned int count, const char *name, int namelen)
-{
- const unsigned char *p = (const unsigned char *)name;
- unsigned int c = 0;
-
- while (--namelen >= 0)
- c += p[namelen];
-
- c = c % count;
-
- return c;
-}
-
-static inline unsigned int
-lmv_hash_fnv1a(unsigned int count, const char *name, int namelen)
-{
- __u64 hash;
-
- hash = lustre_hash_fnv_1a_64(name, namelen);
-
- return do_div(hash, count);
-}
-
-int lmv_name_to_stripe_index(__u32 lmv_hash_type, unsigned int stripe_count,
- const char *name, int namelen)
-{
- __u32 hash_type = lmv_hash_type & LMV_HASH_TYPE_MASK;
- int idx;
-
- LASSERT(namelen > 0);
- if (stripe_count <= 1)
- return 0;
-
- /* for migrating object, always start from 0 stripe */
- if (lmv_hash_type & LMV_HASH_FLAG_MIGRATION)
- return 0;
-
- switch (hash_type) {
- case LMV_HASH_TYPE_ALL_CHARS:
- idx = lmv_hash_all_chars(stripe_count, name, namelen);
- break;
- case LMV_HASH_TYPE_FNV_1A_64:
- idx = lmv_hash_fnv1a(stripe_count, name, namelen);
- break;
- default:
- idx = -EBADFD;
- break;
- }
-
- CDEBUG(D_INFO, "name %.*s hash_type %d idx %d\n", namelen, name,
- hash_type, idx);
-
- return idx;
-}
-
static void lmv_activate_target(struct lmv_obd *lmv,
struct lmv_tgt_desc *tgt,
int activate)
@@ -2531,7 +2474,7 @@ retry_unlink:
* inside lmv_locate_target_for_name(), so we only check
* unknown hash type directory here
*/
- if (!lmv_is_known_hash_type(lsm)) {
+ if (!lmv_is_known_hash_type(lsm->lsm_md_hash_type)) {
struct lmv_oinfo *oinfo;
oinfo = &lsm->lsm_md_oinfo[stripe_index];
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 050/124] staging: lustre: ldlm: per-export lock callback timeout
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Vitaly Fertman, James Simmons
From: Vitaly Fertman <vitaly_fertman@xyratex.com>
The lock callback timeout is calculated as an average per namespace.
This does not reflect individual client behavior.
Instead, we should calculate it on a per-export basis.
This is the client side changes for upstream client.
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4942
Reviewed-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Reviewed-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Xyratex-bug-id: MRP-417
Reviewed-on: http://review.whamcloud.com/9336
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_internal.h | 2 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 2 -
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 54 ++++++++++++--------
3 files changed, 33 insertions(+), 25 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h b/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h
index dc0e4af..6536832 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h
@@ -102,7 +102,7 @@ int ldlm_cancel_lru(struct ldlm_namespace *ns, int nr,
int ldlm_cancel_lru_local(struct ldlm_namespace *ns,
struct list_head *cancels, int count, int max,
enum ldlm_cancel_flags cancel_flags, int flags);
-extern int ldlm_enqueue_min;
+extern unsigned int ldlm_enqueue_min;
/* ldlm_resource.c */
int ldlm_resource_putref_locked(struct ldlm_resource *res);
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
index f516106..3c48b4f 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
@@ -1541,8 +1541,6 @@ enum ldlm_error ldlm_lock_enqueue(struct ldlm_namespace *ns,
struct ldlm_lock *lock = *lockp;
struct ldlm_resource *res = lock->l_resource;
- lock->l_last_activity = ktime_get_real_seconds();
-
lock_res_and_lock(lock);
if (lock->l_req_mode == lock->l_granted_mode) {
/* The server returned a blocked lock, but it was granted
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index cc835cb..46842dd 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -63,8 +63,8 @@
#include "ldlm_internal.h"
-int ldlm_enqueue_min = OBD_TIMEOUT_DEFAULT;
-module_param(ldlm_enqueue_min, int, 0644);
+unsigned int ldlm_enqueue_min = OBD_TIMEOUT_DEFAULT;
+module_param(ldlm_enqueue_min, uint, 0644);
MODULE_PARM_DESC(ldlm_enqueue_min, "lock enqueue timeout minimum");
/* in client side, whether the cached locks will be canceled before replay */
@@ -123,44 +123,56 @@ static int ldlm_expired_completion_wait(void *data)
return 0;
}
+/**
+ * Calculate the Completion timeout (covering enqueue, BL AST, data flush,
+ * lock cancel, and their replies). Used for lock completion timeout on the
+ * client side.
+ *
+ * \param[in] lock lock which is waiting the completion callback
+ *
+ * \retval timeout in seconds to wait for the server reply
+ */
/* We use the same basis for both server side and client side functions
* from a single node.
*/
-static int ldlm_get_enq_timeout(struct ldlm_lock *lock)
+static unsigned int ldlm_cp_timeout(struct ldlm_lock *lock)
{
- int timeout = at_get(ldlm_lock_to_ns_at(lock));
+ unsigned int timeout;
if (AT_OFF)
- return obd_timeout / 2;
- /* Since these are non-updating timeouts, we should be conservative.
- * It would be nice to have some kind of "early reply" mechanism for
- * lock callbacks too...
+ return obd_timeout;
+
+ /*
+ * Wait a long time for enqueue - server may have to callback a
+ * lock from another client. Server will evict the other client if it
+ * doesn't respond reasonably, and then give us the lock.
*/
- timeout = min_t(int, at_max, timeout + (timeout >> 1)); /* 150% */
- return max(timeout, ldlm_enqueue_min);
+ timeout = at_get(ldlm_lock_to_ns_at(lock));
+ return max(3 * timeout, ldlm_enqueue_min);
}
/**
* Helper function for ldlm_completion_ast(), updating timings when lock is
* actually granted.
*/
-static int ldlm_completion_tail(struct ldlm_lock *lock)
+static int ldlm_completion_tail(struct ldlm_lock *lock, void *data)
{
long delay;
- int result;
+ int result = 0;
if (ldlm_is_destroyed(lock) || ldlm_is_failed(lock)) {
LDLM_DEBUG(lock, "client-side enqueue: destroyed");
result = -EIO;
+ } else if (!data) {
+ LDLM_DEBUG(lock, "client-side enqueue: granted");
} else {
+ /* Take into AT only CP RPC, not immediately granted locks */
delay = ktime_get_real_seconds() - lock->l_last_activity;
LDLM_DEBUG(lock, "client-side enqueue: granted after %lds",
delay);
/* Update our time estimate */
- at_measured(ldlm_lock_to_ns_at(lock),
- delay);
- result = 0;
+ at_measured(ldlm_lock_to_ns_at(lock), delay);
}
return result;
}
@@ -179,7 +191,7 @@ int ldlm_completion_ast_async(struct ldlm_lock *lock, __u64 flags, void *data)
if (!(flags & LDLM_FL_BLOCKED_MASK)) {
wake_up(&lock->l_waitq);
- return ldlm_completion_tail(lock);
+ return ldlm_completion_tail(lock, data);
}
LDLM_DEBUG(lock, "client-side enqueue returned a blocked lock, going forward");
@@ -238,13 +250,10 @@ noreproc:
if (obd)
imp = obd->u.cli.cl_import;
- /* Wait a long time for enqueue - server may have to callback a
- * lock from another client. Server will evict the other client if it
- * doesn't respond reasonably, and then give us the lock.
- */
- timeout = ldlm_get_enq_timeout(lock) * 2;
+ timeout = ldlm_cp_timeout(lock);
lwd.lwd_lock = lock;
+ lock->l_last_activity = ktime_get_real_seconds();
if (ldlm_is_no_timeout(lock)) {
LDLM_DEBUG(lock, "waiting indefinitely because of NO_TIMEOUT");
@@ -277,7 +286,7 @@ noreproc:
return rc;
}
- return ldlm_completion_tail(lock);
+ return ldlm_completion_tail(lock, data);
}
EXPORT_SYMBOL(ldlm_completion_ast);
@@ -715,6 +724,7 @@ int ldlm_cli_enqueue(struct obd_export *exp, struct ptlrpc_request **reqp,
lock->l_export = NULL;
lock->l_blocking_ast = einfo->ei_cb_bl;
lock->l_flags |= (*flags & (LDLM_FL_NO_LRU | LDLM_FL_EXCL));
+ lock->l_last_activity = ktime_get_real_seconds();
/* lock not sent to server yet */
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 050/124] staging: lustre: ldlm: per-export lock callback timeout
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Vitaly Fertman, James Simmons
From: Vitaly Fertman <vitaly_fertman@xyratex.com>
The lock callback timeout is calculated as an average per namespace.
This does not reflect individual client behavior.
Instead, we should calculate it on a per-export basis.
This is the client side changes for upstream client.
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4942
Reviewed-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Reviewed-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Xyratex-bug-id: MRP-417
Reviewed-on: http://review.whamcloud.com/9336
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_internal.h | 2 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lock.c | 2 -
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 54 ++++++++++++--------
3 files changed, 33 insertions(+), 25 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h b/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h
index dc0e4af..6536832 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h
@@ -102,7 +102,7 @@ int ldlm_cancel_lru(struct ldlm_namespace *ns, int nr,
int ldlm_cancel_lru_local(struct ldlm_namespace *ns,
struct list_head *cancels, int count, int max,
enum ldlm_cancel_flags cancel_flags, int flags);
-extern int ldlm_enqueue_min;
+extern unsigned int ldlm_enqueue_min;
/* ldlm_resource.c */
int ldlm_resource_putref_locked(struct ldlm_resource *res);
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
index f516106..3c48b4f 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lock.c
@@ -1541,8 +1541,6 @@ enum ldlm_error ldlm_lock_enqueue(struct ldlm_namespace *ns,
struct ldlm_lock *lock = *lockp;
struct ldlm_resource *res = lock->l_resource;
- lock->l_last_activity = ktime_get_real_seconds();
-
lock_res_and_lock(lock);
if (lock->l_req_mode == lock->l_granted_mode) {
/* The server returned a blocked lock, but it was granted
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index cc835cb..46842dd 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -63,8 +63,8 @@
#include "ldlm_internal.h"
-int ldlm_enqueue_min = OBD_TIMEOUT_DEFAULT;
-module_param(ldlm_enqueue_min, int, 0644);
+unsigned int ldlm_enqueue_min = OBD_TIMEOUT_DEFAULT;
+module_param(ldlm_enqueue_min, uint, 0644);
MODULE_PARM_DESC(ldlm_enqueue_min, "lock enqueue timeout minimum");
/* in client side, whether the cached locks will be canceled before replay */
@@ -123,44 +123,56 @@ static int ldlm_expired_completion_wait(void *data)
return 0;
}
+/**
+ * Calculate the Completion timeout (covering enqueue, BL AST, data flush,
+ * lock cancel, and their replies). Used for lock completion timeout on the
+ * client side.
+ *
+ * \param[in] lock lock which is waiting the completion callback
+ *
+ * \retval timeout in seconds to wait for the server reply
+ */
/* We use the same basis for both server side and client side functions
* from a single node.
*/
-static int ldlm_get_enq_timeout(struct ldlm_lock *lock)
+static unsigned int ldlm_cp_timeout(struct ldlm_lock *lock)
{
- int timeout = at_get(ldlm_lock_to_ns_at(lock));
+ unsigned int timeout;
if (AT_OFF)
- return obd_timeout / 2;
- /* Since these are non-updating timeouts, we should be conservative.
- * It would be nice to have some kind of "early reply" mechanism for
- * lock callbacks too...
+ return obd_timeout;
+
+ /*
+ * Wait a long time for enqueue - server may have to callback a
+ * lock from another client. Server will evict the other client if it
+ * doesn't respond reasonably, and then give us the lock.
*/
- timeout = min_t(int, at_max, timeout + (timeout >> 1)); /* 150% */
- return max(timeout, ldlm_enqueue_min);
+ timeout = at_get(ldlm_lock_to_ns_at(lock));
+ return max(3 * timeout, ldlm_enqueue_min);
}
/**
* Helper function for ldlm_completion_ast(), updating timings when lock is
* actually granted.
*/
-static int ldlm_completion_tail(struct ldlm_lock *lock)
+static int ldlm_completion_tail(struct ldlm_lock *lock, void *data)
{
long delay;
- int result;
+ int result = 0;
if (ldlm_is_destroyed(lock) || ldlm_is_failed(lock)) {
LDLM_DEBUG(lock, "client-side enqueue: destroyed");
result = -EIO;
+ } else if (!data) {
+ LDLM_DEBUG(lock, "client-side enqueue: granted");
} else {
+ /* Take into AT only CP RPC, not immediately granted locks */
delay = ktime_get_real_seconds() - lock->l_last_activity;
LDLM_DEBUG(lock, "client-side enqueue: granted after %lds",
delay);
/* Update our time estimate */
- at_measured(ldlm_lock_to_ns_at(lock),
- delay);
- result = 0;
+ at_measured(ldlm_lock_to_ns_at(lock), delay);
}
return result;
}
@@ -179,7 +191,7 @@ int ldlm_completion_ast_async(struct ldlm_lock *lock, __u64 flags, void *data)
if (!(flags & LDLM_FL_BLOCKED_MASK)) {
wake_up(&lock->l_waitq);
- return ldlm_completion_tail(lock);
+ return ldlm_completion_tail(lock, data);
}
LDLM_DEBUG(lock, "client-side enqueue returned a blocked lock, going forward");
@@ -238,13 +250,10 @@ noreproc:
if (obd)
imp = obd->u.cli.cl_import;
- /* Wait a long time for enqueue - server may have to callback a
- * lock from another client. Server will evict the other client if it
- * doesn't respond reasonably, and then give us the lock.
- */
- timeout = ldlm_get_enq_timeout(lock) * 2;
+ timeout = ldlm_cp_timeout(lock);
lwd.lwd_lock = lock;
+ lock->l_last_activity = ktime_get_real_seconds();
if (ldlm_is_no_timeout(lock)) {
LDLM_DEBUG(lock, "waiting indefinitely because of NO_TIMEOUT");
@@ -277,7 +286,7 @@ noreproc:
return rc;
}
- return ldlm_completion_tail(lock);
+ return ldlm_completion_tail(lock, data);
}
EXPORT_SYMBOL(ldlm_completion_ast);
@@ -715,6 +724,7 @@ int ldlm_cli_enqueue(struct obd_export *exp, struct ptlrpc_request **reqp,
lock->l_export = NULL;
lock->l_blocking_ast = einfo->ei_cb_bl;
lock->l_flags |= (*flags & (LDLM_FL_NO_LRU | LDLM_FL_EXCL));
+ lock->l_last_activity = ktime_get_real_seconds();
/* lock not sent to server yet */
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 051/124] staging: lustre: llite: ensure all data flush out when umount
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Yang Sheng,
Jinshan Xiong, James Simmons
From: Yang Sheng <yang.sheng@intel.com>
Write out all extents when clear inode. Otherwise we
may lose data while umount.
Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5584
Reviewed-on: http://review.whamcloud.com/12103
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/llite_lib.c | 15 ++++-----------
1 files changed, 4 insertions(+), 11 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 2aab396..5792354 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -1909,20 +1909,13 @@ void ll_delete_inode(struct inode *inode)
* osc_extent implementation at LU-1030.
*/
cl_sync_file_range(inode, 0, OBD_OBJECT_EOF,
- CL_FSYNC_DISCARD, 1);
+ CL_FSYNC_LOCAL, 1);
truncate_inode_pages_final(&inode->i_data);
- /* Workaround for LU-118 */
- if (inode->i_data.nrpages) {
- spin_lock_irq(&inode->i_data.tree_lock);
- spin_unlock_irq(&inode->i_data.tree_lock);
- LASSERTF(inode->i_data.nrpages == 0,
- "inode="DFID"(%p) nrpages=%lu, see http://jira.whamcloud.com/browse/LU-118\n",
- PFID(ll_inode2fid(inode)), inode,
- inode->i_data.nrpages);
- }
- /* Workaround end */
+ LASSERTF(!inode->i_data.nrpages,
+ "inode=" DFID "(%p) nrpages=%lu, see http://jira.whamcloud.com/browse/LU-118\n",
+ PFID(ll_inode2fid(inode)), inode, inode->i_data.nrpages);
ll_clear_inode(inode);
clear_inode(inode);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 051/124] staging: lustre: llite: ensure all data flush out when umount
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Yang Sheng,
Jinshan Xiong, James Simmons
From: Yang Sheng <yang.sheng@intel.com>
Write out all extents when clear inode. Otherwise we
may lose data while umount.
Signed-off-by: Yang Sheng <yang.sheng@intel.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5584
Reviewed-on: http://review.whamcloud.com/12103
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/llite_lib.c | 15 ++++-----------
1 files changed, 4 insertions(+), 11 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 2aab396..5792354 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -1909,20 +1909,13 @@ void ll_delete_inode(struct inode *inode)
* osc_extent implementation at LU-1030.
*/
cl_sync_file_range(inode, 0, OBD_OBJECT_EOF,
- CL_FSYNC_DISCARD, 1);
+ CL_FSYNC_LOCAL, 1);
truncate_inode_pages_final(&inode->i_data);
- /* Workaround for LU-118 */
- if (inode->i_data.nrpages) {
- spin_lock_irq(&inode->i_data.tree_lock);
- spin_unlock_irq(&inode->i_data.tree_lock);
- LASSERTF(inode->i_data.nrpages == 0,
- "inode="DFID"(%p) nrpages=%lu, see http://jira.whamcloud.com/browse/LU-118\n",
- PFID(ll_inode2fid(inode)), inode,
- inode->i_data.nrpages);
- }
- /* Workaround end */
+ LASSERTF(!inode->i_data.nrpages,
+ "inode=" DFID "(%p) nrpages=%lu, see http://jira.whamcloud.com/browse/LU-118\n",
+ PFID(ll_inode2fid(inode)), inode, inode->i_data.nrpages);
ll_clear_inode(inode);
clear_inode(inode);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 052/124] staging: lustre: lmv: add testing for bad name hash
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Fan Yong,
James Simmons
From: Fan Yong <fan.yong@intel.com>
Enable testing of the lfsck recovery feature in the
client code for the case when name hash for some
entry becomes corrupt.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5519
Reviewed-on: http://review.whamcloud.com/11846
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/obd_support.h | 1 +
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 10 +++++++---
2 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h
index d4c41d0..3144cca 100644
--- a/drivers/staging/lustre/lustre/include/obd_support.h
+++ b/drivers/staging/lustre/lustre/include/obd_support.h
@@ -479,6 +479,7 @@ extern char obd_jobid_var[];
#define OBD_FAIL_LFSCK_NO_AUTO 0x160b
#define OBD_FAIL_LFSCK_NO_DOUBLESCAN 0x160c
#define OBD_FAIL_LFSCK_INVALID_PFID 0x1619
+#define OBD_FAIL_LFSCK_BAD_NAME_HASH 0x1628
/* UPDATE */
#define OBD_FAIL_UPDATE_OBJ_NET 0x1700
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index d9f8003..fde71c9 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1575,9 +1575,13 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
const struct lmv_oinfo *oinfo;
struct lmv_tgt_desc *tgt;
- oinfo = lsm_name_to_stripe_info(lsm, name, namelen);
- if (IS_ERR(oinfo))
- return ERR_CAST(oinfo);
+ if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_BAD_NAME_HASH)) {
+ oinfo = &lsm->lsm_md_oinfo[cfs_fail_val];
+ } else {
+ oinfo = lsm_name_to_stripe_info(lsm, name, namelen);
+ if (IS_ERR(oinfo))
+ return ERR_CAST(oinfo);
+ }
*fid = oinfo->lmo_fid;
*mds = oinfo->lmo_mds;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 052/124] staging: lustre: lmv: add testing for bad name hash
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Fan Yong,
James Simmons
From: Fan Yong <fan.yong@intel.com>
Enable testing of the lfsck recovery feature in the
client code for the case when name hash for some
entry becomes corrupt.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5519
Reviewed-on: http://review.whamcloud.com/11846
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/obd_support.h | 1 +
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 10 +++++++---
2 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h
index d4c41d0..3144cca 100644
--- a/drivers/staging/lustre/lustre/include/obd_support.h
+++ b/drivers/staging/lustre/lustre/include/obd_support.h
@@ -479,6 +479,7 @@ extern char obd_jobid_var[];
#define OBD_FAIL_LFSCK_NO_AUTO 0x160b
#define OBD_FAIL_LFSCK_NO_DOUBLESCAN 0x160c
#define OBD_FAIL_LFSCK_INVALID_PFID 0x1619
+#define OBD_FAIL_LFSCK_BAD_NAME_HASH 0x1628
/* UPDATE */
#define OBD_FAIL_UPDATE_OBJ_NET 0x1700
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index d9f8003..fde71c9 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1575,9 +1575,13 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
const struct lmv_oinfo *oinfo;
struct lmv_tgt_desc *tgt;
- oinfo = lsm_name_to_stripe_info(lsm, name, namelen);
- if (IS_ERR(oinfo))
- return ERR_CAST(oinfo);
+ if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_BAD_NAME_HASH)) {
+ oinfo = &lsm->lsm_md_oinfo[cfs_fail_val];
+ } else {
+ oinfo = lsm_name_to_stripe_info(lsm, name, namelen);
+ if (IS_ERR(oinfo))
+ return ERR_CAST(oinfo);
+ }
*fid = oinfo->lmo_fid;
*mds = oinfo->lmo_mds;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 053/124] staging: lustre: obd: restore linkea support
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
Original linkea was only used for the lustre server code
so it was removed from the upstream client. Now it needs
to be restored for client work that uses this infrastructure.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 7 +
.../staging/lustre/lustre/include/lustre_linkea.h | 79 ++++++++
drivers/staging/lustre/lustre/obdclass/Makefile | 2 +-
drivers/staging/lustre/lustre/obdclass/linkea.c | 201 ++++++++++++++++++++
drivers/staging/lustre/lustre/obdclass/lu_object.c | 70 +++++++
5 files changed, 358 insertions(+), 1 deletions(-)
create mode 100644 drivers/staging/lustre/lustre/include/lustre_linkea.h
create mode 100644 drivers/staging/lustre/lustre/obdclass/linkea.c
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index 66ab98a..7b37cae 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -1319,5 +1319,12 @@ struct lu_kmem_descr {
int lu_kmem_init(struct lu_kmem_descr *caches);
void lu_kmem_fini(struct lu_kmem_descr *caches);
+void lu_buf_free(struct lu_buf *buf);
+void lu_buf_alloc(struct lu_buf *buf, size_t size);
+void lu_buf_realloc(struct lu_buf *buf, size_t size);
+
+int lu_buf_check_and_grow(struct lu_buf *buf, size_t len);
+struct lu_buf *lu_buf_check_and_alloc(struct lu_buf *buf, size_t len);
+
/** @} lu */
#endif /* __LUSTRE_LU_OBJECT_H */
diff --git a/drivers/staging/lustre/lustre/include/lustre_linkea.h b/drivers/staging/lustre/lustre/include/lustre_linkea.h
new file mode 100644
index 0000000..249e8bf
--- /dev/null
+++ b/drivers/staging/lustre/lustre/include/lustre_linkea.h
@@ -0,0 +1,79 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2013, 2014, Intel Corporation.
+ * Use is subject to license terms.
+ *
+ * Author: di wang <di.wang@intel.com>
+ */
+
+#define DEFAULT_LINKEA_SIZE 4096
+
+struct linkea_data {
+ /**
+ * Buffer to keep link EA body.
+ */
+ struct lu_buf *ld_buf;
+ /**
+ * The matched header, entry and its length in the EA
+ */
+ struct link_ea_header *ld_leh;
+ struct link_ea_entry *ld_lee;
+ int ld_reclen;
+};
+
+int linkea_data_new(struct linkea_data *ldata, struct lu_buf *buf);
+int linkea_init(struct linkea_data *ldata);
+void linkea_entry_unpack(const struct link_ea_entry *lee, int *reclen,
+ struct lu_name *lname, struct lu_fid *pfid);
+int linkea_entry_pack(struct link_ea_entry *lee, const struct lu_name *lname,
+ const struct lu_fid *pfid);
+int linkea_add_buf(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid);
+void linkea_del_buf(struct linkea_data *ldata, const struct lu_name *lname);
+int linkea_links_find(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid);
+
+static inline void linkea_first_entry(struct linkea_data *ldata)
+{
+ LASSERT(ldata);
+ LASSERT(ldata->ld_leh);
+
+ if (ldata->ld_leh->leh_reccount == 0)
+ ldata->ld_lee = NULL;
+ else
+ ldata->ld_lee = (struct link_ea_entry *)(ldata->ld_leh + 1);
+}
+
+static inline void linkea_next_entry(struct linkea_data *ldata)
+{
+ LASSERT(ldata);
+ LASSERT(ldata->ld_leh);
+
+ if (ldata->ld_lee) {
+ ldata->ld_lee = (struct link_ea_entry *)((char *)ldata->ld_lee +
+ ldata->ld_reclen);
+ if ((char *)ldata->ld_lee >= ((char *)ldata->ld_leh +
+ ldata->ld_leh->leh_len))
+ ldata->ld_lee = NULL;
+ }
+}
diff --git a/drivers/staging/lustre/lustre/obdclass/Makefile b/drivers/staging/lustre/lustre/obdclass/Makefile
index df7e47f..b42e109 100644
--- a/drivers/staging/lustre/lustre/obdclass/Makefile
+++ b/drivers/staging/lustre/lustre/obdclass/Makefile
@@ -3,6 +3,6 @@ obj-$(CONFIG_LUSTRE_FS) += obdclass.o
obdclass-y := linux/linux-module.o linux/linux-obdo.o linux/linux-sysctl.o \
llog.o llog_cat.o llog_obd.o llog_swab.o class_obd.o debug.o \
genops.o uuid.o lprocfs_status.o lprocfs_counters.o \
- lustre_handles.o lustre_peer.o statfs_pack.o \
+ lustre_handles.o lustre_peer.o statfs_pack.o linkea.o \
obdo.o obd_config.o obd_mount.o lu_object.o lu_ref.o \
cl_object.o cl_page.o cl_lock.o cl_io.o kernelcomm.o
diff --git a/drivers/staging/lustre/lustre/obdclass/linkea.c b/drivers/staging/lustre/lustre/obdclass/linkea.c
new file mode 100644
index 0000000..ee0a6e7
--- /dev/null
+++ b/drivers/staging/lustre/lustre/obdclass/linkea.c
@@ -0,0 +1,201 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2013, 2014, Intel Corporation.
+ * Use is subject to license terms.
+ *
+ * Author: Di Wang <di.wang@intel.com>
+ */
+
+#include "../include/lustre/lustre_idl.h"
+#include "../include/obd.h"
+#include "../include/lustre_linkea.h"
+
+int linkea_data_new(struct linkea_data *ldata, struct lu_buf *buf)
+{
+ ldata->ld_buf = lu_buf_check_and_alloc(buf, PAGE_SIZE);
+ if (!ldata->ld_buf->lb_buf)
+ return -ENOMEM;
+ ldata->ld_leh = ldata->ld_buf->lb_buf;
+ ldata->ld_leh->leh_magic = LINK_EA_MAGIC;
+ ldata->ld_leh->leh_len = sizeof(struct link_ea_header);
+ ldata->ld_leh->leh_reccount = 0;
+ return 0;
+}
+EXPORT_SYMBOL(linkea_data_new);
+
+int linkea_init(struct linkea_data *ldata)
+{
+ struct link_ea_header *leh;
+
+ LASSERT(ldata->ld_buf);
+ leh = ldata->ld_buf->lb_buf;
+ if (leh->leh_magic == __swab32(LINK_EA_MAGIC)) {
+ leh->leh_magic = LINK_EA_MAGIC;
+ leh->leh_reccount = __swab32(leh->leh_reccount);
+ leh->leh_len = __swab64(leh->leh_len);
+ /* entries are swabbed by linkea_entry_unpack */
+ }
+ if (leh->leh_magic != LINK_EA_MAGIC)
+ return -EINVAL;
+ if (leh->leh_reccount == 0)
+ return -ENODATA;
+
+ ldata->ld_leh = leh;
+ return 0;
+}
+EXPORT_SYMBOL(linkea_init);
+
+/**
+ * Pack a link_ea_entry.
+ * All elements are stored as chars to avoid alignment issues.
+ * Numbers are always big-endian
+ * \retval record length
+ */
+int linkea_entry_pack(struct link_ea_entry *lee, const struct lu_name *lname,
+ const struct lu_fid *pfid)
+{
+ struct lu_fid tmpfid;
+ int reclen;
+
+ tmpfid = *pfid;
+ if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_LINKEA_CRASH))
+ tmpfid.f_ver = ~0;
+ fid_cpu_to_be(&tmpfid, &tmpfid);
+ memcpy(&lee->lee_parent_fid, &tmpfid, sizeof(tmpfid));
+ memcpy(lee->lee_name, lname->ln_name, lname->ln_namelen);
+ reclen = sizeof(struct link_ea_entry) + lname->ln_namelen;
+
+ lee->lee_reclen[0] = (reclen >> 8) & 0xff;
+ lee->lee_reclen[1] = reclen & 0xff;
+ return reclen;
+}
+EXPORT_SYMBOL(linkea_entry_pack);
+
+void linkea_entry_unpack(const struct link_ea_entry *lee, int *reclen,
+ struct lu_name *lname, struct lu_fid *pfid)
+{
+ *reclen = (lee->lee_reclen[0] << 8) | lee->lee_reclen[1];
+ memcpy(pfid, &lee->lee_parent_fid, sizeof(*pfid));
+ fid_be_to_cpu(pfid, pfid);
+ if (lname) {
+ lname->ln_name = lee->lee_name;
+ lname->ln_namelen = *reclen - sizeof(struct link_ea_entry);
+ }
+}
+EXPORT_SYMBOL(linkea_entry_unpack);
+
+/**
+ * Add a record to the end of link ea buf
+ **/
+int linkea_add_buf(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid)
+{
+ LASSERT(ldata->ld_leh);
+
+ if (!lname || !pfid)
+ return -EINVAL;
+
+ ldata->ld_reclen = lname->ln_namelen + sizeof(struct link_ea_entry);
+ if (ldata->ld_leh->leh_len + ldata->ld_reclen >
+ ldata->ld_buf->lb_len) {
+ if (lu_buf_check_and_grow(ldata->ld_buf,
+ ldata->ld_leh->leh_len +
+ ldata->ld_reclen) < 0)
+ return -ENOMEM;
+ }
+
+ ldata->ld_leh = ldata->ld_buf->lb_buf;
+ ldata->ld_lee = ldata->ld_buf->lb_buf + ldata->ld_leh->leh_len;
+ ldata->ld_reclen = linkea_entry_pack(ldata->ld_lee, lname, pfid);
+ ldata->ld_leh->leh_len += ldata->ld_reclen;
+ ldata->ld_leh->leh_reccount++;
+ CDEBUG(D_INODE, "New link_ea name '" DFID ":%.*s' is added\n",
+ PFID(pfid), lname->ln_namelen, lname->ln_name);
+ return 0;
+}
+EXPORT_SYMBOL(linkea_add_buf);
+
+/** Del the current record from the link ea buf */
+void linkea_del_buf(struct linkea_data *ldata, const struct lu_name *lname)
+{
+ LASSERT(ldata->ld_leh && ldata->ld_lee);
+
+ ldata->ld_leh->leh_reccount--;
+ ldata->ld_leh->leh_len -= ldata->ld_reclen;
+ memmove(ldata->ld_lee, (char *)ldata->ld_lee + ldata->ld_reclen,
+ (char *)ldata->ld_leh + ldata->ld_leh->leh_len -
+ (char *)ldata->ld_lee);
+ CDEBUG(D_INODE, "Old link_ea name '%.*s' is removed\n",
+ lname->ln_namelen, lname->ln_name);
+
+ if ((char *)ldata->ld_lee >= ((char *)ldata->ld_leh +
+ ldata->ld_leh->leh_len))
+ ldata->ld_lee = NULL;
+}
+EXPORT_SYMBOL(linkea_del_buf);
+
+/**
+ * Check if such a link exists in linkEA.
+ *
+ * \param ldata link data the search to be done on
+ * \param lname name in the parent's directory entry pointing to this object
+ * \param pfid parent fid the link to be found for
+ *
+ * \retval 0 success
+ * \retval -ENOENT link does not exist
+ * \retval -ve on error
+ */
+int linkea_links_find(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid)
+{
+ struct lu_name tmpname;
+ struct lu_fid tmpfid;
+ int count;
+
+ LASSERT(ldata->ld_leh);
+
+ /* link #0 */
+ ldata->ld_lee = (struct link_ea_entry *)(ldata->ld_leh + 1);
+
+ for (count = 0; count < ldata->ld_leh->leh_reccount; count++) {
+ linkea_entry_unpack(ldata->ld_lee, &ldata->ld_reclen,
+ &tmpname, &tmpfid);
+ if (tmpname.ln_namelen == lname->ln_namelen &&
+ lu_fid_eq(&tmpfid, pfid) &&
+ (strncmp(tmpname.ln_name, lname->ln_name,
+ tmpname.ln_namelen) == 0))
+ break;
+ ldata->ld_lee = (struct link_ea_entry *)((char *)ldata->ld_lee +
+ ldata->ld_reclen);
+ }
+
+ if (count == ldata->ld_leh->leh_reccount) {
+ CDEBUG(D_INODE, "Old link_ea name '%.*s' not found\n",
+ lname->ln_namelen, lname->ln_name);
+ ldata->ld_lee = NULL;
+ ldata->ld_reclen = 0;
+ return -ENOENT;
+ }
+ return 0;
+}
+EXPORT_SYMBOL(linkea_links_find);
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 44a854c..e284e5d 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -1984,3 +1984,73 @@ void lu_kmem_fini(struct lu_kmem_descr *caches)
}
}
EXPORT_SYMBOL(lu_kmem_fini);
+
+void lu_buf_free(struct lu_buf *buf)
+{
+ LASSERT(buf);
+ if (buf->lb_buf) {
+ LASSERT(buf->lb_len > 0);
+ kvfree(buf->lb_buf);
+ buf->lb_buf = NULL;
+ buf->lb_len = 0;
+ }
+}
+EXPORT_SYMBOL(lu_buf_free);
+
+void lu_buf_alloc(struct lu_buf *buf, size_t size)
+{
+ LASSERT(buf);
+ LASSERT(!buf->lb_buf);
+ LASSERT(!buf->lb_len);
+ buf->lb_buf = libcfs_kvzalloc(size, GFP_NOFS);
+ if (likely(buf->lb_buf))
+ buf->lb_len = size;
+}
+EXPORT_SYMBOL(lu_buf_alloc);
+
+void lu_buf_realloc(struct lu_buf *buf, size_t size)
+{
+ lu_buf_free(buf);
+ lu_buf_alloc(buf, size);
+}
+EXPORT_SYMBOL(lu_buf_realloc);
+
+struct lu_buf *lu_buf_check_and_alloc(struct lu_buf *buf, size_t len)
+{
+ if (!buf->lb_buf && !buf->lb_len)
+ lu_buf_alloc(buf, len);
+
+ if ((len > buf->lb_len) && buf->lb_buf)
+ lu_buf_realloc(buf, len);
+
+ return buf;
+}
+EXPORT_SYMBOL(lu_buf_check_and_alloc);
+
+/**
+ * Increase the size of the \a buf.
+ * preserves old data in buffer
+ * old buffer remains unchanged on error
+ * \retval 0 or -ENOMEM
+ */
+int lu_buf_check_and_grow(struct lu_buf *buf, size_t len)
+{
+ char *ptr;
+
+ if (len <= buf->lb_len)
+ return 0;
+
+ ptr = libcfs_kvzalloc(len, GFP_NOFS);
+ if (!ptr)
+ return -ENOMEM;
+
+ /* Free the old buf */
+ if (buf->lb_buf) {
+ memcpy(ptr, buf->lb_buf, buf->lb_len);
+ kvfree(buf->lb_buf);
+ }
+
+ buf->lb_buf = ptr;
+ buf->lb_len = len;
+ return 0;
+}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 053/124] staging: lustre: obd: restore linkea support
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
Original linkea was only used for the lustre server code
so it was removed from the upstream client. Now it needs
to be restored for client work that uses this infrastructure.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 7 +
.../staging/lustre/lustre/include/lustre_linkea.h | 79 ++++++++
drivers/staging/lustre/lustre/obdclass/Makefile | 2 +-
drivers/staging/lustre/lustre/obdclass/linkea.c | 201 ++++++++++++++++++++
drivers/staging/lustre/lustre/obdclass/lu_object.c | 70 +++++++
5 files changed, 358 insertions(+), 1 deletions(-)
create mode 100644 drivers/staging/lustre/lustre/include/lustre_linkea.h
create mode 100644 drivers/staging/lustre/lustre/obdclass/linkea.c
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index 66ab98a..7b37cae 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -1319,5 +1319,12 @@ struct lu_kmem_descr {
int lu_kmem_init(struct lu_kmem_descr *caches);
void lu_kmem_fini(struct lu_kmem_descr *caches);
+void lu_buf_free(struct lu_buf *buf);
+void lu_buf_alloc(struct lu_buf *buf, size_t size);
+void lu_buf_realloc(struct lu_buf *buf, size_t size);
+
+int lu_buf_check_and_grow(struct lu_buf *buf, size_t len);
+struct lu_buf *lu_buf_check_and_alloc(struct lu_buf *buf, size_t len);
+
/** @} lu */
#endif /* __LUSTRE_LU_OBJECT_H */
diff --git a/drivers/staging/lustre/lustre/include/lustre_linkea.h b/drivers/staging/lustre/lustre/include/lustre_linkea.h
new file mode 100644
index 0000000..249e8bf
--- /dev/null
+++ b/drivers/staging/lustre/lustre/include/lustre_linkea.h
@@ -0,0 +1,79 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2013, 2014, Intel Corporation.
+ * Use is subject to license terms.
+ *
+ * Author: di wang <di.wang@intel.com>
+ */
+
+#define DEFAULT_LINKEA_SIZE 4096
+
+struct linkea_data {
+ /**
+ * Buffer to keep link EA body.
+ */
+ struct lu_buf *ld_buf;
+ /**
+ * The matched header, entry and its length in the EA
+ */
+ struct link_ea_header *ld_leh;
+ struct link_ea_entry *ld_lee;
+ int ld_reclen;
+};
+
+int linkea_data_new(struct linkea_data *ldata, struct lu_buf *buf);
+int linkea_init(struct linkea_data *ldata);
+void linkea_entry_unpack(const struct link_ea_entry *lee, int *reclen,
+ struct lu_name *lname, struct lu_fid *pfid);
+int linkea_entry_pack(struct link_ea_entry *lee, const struct lu_name *lname,
+ const struct lu_fid *pfid);
+int linkea_add_buf(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid);
+void linkea_del_buf(struct linkea_data *ldata, const struct lu_name *lname);
+int linkea_links_find(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid);
+
+static inline void linkea_first_entry(struct linkea_data *ldata)
+{
+ LASSERT(ldata);
+ LASSERT(ldata->ld_leh);
+
+ if (ldata->ld_leh->leh_reccount == 0)
+ ldata->ld_lee = NULL;
+ else
+ ldata->ld_lee = (struct link_ea_entry *)(ldata->ld_leh + 1);
+}
+
+static inline void linkea_next_entry(struct linkea_data *ldata)
+{
+ LASSERT(ldata);
+ LASSERT(ldata->ld_leh);
+
+ if (ldata->ld_lee) {
+ ldata->ld_lee = (struct link_ea_entry *)((char *)ldata->ld_lee +
+ ldata->ld_reclen);
+ if ((char *)ldata->ld_lee >= ((char *)ldata->ld_leh +
+ ldata->ld_leh->leh_len))
+ ldata->ld_lee = NULL;
+ }
+}
diff --git a/drivers/staging/lustre/lustre/obdclass/Makefile b/drivers/staging/lustre/lustre/obdclass/Makefile
index df7e47f..b42e109 100644
--- a/drivers/staging/lustre/lustre/obdclass/Makefile
+++ b/drivers/staging/lustre/lustre/obdclass/Makefile
@@ -3,6 +3,6 @@ obj-$(CONFIG_LUSTRE_FS) += obdclass.o
obdclass-y := linux/linux-module.o linux/linux-obdo.o linux/linux-sysctl.o \
llog.o llog_cat.o llog_obd.o llog_swab.o class_obd.o debug.o \
genops.o uuid.o lprocfs_status.o lprocfs_counters.o \
- lustre_handles.o lustre_peer.o statfs_pack.o \
+ lustre_handles.o lustre_peer.o statfs_pack.o linkea.o \
obdo.o obd_config.o obd_mount.o lu_object.o lu_ref.o \
cl_object.o cl_page.o cl_lock.o cl_io.o kernelcomm.o
diff --git a/drivers/staging/lustre/lustre/obdclass/linkea.c b/drivers/staging/lustre/lustre/obdclass/linkea.c
new file mode 100644
index 0000000..ee0a6e7
--- /dev/null
+++ b/drivers/staging/lustre/lustre/obdclass/linkea.c
@@ -0,0 +1,201 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2013, 2014, Intel Corporation.
+ * Use is subject to license terms.
+ *
+ * Author: Di Wang <di.wang@intel.com>
+ */
+
+#include "../include/lustre/lustre_idl.h"
+#include "../include/obd.h"
+#include "../include/lustre_linkea.h"
+
+int linkea_data_new(struct linkea_data *ldata, struct lu_buf *buf)
+{
+ ldata->ld_buf = lu_buf_check_and_alloc(buf, PAGE_SIZE);
+ if (!ldata->ld_buf->lb_buf)
+ return -ENOMEM;
+ ldata->ld_leh = ldata->ld_buf->lb_buf;
+ ldata->ld_leh->leh_magic = LINK_EA_MAGIC;
+ ldata->ld_leh->leh_len = sizeof(struct link_ea_header);
+ ldata->ld_leh->leh_reccount = 0;
+ return 0;
+}
+EXPORT_SYMBOL(linkea_data_new);
+
+int linkea_init(struct linkea_data *ldata)
+{
+ struct link_ea_header *leh;
+
+ LASSERT(ldata->ld_buf);
+ leh = ldata->ld_buf->lb_buf;
+ if (leh->leh_magic == __swab32(LINK_EA_MAGIC)) {
+ leh->leh_magic = LINK_EA_MAGIC;
+ leh->leh_reccount = __swab32(leh->leh_reccount);
+ leh->leh_len = __swab64(leh->leh_len);
+ /* entries are swabbed by linkea_entry_unpack */
+ }
+ if (leh->leh_magic != LINK_EA_MAGIC)
+ return -EINVAL;
+ if (leh->leh_reccount == 0)
+ return -ENODATA;
+
+ ldata->ld_leh = leh;
+ return 0;
+}
+EXPORT_SYMBOL(linkea_init);
+
+/**
+ * Pack a link_ea_entry.
+ * All elements are stored as chars to avoid alignment issues.
+ * Numbers are always big-endian
+ * \retval record length
+ */
+int linkea_entry_pack(struct link_ea_entry *lee, const struct lu_name *lname,
+ const struct lu_fid *pfid)
+{
+ struct lu_fid tmpfid;
+ int reclen;
+
+ tmpfid = *pfid;
+ if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_LINKEA_CRASH))
+ tmpfid.f_ver = ~0;
+ fid_cpu_to_be(&tmpfid, &tmpfid);
+ memcpy(&lee->lee_parent_fid, &tmpfid, sizeof(tmpfid));
+ memcpy(lee->lee_name, lname->ln_name, lname->ln_namelen);
+ reclen = sizeof(struct link_ea_entry) + lname->ln_namelen;
+
+ lee->lee_reclen[0] = (reclen >> 8) & 0xff;
+ lee->lee_reclen[1] = reclen & 0xff;
+ return reclen;
+}
+EXPORT_SYMBOL(linkea_entry_pack);
+
+void linkea_entry_unpack(const struct link_ea_entry *lee, int *reclen,
+ struct lu_name *lname, struct lu_fid *pfid)
+{
+ *reclen = (lee->lee_reclen[0] << 8) | lee->lee_reclen[1];
+ memcpy(pfid, &lee->lee_parent_fid, sizeof(*pfid));
+ fid_be_to_cpu(pfid, pfid);
+ if (lname) {
+ lname->ln_name = lee->lee_name;
+ lname->ln_namelen = *reclen - sizeof(struct link_ea_entry);
+ }
+}
+EXPORT_SYMBOL(linkea_entry_unpack);
+
+/**
+ * Add a record to the end of link ea buf
+ **/
+int linkea_add_buf(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid)
+{
+ LASSERT(ldata->ld_leh);
+
+ if (!lname || !pfid)
+ return -EINVAL;
+
+ ldata->ld_reclen = lname->ln_namelen + sizeof(struct link_ea_entry);
+ if (ldata->ld_leh->leh_len + ldata->ld_reclen >
+ ldata->ld_buf->lb_len) {
+ if (lu_buf_check_and_grow(ldata->ld_buf,
+ ldata->ld_leh->leh_len +
+ ldata->ld_reclen) < 0)
+ return -ENOMEM;
+ }
+
+ ldata->ld_leh = ldata->ld_buf->lb_buf;
+ ldata->ld_lee = ldata->ld_buf->lb_buf + ldata->ld_leh->leh_len;
+ ldata->ld_reclen = linkea_entry_pack(ldata->ld_lee, lname, pfid);
+ ldata->ld_leh->leh_len += ldata->ld_reclen;
+ ldata->ld_leh->leh_reccount++;
+ CDEBUG(D_INODE, "New link_ea name '" DFID ":%.*s' is added\n",
+ PFID(pfid), lname->ln_namelen, lname->ln_name);
+ return 0;
+}
+EXPORT_SYMBOL(linkea_add_buf);
+
+/** Del the current record from the link ea buf */
+void linkea_del_buf(struct linkea_data *ldata, const struct lu_name *lname)
+{
+ LASSERT(ldata->ld_leh && ldata->ld_lee);
+
+ ldata->ld_leh->leh_reccount--;
+ ldata->ld_leh->leh_len -= ldata->ld_reclen;
+ memmove(ldata->ld_lee, (char *)ldata->ld_lee + ldata->ld_reclen,
+ (char *)ldata->ld_leh + ldata->ld_leh->leh_len -
+ (char *)ldata->ld_lee);
+ CDEBUG(D_INODE, "Old link_ea name '%.*s' is removed\n",
+ lname->ln_namelen, lname->ln_name);
+
+ if ((char *)ldata->ld_lee >= ((char *)ldata->ld_leh +
+ ldata->ld_leh->leh_len))
+ ldata->ld_lee = NULL;
+}
+EXPORT_SYMBOL(linkea_del_buf);
+
+/**
+ * Check if such a link exists in linkEA.
+ *
+ * \param ldata link data the search to be done on
+ * \param lname name in the parent's directory entry pointing to this object
+ * \param pfid parent fid the link to be found for
+ *
+ * \retval 0 success
+ * \retval -ENOENT link does not exist
+ * \retval -ve on error
+ */
+int linkea_links_find(struct linkea_data *ldata, const struct lu_name *lname,
+ const struct lu_fid *pfid)
+{
+ struct lu_name tmpname;
+ struct lu_fid tmpfid;
+ int count;
+
+ LASSERT(ldata->ld_leh);
+
+ /* link #0 */
+ ldata->ld_lee = (struct link_ea_entry *)(ldata->ld_leh + 1);
+
+ for (count = 0; count < ldata->ld_leh->leh_reccount; count++) {
+ linkea_entry_unpack(ldata->ld_lee, &ldata->ld_reclen,
+ &tmpname, &tmpfid);
+ if (tmpname.ln_namelen == lname->ln_namelen &&
+ lu_fid_eq(&tmpfid, pfid) &&
+ (strncmp(tmpname.ln_name, lname->ln_name,
+ tmpname.ln_namelen) == 0))
+ break;
+ ldata->ld_lee = (struct link_ea_entry *)((char *)ldata->ld_lee +
+ ldata->ld_reclen);
+ }
+
+ if (count == ldata->ld_leh->leh_reccount) {
+ CDEBUG(D_INODE, "Old link_ea name '%.*s' not found\n",
+ lname->ln_namelen, lname->ln_name);
+ ldata->ld_lee = NULL;
+ ldata->ld_reclen = 0;
+ return -ENOENT;
+ }
+ return 0;
+}
+EXPORT_SYMBOL(linkea_links_find);
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 44a854c..e284e5d 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -1984,3 +1984,73 @@ void lu_kmem_fini(struct lu_kmem_descr *caches)
}
}
EXPORT_SYMBOL(lu_kmem_fini);
+
+void lu_buf_free(struct lu_buf *buf)
+{
+ LASSERT(buf);
+ if (buf->lb_buf) {
+ LASSERT(buf->lb_len > 0);
+ kvfree(buf->lb_buf);
+ buf->lb_buf = NULL;
+ buf->lb_len = 0;
+ }
+}
+EXPORT_SYMBOL(lu_buf_free);
+
+void lu_buf_alloc(struct lu_buf *buf, size_t size)
+{
+ LASSERT(buf);
+ LASSERT(!buf->lb_buf);
+ LASSERT(!buf->lb_len);
+ buf->lb_buf = libcfs_kvzalloc(size, GFP_NOFS);
+ if (likely(buf->lb_buf))
+ buf->lb_len = size;
+}
+EXPORT_SYMBOL(lu_buf_alloc);
+
+void lu_buf_realloc(struct lu_buf *buf, size_t size)
+{
+ lu_buf_free(buf);
+ lu_buf_alloc(buf, size);
+}
+EXPORT_SYMBOL(lu_buf_realloc);
+
+struct lu_buf *lu_buf_check_and_alloc(struct lu_buf *buf, size_t len)
+{
+ if (!buf->lb_buf && !buf->lb_len)
+ lu_buf_alloc(buf, len);
+
+ if ((len > buf->lb_len) && buf->lb_buf)
+ lu_buf_realloc(buf, len);
+
+ return buf;
+}
+EXPORT_SYMBOL(lu_buf_check_and_alloc);
+
+/**
+ * Increase the size of the \a buf.
+ * preserves old data in buffer
+ * old buffer remains unchanged on error
+ * \retval 0 or -ENOMEM
+ */
+int lu_buf_check_and_grow(struct lu_buf *buf, size_t len)
+{
+ char *ptr;
+
+ if (len <= buf->lb_len)
+ return 0;
+
+ ptr = libcfs_kvzalloc(len, GFP_NOFS);
+ if (!ptr)
+ return -ENOMEM;
+
+ /* Free the old buf */
+ if (buf->lb_buf) {
+ memcpy(ptr, buf->lb_buf, buf->lb_len);
+ kvfree(buf->lb_buf);
+ }
+
+ buf->lb_buf = ptr;
+ buf->lb_len = len;
+ return 0;
+}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 054/124] staging: lustre: llite: Add ioctl to get parent fids from link EA.
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
Thomas Leibovici, James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Added LL_IOC_GETPARENT to retrieve the <parent_fid>/name(s) of a given
entry, based on its link EA. This saves multiple calls to
path2fid/fid2path.
Merged with second later patch that does various cleanups.
Avoid unneeded allocation. Get read-only attributes from the user
getparent structure and write the modified attributes only, instead
of populating a whole structure in kernel and copying it back.
Signed-off-by: Thomas Leibovici <thomas.leibovici@cea.fr>
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3613
Reviewed-on: http://review.whamcloud.com/7069
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5837
Reviewed-on: http://review.whamcloud.com/12527
Reviewed-by: Ned Bass <bass6@llnl.gov>
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 8 ++
.../lustre/lustre/include/lustre/lustre_user.h | 1 +
drivers/staging/lustre/lustre/llite/dir.c | 2 +
drivers/staging/lustre/lustre/llite/file.c | 2 +
.../staging/lustre/lustre/llite/llite_internal.h | 13 ++
drivers/staging/lustre/lustre/llite/llite_lib.c | 125 ++++++++++++++++++++
drivers/staging/lustre/lustre/llite/xattr.c | 10 +--
7 files changed, 152 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 2dc550a..0cc47bc 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -3468,6 +3468,14 @@ struct getinfo_fid2path {
void lustre_swab_fid2path(struct getinfo_fid2path *gf);
+/** path2parent request/reply structures */
+struct getparent {
+ struct lu_fid gp_fid; /**< parent FID */
+ __u32 gp_linkno; /**< hardlink number */
+ __u32 gp_name_size; /**< size of the name field */
+ char gp_name[0]; /**< zero-terminated link name */
+} __packed;
+
enum {
LAYOUT_INTENT_ACCESS = 0,
LAYOUT_INTENT_READ = 1,
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index cc0a786..08ac6e4 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -245,6 +245,7 @@ struct ost_id {
#define LL_IOC_LMV_SET_DEFAULT_STRIPE _IOWR('f', 246, struct lmv_user_md)
#define LL_IOC_MIGRATE _IOR('f', 247, int)
#define LL_IOC_FID2MDTIDX _IOWR('f', 248, struct lu_fid)
+#define LL_IOC_GETPARENT _IOWR('f', 249, struct getparent)
/* Lease types for use as arg and return of LL_IOC_{GET,SET}_LEASE ioctl. */
enum ll_lease_type {
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index 234095e..705de1c 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -1562,6 +1562,8 @@ out_quotactl:
return rc;
case OBD_IOC_FID2PATH:
return ll_fid2path(inode, (void __user *)arg);
+ case LL_IOC_GETPARENT:
+ return ll_getparent(file, (void __user *)arg);
case LL_IOC_FID2MDTIDX: {
struct obd_export *exp = ll_i2mdexp(inode);
struct lu_fid fid;
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 127d0e0..45acc5d 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -2362,6 +2362,8 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
return 0;
}
+ case LL_IOC_GETPARENT:
+ return ll_getparent(file, (struct getparent __user *)arg);
case OBD_IOC_FID2PATH:
return ll_fid2path(inode, (void __user *)arg);
case LL_IOC_DATA_VERSION: {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index fc707c2..1cc427c 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -36,6 +36,7 @@
#include "../include/lustre_ver.h"
#include "../include/lustre_disk.h" /* for s2sbi */
#include "../include/lustre_eacl.h"
+#include "../include/lustre_linkea.h"
/* for struct cl_lock_descr and struct cl_io */
#include "../include/cl_object.h"
@@ -1038,7 +1039,17 @@ static inline __u64 ll_file_maxbytes(struct inode *inode)
/* llite/xattr.c */
extern const struct xattr_handler *ll_xattr_handlers[];
+#define XATTR_USER_T 1
+#define XATTR_TRUSTED_T 2
+#define XATTR_SECURITY_T 3
+#define XATTR_ACL_ACCESS_T 4
+#define XATTR_ACL_DEFAULT_T 5
+#define XATTR_LUSTRE_T 6
+#define XATTR_OTHER_T 7
+
ssize_t ll_listxattr(struct dentry *dentry, char *buffer, size_t size);
+int ll_xattr_list(struct inode *inode, const char *name, int type,
+ void *buffer, size_t size, __u64 valid);
/**
* Common IO arguments for various VFS I/O interfaces.
@@ -1399,6 +1410,8 @@ void ll_xattr_fini(void);
int ll_page_sync_io(const struct lu_env *env, struct cl_io *io,
struct cl_page *page, enum cl_req_type crt);
+int ll_getparent(struct file *file, struct getparent __user *arg);
+
/* lcommon_cl.c */
int cl_setattr_ost(struct inode *inode, const struct iattr *attr);
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 5792354..5cbce1b 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2545,3 +2545,128 @@ void ll_compute_rootsquash_state(struct ll_sb_info *sbi)
}
up_write(&squash->rsi_sem);
}
+
+/**
+ * Parse linkea content to extract information about a given hardlink
+ *
+ * \param[in] ldata - Initialized linkea data
+ * \param[in] linkno - Link identifier
+ * \param[out] parent_fid - The entry's parent FID
+ * \param[in] size - Entry name destination buffer
+ *
+ * \retval 0 on success
+ * \retval Appropriate negative error code on failure
+ */
+static int ll_linkea_decode(struct linkea_data *ldata, unsigned int linkno,
+ struct lu_fid *parent_fid, struct lu_name *ln)
+{
+ unsigned int idx;
+ int rc;
+
+ rc = linkea_init(ldata);
+ if (rc < 0)
+ return rc;
+
+ if (linkno >= ldata->ld_leh->leh_reccount)
+ /* beyond last link */
+ return -ENODATA;
+
+ linkea_first_entry(ldata);
+ for (idx = 0; ldata->ld_lee; idx++) {
+ linkea_entry_unpack(ldata->ld_lee, &ldata->ld_reclen, ln,
+ parent_fid);
+ if (idx == linkno)
+ break;
+
+ linkea_next_entry(ldata);
+ }
+
+ if (idx < linkno)
+ return -ENODATA;
+
+ return 0;
+}
+
+/**
+ * Get parent FID and name of an identified link. Operation is performed for
+ * a given link number, letting the caller iterate over linkno to list one or
+ * all links of an entry.
+ *
+ * \param[in] file - File descriptor against which to perform the operation
+ * \param[in,out] arg - User-filled structure containing the linkno to operate
+ * on and the available size. It is eventually filled with
+ * the requested information or left untouched on error
+ *
+ * \retval - 0 on success
+ * \retval - Appropriate negative error code on failure
+ */
+int ll_getparent(struct file *file, struct getparent __user *arg)
+{
+ struct inode *inode = file_inode(file);
+ struct linkea_data *ldata;
+ struct lu_fid parent_fid;
+ struct lu_buf buf = {
+ .lb_buf = NULL,
+ .lb_len = 0
+ };
+ struct lu_name ln;
+ u32 name_size;
+ u32 linkno;
+ int rc;
+
+ if (!capable(CFS_CAP_DAC_READ_SEARCH) &&
+ !(ll_i2sbi(inode)->ll_flags & LL_SBI_USER_FID2PATH))
+ return -EPERM;
+
+ if (get_user(name_size, &arg->gp_name_size))
+ return -EFAULT;
+
+ if (get_user(linkno, &arg->gp_linkno))
+ return -EFAULT;
+
+ if (name_size > PATH_MAX)
+ return -EINVAL;
+
+ ldata = kzalloc(sizeof(*ldata), GFP_NOFS);
+ if (!ldata)
+ return -ENOMEM;
+
+ rc = linkea_data_new(ldata, &buf);
+ if (rc < 0)
+ goto ldata_free;
+
+ rc = ll_xattr_list(inode, XATTR_NAME_LINK, XATTR_TRUSTED_T, buf.lb_buf,
+ buf.lb_len, OBD_MD_FLXATTR);
+ if (rc < 0)
+ goto lb_free;
+
+ rc = ll_linkea_decode(ldata, linkno, &parent_fid, &ln);
+ if (rc < 0)
+ goto lb_free;
+
+ if (ln.ln_namelen >= name_size) {
+ rc = -EOVERFLOW;
+ goto lb_free;
+ }
+
+ if (copy_to_user(&arg->gp_fid, &parent_fid, sizeof(arg->gp_fid))) {
+ rc = -EFAULT;
+ goto lb_free;
+ }
+
+ if (copy_to_user(&arg->gp_name, ln.ln_name, ln.ln_namelen)) {
+ rc = -EFAULT;
+ goto lb_free;
+ }
+
+ if (put_user('\0', arg->gp_name + ln.ln_namelen)) {
+ rc = -EFAULT;
+ goto lb_free;
+ }
+
+lb_free:
+ lu_buf_free(&buf);
+ldata_free:
+ kfree(ldata);
+ return rc;
+}
diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c
index 17ad04f..0e6a559 100644
--- a/drivers/staging/lustre/lustre/llite/xattr.c
+++ b/drivers/staging/lustre/lustre/llite/xattr.c
@@ -44,14 +44,6 @@
#include "llite_internal.h"
-#define XATTR_USER_T (1)
-#define XATTR_TRUSTED_T (2)
-#define XATTR_SECURITY_T (3)
-#define XATTR_ACL_ACCESS_T (4)
-#define XATTR_ACL_DEFAULT_T (5)
-#define XATTR_LUSTRE_T (6)
-#define XATTR_OTHER_T (7)
-
static
int get_xattr_type(const char *name)
{
@@ -219,7 +211,7 @@ static int ll_xattr_set(const struct xattr_handler *handler,
flags);
}
-static int
+int
ll_xattr_list(struct inode *inode, const char *name, int type, void *buffer,
size_t size, __u64 valid)
{
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 054/124] staging: lustre: llite: Add ioctl to get parent fids from link EA.
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
Thomas Leibovici, James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Added LL_IOC_GETPARENT to retrieve the <parent_fid>/name(s) of a given
entry, based on its link EA. This saves multiple calls to
path2fid/fid2path.
Merged with second later patch that does various cleanups.
Avoid unneeded allocation. Get read-only attributes from the user
getparent structure and write the modified attributes only, instead
of populating a whole structure in kernel and copying it back.
Signed-off-by: Thomas Leibovici <thomas.leibovici@cea.fr>
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3613
Reviewed-on: http://review.whamcloud.com/7069
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5837
Reviewed-on: http://review.whamcloud.com/12527
Reviewed-by: Ned Bass <bass6@llnl.gov>
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 8 ++
.../lustre/lustre/include/lustre/lustre_user.h | 1 +
drivers/staging/lustre/lustre/llite/dir.c | 2 +
drivers/staging/lustre/lustre/llite/file.c | 2 +
.../staging/lustre/lustre/llite/llite_internal.h | 13 ++
drivers/staging/lustre/lustre/llite/llite_lib.c | 125 ++++++++++++++++++++
drivers/staging/lustre/lustre/llite/xattr.c | 10 +--
7 files changed, 152 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 2dc550a..0cc47bc 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -3468,6 +3468,14 @@ struct getinfo_fid2path {
void lustre_swab_fid2path(struct getinfo_fid2path *gf);
+/** path2parent request/reply structures */
+struct getparent {
+ struct lu_fid gp_fid; /**< parent FID */
+ __u32 gp_linkno; /**< hardlink number */
+ __u32 gp_name_size; /**< size of the name field */
+ char gp_name[0]; /**< zero-terminated link name */
+} __packed;
+
enum {
LAYOUT_INTENT_ACCESS = 0,
LAYOUT_INTENT_READ = 1,
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index cc0a786..08ac6e4 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -245,6 +245,7 @@ struct ost_id {
#define LL_IOC_LMV_SET_DEFAULT_STRIPE _IOWR('f', 246, struct lmv_user_md)
#define LL_IOC_MIGRATE _IOR('f', 247, int)
#define LL_IOC_FID2MDTIDX _IOWR('f', 248, struct lu_fid)
+#define LL_IOC_GETPARENT _IOWR('f', 249, struct getparent)
/* Lease types for use as arg and return of LL_IOC_{GET,SET}_LEASE ioctl. */
enum ll_lease_type {
diff --git a/drivers/staging/lustre/lustre/llite/dir.c b/drivers/staging/lustre/lustre/llite/dir.c
index 234095e..705de1c 100644
--- a/drivers/staging/lustre/lustre/llite/dir.c
+++ b/drivers/staging/lustre/lustre/llite/dir.c
@@ -1562,6 +1562,8 @@ out_quotactl:
return rc;
case OBD_IOC_FID2PATH:
return ll_fid2path(inode, (void __user *)arg);
+ case LL_IOC_GETPARENT:
+ return ll_getparent(file, (void __user *)arg);
case LL_IOC_FID2MDTIDX: {
struct obd_export *exp = ll_i2mdexp(inode);
struct lu_fid fid;
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 127d0e0..45acc5d 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -2362,6 +2362,8 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
return 0;
}
+ case LL_IOC_GETPARENT:
+ return ll_getparent(file, (struct getparent __user *)arg);
case OBD_IOC_FID2PATH:
return ll_fid2path(inode, (void __user *)arg);
case LL_IOC_DATA_VERSION: {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index fc707c2..1cc427c 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -36,6 +36,7 @@
#include "../include/lustre_ver.h"
#include "../include/lustre_disk.h" /* for s2sbi */
#include "../include/lustre_eacl.h"
+#include "../include/lustre_linkea.h"
/* for struct cl_lock_descr and struct cl_io */
#include "../include/cl_object.h"
@@ -1038,7 +1039,17 @@ static inline __u64 ll_file_maxbytes(struct inode *inode)
/* llite/xattr.c */
extern const struct xattr_handler *ll_xattr_handlers[];
+#define XATTR_USER_T 1
+#define XATTR_TRUSTED_T 2
+#define XATTR_SECURITY_T 3
+#define XATTR_ACL_ACCESS_T 4
+#define XATTR_ACL_DEFAULT_T 5
+#define XATTR_LUSTRE_T 6
+#define XATTR_OTHER_T 7
+
ssize_t ll_listxattr(struct dentry *dentry, char *buffer, size_t size);
+int ll_xattr_list(struct inode *inode, const char *name, int type,
+ void *buffer, size_t size, __u64 valid);
/**
* Common IO arguments for various VFS I/O interfaces.
@@ -1399,6 +1410,8 @@ void ll_xattr_fini(void);
int ll_page_sync_io(const struct lu_env *env, struct cl_io *io,
struct cl_page *page, enum cl_req_type crt);
+int ll_getparent(struct file *file, struct getparent __user *arg);
+
/* lcommon_cl.c */
int cl_setattr_ost(struct inode *inode, const struct iattr *attr);
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 5792354..5cbce1b 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2545,3 +2545,128 @@ void ll_compute_rootsquash_state(struct ll_sb_info *sbi)
}
up_write(&squash->rsi_sem);
}
+
+/**
+ * Parse linkea content to extract information about a given hardlink
+ *
+ * \param[in] ldata - Initialized linkea data
+ * \param[in] linkno - Link identifier
+ * \param[out] parent_fid - The entry's parent FID
+ * \param[in] size - Entry name destination buffer
+ *
+ * \retval 0 on success
+ * \retval Appropriate negative error code on failure
+ */
+static int ll_linkea_decode(struct linkea_data *ldata, unsigned int linkno,
+ struct lu_fid *parent_fid, struct lu_name *ln)
+{
+ unsigned int idx;
+ int rc;
+
+ rc = linkea_init(ldata);
+ if (rc < 0)
+ return rc;
+
+ if (linkno >= ldata->ld_leh->leh_reccount)
+ /* beyond last link */
+ return -ENODATA;
+
+ linkea_first_entry(ldata);
+ for (idx = 0; ldata->ld_lee; idx++) {
+ linkea_entry_unpack(ldata->ld_lee, &ldata->ld_reclen, ln,
+ parent_fid);
+ if (idx == linkno)
+ break;
+
+ linkea_next_entry(ldata);
+ }
+
+ if (idx < linkno)
+ return -ENODATA;
+
+ return 0;
+}
+
+/**
+ * Get parent FID and name of an identified link. Operation is performed for
+ * a given link number, letting the caller iterate over linkno to list one or
+ * all links of an entry.
+ *
+ * \param[in] file - File descriptor against which to perform the operation
+ * \param[in,out] arg - User-filled structure containing the linkno to operate
+ * on and the available size. It is eventually filled with
+ * the requested information or left untouched on error
+ *
+ * \retval - 0 on success
+ * \retval - Appropriate negative error code on failure
+ */
+int ll_getparent(struct file *file, struct getparent __user *arg)
+{
+ struct inode *inode = file_inode(file);
+ struct linkea_data *ldata;
+ struct lu_fid parent_fid;
+ struct lu_buf buf = {
+ .lb_buf = NULL,
+ .lb_len = 0
+ };
+ struct lu_name ln;
+ u32 name_size;
+ u32 linkno;
+ int rc;
+
+ if (!capable(CFS_CAP_DAC_READ_SEARCH) &&
+ !(ll_i2sbi(inode)->ll_flags & LL_SBI_USER_FID2PATH))
+ return -EPERM;
+
+ if (get_user(name_size, &arg->gp_name_size))
+ return -EFAULT;
+
+ if (get_user(linkno, &arg->gp_linkno))
+ return -EFAULT;
+
+ if (name_size > PATH_MAX)
+ return -EINVAL;
+
+ ldata = kzalloc(sizeof(*ldata), GFP_NOFS);
+ if (!ldata)
+ return -ENOMEM;
+
+ rc = linkea_data_new(ldata, &buf);
+ if (rc < 0)
+ goto ldata_free;
+
+ rc = ll_xattr_list(inode, XATTR_NAME_LINK, XATTR_TRUSTED_T, buf.lb_buf,
+ buf.lb_len, OBD_MD_FLXATTR);
+ if (rc < 0)
+ goto lb_free;
+
+ rc = ll_linkea_decode(ldata, linkno, &parent_fid, &ln);
+ if (rc < 0)
+ goto lb_free;
+
+ if (ln.ln_namelen >= name_size) {
+ rc = -EOVERFLOW;
+ goto lb_free;
+ }
+
+ if (copy_to_user(&arg->gp_fid, &parent_fid, sizeof(arg->gp_fid))) {
+ rc = -EFAULT;
+ goto lb_free;
+ }
+
+ if (copy_to_user(&arg->gp_name, ln.ln_name, ln.ln_namelen)) {
+ rc = -EFAULT;
+ goto lb_free;
+ }
+
+ if (put_user('\0', arg->gp_name + ln.ln_namelen)) {
+ rc = -EFAULT;
+ goto lb_free;
+ }
+
+lb_free:
+ lu_buf_free(&buf);
+ldata_free:
+ kfree(ldata);
+ return rc;
+}
diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c
index 17ad04f..0e6a559 100644
--- a/drivers/staging/lustre/lustre/llite/xattr.c
+++ b/drivers/staging/lustre/lustre/llite/xattr.c
@@ -44,14 +44,6 @@
#include "llite_internal.h"
-#define XATTR_USER_T (1)
-#define XATTR_TRUSTED_T (2)
-#define XATTR_SECURITY_T (3)
-#define XATTR_ACL_ACCESS_T (4)
-#define XATTR_ACL_DEFAULT_T (5)
-#define XATTR_LUSTRE_T (6)
-#define XATTR_OTHER_T (7)
-
static
int get_xattr_type(const char *name)
{
@@ -219,7 +211,7 @@ static int ll_xattr_set(const struct xattr_handler *handler,
flags);
}
-static int
+int
ll_xattr_list(struct inode *inode, const char *name, int type, void *buffer,
size_t size, __u64 valid)
{
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 055/124] staging: lustre: llite: allow setting stripes to specify OSTs
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, Jian Yu, James Simmons, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
Extend the llite layer to support specifying individual target
OSTs. Support specifying OSTs for regular files only. Directory
support will be implemented later in a separate project. With
this a file could have for example a OST index layout of
2,4,5,9,11. In addition, duplicate indices will be eliminated
automatically. Calculate the max easize by ld_active_tgt_count
instead of ld_tgt_count. However this may introduce problems
when the OSTs are in recovery because non sufficient buffer
may be allocated to store EA.
Signed-off-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4665
Reviewed-on: http://review.whamcloud.com/9383
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 2 +
.../lustre/lustre/include/lustre/lustre_user.h | 17 ++++++-----
drivers/staging/lustre/lustre/llite/file.c | 28 +++++++-----------
.../staging/lustre/lustre/llite/llite_internal.h | 20 +++++++++++++
drivers/staging/lustre/lustre/llite/llite_lib.c | 30 ++++++++++++++++++++
drivers/staging/lustre/lustre/llite/xattr.c | 9 ++++--
drivers/staging/lustre/lustre/lov/lov_pack.c | 18 ++++-------
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 14 ---------
8 files changed, 85 insertions(+), 53 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 0cc47bc..a0ca571 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1483,6 +1483,8 @@ enum obdo_flags {
#define LOV_MAGIC_JOIN_V1 (0x0BD20000 | LOV_MAGIC_MAGIC)
#define LOV_MAGIC_V3 (0x0BD30000 | LOV_MAGIC_MAGIC)
#define LOV_MAGIC_MIGRATE (0x0BD40000 | LOV_MAGIC_MAGIC)
+/* reserved for specifying OSTs */
+#define LOV_MAGIC_SPECIFIC (0x0BD50000 | LOV_MAGIC_MAGIC)
#define LOV_MAGIC LOV_MAGIC_V1
/*
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 08ac6e4..3ef5db0 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -280,10 +280,12 @@ enum ll_lease_type {
#define LL_FILE_LOCKLESS_IO 0x00000010 /* server-side locks with cio */
#define LL_FILE_RMTACL 0x00000020
-#define LOV_USER_MAGIC_V1 0x0BD10BD0
-#define LOV_USER_MAGIC LOV_USER_MAGIC_V1
-#define LOV_USER_MAGIC_JOIN_V1 0x0BD20BD0
-#define LOV_USER_MAGIC_V3 0x0BD30BD0
+#define LOV_USER_MAGIC_V1 0x0BD10BD0
+#define LOV_USER_MAGIC LOV_USER_MAGIC_V1
+#define LOV_USER_MAGIC_JOIN_V1 0x0BD20BD0
+#define LOV_USER_MAGIC_V3 0x0BD30BD0
+/* 0x0BD40BD0 is occupied by LOV_MAGIC_MIGRATE */
+#define LOV_USER_MAGIC_SPECIFIC 0x0BD50BD0 /* for specific OSTs */
#define LMV_USER_MAGIC 0x0CD30CD0 /*default lmv magic*/
@@ -361,12 +363,11 @@ struct lov_user_md_v3 { /* LOV EA user data (host-endian) */
static inline __u32 lov_user_md_size(__u16 stripes, __u32 lmm_magic)
{
- if (lmm_magic == LOV_USER_MAGIC_V3)
- return sizeof(struct lov_user_md_v3) +
- stripes * sizeof(struct lov_user_ost_data_v1);
- else
+ if (lmm_magic == LOV_USER_MAGIC_V1)
return sizeof(struct lov_user_md_v1) +
stripes * sizeof(struct lov_user_ost_data_v1);
+ return sizeof(struct lov_user_md_v3) +
+ stripes * sizeof(struct lov_user_ost_data_v1);
}
/* Compile with -D_LARGEFILE64_SOURCE or -D_GNU_SOURCE (or #define) to
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 45acc5d..fc81551 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1540,39 +1540,33 @@ static int ll_lov_setea(struct inode *inode, struct file *file,
static int ll_lov_setstripe(struct inode *inode, struct file *file,
unsigned long arg)
{
- struct lov_user_md_v3 lumv3;
- struct lov_user_md_v1 *lumv1 = (struct lov_user_md_v1 *)&lumv3;
- struct lov_user_md_v1 __user *lumv1p = (void __user *)arg;
- struct lov_user_md_v3 __user *lumv3p = (void __user *)arg;
+ struct lov_user_md __user *lum = (struct lov_user_md __user *)arg;
+ struct lov_user_md *klum;
int lum_size, rc;
__u64 flags = FMODE_WRITE;
- /* first try with v1 which is smaller than v3 */
- lum_size = sizeof(struct lov_user_md_v1);
- if (copy_from_user(lumv1, lumv1p, lum_size))
- return -EFAULT;
-
- if (lumv1->lmm_magic == LOV_USER_MAGIC_V3) {
- lum_size = sizeof(struct lov_user_md_v3);
- if (copy_from_user(&lumv3, lumv3p, lum_size))
- return -EFAULT;
- }
+ rc = ll_copy_user_md(lum, &klum);
+ if (rc < 0)
+ return rc;
- rc = ll_lov_setstripe_ea_info(inode, file->f_path.dentry, flags, lumv1,
+ lum_size = rc;
+ rc = ll_lov_setstripe_ea_info(inode, file->f_path.dentry, flags, klum,
lum_size);
cl_lov_delay_create_clear(&file->f_flags);
if (rc == 0) {
struct lov_stripe_md *lsm;
__u32 gen;
- put_user(0, &lumv1p->lmm_stripe_count);
+ put_user(0, &lum->lmm_stripe_count);
ll_layout_refresh(inode, &gen);
lsm = ccc_inode_lsm_get(inode);
rc = obd_iocontrol(LL_IOC_LOV_GETSTRIPE, ll_i2dtexp(inode),
- 0, lsm, (void __user *)arg);
+ 0, lsm, lum);
ccc_inode_lsm_put(inode, lsm);
}
+
+ kfree(klum);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 1cc427c..9a9fefd 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -875,6 +875,26 @@ int ll_get_obd_name(struct inode *inode, unsigned int cmd, unsigned long arg);
char *ll_get_fsname(struct super_block *sb, char *buf, int buflen);
void ll_compute_rootsquash_state(struct ll_sb_info *sbi);
void ll_open_cleanup(struct super_block *sb, struct ptlrpc_request *open_req);
+ssize_t ll_copy_user_md(const struct lov_user_md __user *md,
+ struct lov_user_md **kbuf);
+
+/* Compute expected user md size when passing in a md from user space */
+static inline ssize_t ll_lov_user_md_size(const struct lov_user_md *lum)
+{
+ switch (lum->lmm_magic) {
+ case LOV_USER_MAGIC_V1:
+ return sizeof(struct lov_user_md_v1);
+ case LOV_USER_MAGIC_V3:
+ return sizeof(struct lov_user_md_v3);
+ case LOV_USER_MAGIC_SPECIFIC:
+ if (lum->lmm_stripe_count > LOV_MAX_STRIPE_COUNT)
+ return -EINVAL;
+
+ return lov_user_md_size(lum->lmm_stripe_count,
+ LOV_USER_MAGIC_SPECIFIC);
+ }
+ return -EINVAL;
+}
/* llite/llite_nfs.c */
extern const struct export_operations lustre_export_operations;
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 5cbce1b..3892d06 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2507,6 +2507,36 @@ void ll_dirty_page_discard_warn(struct page *page, int ioret)
free_page((unsigned long)buf);
}
+ssize_t ll_copy_user_md(const struct lov_user_md __user *md,
+ struct lov_user_md **kbuf)
+{
+ struct lov_user_md lum;
+ ssize_t lum_size;
+
+ if (copy_from_user(&lum, md, sizeof(lum))) {
+ lum_size = -EFAULT;
+ goto no_kbuf;
+ }
+
+ lum_size = ll_lov_user_md_size(&lum);
+ if (lum_size < 0)
+ goto no_kbuf;
+
+ *kbuf = kzalloc(lum_size, GFP_NOFS);
+ if (!*kbuf) {
+ lum_size = -ENOMEM;
+ goto no_kbuf;
+ }
+
+ if (copy_from_user(*kbuf, md, lum_size) != 0) {
+ kfree(*kbuf);
+ *kbuf = NULL;
+ lum_size = -EFAULT;
+ }
+no_kbuf:
+ return lum_size;
+}
+
/*
* Compute llite root squash state after a change of root squash
* configuration setting or add/remove of a lnet nid
diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c
index 0e6a559..e070adb 100644
--- a/drivers/staging/lustre/lustre/llite/xattr.c
+++ b/drivers/staging/lustre/lustre/llite/xattr.c
@@ -189,12 +189,15 @@ static int ll_xattr_set(const struct xattr_handler *handler,
if (lump && S_ISREG(inode->i_mode)) {
__u64 it_flags = FMODE_WRITE;
- int lum_size = (lump->lmm_magic == LOV_USER_MAGIC_V1) ?
- sizeof(*lump) : sizeof(struct lov_user_md_v3);
+ int lum_size;
+
+ lum_size = ll_lov_user_md_size(lump);
+ if (lum_size < 0 || size < lum_size)
+ return 0; /* b=10667: ignore error */
rc = ll_lov_setstripe_ea_info(inode, dentry, it_flags,
lump, lum_size);
- /* b10667: rc always be 0 here for now */
+ /* b=10667: rc always be 0 here for now */
rc = 0;
} else if (S_ISDIR(inode->i_mode)) {
rc = ll_dir_setstripe(inode, lump, 0);
diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
index c654810..595cf16 100644
--- a/drivers/staging/lustre/lustre/lov/lov_pack.c
+++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
@@ -148,16 +148,11 @@ int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
stripe_count = 0;
}
} else {
- /* No need to allocate more than maximum supported stripes.
- * Anyway, this is pretty inaccurate since ld_tgt_count now
- * represents max index and we should rely on the actual number
- * of OSTs instead
+ /*
+ * To calculate maximum easize by active targets at present,
+ * which is exactly the maximum easize to be seen by LOV
*/
- stripe_count = lov_mds_md_max_stripe_count(
- lov->lov_ocd.ocd_max_easize, lmm_magic);
-
- if (stripe_count > lov->desc.ld_tgt_count)
- stripe_count = lov->desc.ld_tgt_count;
+ stripe_count = lov->desc.ld_active_tgt_count;
}
/* XXX LOV STACKING call into osc for sizes */
@@ -403,8 +398,9 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
rc = -EFAULT;
goto out_set;
}
- if ((lum.lmm_magic != LOV_USER_MAGIC) &&
- (lum.lmm_magic != LOV_USER_MAGIC_V3)) {
+ if (lum.lmm_magic != LOV_USER_MAGIC_V1 &&
+ lum.lmm_magic != LOV_USER_MAGIC_V3 &&
+ lum.lmm_magic != LOV_USER_MAGIC_SPECIFIC) {
rc = -EINVAL;
goto out_set;
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index e25596f..87027c5 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1916,19 +1916,6 @@ void lustre_swab_lmv_user_md(struct lmv_user_md *lum)
}
EXPORT_SYMBOL(lustre_swab_lmv_user_md);
-static void print_lum(struct lov_user_md *lum)
-{
- CDEBUG(D_OTHER, "lov_user_md %p:\n", lum);
- CDEBUG(D_OTHER, "\tlmm_magic: %#x\n", lum->lmm_magic);
- CDEBUG(D_OTHER, "\tlmm_pattern: %#x\n", lum->lmm_pattern);
- CDEBUG(D_OTHER, "\tlmm_object_id: %llu\n", lmm_oi_id(&lum->lmm_oi));
- CDEBUG(D_OTHER, "\tlmm_object_gr: %llu\n", lmm_oi_seq(&lum->lmm_oi));
- CDEBUG(D_OTHER, "\tlmm_stripe_size: %#x\n", lum->lmm_stripe_size);
- CDEBUG(D_OTHER, "\tlmm_stripe_count: %#x\n", lum->lmm_stripe_count);
- CDEBUG(D_OTHER, "\tlmm_stripe_offset/lmm_layout_gen: %#x\n",
- lum->lmm_stripe_offset);
-}
-
static void lustre_swab_lmm_oi(struct ost_id *oi)
{
__swab64s(&oi->oi.oi_id);
@@ -1943,7 +1930,6 @@ static void lustre_swab_lov_user_md_common(struct lov_user_md_v1 *lum)
__swab32s(&lum->lmm_stripe_size);
__swab16s(&lum->lmm_stripe_count);
__swab16s(&lum->lmm_stripe_offset);
- print_lum(lum);
}
void lustre_swab_lov_user_md_v1(struct lov_user_md_v1 *lum)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 055/124] staging: lustre: llite: allow setting stripes to specify OSTs
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, Jian Yu, James Simmons, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
Extend the llite layer to support specifying individual target
OSTs. Support specifying OSTs for regular files only. Directory
support will be implemented later in a separate project. With
this a file could have for example a OST index layout of
2,4,5,9,11. In addition, duplicate indices will be eliminated
automatically. Calculate the max easize by ld_active_tgt_count
instead of ld_tgt_count. However this may introduce problems
when the OSTs are in recovery because non sufficient buffer
may be allocated to store EA.
Signed-off-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4665
Reviewed-on: http://review.whamcloud.com/9383
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 2 +
.../lustre/lustre/include/lustre/lustre_user.h | 17 ++++++-----
drivers/staging/lustre/lustre/llite/file.c | 28 +++++++-----------
.../staging/lustre/lustre/llite/llite_internal.h | 20 +++++++++++++
drivers/staging/lustre/lustre/llite/llite_lib.c | 30 ++++++++++++++++++++
drivers/staging/lustre/lustre/llite/xattr.c | 9 ++++--
drivers/staging/lustre/lustre/lov/lov_pack.c | 18 ++++-------
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 14 ---------
8 files changed, 85 insertions(+), 53 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 0cc47bc..a0ca571 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1483,6 +1483,8 @@ enum obdo_flags {
#define LOV_MAGIC_JOIN_V1 (0x0BD20000 | LOV_MAGIC_MAGIC)
#define LOV_MAGIC_V3 (0x0BD30000 | LOV_MAGIC_MAGIC)
#define LOV_MAGIC_MIGRATE (0x0BD40000 | LOV_MAGIC_MAGIC)
+/* reserved for specifying OSTs */
+#define LOV_MAGIC_SPECIFIC (0x0BD50000 | LOV_MAGIC_MAGIC)
#define LOV_MAGIC LOV_MAGIC_V1
/*
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 08ac6e4..3ef5db0 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -280,10 +280,12 @@ enum ll_lease_type {
#define LL_FILE_LOCKLESS_IO 0x00000010 /* server-side locks with cio */
#define LL_FILE_RMTACL 0x00000020
-#define LOV_USER_MAGIC_V1 0x0BD10BD0
-#define LOV_USER_MAGIC LOV_USER_MAGIC_V1
-#define LOV_USER_MAGIC_JOIN_V1 0x0BD20BD0
-#define LOV_USER_MAGIC_V3 0x0BD30BD0
+#define LOV_USER_MAGIC_V1 0x0BD10BD0
+#define LOV_USER_MAGIC LOV_USER_MAGIC_V1
+#define LOV_USER_MAGIC_JOIN_V1 0x0BD20BD0
+#define LOV_USER_MAGIC_V3 0x0BD30BD0
+/* 0x0BD40BD0 is occupied by LOV_MAGIC_MIGRATE */
+#define LOV_USER_MAGIC_SPECIFIC 0x0BD50BD0 /* for specific OSTs */
#define LMV_USER_MAGIC 0x0CD30CD0 /*default lmv magic*/
@@ -361,12 +363,11 @@ struct lov_user_md_v3 { /* LOV EA user data (host-endian) */
static inline __u32 lov_user_md_size(__u16 stripes, __u32 lmm_magic)
{
- if (lmm_magic == LOV_USER_MAGIC_V3)
- return sizeof(struct lov_user_md_v3) +
- stripes * sizeof(struct lov_user_ost_data_v1);
- else
+ if (lmm_magic == LOV_USER_MAGIC_V1)
return sizeof(struct lov_user_md_v1) +
stripes * sizeof(struct lov_user_ost_data_v1);
+ return sizeof(struct lov_user_md_v3) +
+ stripes * sizeof(struct lov_user_ost_data_v1);
}
/* Compile with -D_LARGEFILE64_SOURCE or -D_GNU_SOURCE (or #define) to
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 45acc5d..fc81551 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1540,39 +1540,33 @@ static int ll_lov_setea(struct inode *inode, struct file *file,
static int ll_lov_setstripe(struct inode *inode, struct file *file,
unsigned long arg)
{
- struct lov_user_md_v3 lumv3;
- struct lov_user_md_v1 *lumv1 = (struct lov_user_md_v1 *)&lumv3;
- struct lov_user_md_v1 __user *lumv1p = (void __user *)arg;
- struct lov_user_md_v3 __user *lumv3p = (void __user *)arg;
+ struct lov_user_md __user *lum = (struct lov_user_md __user *)arg;
+ struct lov_user_md *klum;
int lum_size, rc;
__u64 flags = FMODE_WRITE;
- /* first try with v1 which is smaller than v3 */
- lum_size = sizeof(struct lov_user_md_v1);
- if (copy_from_user(lumv1, lumv1p, lum_size))
- return -EFAULT;
-
- if (lumv1->lmm_magic == LOV_USER_MAGIC_V3) {
- lum_size = sizeof(struct lov_user_md_v3);
- if (copy_from_user(&lumv3, lumv3p, lum_size))
- return -EFAULT;
- }
+ rc = ll_copy_user_md(lum, &klum);
+ if (rc < 0)
+ return rc;
- rc = ll_lov_setstripe_ea_info(inode, file->f_path.dentry, flags, lumv1,
+ lum_size = rc;
+ rc = ll_lov_setstripe_ea_info(inode, file->f_path.dentry, flags, klum,
lum_size);
cl_lov_delay_create_clear(&file->f_flags);
if (rc == 0) {
struct lov_stripe_md *lsm;
__u32 gen;
- put_user(0, &lumv1p->lmm_stripe_count);
+ put_user(0, &lum->lmm_stripe_count);
ll_layout_refresh(inode, &gen);
lsm = ccc_inode_lsm_get(inode);
rc = obd_iocontrol(LL_IOC_LOV_GETSTRIPE, ll_i2dtexp(inode),
- 0, lsm, (void __user *)arg);
+ 0, lsm, lum);
ccc_inode_lsm_put(inode, lsm);
}
+
+ kfree(klum);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 1cc427c..9a9fefd 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -875,6 +875,26 @@ int ll_get_obd_name(struct inode *inode, unsigned int cmd, unsigned long arg);
char *ll_get_fsname(struct super_block *sb, char *buf, int buflen);
void ll_compute_rootsquash_state(struct ll_sb_info *sbi);
void ll_open_cleanup(struct super_block *sb, struct ptlrpc_request *open_req);
+ssize_t ll_copy_user_md(const struct lov_user_md __user *md,
+ struct lov_user_md **kbuf);
+
+/* Compute expected user md size when passing in a md from user space */
+static inline ssize_t ll_lov_user_md_size(const struct lov_user_md *lum)
+{
+ switch (lum->lmm_magic) {
+ case LOV_USER_MAGIC_V1:
+ return sizeof(struct lov_user_md_v1);
+ case LOV_USER_MAGIC_V3:
+ return sizeof(struct lov_user_md_v3);
+ case LOV_USER_MAGIC_SPECIFIC:
+ if (lum->lmm_stripe_count > LOV_MAX_STRIPE_COUNT)
+ return -EINVAL;
+
+ return lov_user_md_size(lum->lmm_stripe_count,
+ LOV_USER_MAGIC_SPECIFIC);
+ }
+ return -EINVAL;
+}
/* llite/llite_nfs.c */
extern const struct export_operations lustre_export_operations;
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 5cbce1b..3892d06 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2507,6 +2507,36 @@ void ll_dirty_page_discard_warn(struct page *page, int ioret)
free_page((unsigned long)buf);
}
+ssize_t ll_copy_user_md(const struct lov_user_md __user *md,
+ struct lov_user_md **kbuf)
+{
+ struct lov_user_md lum;
+ ssize_t lum_size;
+
+ if (copy_from_user(&lum, md, sizeof(lum))) {
+ lum_size = -EFAULT;
+ goto no_kbuf;
+ }
+
+ lum_size = ll_lov_user_md_size(&lum);
+ if (lum_size < 0)
+ goto no_kbuf;
+
+ *kbuf = kzalloc(lum_size, GFP_NOFS);
+ if (!*kbuf) {
+ lum_size = -ENOMEM;
+ goto no_kbuf;
+ }
+
+ if (copy_from_user(*kbuf, md, lum_size) != 0) {
+ kfree(*kbuf);
+ *kbuf = NULL;
+ lum_size = -EFAULT;
+ }
+no_kbuf:
+ return lum_size;
+}
+
/*
* Compute llite root squash state after a change of root squash
* configuration setting or add/remove of a lnet nid
diff --git a/drivers/staging/lustre/lustre/llite/xattr.c b/drivers/staging/lustre/lustre/llite/xattr.c
index 0e6a559..e070adb 100644
--- a/drivers/staging/lustre/lustre/llite/xattr.c
+++ b/drivers/staging/lustre/lustre/llite/xattr.c
@@ -189,12 +189,15 @@ static int ll_xattr_set(const struct xattr_handler *handler,
if (lump && S_ISREG(inode->i_mode)) {
__u64 it_flags = FMODE_WRITE;
- int lum_size = (lump->lmm_magic == LOV_USER_MAGIC_V1) ?
- sizeof(*lump) : sizeof(struct lov_user_md_v3);
+ int lum_size;
+
+ lum_size = ll_lov_user_md_size(lump);
+ if (lum_size < 0 || size < lum_size)
+ return 0; /* b=10667: ignore error */
rc = ll_lov_setstripe_ea_info(inode, dentry, it_flags,
lump, lum_size);
- /* b10667: rc always be 0 here for now */
+ /* b=10667: rc always be 0 here for now */
rc = 0;
} else if (S_ISDIR(inode->i_mode)) {
rc = ll_dir_setstripe(inode, lump, 0);
diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
index c654810..595cf16 100644
--- a/drivers/staging/lustre/lustre/lov/lov_pack.c
+++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
@@ -148,16 +148,11 @@ int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
stripe_count = 0;
}
} else {
- /* No need to allocate more than maximum supported stripes.
- * Anyway, this is pretty inaccurate since ld_tgt_count now
- * represents max index and we should rely on the actual number
- * of OSTs instead
+ /*
+ * To calculate maximum easize by active targets at present,
+ * which is exactly the maximum easize to be seen by LOV
*/
- stripe_count = lov_mds_md_max_stripe_count(
- lov->lov_ocd.ocd_max_easize, lmm_magic);
-
- if (stripe_count > lov->desc.ld_tgt_count)
- stripe_count = lov->desc.ld_tgt_count;
+ stripe_count = lov->desc.ld_active_tgt_count;
}
/* XXX LOV STACKING call into osc for sizes */
@@ -403,8 +398,9 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
rc = -EFAULT;
goto out_set;
}
- if ((lum.lmm_magic != LOV_USER_MAGIC) &&
- (lum.lmm_magic != LOV_USER_MAGIC_V3)) {
+ if (lum.lmm_magic != LOV_USER_MAGIC_V1 &&
+ lum.lmm_magic != LOV_USER_MAGIC_V3 &&
+ lum.lmm_magic != LOV_USER_MAGIC_SPECIFIC) {
rc = -EINVAL;
goto out_set;
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index e25596f..87027c5 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -1916,19 +1916,6 @@ void lustre_swab_lmv_user_md(struct lmv_user_md *lum)
}
EXPORT_SYMBOL(lustre_swab_lmv_user_md);
-static void print_lum(struct lov_user_md *lum)
-{
- CDEBUG(D_OTHER, "lov_user_md %p:\n", lum);
- CDEBUG(D_OTHER, "\tlmm_magic: %#x\n", lum->lmm_magic);
- CDEBUG(D_OTHER, "\tlmm_pattern: %#x\n", lum->lmm_pattern);
- CDEBUG(D_OTHER, "\tlmm_object_id: %llu\n", lmm_oi_id(&lum->lmm_oi));
- CDEBUG(D_OTHER, "\tlmm_object_gr: %llu\n", lmm_oi_seq(&lum->lmm_oi));
- CDEBUG(D_OTHER, "\tlmm_stripe_size: %#x\n", lum->lmm_stripe_size);
- CDEBUG(D_OTHER, "\tlmm_stripe_count: %#x\n", lum->lmm_stripe_count);
- CDEBUG(D_OTHER, "\tlmm_stripe_offset/lmm_layout_gen: %#x\n",
- lum->lmm_stripe_offset);
-}
-
static void lustre_swab_lmm_oi(struct ost_id *oi)
{
__swab64s(&oi->oi.oi_id);
@@ -1943,7 +1930,6 @@ static void lustre_swab_lov_user_md_common(struct lov_user_md_v1 *lum)
__swab32s(&lum->lmm_stripe_size);
__swab16s(&lum->lmm_stripe_count);
__swab16s(&lum->lmm_stripe_offset);
- print_lum(lum);
}
void lustre_swab_lov_user_md_v1(struct lov_user_md_v1 *lum)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 056/124] staging: lustre: statahead: use dcache-like interface for sa entry
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
Rename ll_sa_entry to sa_entry, and manage sa_entry cache with
dcache-like interfaces.
sa_entry is not needed to be refcounted, because only scanner
can free it, so after it's put in stat list, statahead thread
shouldn't access it any longer.
ll_statahead_interpret() doesn't need to take sai refcount,
because statahead thread will wait for all inflight RPC to
finish.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9664
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
.../staging/lustre/lustre/llite/llite_internal.h | 10 +-
drivers/staging/lustre/lustre/llite/statahead.c | 461 +++++++++-----------
3 files changed, 207 insertions(+), 266 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 1c884c0..6758bf4 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -812,7 +812,7 @@ struct md_enqueue_info {
struct inode *mi_dir;
int (*mi_cb)(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc);
- __u64 mi_cbdata;
+ void *mi_cbdata;
};
struct obd_ops {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 9a9fefd..cd639f4 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -1131,10 +1131,12 @@ struct ll_statahead_info {
wait_queue_head_t sai_waitq; /* stat-ahead wait queue */
struct ptlrpc_thread sai_thread; /* stat-ahead thread */
struct ptlrpc_thread sai_agl_thread; /* AGL thread */
- struct list_head sai_entries; /* entry list */
- struct list_head sai_entries_received; /* entries returned */
- struct list_head sai_entries_stated; /* entries stated */
- struct list_head sai_entries_agl; /* AGL entries to be sent */
+ struct list_head sai_interim_entries; /* entries which got async
+ * stat reply, but not
+ * instantiated
+ */
+ struct list_head sai_entries; /* completed entries */
+ struct list_head sai_agls; /* AGLs to be sent */
struct list_head sai_cache[LL_SA_CACHE_SIZE];
spinlock_t sai_cache_lock[LL_SA_CACHE_SIZE];
atomic_t sai_cache_count; /* entry count in cache */
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 5760974..d041720 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -49,24 +49,26 @@ enum se_stat {
SA_ENTRY_INIT = 0, /** init entry */
SA_ENTRY_SUCC = 1, /** stat succeed */
SA_ENTRY_INVA = 2, /** invalid entry */
- SA_ENTRY_DEST = 3, /** entry to be destroyed */
};
-struct ll_sa_entry {
- /* link into sai->sai_entries */
- struct list_head se_link;
- /* link into sai->sai_entries_{received,stated} */
+/*
+ * sa_entry is not refcounted: statahead thread allocates it and do async stat,
+ * and in async stat callback ll_statahead_interpret() will add it into
+ * sai_cb_entries, later statahead thread will call sa_handle_callback() to
+ * instantiate entry and move it into sai_entries, and then only scanner process
+ * can access and free it.
+ */
+struct sa_entry {
+ /* link into sai_cb_entries or sai_entries */
struct list_head se_list;
/* link into sai hash table locally */
struct list_head se_hash;
- /* entry reference count */
- atomic_t se_refcount;
/* entry index in the sai */
__u64 se_index;
/* low layer ldlm lock handle */
__u64 se_handle;
/* entry status */
- enum se_stat se_stat;
+ enum se_stat se_state;
/* entry size, contains name */
int se_size;
/* pointer to async getattr enqueue info */
@@ -85,13 +87,13 @@ static DEFINE_SPINLOCK(sai_generation_lock);
/*
* The entry only can be released by the caller, it is necessary to hold lock.
*/
-static inline int ll_sa_entry_stated(struct ll_sa_entry *entry)
+static inline int sa_ready(struct sa_entry *entry)
{
smp_rmb();
- return (entry->se_stat != SA_ENTRY_INIT);
+ return (entry->se_state != SA_ENTRY_INIT);
}
-static inline int ll_sa_entry_hash(int val)
+static inline int sa_hash(int val)
{
return val & LL_SA_CACHE_MASK;
}
@@ -100,9 +102,9 @@ static inline int ll_sa_entry_hash(int val)
* Insert entry to hash SA table.
*/
static inline void
-ll_sa_entry_enhash(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_rehash(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- int i = ll_sa_entry_hash(entry->se_qstr.hash);
+ int i = sa_hash(entry->se_qstr.hash);
spin_lock(&sai->sai_cache_lock[i]);
list_add_tail(&entry->se_hash, &sai->sai_cache[i]);
@@ -113,9 +115,9 @@ ll_sa_entry_enhash(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
* Remove entry from SA table.
*/
static inline void
-ll_sa_entry_unhash(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_unhash(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- int i = ll_sa_entry_hash(entry->se_qstr.hash);
+ int i = sa_hash(entry->se_qstr.hash);
spin_lock(&sai->sai_cache_lock[i]);
list_del_init(&entry->se_hash);
@@ -133,14 +135,14 @@ static inline int sa_sent_full(struct ll_statahead_info *sai)
return atomic_read(&sai->sai_cache_count) >= sai->sai_max;
}
-static inline int sa_received_empty(struct ll_statahead_info *sai)
+static inline int sa_has_callback(struct ll_statahead_info *sai)
{
- return list_empty(&sai->sai_entries_received);
+ return !list_empty(&sai->sai_interim_entries);
}
static inline int agl_list_empty(struct ll_statahead_info *sai)
{
- return list_empty(&sai->sai_entries_agl);
+ return list_empty(&sai->sai_agls);
}
/**
@@ -168,17 +170,16 @@ static inline int is_omitted_entry(struct ll_statahead_info *sai, __u64 index)
/*
* Insert it into sai_entries tail when init.
*/
-static struct ll_sa_entry *
-ll_sa_entry_alloc(struct dentry *parent,
- struct ll_statahead_info *sai, __u64 index,
- const char *name, int len)
+static struct sa_entry *
+sa_alloc(struct dentry *parent, struct ll_statahead_info *sai, __u64 index,
+ const char *name, int len)
{
struct ll_inode_info *lli;
- struct ll_sa_entry *entry;
+ struct sa_entry *entry;
int entry_size;
char *dname;
- entry_size = sizeof(struct ll_sa_entry) + (len & ~3) + 4;
+ entry_size = sizeof(struct sa_entry) + (len & ~3) + 4;
entry = kzalloc(entry_size, GFP_NOFS);
if (unlikely(!entry))
return ERR_PTR(-ENOMEM);
@@ -187,34 +188,9 @@ ll_sa_entry_alloc(struct dentry *parent,
len, name, entry, index);
entry->se_index = index;
-
- /*
- * Statahead entry reference rules:
- *
- * 1) When statahead entry is initialized, its reference is set as 2.
- * One reference is used by the directory scanner. When the scanner
- * searches the statahead cache for the given name, it can perform
- * lockless hash lookup (only the scanner can remove entry from hash
- * list), and once found, it needn't to call "atomic_inc()" for the
- * entry reference. So the performance is improved. After using the
- * statahead entry, the scanner will call "atomic_dec()" to drop the
- * reference held when initialization. If it is the last reference,
- * the statahead entry will be freed.
- *
- * 2) All other threads, including statahead thread and ptlrpcd thread,
- * when they process the statahead entry, the reference for target
- * should be held to guarantee the entry will not be released by the
- * directory scanner. After processing the entry, these threads will
- * drop the entry reference. If it is the last reference, the entry
- * will be freed.
- *
- * The second reference when initializes the statahead entry is used
- * by the statahead thread, following the rule 2).
- */
- atomic_set(&entry->se_refcount, 2);
- entry->se_stat = SA_ENTRY_INIT;
+ entry->se_state = SA_ENTRY_INIT;
entry->se_size = entry_size;
- dname = (char *)entry + sizeof(struct ll_sa_entry);
+ dname = (char *)entry + sizeof(struct sa_entry);
memcpy(dname, name, len);
dname[len] = 0;
@@ -224,9 +200,8 @@ ll_sa_entry_alloc(struct dentry *parent,
lli = ll_i2info(sai->sai_inode);
spin_lock(&lli->lli_sa_lock);
- list_add_tail(&entry->se_link, &sai->sai_entries);
INIT_LIST_HEAD(&entry->se_list);
- ll_sa_entry_enhash(sai, entry);
+ sa_rehash(sai, entry);
spin_unlock(&lli->lli_sa_lock);
atomic_inc(&sai->sai_cache_count);
@@ -234,151 +209,130 @@ ll_sa_entry_alloc(struct dentry *parent,
return entry;
}
-/*
- * Used by the directory scanner to search entry with name.
- *
- * Only the caller can remove the entry from hash, so it is unnecessary to hold
- * hash lock. It is caller's duty to release the init refcount on the entry, so
- * it is also unnecessary to increase refcount on the entry.
- */
-static struct ll_sa_entry *
-ll_sa_entry_get_byname(struct ll_statahead_info *sai, const struct qstr *qstr)
+/* free sa_entry, which should have been unhashed and not in any list */
+static void sa_free(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- struct ll_sa_entry *entry;
- int i = ll_sa_entry_hash(qstr->hash);
+ CDEBUG(D_READA, "free sa entry %.*s(%p) index %llu\n",
+ entry->se_qstr.len, entry->se_qstr.name, entry,
+ entry->se_index);
- list_for_each_entry(entry, &sai->sai_cache[i], se_hash) {
- if (entry->se_qstr.hash == qstr->hash &&
- entry->se_qstr.len == qstr->len &&
- memcmp(entry->se_qstr.name, qstr->name, qstr->len) == 0)
- return entry;
- }
- return NULL;
+ LASSERT(list_empty(&entry->se_list));
+ LASSERT(list_empty(&entry->se_hash));
+
+ kfree(entry);
+ atomic_dec(&sai->sai_cache_count);
}
/*
- * Used by the async getattr request callback to find entry with index.
- *
- * Inside lli_sa_lock to prevent others to change the list during the search.
- * It needs to increase entry refcount before returning to guarantee that the
- * entry cannot be freed by others.
+ * find sa_entry by name, used by directory scanner, lock is not needed because
+ * only scanner can remove the entry from cache.
*/
-static struct ll_sa_entry *
-ll_sa_entry_get_byindex(struct ll_statahead_info *sai, __u64 index)
+static struct sa_entry *
+sa_get(struct ll_statahead_info *sai, const struct qstr *qstr)
{
- struct ll_sa_entry *entry;
+ struct sa_entry *entry;
+ int i = sa_hash(qstr->hash);
- list_for_each_entry(entry, &sai->sai_entries, se_link) {
- if (entry->se_index == index) {
- LASSERT(atomic_read(&entry->se_refcount) > 0);
- atomic_inc(&entry->se_refcount);
+ list_for_each_entry(entry, &sai->sai_cache[i], se_hash) {
+ if (entry->se_qstr.hash == qstr->hash &&
+ entry->se_qstr.len == qstr->len &&
+ memcmp(entry->se_qstr.name, qstr->name, qstr->len) == 0)
return entry;
- }
- if (entry->se_index > index)
- break;
}
return NULL;
}
-static void ll_sa_entry_put(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry)
-{
- if (atomic_dec_and_test(&entry->se_refcount)) {
- CDEBUG(D_READA, "free sa entry %.*s(%p) index %llu\n",
- entry->se_qstr.len, entry->se_qstr.name, entry,
- entry->se_index);
-
- LASSERT(list_empty(&entry->se_link));
- LASSERT(list_empty(&entry->se_list));
- LASSERT(list_empty(&entry->se_hash));
-
- iput(entry->se_inode);
-
- kfree(entry);
- atomic_dec(&sai->sai_cache_count);
- }
-}
-
+/* unhash and unlink sa_entry, and then free it */
static inline void
-do_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_kill(struct ll_statahead_info *sai, struct sa_entry *entry)
{
struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
LASSERT(!list_empty(&entry->se_hash));
- LASSERT(!list_empty(&entry->se_link));
+ LASSERT(!list_empty(&entry->se_list));
+ LASSERT(sa_ready(entry));
- ll_sa_entry_unhash(sai, entry);
+ sa_unhash(sai, entry);
spin_lock(&lli->lli_sa_lock);
- entry->se_stat = SA_ENTRY_DEST;
- list_del_init(&entry->se_link);
- if (likely(!list_empty(&entry->se_list)))
- list_del_init(&entry->se_list);
+ list_del_init(&entry->se_list);
spin_unlock(&lli->lli_sa_lock);
- ll_sa_entry_put(sai, entry);
+ if (entry->se_inode)
+ iput(entry->se_inode);
+
+ sa_free(sai, entry);
}
-/*
- * Delete it from sai_entries_stated list when fini.
- */
+/* called by scanner after use, sa_entry will be killed */
static void
-ll_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_put(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- struct ll_sa_entry *pos, *next;
+ struct sa_entry *tmp, *next;
+
+ if (entry && entry->se_state == SA_ENTRY_SUCC) {
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
+
+ sai->sai_hit++;
+ sai->sai_consecutive_miss = 0;
+ sai->sai_max = min(2 * sai->sai_max, sbi->ll_sa_max);
+ } else {
+ sai->sai_miss++;
+ sai->sai_consecutive_miss++;
+ }
if (entry)
- do_sa_entry_fini(sai, entry);
+ sa_kill(sai, entry);
- /* drop old entry, only 'scanner' process does this, no need to lock */
- list_for_each_entry_safe(pos, next, &sai->sai_entries, se_link) {
- if (!is_omitted_entry(sai, pos->se_index))
+ /*
+ * kill old completed entries, only scanner process does this, no need
+ * to lock
+ */
+ list_for_each_entry_safe(tmp, next, &sai->sai_entries, se_list) {
+ if (!is_omitted_entry(sai, tmp->se_index))
break;
- /* keep those whose statahead RPC not finished */
- if (pos->se_stat == SA_ENTRY_SUCC ||
- pos->se_stat == SA_ENTRY_INVA)
- do_sa_entry_fini(sai, pos);
+ sa_kill(sai, tmp);
}
+ wake_up(&sai->sai_thread.t_ctl_waitq);
}
/*
- * Inside lli_sa_lock.
+ * update state and sort add entry to sai_entries by index, return true if
+ * scanner is waiting on this entry.
*/
-static void
-__sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
- enum se_stat stat)
+static bool
+__sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
{
- struct ll_sa_entry *se;
- struct list_head *pos = &sai->sai_entries_stated;
-
- LASSERT(entry->se_stat == SA_ENTRY_INIT);
+ struct list_head *pos = &sai->sai_entries;
+ __u64 index = entry->se_index;
+ struct sa_entry *se;
- if (!list_empty(&entry->se_list))
- list_del_init(&entry->se_list);
+ LASSERT(!sa_ready(entry));
+ LASSERT(list_empty(&entry->se_list));
- list_for_each_entry_reverse(se, &sai->sai_entries_stated, se_list) {
+ list_for_each_entry_reverse(se, &sai->sai_entries, se_list) {
if (se->se_index < entry->se_index) {
pos = &se->se_list;
break;
}
}
-
list_add(&entry->se_list, pos);
- entry->se_stat = stat;
+ entry->se_state = ret < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC;
+
+ return (index == sai->sai_index_wait);
}
/*
- * Move entry to sai_entries_stated and sort with the index.
- * \retval 1 -- entry to be destroyed.
- * \retval 0 -- entry is inserted into stated list.
+ * release resources used in async stat RPC, update entry state and wakeup if
+ * scanner process it waiting on this entry.
*/
static void
-sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
- enum se_stat stat)
+sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
{
struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
struct md_enqueue_info *minfo = entry->se_minfo;
struct ptlrpc_request *req = entry->se_req;
+ bool wakeup;
/* release resources used in RPC */
if (minfo) {
@@ -394,12 +348,15 @@ sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
}
spin_lock(&lli->lli_sa_lock);
- __sa_entry_post_stat(sai, entry, stat);
+ wakeup = __sa_make_ready(sai, entry, ret);
spin_unlock(&lli->lli_sa_lock);
+
+ if (wakeup)
+ wake_up(&sai->sai_waitq);
}
/*
- * Insert inode into the list of sai_entries_agl.
+ * Insert inode into the list of sai_agls.
*/
static void ll_agl_add(struct ll_statahead_info *sai,
struct inode *inode, int index)
@@ -417,9 +374,9 @@ static void ll_agl_add(struct ll_statahead_info *sai,
igrab(inode);
spin_lock(&parent->lli_agl_lock);
- if (list_empty(&sai->sai_entries_agl))
+ if (list_empty(&sai->sai_agls))
added = 1;
- list_add_tail(&child->lli_agl_list, &sai->sai_entries_agl);
+ list_add_tail(&child->lli_agl_list, &sai->sai_agls);
spin_unlock(&parent->lli_agl_lock);
} else {
spin_unlock(&child->lli_agl_lock);
@@ -429,6 +386,7 @@ static void ll_agl_add(struct ll_statahead_info *sai,
wake_up(&sai->sai_agl_thread.t_ctl_waitq);
}
+/* allocate sai */
static struct ll_statahead_info *ll_sai_alloc(void)
{
struct ll_statahead_info *sai;
@@ -452,10 +410,9 @@ static struct ll_statahead_info *ll_sai_alloc(void)
init_waitqueue_head(&sai->sai_thread.t_ctl_waitq);
init_waitqueue_head(&sai->sai_agl_thread.t_ctl_waitq);
+ INIT_LIST_HEAD(&sai->sai_interim_entries);
INIT_LIST_HEAD(&sai->sai_entries);
- INIT_LIST_HEAD(&sai->sai_entries_received);
- INIT_LIST_HEAD(&sai->sai_entries_stated);
- INIT_LIST_HEAD(&sai->sai_entries_agl);
+ INIT_LIST_HEAD(&sai->sai_agls);
for (i = 0; i < LL_SA_CACHE_SIZE; i++) {
INIT_LIST_HEAD(&sai->sai_cache[i]);
@@ -486,7 +443,7 @@ static void ll_sai_put(struct ll_statahead_info *sai)
if (atomic_dec_and_lock(&sai->sai_refcount, &lli->lli_sa_lock)) {
struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
- struct ll_sa_entry *entry, *next;
+ struct sa_entry *entry, *next;
lli->lli_sai = NULL;
spin_unlock(&lli->lli_sa_lock);
@@ -494,14 +451,14 @@ static void ll_sai_put(struct ll_statahead_info *sai)
LASSERT(thread_is_stopped(&sai->sai_thread));
LASSERT(thread_is_stopped(&sai->sai_agl_thread));
LASSERT(sai->sai_sent == sai->sai_replied);
+ LASSERT(!sa_has_callback(sai));
list_for_each_entry_safe(entry, next, &sai->sai_entries,
- se_link)
- do_sa_entry_fini(sai, entry);
+ se_list)
+ sa_kill(sai, entry);
LASSERT(atomic_read(&sai->sai_cache_count) == 0);
- LASSERT(list_empty(&sai->sai_entries_agl));
- LASSERT(atomic_read(&sai->sai_refcount) == 0);
+ LASSERT(list_empty(&sai->sai_agls));
iput(sai->sai_inode);
kfree(sai);
@@ -509,7 +466,7 @@ static void ll_sai_put(struct ll_statahead_info *sai)
}
}
-/* Do NOT forget to drop inode refcount when into sai_entries_agl. */
+/* Do NOT forget to drop inode refcount when into sai_agls. */
static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
{
struct ll_inode_info *lli = ll_i2info(inode);
@@ -569,9 +526,12 @@ static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
iput(inode);
}
-/* prepare inode for received statahead entry, and add it into agl list */
-static void sa_post_one(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry)
+/*
+ * prepare inode for sa entry, add it into agl list, now sa_entry is ready
+ * to be used by scanner process.
+ */
+static void sa_instantiate(struct ll_statahead_info *sai,
+ struct sa_entry *entry)
{
struct inode *dir = sai->sai_inode;
struct inode *child;
@@ -629,8 +589,9 @@ static void sa_post_one(struct ll_statahead_info *sai,
if (rc)
goto out;
- CDEBUG(D_DLMTRACE, "%s: setting l_data to inode "DFID"%p\n",
+ CDEBUG(D_READA, "%s: setting %.*s" DFID " l_data to inode %p\n",
ll_get_fsname(child->i_sb, NULL, 0),
+ entry->se_qstr.len, entry->se_qstr.name,
PFID(ll_inode2fid(child)), child);
ll_set_lock_data(ll_i2sbi(dir)->ll_md_exp, child, it, NULL);
@@ -640,45 +601,42 @@ static void sa_post_one(struct ll_statahead_info *sai,
ll_agl_add(sai, child, entry->se_index);
out:
- /* The "sa_entry_post_stat()" will drop related ldlm ibits lock
- * reference count by calling "ll_intent_drop_lock()" in spite of the
- * above operations failed or not. Do not worry about calling
- * "ll_intent_drop_lock()" more than once.
+ /*
+ * sa_make_ready() will drop ldlm ibits lock refcount by calling
+ * ll_intent_drop_lock() in spite of failures. Do not worry about
+ * calling ll_intent_drop_lock() more than once.
*/
- sa_entry_post_stat(sai, entry, rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
- if (entry->se_index == sai->sai_index_wait)
- wake_up(&sai->sai_waitq);
- ll_sa_entry_put(sai, entry);
+ sa_make_ready(sai, entry, rc);
}
-static void ll_post_statahead(struct ll_statahead_info *sai)
+/* once there are async stat replies, instantiate sa_entry */
+static void sa_handle_callback(struct ll_statahead_info *sai)
{
struct ll_inode_info *lli;
lli = ll_i2info(sai->sai_inode);
- while (!sa_received_empty(sai)) {
- struct ll_sa_entry *entry;
+ while (sa_has_callback(sai)) {
+ struct sa_entry *entry;
spin_lock(&lli->lli_sa_lock);
- if (unlikely(sa_received_empty(sai))) {
+ if (unlikely(!sa_has_callback(sai))) {
spin_unlock(&lli->lli_sa_lock);
break;
}
- entry = list_entry(sai->sai_entries_received.next,
- struct ll_sa_entry, se_list);
- atomic_inc(&entry->se_refcount);
+ entry = list_entry(sai->sai_interim_entries.next,
+ struct sa_entry, se_list);
list_del_init(&entry->se_list);
spin_unlock(&lli->lli_sa_lock);
- sa_post_one(sai, entry);
+ sa_instantiate(sai, entry);
}
spin_lock(&lli->lli_agl_lock);
while (!agl_list_empty(sai)) {
struct ll_inode_info *clli;
- clli = list_entry(sai->sai_entries_agl.next,
+ clli = list_entry(sai->sai_agls.next,
struct ll_inode_info, lli_agl_list);
list_del_init(&clli->lli_agl_list);
spin_unlock(&lli->lli_agl_lock);
@@ -690,29 +648,38 @@ static void ll_post_statahead(struct ll_statahead_info *sai)
spin_unlock(&lli->lli_agl_lock);
}
+/*
+ * callback for async stat, because this is called in ptlrpcd context, we only
+ * put sa_entry in sai_cb_entries list, and let sa_handle_callback() to really
+ * prepare inode and instantiate sa_entry later.
+ */
static int ll_statahead_interpret(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc)
{
struct lookup_intent *it = &minfo->mi_it;
struct inode *dir = minfo->mi_dir;
struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_statahead_info *sai = NULL;
- struct ll_sa_entry *entry;
- int wakeup;
+ struct ll_statahead_info *sai = lli->lli_sai;
+ struct sa_entry *entry = (struct sa_entry *)minfo->mi_cbdata;
+ bool wakeup;
if (it_disposition(it, DISP_LOOKUP_NEG))
rc = -ENOENT;
- sai = ll_sai_get(dir);
+ /*
+ * because statahead thread will wait for all inflight RPC to finish,
+ * sai should be always valid, no need to refcount
+ */
LASSERT(sai);
LASSERT(!thread_is_stopped(&sai->sai_thread));
+ LASSERT(entry);
+
+ CDEBUG(D_READA, "sa_entry %.*s rc %d\n",
+ entry->se_qstr.len, entry->se_qstr.name, rc);
spin_lock(&lli->lli_sa_lock);
- entry = ll_sa_entry_get_byindex(sai, minfo->mi_cbdata);
- LASSERT(entry);
if (rc) {
- __sa_entry_post_stat(sai, entry, SA_ENTRY_INVA);
- wakeup = (entry->se_index == sai->sai_index_wait);
+ wakeup = __sa_make_ready(sai, entry, rc);
} else {
entry->se_minfo = minfo;
entry->se_req = ptlrpc_request_addref(req);
@@ -724,27 +691,24 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
*/
entry->se_handle = it->it_lock_handle;
ll_intent_drop_lock(it);
- wakeup = sa_received_empty(sai);
- list_add_tail(&entry->se_list, &sai->sai_entries_received);
+ wakeup = !sa_has_callback(sai);
+ list_add_tail(&entry->se_list, &sai->sai_interim_entries);
}
sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
- ll_sa_entry_put(sai, entry);
if (wakeup)
wake_up(&sai->sai_thread.t_ctl_waitq);
+ spin_unlock(&lli->lli_sa_lock);
if (rc) {
ll_intent_release(it);
iput(dir);
kfree(minfo);
}
- if (sai)
- ll_sai_put(sai);
return rc;
}
-static void sa_args_fini(struct md_enqueue_info *minfo,
+static void sa_fini_data(struct md_enqueue_info *minfo,
struct ldlm_enqueue_info *einfo)
{
LASSERT(minfo && einfo);
@@ -756,8 +720,8 @@ static void sa_args_fini(struct md_enqueue_info *minfo,
/**
* prepare arguments for async stat RPC.
*/
-static int sa_args_init(struct inode *dir, struct inode *child,
- struct ll_sa_entry *entry, struct md_enqueue_info **pmi,
+static int sa_prep_data(struct inode *dir, struct inode *child,
+ struct sa_entry *entry, struct md_enqueue_info **pmi,
struct ldlm_enqueue_info **pei)
{
const struct qstr *qstr = &entry->se_qstr;
@@ -786,7 +750,7 @@ static int sa_args_init(struct inode *dir, struct inode *child,
minfo->mi_it.it_op = IT_GETATTR;
minfo->mi_dir = igrab(dir);
minfo->mi_cb = ll_statahead_interpret;
- minfo->mi_cbdata = entry->se_index;
+ minfo->mi_cbdata = entry;
einfo->ei_type = LDLM_IBITS;
einfo->ei_mode = it_to_lock_mode(&minfo->mi_it);
@@ -801,19 +765,19 @@ static int sa_args_init(struct inode *dir, struct inode *child,
return 0;
}
-static int do_sa_lookup(struct inode *dir, struct ll_sa_entry *entry)
+static int sa_lookup(struct inode *dir, struct sa_entry *entry)
{
struct md_enqueue_info *minfo;
struct ldlm_enqueue_info *einfo;
int rc;
- rc = sa_args_init(dir, NULL, entry, &minfo, &einfo);
+ rc = sa_prep_data(dir, NULL, entry, &minfo, &einfo);
if (rc)
return rc;
rc = md_intent_getattr_async(ll_i2mdexp(dir), minfo, einfo);
if (rc < 0)
- sa_args_fini(minfo, einfo);
+ sa_fini_data(minfo, einfo);
return rc;
}
@@ -824,8 +788,8 @@ static int do_sa_lookup(struct inode *dir, struct ll_sa_entry *entry)
* \retval 0 -- will send stat-ahead request
* \retval others -- prepare stat-ahead request failed
*/
-static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
- struct dentry *dentry)
+static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
+ struct dentry *dentry)
{
struct inode *inode = d_inode(dentry);
struct lookup_intent it = { .it_op = IT_GETATTR,
@@ -849,7 +813,7 @@ static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
return 1;
}
- rc = sa_args_init(dir, inode, entry, &minfo, &einfo);
+ rc = sa_prep_data(dir, inode, entry, &minfo, &einfo);
if (rc) {
entry->se_inode = NULL;
iput(inode);
@@ -860,32 +824,30 @@ static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
if (rc < 0) {
entry->se_inode = NULL;
iput(inode);
- sa_args_fini(minfo, einfo);
+ sa_fini_data(minfo, einfo);
}
return rc;
}
-static void ll_statahead_one(struct dentry *parent, const char *name,
- const int name_len)
+static void sa_statahead(struct dentry *parent, const char *name, int len)
{
struct inode *dir = d_inode(parent);
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_statahead_info *sai = lli->lli_sai;
struct dentry *dentry = NULL;
- struct ll_sa_entry *entry;
+ struct sa_entry *entry;
int rc;
- entry = ll_sa_entry_alloc(parent, sai, sai->sai_index, name,
- name_len);
+ entry = sa_alloc(parent, sai, sai->sai_index, name, len);
if (IS_ERR(entry))
return;
dentry = d_lookup(parent, &entry->se_qstr);
if (!dentry) {
- rc = do_sa_lookup(dir, entry);
+ rc = sa_lookup(dir, entry);
} else {
- rc = do_sa_revalidate(dir, entry, dentry);
+ rc = sa_revalidate(dir, entry, dentry);
if (rc == 1 && agl_should_run(sai, d_inode(dentry)))
ll_agl_add(sai, d_inode(dentry), entry->se_index);
}
@@ -893,18 +855,12 @@ static void ll_statahead_one(struct dentry *parent, const char *name,
if (dentry)
dput(dentry);
- if (rc) {
- sa_entry_post_stat(sai, entry,
- rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
- if (entry->se_index == sai->sai_index_wait)
- wake_up(&sai->sai_waitq);
- } else {
+ if (rc)
+ sa_make_ready(sai, entry, rc);
+ else
sai->sai_sent++;
- }
sai->sai_index++;
- /* drop one refcount on entry by ll_sa_entry_alloc */
- ll_sa_entry_put(sai, entry);
}
static int ll_agl_thread(void *arg)
@@ -938,7 +894,7 @@ static int ll_agl_thread(void *arg)
while (1) {
l_wait_event(thread->t_ctl_waitq,
- !list_empty(&sai->sai_entries_agl) ||
+ !list_empty(&sai->sai_agls) ||
!thread_is_running(thread),
&lwi);
@@ -949,8 +905,8 @@ static int ll_agl_thread(void *arg)
/* The statahead thread maybe help to process AGL entries,
* so check whether list empty again.
*/
- if (!list_empty(&sai->sai_entries_agl)) {
- clli = list_entry(sai->sai_entries_agl.next,
+ if (!list_empty(&sai->sai_agls)) {
+ clli = list_entry(sai->sai_agls.next,
struct ll_inode_info, lli_agl_list);
list_del_init(&clli->lli_agl_list);
spin_unlock(&plli->lli_agl_lock);
@@ -962,8 +918,8 @@ static int ll_agl_thread(void *arg)
spin_lock(&plli->lli_agl_lock);
sai->sai_agl_valid = 0;
- while (!list_empty(&sai->sai_entries_agl)) {
- clli = list_entry(sai->sai_entries_agl.next,
+ while (!list_empty(&sai->sai_agls)) {
+ clli = list_entry(sai->sai_agls.next,
struct ll_inode_info, lli_agl_list);
list_del_init(&clli->lli_agl_list);
spin_unlock(&plli->lli_agl_lock);
@@ -1118,15 +1074,15 @@ static int ll_statahead_thread(void *arg)
do {
l_wait_event(thread->t_ctl_waitq,
!sa_sent_full(sai) ||
- !list_empty(&sai->sai_entries_received) ||
- !list_empty(&sai->sai_entries_agl) ||
+ sa_has_callback(sai) ||
+ !list_empty(&sai->sai_agls) ||
!thread_is_running(thread),
&lwi);
- ll_post_statahead(sai);
+ sa_handle_callback(sai);
} while (sa_sent_full(sai) &&
thread_is_running(thread));
- ll_statahead_one(parent, name, namelen);
+ sa_statahead(parent, name, namelen);
}
pos = le64_to_cpu(dp->ldp_hash_end);
@@ -1158,12 +1114,12 @@ static int ll_statahead_thread(void *arg)
*/
while (thread_is_running(thread)) {
l_wait_event(thread->t_ctl_waitq,
- !sa_received_empty(sai) ||
+ sa_has_callback(sai) ||
!agl_list_empty(sai) ||
!thread_is_running(thread),
&lwi);
- ll_post_statahead(sai);
+ sa_handle_callback(sai);
}
out:
if (sai->sai_agl_valid) {
@@ -1193,8 +1149,8 @@ out:
sai->sai_sent == sai->sai_replied, &lwi);
}
- /* release resources held by received entries. */
- ll_post_statahead(sai);
+ /* release resources held by statahead RPCs */
+ sa_handle_callback(sai);
spin_lock(&lli->lli_sa_lock);
thread_set_flags(thread, SVC_STOPPED);
@@ -1251,7 +1207,7 @@ void ll_deauthorize_statahead(struct inode *dir, void *key)
if (sai && thread_is_running(&sai->sai_thread)) {
/*
* statahead thread may not quit yet because it needs to cache
- * stated entries, now it's time to tell it to quit.
+ * entries, now it's time to tell it to quit.
*/
thread_set_flags(&sai->sai_thread, SVC_STOPPING);
wake_up(&sai->sai_thread.t_ctl_waitq);
@@ -1389,29 +1345,12 @@ out:
return rc;
}
-static void
-ll_sai_unplug(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
-{
- if (entry && entry->se_stat == SA_ENTRY_SUCC) {
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
-
- sai->sai_hit++;
- sai->sai_consecutive_miss = 0;
- sai->sai_max = min(2 * sai->sai_max, sbi->ll_sa_max);
- } else {
- sai->sai_miss++;
- sai->sai_consecutive_miss++;
- }
- ll_sa_entry_fini(sai, entry);
- wake_up(&sai->sai_thread.t_ctl_waitq);
-}
-
static int revalidate_statahead_dentry(struct inode *dir,
struct ll_statahead_info *sai,
struct dentry **dentryp,
int only_unplug)
{
- struct ll_sa_entry *entry = NULL;
+ struct sa_entry *entry = NULL;
struct l_wait_info lwi = { 0 };
int rc = 0;
@@ -1442,31 +1381,31 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
- entry = ll_sa_entry_get_byname(sai, &(*dentryp)->d_name);
+ entry = sa_get(sai, &(*dentryp)->d_name);
if (!entry || only_unplug) {
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return entry ? 1 : -EAGAIN;
}
/* if statahead is busy in readdir, help it do post-work */
- if (!ll_sa_entry_stated(entry) && sai->sai_in_readpage)
- ll_post_statahead(sai);
+ if (!sa_ready(entry) && sai->sai_in_readpage)
+ sa_handle_callback(sai);
- if (!ll_sa_entry_stated(entry)) {
+ if (!sa_ready(entry)) {
sai->sai_index_wait = entry->se_index;
lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
LWI_ON_SIGNAL_NOOP, NULL);
rc = l_wait_event(sai->sai_waitq,
- ll_sa_entry_stated(entry) ||
+ sa_ready(entry) ||
thread_is_stopped(&sai->sai_thread),
&lwi);
if (rc < 0) {
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return -EAGAIN;
}
}
- if (entry->se_stat == SA_ENTRY_SUCC && entry->se_inode) {
+ if (entry->se_state == SA_ENTRY_SUCC && entry->se_inode) {
struct inode *inode = entry->se_inode;
struct lookup_intent it = { .it_op = IT_GETATTR,
.it_lock_handle = entry->se_handle };
@@ -1480,7 +1419,7 @@ static int revalidate_statahead_dentry(struct inode *dir,
alias = ll_splice_alias(inode, *dentryp);
if (IS_ERR(alias)) {
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return PTR_ERR(alias);
}
*dentryp = alias;
@@ -1507,7 +1446,7 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
out_unplug:
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 056/124] staging: lustre: statahead: use dcache-like interface for sa entry
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
Rename ll_sa_entry to sa_entry, and manage sa_entry cache with
dcache-like interfaces.
sa_entry is not needed to be refcounted, because only scanner
can free it, so after it's put in stat list, statahead thread
shouldn't access it any longer.
ll_statahead_interpret() doesn't need to take sai refcount,
because statahead thread will wait for all inflight RPC to
finish.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9664
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
.../staging/lustre/lustre/llite/llite_internal.h | 10 +-
drivers/staging/lustre/lustre/llite/statahead.c | 461 +++++++++-----------
3 files changed, 207 insertions(+), 266 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 1c884c0..6758bf4 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -812,7 +812,7 @@ struct md_enqueue_info {
struct inode *mi_dir;
int (*mi_cb)(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc);
- __u64 mi_cbdata;
+ void *mi_cbdata;
};
struct obd_ops {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 9a9fefd..cd639f4 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -1131,10 +1131,12 @@ struct ll_statahead_info {
wait_queue_head_t sai_waitq; /* stat-ahead wait queue */
struct ptlrpc_thread sai_thread; /* stat-ahead thread */
struct ptlrpc_thread sai_agl_thread; /* AGL thread */
- struct list_head sai_entries; /* entry list */
- struct list_head sai_entries_received; /* entries returned */
- struct list_head sai_entries_stated; /* entries stated */
- struct list_head sai_entries_agl; /* AGL entries to be sent */
+ struct list_head sai_interim_entries; /* entries which got async
+ * stat reply, but not
+ * instantiated
+ */
+ struct list_head sai_entries; /* completed entries */
+ struct list_head sai_agls; /* AGLs to be sent */
struct list_head sai_cache[LL_SA_CACHE_SIZE];
spinlock_t sai_cache_lock[LL_SA_CACHE_SIZE];
atomic_t sai_cache_count; /* entry count in cache */
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 5760974..d041720 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -49,24 +49,26 @@ enum se_stat {
SA_ENTRY_INIT = 0, /** init entry */
SA_ENTRY_SUCC = 1, /** stat succeed */
SA_ENTRY_INVA = 2, /** invalid entry */
- SA_ENTRY_DEST = 3, /** entry to be destroyed */
};
-struct ll_sa_entry {
- /* link into sai->sai_entries */
- struct list_head se_link;
- /* link into sai->sai_entries_{received,stated} */
+/*
+ * sa_entry is not refcounted: statahead thread allocates it and do async stat,
+ * and in async stat callback ll_statahead_interpret() will add it into
+ * sai_cb_entries, later statahead thread will call sa_handle_callback() to
+ * instantiate entry and move it into sai_entries, and then only scanner process
+ * can access and free it.
+ */
+struct sa_entry {
+ /* link into sai_cb_entries or sai_entries */
struct list_head se_list;
/* link into sai hash table locally */
struct list_head se_hash;
- /* entry reference count */
- atomic_t se_refcount;
/* entry index in the sai */
__u64 se_index;
/* low layer ldlm lock handle */
__u64 se_handle;
/* entry status */
- enum se_stat se_stat;
+ enum se_stat se_state;
/* entry size, contains name */
int se_size;
/* pointer to async getattr enqueue info */
@@ -85,13 +87,13 @@ static DEFINE_SPINLOCK(sai_generation_lock);
/*
* The entry only can be released by the caller, it is necessary to hold lock.
*/
-static inline int ll_sa_entry_stated(struct ll_sa_entry *entry)
+static inline int sa_ready(struct sa_entry *entry)
{
smp_rmb();
- return (entry->se_stat != SA_ENTRY_INIT);
+ return (entry->se_state != SA_ENTRY_INIT);
}
-static inline int ll_sa_entry_hash(int val)
+static inline int sa_hash(int val)
{
return val & LL_SA_CACHE_MASK;
}
@@ -100,9 +102,9 @@ static inline int ll_sa_entry_hash(int val)
* Insert entry to hash SA table.
*/
static inline void
-ll_sa_entry_enhash(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_rehash(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- int i = ll_sa_entry_hash(entry->se_qstr.hash);
+ int i = sa_hash(entry->se_qstr.hash);
spin_lock(&sai->sai_cache_lock[i]);
list_add_tail(&entry->se_hash, &sai->sai_cache[i]);
@@ -113,9 +115,9 @@ ll_sa_entry_enhash(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
* Remove entry from SA table.
*/
static inline void
-ll_sa_entry_unhash(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_unhash(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- int i = ll_sa_entry_hash(entry->se_qstr.hash);
+ int i = sa_hash(entry->se_qstr.hash);
spin_lock(&sai->sai_cache_lock[i]);
list_del_init(&entry->se_hash);
@@ -133,14 +135,14 @@ static inline int sa_sent_full(struct ll_statahead_info *sai)
return atomic_read(&sai->sai_cache_count) >= sai->sai_max;
}
-static inline int sa_received_empty(struct ll_statahead_info *sai)
+static inline int sa_has_callback(struct ll_statahead_info *sai)
{
- return list_empty(&sai->sai_entries_received);
+ return !list_empty(&sai->sai_interim_entries);
}
static inline int agl_list_empty(struct ll_statahead_info *sai)
{
- return list_empty(&sai->sai_entries_agl);
+ return list_empty(&sai->sai_agls);
}
/**
@@ -168,17 +170,16 @@ static inline int is_omitted_entry(struct ll_statahead_info *sai, __u64 index)
/*
* Insert it into sai_entries tail when init.
*/
-static struct ll_sa_entry *
-ll_sa_entry_alloc(struct dentry *parent,
- struct ll_statahead_info *sai, __u64 index,
- const char *name, int len)
+static struct sa_entry *
+sa_alloc(struct dentry *parent, struct ll_statahead_info *sai, __u64 index,
+ const char *name, int len)
{
struct ll_inode_info *lli;
- struct ll_sa_entry *entry;
+ struct sa_entry *entry;
int entry_size;
char *dname;
- entry_size = sizeof(struct ll_sa_entry) + (len & ~3) + 4;
+ entry_size = sizeof(struct sa_entry) + (len & ~3) + 4;
entry = kzalloc(entry_size, GFP_NOFS);
if (unlikely(!entry))
return ERR_PTR(-ENOMEM);
@@ -187,34 +188,9 @@ ll_sa_entry_alloc(struct dentry *parent,
len, name, entry, index);
entry->se_index = index;
-
- /*
- * Statahead entry reference rules:
- *
- * 1) When statahead entry is initialized, its reference is set as 2.
- * One reference is used by the directory scanner. When the scanner
- * searches the statahead cache for the given name, it can perform
- * lockless hash lookup (only the scanner can remove entry from hash
- * list), and once found, it needn't to call "atomic_inc()" for the
- * entry reference. So the performance is improved. After using the
- * statahead entry, the scanner will call "atomic_dec()" to drop the
- * reference held when initialization. If it is the last reference,
- * the statahead entry will be freed.
- *
- * 2) All other threads, including statahead thread and ptlrpcd thread,
- * when they process the statahead entry, the reference for target
- * should be held to guarantee the entry will not be released by the
- * directory scanner. After processing the entry, these threads will
- * drop the entry reference. If it is the last reference, the entry
- * will be freed.
- *
- * The second reference when initializes the statahead entry is used
- * by the statahead thread, following the rule 2).
- */
- atomic_set(&entry->se_refcount, 2);
- entry->se_stat = SA_ENTRY_INIT;
+ entry->se_state = SA_ENTRY_INIT;
entry->se_size = entry_size;
- dname = (char *)entry + sizeof(struct ll_sa_entry);
+ dname = (char *)entry + sizeof(struct sa_entry);
memcpy(dname, name, len);
dname[len] = 0;
@@ -224,9 +200,8 @@ ll_sa_entry_alloc(struct dentry *parent,
lli = ll_i2info(sai->sai_inode);
spin_lock(&lli->lli_sa_lock);
- list_add_tail(&entry->se_link, &sai->sai_entries);
INIT_LIST_HEAD(&entry->se_list);
- ll_sa_entry_enhash(sai, entry);
+ sa_rehash(sai, entry);
spin_unlock(&lli->lli_sa_lock);
atomic_inc(&sai->sai_cache_count);
@@ -234,151 +209,130 @@ ll_sa_entry_alloc(struct dentry *parent,
return entry;
}
-/*
- * Used by the directory scanner to search entry with name.
- *
- * Only the caller can remove the entry from hash, so it is unnecessary to hold
- * hash lock. It is caller's duty to release the init refcount on the entry, so
- * it is also unnecessary to increase refcount on the entry.
- */
-static struct ll_sa_entry *
-ll_sa_entry_get_byname(struct ll_statahead_info *sai, const struct qstr *qstr)
+/* free sa_entry, which should have been unhashed and not in any list */
+static void sa_free(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- struct ll_sa_entry *entry;
- int i = ll_sa_entry_hash(qstr->hash);
+ CDEBUG(D_READA, "free sa entry %.*s(%p) index %llu\n",
+ entry->se_qstr.len, entry->se_qstr.name, entry,
+ entry->se_index);
- list_for_each_entry(entry, &sai->sai_cache[i], se_hash) {
- if (entry->se_qstr.hash == qstr->hash &&
- entry->se_qstr.len == qstr->len &&
- memcmp(entry->se_qstr.name, qstr->name, qstr->len) == 0)
- return entry;
- }
- return NULL;
+ LASSERT(list_empty(&entry->se_list));
+ LASSERT(list_empty(&entry->se_hash));
+
+ kfree(entry);
+ atomic_dec(&sai->sai_cache_count);
}
/*
- * Used by the async getattr request callback to find entry with index.
- *
- * Inside lli_sa_lock to prevent others to change the list during the search.
- * It needs to increase entry refcount before returning to guarantee that the
- * entry cannot be freed by others.
+ * find sa_entry by name, used by directory scanner, lock is not needed because
+ * only scanner can remove the entry from cache.
*/
-static struct ll_sa_entry *
-ll_sa_entry_get_byindex(struct ll_statahead_info *sai, __u64 index)
+static struct sa_entry *
+sa_get(struct ll_statahead_info *sai, const struct qstr *qstr)
{
- struct ll_sa_entry *entry;
+ struct sa_entry *entry;
+ int i = sa_hash(qstr->hash);
- list_for_each_entry(entry, &sai->sai_entries, se_link) {
- if (entry->se_index == index) {
- LASSERT(atomic_read(&entry->se_refcount) > 0);
- atomic_inc(&entry->se_refcount);
+ list_for_each_entry(entry, &sai->sai_cache[i], se_hash) {
+ if (entry->se_qstr.hash == qstr->hash &&
+ entry->se_qstr.len == qstr->len &&
+ memcmp(entry->se_qstr.name, qstr->name, qstr->len) == 0)
return entry;
- }
- if (entry->se_index > index)
- break;
}
return NULL;
}
-static void ll_sa_entry_put(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry)
-{
- if (atomic_dec_and_test(&entry->se_refcount)) {
- CDEBUG(D_READA, "free sa entry %.*s(%p) index %llu\n",
- entry->se_qstr.len, entry->se_qstr.name, entry,
- entry->se_index);
-
- LASSERT(list_empty(&entry->se_link));
- LASSERT(list_empty(&entry->se_list));
- LASSERT(list_empty(&entry->se_hash));
-
- iput(entry->se_inode);
-
- kfree(entry);
- atomic_dec(&sai->sai_cache_count);
- }
-}
-
+/* unhash and unlink sa_entry, and then free it */
static inline void
-do_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_kill(struct ll_statahead_info *sai, struct sa_entry *entry)
{
struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
LASSERT(!list_empty(&entry->se_hash));
- LASSERT(!list_empty(&entry->se_link));
+ LASSERT(!list_empty(&entry->se_list));
+ LASSERT(sa_ready(entry));
- ll_sa_entry_unhash(sai, entry);
+ sa_unhash(sai, entry);
spin_lock(&lli->lli_sa_lock);
- entry->se_stat = SA_ENTRY_DEST;
- list_del_init(&entry->se_link);
- if (likely(!list_empty(&entry->se_list)))
- list_del_init(&entry->se_list);
+ list_del_init(&entry->se_list);
spin_unlock(&lli->lli_sa_lock);
- ll_sa_entry_put(sai, entry);
+ if (entry->se_inode)
+ iput(entry->se_inode);
+
+ sa_free(sai, entry);
}
-/*
- * Delete it from sai_entries_stated list when fini.
- */
+/* called by scanner after use, sa_entry will be killed */
static void
-ll_sa_entry_fini(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
+sa_put(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- struct ll_sa_entry *pos, *next;
+ struct sa_entry *tmp, *next;
+
+ if (entry && entry->se_state == SA_ENTRY_SUCC) {
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
+
+ sai->sai_hit++;
+ sai->sai_consecutive_miss = 0;
+ sai->sai_max = min(2 * sai->sai_max, sbi->ll_sa_max);
+ } else {
+ sai->sai_miss++;
+ sai->sai_consecutive_miss++;
+ }
if (entry)
- do_sa_entry_fini(sai, entry);
+ sa_kill(sai, entry);
- /* drop old entry, only 'scanner' process does this, no need to lock */
- list_for_each_entry_safe(pos, next, &sai->sai_entries, se_link) {
- if (!is_omitted_entry(sai, pos->se_index))
+ /*
+ * kill old completed entries, only scanner process does this, no need
+ * to lock
+ */
+ list_for_each_entry_safe(tmp, next, &sai->sai_entries, se_list) {
+ if (!is_omitted_entry(sai, tmp->se_index))
break;
- /* keep those whose statahead RPC not finished */
- if (pos->se_stat == SA_ENTRY_SUCC ||
- pos->se_stat == SA_ENTRY_INVA)
- do_sa_entry_fini(sai, pos);
+ sa_kill(sai, tmp);
}
+ wake_up(&sai->sai_thread.t_ctl_waitq);
}
/*
- * Inside lli_sa_lock.
+ * update state and sort add entry to sai_entries by index, return true if
+ * scanner is waiting on this entry.
*/
-static void
-__sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
- enum se_stat stat)
+static bool
+__sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
{
- struct ll_sa_entry *se;
- struct list_head *pos = &sai->sai_entries_stated;
-
- LASSERT(entry->se_stat == SA_ENTRY_INIT);
+ struct list_head *pos = &sai->sai_entries;
+ __u64 index = entry->se_index;
+ struct sa_entry *se;
- if (!list_empty(&entry->se_list))
- list_del_init(&entry->se_list);
+ LASSERT(!sa_ready(entry));
+ LASSERT(list_empty(&entry->se_list));
- list_for_each_entry_reverse(se, &sai->sai_entries_stated, se_list) {
+ list_for_each_entry_reverse(se, &sai->sai_entries, se_list) {
if (se->se_index < entry->se_index) {
pos = &se->se_list;
break;
}
}
-
list_add(&entry->se_list, pos);
- entry->se_stat = stat;
+ entry->se_state = ret < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC;
+
+ return (index == sai->sai_index_wait);
}
/*
- * Move entry to sai_entries_stated and sort with the index.
- * \retval 1 -- entry to be destroyed.
- * \retval 0 -- entry is inserted into stated list.
+ * release resources used in async stat RPC, update entry state and wakeup if
+ * scanner process it waiting on this entry.
*/
static void
-sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
- enum se_stat stat)
+sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
{
struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
struct md_enqueue_info *minfo = entry->se_minfo;
struct ptlrpc_request *req = entry->se_req;
+ bool wakeup;
/* release resources used in RPC */
if (minfo) {
@@ -394,12 +348,15 @@ sa_entry_post_stat(struct ll_statahead_info *sai, struct ll_sa_entry *entry,
}
spin_lock(&lli->lli_sa_lock);
- __sa_entry_post_stat(sai, entry, stat);
+ wakeup = __sa_make_ready(sai, entry, ret);
spin_unlock(&lli->lli_sa_lock);
+
+ if (wakeup)
+ wake_up(&sai->sai_waitq);
}
/*
- * Insert inode into the list of sai_entries_agl.
+ * Insert inode into the list of sai_agls.
*/
static void ll_agl_add(struct ll_statahead_info *sai,
struct inode *inode, int index)
@@ -417,9 +374,9 @@ static void ll_agl_add(struct ll_statahead_info *sai,
igrab(inode);
spin_lock(&parent->lli_agl_lock);
- if (list_empty(&sai->sai_entries_agl))
+ if (list_empty(&sai->sai_agls))
added = 1;
- list_add_tail(&child->lli_agl_list, &sai->sai_entries_agl);
+ list_add_tail(&child->lli_agl_list, &sai->sai_agls);
spin_unlock(&parent->lli_agl_lock);
} else {
spin_unlock(&child->lli_agl_lock);
@@ -429,6 +386,7 @@ static void ll_agl_add(struct ll_statahead_info *sai,
wake_up(&sai->sai_agl_thread.t_ctl_waitq);
}
+/* allocate sai */
static struct ll_statahead_info *ll_sai_alloc(void)
{
struct ll_statahead_info *sai;
@@ -452,10 +410,9 @@ static struct ll_statahead_info *ll_sai_alloc(void)
init_waitqueue_head(&sai->sai_thread.t_ctl_waitq);
init_waitqueue_head(&sai->sai_agl_thread.t_ctl_waitq);
+ INIT_LIST_HEAD(&sai->sai_interim_entries);
INIT_LIST_HEAD(&sai->sai_entries);
- INIT_LIST_HEAD(&sai->sai_entries_received);
- INIT_LIST_HEAD(&sai->sai_entries_stated);
- INIT_LIST_HEAD(&sai->sai_entries_agl);
+ INIT_LIST_HEAD(&sai->sai_agls);
for (i = 0; i < LL_SA_CACHE_SIZE; i++) {
INIT_LIST_HEAD(&sai->sai_cache[i]);
@@ -486,7 +443,7 @@ static void ll_sai_put(struct ll_statahead_info *sai)
if (atomic_dec_and_lock(&sai->sai_refcount, &lli->lli_sa_lock)) {
struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
- struct ll_sa_entry *entry, *next;
+ struct sa_entry *entry, *next;
lli->lli_sai = NULL;
spin_unlock(&lli->lli_sa_lock);
@@ -494,14 +451,14 @@ static void ll_sai_put(struct ll_statahead_info *sai)
LASSERT(thread_is_stopped(&sai->sai_thread));
LASSERT(thread_is_stopped(&sai->sai_agl_thread));
LASSERT(sai->sai_sent == sai->sai_replied);
+ LASSERT(!sa_has_callback(sai));
list_for_each_entry_safe(entry, next, &sai->sai_entries,
- se_link)
- do_sa_entry_fini(sai, entry);
+ se_list)
+ sa_kill(sai, entry);
LASSERT(atomic_read(&sai->sai_cache_count) == 0);
- LASSERT(list_empty(&sai->sai_entries_agl));
- LASSERT(atomic_read(&sai->sai_refcount) == 0);
+ LASSERT(list_empty(&sai->sai_agls));
iput(sai->sai_inode);
kfree(sai);
@@ -509,7 +466,7 @@ static void ll_sai_put(struct ll_statahead_info *sai)
}
}
-/* Do NOT forget to drop inode refcount when into sai_entries_agl. */
+/* Do NOT forget to drop inode refcount when into sai_agls. */
static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
{
struct ll_inode_info *lli = ll_i2info(inode);
@@ -569,9 +526,12 @@ static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
iput(inode);
}
-/* prepare inode for received statahead entry, and add it into agl list */
-static void sa_post_one(struct ll_statahead_info *sai,
- struct ll_sa_entry *entry)
+/*
+ * prepare inode for sa entry, add it into agl list, now sa_entry is ready
+ * to be used by scanner process.
+ */
+static void sa_instantiate(struct ll_statahead_info *sai,
+ struct sa_entry *entry)
{
struct inode *dir = sai->sai_inode;
struct inode *child;
@@ -629,8 +589,9 @@ static void sa_post_one(struct ll_statahead_info *sai,
if (rc)
goto out;
- CDEBUG(D_DLMTRACE, "%s: setting l_data to inode "DFID"%p\n",
+ CDEBUG(D_READA, "%s: setting %.*s" DFID " l_data to inode %p\n",
ll_get_fsname(child->i_sb, NULL, 0),
+ entry->se_qstr.len, entry->se_qstr.name,
PFID(ll_inode2fid(child)), child);
ll_set_lock_data(ll_i2sbi(dir)->ll_md_exp, child, it, NULL);
@@ -640,45 +601,42 @@ static void sa_post_one(struct ll_statahead_info *sai,
ll_agl_add(sai, child, entry->se_index);
out:
- /* The "sa_entry_post_stat()" will drop related ldlm ibits lock
- * reference count by calling "ll_intent_drop_lock()" in spite of the
- * above operations failed or not. Do not worry about calling
- * "ll_intent_drop_lock()" more than once.
+ /*
+ * sa_make_ready() will drop ldlm ibits lock refcount by calling
+ * ll_intent_drop_lock() in spite of failures. Do not worry about
+ * calling ll_intent_drop_lock() more than once.
*/
- sa_entry_post_stat(sai, entry, rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
- if (entry->se_index == sai->sai_index_wait)
- wake_up(&sai->sai_waitq);
- ll_sa_entry_put(sai, entry);
+ sa_make_ready(sai, entry, rc);
}
-static void ll_post_statahead(struct ll_statahead_info *sai)
+/* once there are async stat replies, instantiate sa_entry */
+static void sa_handle_callback(struct ll_statahead_info *sai)
{
struct ll_inode_info *lli;
lli = ll_i2info(sai->sai_inode);
- while (!sa_received_empty(sai)) {
- struct ll_sa_entry *entry;
+ while (sa_has_callback(sai)) {
+ struct sa_entry *entry;
spin_lock(&lli->lli_sa_lock);
- if (unlikely(sa_received_empty(sai))) {
+ if (unlikely(!sa_has_callback(sai))) {
spin_unlock(&lli->lli_sa_lock);
break;
}
- entry = list_entry(sai->sai_entries_received.next,
- struct ll_sa_entry, se_list);
- atomic_inc(&entry->se_refcount);
+ entry = list_entry(sai->sai_interim_entries.next,
+ struct sa_entry, se_list);
list_del_init(&entry->se_list);
spin_unlock(&lli->lli_sa_lock);
- sa_post_one(sai, entry);
+ sa_instantiate(sai, entry);
}
spin_lock(&lli->lli_agl_lock);
while (!agl_list_empty(sai)) {
struct ll_inode_info *clli;
- clli = list_entry(sai->sai_entries_agl.next,
+ clli = list_entry(sai->sai_agls.next,
struct ll_inode_info, lli_agl_list);
list_del_init(&clli->lli_agl_list);
spin_unlock(&lli->lli_agl_lock);
@@ -690,29 +648,38 @@ static void ll_post_statahead(struct ll_statahead_info *sai)
spin_unlock(&lli->lli_agl_lock);
}
+/*
+ * callback for async stat, because this is called in ptlrpcd context, we only
+ * put sa_entry in sai_cb_entries list, and let sa_handle_callback() to really
+ * prepare inode and instantiate sa_entry later.
+ */
static int ll_statahead_interpret(struct ptlrpc_request *req,
struct md_enqueue_info *minfo, int rc)
{
struct lookup_intent *it = &minfo->mi_it;
struct inode *dir = minfo->mi_dir;
struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_statahead_info *sai = NULL;
- struct ll_sa_entry *entry;
- int wakeup;
+ struct ll_statahead_info *sai = lli->lli_sai;
+ struct sa_entry *entry = (struct sa_entry *)minfo->mi_cbdata;
+ bool wakeup;
if (it_disposition(it, DISP_LOOKUP_NEG))
rc = -ENOENT;
- sai = ll_sai_get(dir);
+ /*
+ * because statahead thread will wait for all inflight RPC to finish,
+ * sai should be always valid, no need to refcount
+ */
LASSERT(sai);
LASSERT(!thread_is_stopped(&sai->sai_thread));
+ LASSERT(entry);
+
+ CDEBUG(D_READA, "sa_entry %.*s rc %d\n",
+ entry->se_qstr.len, entry->se_qstr.name, rc);
spin_lock(&lli->lli_sa_lock);
- entry = ll_sa_entry_get_byindex(sai, minfo->mi_cbdata);
- LASSERT(entry);
if (rc) {
- __sa_entry_post_stat(sai, entry, SA_ENTRY_INVA);
- wakeup = (entry->se_index == sai->sai_index_wait);
+ wakeup = __sa_make_ready(sai, entry, rc);
} else {
entry->se_minfo = minfo;
entry->se_req = ptlrpc_request_addref(req);
@@ -724,27 +691,24 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
*/
entry->se_handle = it->it_lock_handle;
ll_intent_drop_lock(it);
- wakeup = sa_received_empty(sai);
- list_add_tail(&entry->se_list, &sai->sai_entries_received);
+ wakeup = !sa_has_callback(sai);
+ list_add_tail(&entry->se_list, &sai->sai_interim_entries);
}
sai->sai_replied++;
- spin_unlock(&lli->lli_sa_lock);
- ll_sa_entry_put(sai, entry);
if (wakeup)
wake_up(&sai->sai_thread.t_ctl_waitq);
+ spin_unlock(&lli->lli_sa_lock);
if (rc) {
ll_intent_release(it);
iput(dir);
kfree(minfo);
}
- if (sai)
- ll_sai_put(sai);
return rc;
}
-static void sa_args_fini(struct md_enqueue_info *minfo,
+static void sa_fini_data(struct md_enqueue_info *minfo,
struct ldlm_enqueue_info *einfo)
{
LASSERT(minfo && einfo);
@@ -756,8 +720,8 @@ static void sa_args_fini(struct md_enqueue_info *minfo,
/**
* prepare arguments for async stat RPC.
*/
-static int sa_args_init(struct inode *dir, struct inode *child,
- struct ll_sa_entry *entry, struct md_enqueue_info **pmi,
+static int sa_prep_data(struct inode *dir, struct inode *child,
+ struct sa_entry *entry, struct md_enqueue_info **pmi,
struct ldlm_enqueue_info **pei)
{
const struct qstr *qstr = &entry->se_qstr;
@@ -786,7 +750,7 @@ static int sa_args_init(struct inode *dir, struct inode *child,
minfo->mi_it.it_op = IT_GETATTR;
minfo->mi_dir = igrab(dir);
minfo->mi_cb = ll_statahead_interpret;
- minfo->mi_cbdata = entry->se_index;
+ minfo->mi_cbdata = entry;
einfo->ei_type = LDLM_IBITS;
einfo->ei_mode = it_to_lock_mode(&minfo->mi_it);
@@ -801,19 +765,19 @@ static int sa_args_init(struct inode *dir, struct inode *child,
return 0;
}
-static int do_sa_lookup(struct inode *dir, struct ll_sa_entry *entry)
+static int sa_lookup(struct inode *dir, struct sa_entry *entry)
{
struct md_enqueue_info *minfo;
struct ldlm_enqueue_info *einfo;
int rc;
- rc = sa_args_init(dir, NULL, entry, &minfo, &einfo);
+ rc = sa_prep_data(dir, NULL, entry, &minfo, &einfo);
if (rc)
return rc;
rc = md_intent_getattr_async(ll_i2mdexp(dir), minfo, einfo);
if (rc < 0)
- sa_args_fini(minfo, einfo);
+ sa_fini_data(minfo, einfo);
return rc;
}
@@ -824,8 +788,8 @@ static int do_sa_lookup(struct inode *dir, struct ll_sa_entry *entry)
* \retval 0 -- will send stat-ahead request
* \retval others -- prepare stat-ahead request failed
*/
-static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
- struct dentry *dentry)
+static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
+ struct dentry *dentry)
{
struct inode *inode = d_inode(dentry);
struct lookup_intent it = { .it_op = IT_GETATTR,
@@ -849,7 +813,7 @@ static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
return 1;
}
- rc = sa_args_init(dir, inode, entry, &minfo, &einfo);
+ rc = sa_prep_data(dir, inode, entry, &minfo, &einfo);
if (rc) {
entry->se_inode = NULL;
iput(inode);
@@ -860,32 +824,30 @@ static int do_sa_revalidate(struct inode *dir, struct ll_sa_entry *entry,
if (rc < 0) {
entry->se_inode = NULL;
iput(inode);
- sa_args_fini(minfo, einfo);
+ sa_fini_data(minfo, einfo);
}
return rc;
}
-static void ll_statahead_one(struct dentry *parent, const char *name,
- const int name_len)
+static void sa_statahead(struct dentry *parent, const char *name, int len)
{
struct inode *dir = d_inode(parent);
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_statahead_info *sai = lli->lli_sai;
struct dentry *dentry = NULL;
- struct ll_sa_entry *entry;
+ struct sa_entry *entry;
int rc;
- entry = ll_sa_entry_alloc(parent, sai, sai->sai_index, name,
- name_len);
+ entry = sa_alloc(parent, sai, sai->sai_index, name, len);
if (IS_ERR(entry))
return;
dentry = d_lookup(parent, &entry->se_qstr);
if (!dentry) {
- rc = do_sa_lookup(dir, entry);
+ rc = sa_lookup(dir, entry);
} else {
- rc = do_sa_revalidate(dir, entry, dentry);
+ rc = sa_revalidate(dir, entry, dentry);
if (rc == 1 && agl_should_run(sai, d_inode(dentry)))
ll_agl_add(sai, d_inode(dentry), entry->se_index);
}
@@ -893,18 +855,12 @@ static void ll_statahead_one(struct dentry *parent, const char *name,
if (dentry)
dput(dentry);
- if (rc) {
- sa_entry_post_stat(sai, entry,
- rc < 0 ? SA_ENTRY_INVA : SA_ENTRY_SUCC);
- if (entry->se_index == sai->sai_index_wait)
- wake_up(&sai->sai_waitq);
- } else {
+ if (rc)
+ sa_make_ready(sai, entry, rc);
+ else
sai->sai_sent++;
- }
sai->sai_index++;
- /* drop one refcount on entry by ll_sa_entry_alloc */
- ll_sa_entry_put(sai, entry);
}
static int ll_agl_thread(void *arg)
@@ -938,7 +894,7 @@ static int ll_agl_thread(void *arg)
while (1) {
l_wait_event(thread->t_ctl_waitq,
- !list_empty(&sai->sai_entries_agl) ||
+ !list_empty(&sai->sai_agls) ||
!thread_is_running(thread),
&lwi);
@@ -949,8 +905,8 @@ static int ll_agl_thread(void *arg)
/* The statahead thread maybe help to process AGL entries,
* so check whether list empty again.
*/
- if (!list_empty(&sai->sai_entries_agl)) {
- clli = list_entry(sai->sai_entries_agl.next,
+ if (!list_empty(&sai->sai_agls)) {
+ clli = list_entry(sai->sai_agls.next,
struct ll_inode_info, lli_agl_list);
list_del_init(&clli->lli_agl_list);
spin_unlock(&plli->lli_agl_lock);
@@ -962,8 +918,8 @@ static int ll_agl_thread(void *arg)
spin_lock(&plli->lli_agl_lock);
sai->sai_agl_valid = 0;
- while (!list_empty(&sai->sai_entries_agl)) {
- clli = list_entry(sai->sai_entries_agl.next,
+ while (!list_empty(&sai->sai_agls)) {
+ clli = list_entry(sai->sai_agls.next,
struct ll_inode_info, lli_agl_list);
list_del_init(&clli->lli_agl_list);
spin_unlock(&plli->lli_agl_lock);
@@ -1118,15 +1074,15 @@ static int ll_statahead_thread(void *arg)
do {
l_wait_event(thread->t_ctl_waitq,
!sa_sent_full(sai) ||
- !list_empty(&sai->sai_entries_received) ||
- !list_empty(&sai->sai_entries_agl) ||
+ sa_has_callback(sai) ||
+ !list_empty(&sai->sai_agls) ||
!thread_is_running(thread),
&lwi);
- ll_post_statahead(sai);
+ sa_handle_callback(sai);
} while (sa_sent_full(sai) &&
thread_is_running(thread));
- ll_statahead_one(parent, name, namelen);
+ sa_statahead(parent, name, namelen);
}
pos = le64_to_cpu(dp->ldp_hash_end);
@@ -1158,12 +1114,12 @@ static int ll_statahead_thread(void *arg)
*/
while (thread_is_running(thread)) {
l_wait_event(thread->t_ctl_waitq,
- !sa_received_empty(sai) ||
+ sa_has_callback(sai) ||
!agl_list_empty(sai) ||
!thread_is_running(thread),
&lwi);
- ll_post_statahead(sai);
+ sa_handle_callback(sai);
}
out:
if (sai->sai_agl_valid) {
@@ -1193,8 +1149,8 @@ out:
sai->sai_sent == sai->sai_replied, &lwi);
}
- /* release resources held by received entries. */
- ll_post_statahead(sai);
+ /* release resources held by statahead RPCs */
+ sa_handle_callback(sai);
spin_lock(&lli->lli_sa_lock);
thread_set_flags(thread, SVC_STOPPED);
@@ -1251,7 +1207,7 @@ void ll_deauthorize_statahead(struct inode *dir, void *key)
if (sai && thread_is_running(&sai->sai_thread)) {
/*
* statahead thread may not quit yet because it needs to cache
- * stated entries, now it's time to tell it to quit.
+ * entries, now it's time to tell it to quit.
*/
thread_set_flags(&sai->sai_thread, SVC_STOPPING);
wake_up(&sai->sai_thread.t_ctl_waitq);
@@ -1389,29 +1345,12 @@ out:
return rc;
}
-static void
-ll_sai_unplug(struct ll_statahead_info *sai, struct ll_sa_entry *entry)
-{
- if (entry && entry->se_stat == SA_ENTRY_SUCC) {
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
-
- sai->sai_hit++;
- sai->sai_consecutive_miss = 0;
- sai->sai_max = min(2 * sai->sai_max, sbi->ll_sa_max);
- } else {
- sai->sai_miss++;
- sai->sai_consecutive_miss++;
- }
- ll_sa_entry_fini(sai, entry);
- wake_up(&sai->sai_thread.t_ctl_waitq);
-}
-
static int revalidate_statahead_dentry(struct inode *dir,
struct ll_statahead_info *sai,
struct dentry **dentryp,
int only_unplug)
{
- struct ll_sa_entry *entry = NULL;
+ struct sa_entry *entry = NULL;
struct l_wait_info lwi = { 0 };
int rc = 0;
@@ -1442,31 +1381,31 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
- entry = ll_sa_entry_get_byname(sai, &(*dentryp)->d_name);
+ entry = sa_get(sai, &(*dentryp)->d_name);
if (!entry || only_unplug) {
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return entry ? 1 : -EAGAIN;
}
/* if statahead is busy in readdir, help it do post-work */
- if (!ll_sa_entry_stated(entry) && sai->sai_in_readpage)
- ll_post_statahead(sai);
+ if (!sa_ready(entry) && sai->sai_in_readpage)
+ sa_handle_callback(sai);
- if (!ll_sa_entry_stated(entry)) {
+ if (!sa_ready(entry)) {
sai->sai_index_wait = entry->se_index;
lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
LWI_ON_SIGNAL_NOOP, NULL);
rc = l_wait_event(sai->sai_waitq,
- ll_sa_entry_stated(entry) ||
+ sa_ready(entry) ||
thread_is_stopped(&sai->sai_thread),
&lwi);
if (rc < 0) {
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return -EAGAIN;
}
}
- if (entry->se_stat == SA_ENTRY_SUCC && entry->se_inode) {
+ if (entry->se_state == SA_ENTRY_SUCC && entry->se_inode) {
struct inode *inode = entry->se_inode;
struct lookup_intent it = { .it_op = IT_GETATTR,
.it_lock_handle = entry->se_handle };
@@ -1480,7 +1419,7 @@ static int revalidate_statahead_dentry(struct inode *dir,
alias = ll_splice_alias(inode, *dentryp);
if (IS_ERR(alias)) {
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return PTR_ERR(alias);
}
*dentryp = alias;
@@ -1507,7 +1446,7 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
out_unplug:
- ll_sai_unplug(sai, entry);
+ sa_put(sai, entry);
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 057/124] staging: lustre: statahead: ll_intent_drop_lock() called in spinlock
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
ll_intent_drop_lock() may sleep, which should not be called inside
spinlock.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2272
Reviewed-on: http://review.whamcloud.com/9665
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/statahead.c | 24 ++++++++++++++++------
1 files changed, 17 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index d041720..5be955d 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -661,6 +661,7 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_statahead_info *sai = lli->lli_sai;
struct sa_entry *entry = (struct sa_entry *)minfo->mi_cbdata;
+ __u64 handle = 0;
bool wakeup;
if (it_disposition(it, DISP_LOOKUP_NEG))
@@ -677,6 +678,21 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
CDEBUG(D_READA, "sa_entry %.*s rc %d\n",
entry->se_qstr.len, entry->se_qstr.name, rc);
+ if (rc) {
+ ll_intent_release(it);
+ iput(dir);
+ kfree(minfo);
+ } else {
+ /*
+ * release ibits lock ASAP to avoid deadlock when statahead
+ * thread enqueues lock on parent in readdir and another
+ * process enqueues lock on child with parent lock held, eg.
+ * unlink.
+ */
+ handle = it->it_lock_handle;
+ ll_intent_drop_lock(it);
+ }
+
spin_lock(&lli->lli_sa_lock);
if (rc) {
wakeup = __sa_make_ready(sai, entry, rc);
@@ -689,8 +705,7 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
* for readpage and other tries to enqueue lock on child
* with parent's lock held, for example: unlink.
*/
- entry->se_handle = it->it_lock_handle;
- ll_intent_drop_lock(it);
+ entry->se_handle = handle;
wakeup = !sa_has_callback(sai);
list_add_tail(&entry->se_list, &sai->sai_interim_entries);
}
@@ -700,11 +715,6 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
wake_up(&sai->sai_thread.t_ctl_waitq);
spin_unlock(&lli->lli_sa_lock);
- if (rc) {
- ll_intent_release(it);
- iput(dir);
- kfree(minfo);
- }
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 057/124] staging: lustre: statahead: ll_intent_drop_lock() called in spinlock
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
ll_intent_drop_lock() may sleep, which should not be called inside
spinlock.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2272
Reviewed-on: http://review.whamcloud.com/9665
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/statahead.c | 24 ++++++++++++++++------
1 files changed, 17 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index d041720..5be955d 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -661,6 +661,7 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_statahead_info *sai = lli->lli_sai;
struct sa_entry *entry = (struct sa_entry *)minfo->mi_cbdata;
+ __u64 handle = 0;
bool wakeup;
if (it_disposition(it, DISP_LOOKUP_NEG))
@@ -677,6 +678,21 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
CDEBUG(D_READA, "sa_entry %.*s rc %d\n",
entry->se_qstr.len, entry->se_qstr.name, rc);
+ if (rc) {
+ ll_intent_release(it);
+ iput(dir);
+ kfree(minfo);
+ } else {
+ /*
+ * release ibits lock ASAP to avoid deadlock when statahead
+ * thread enqueues lock on parent in readdir and another
+ * process enqueues lock on child with parent lock held, eg.
+ * unlink.
+ */
+ handle = it->it_lock_handle;
+ ll_intent_drop_lock(it);
+ }
+
spin_lock(&lli->lli_sa_lock);
if (rc) {
wakeup = __sa_make_ready(sai, entry, rc);
@@ -689,8 +705,7 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
* for readpage and other tries to enqueue lock on child
* with parent's lock held, for example: unlink.
*/
- entry->se_handle = it->it_lock_handle;
- ll_intent_drop_lock(it);
+ entry->se_handle = handle;
wakeup = !sa_has_callback(sai);
list_add_tail(&entry->se_list, &sai->sai_interim_entries);
}
@@ -700,11 +715,6 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
wake_up(&sai->sai_thread.t_ctl_waitq);
spin_unlock(&lli->lli_sa_lock);
- if (rc) {
- ll_intent_release(it);
- iput(dir);
- kfree(minfo);
- }
return rc;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 058/124] staging: lustre: statahead: race in start/stop statahead
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
When starting statahead thread, it should check whether current
lli_opendir_key was deauthorized in the mean time by another
process.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9666
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/statahead.c | 29 +++++++++++++++++++---
1 files changed, 25 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 5be955d..70ca252 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -1509,7 +1509,24 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
CDEBUG(D_READA, "start statahead thread: sai %p, parent %pd\n",
sai, parent);
+ /*
+ * if another process started statahead thread, or deauthorized current
+ * lli_opendir_key, don't start statahead.
+ */
+ spin_lock(&lli->lli_sa_lock);
+ if (unlikely(lli->lli_sai || lli->lli_opendir_key ||
+ lli->lli_opendir_pid != current->pid)) {
+ spin_unlock(&lli->lli_sa_lock);
+
+ dput(parent);
+ iput(sai->sai_inode);
+ rc = -EAGAIN;
+ goto out;
+ }
lli->lli_sai = sai;
+ spin_unlock(&lli->lli_sa_lock);
+
+ atomic_inc(&ll_i2sbi(parent->d_inode)->ll_sa_running);
task = kthread_run(ll_statahead_thread, parent, "ll_sa_%u",
lli->lli_opendir_pid);
@@ -1518,9 +1535,12 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
rc = PTR_ERR(task);
CERROR("cannot start ll_sa thread: rc = %d\n", rc);
dput(parent);
- lli->lli_opendir_key = NULL;
+
+ spin_lock(&lli->lli_sa_lock);
thread_set_flags(thread, SVC_STOPPED);
thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
+ spin_unlock(&lli->lli_sa_lock);
+
ll_sai_put(sai);
LASSERT(!lli->lli_sai);
return -EAGAIN;
@@ -1529,7 +1549,6 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
l_wait_event(thread->t_ctl_waitq,
thread_is_running(thread) || thread_is_stopped(thread),
&lwi);
- atomic_inc(&ll_i2sbi(d_inode(parent))->ll_sa_running);
ll_sai_put(sai);
/*
@@ -1540,9 +1559,11 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
out:
kfree(sai);
+ /*
+ * once we start statahead thread failed, disable statahead so
+ * subsequent won't waste time to try it.
+ */
spin_lock(&lli->lli_sa_lock);
- lli->lli_opendir_key = NULL;
- lli->lli_opendir_pid = 0;
lli->lli_sa_enabled = 0;
spin_unlock(&lli->lli_sa_lock);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 058/124] staging: lustre: statahead: race in start/stop statahead
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
When starting statahead thread, it should check whether current
lli_opendir_key was deauthorized in the mean time by another
process.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9666
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/statahead.c | 29 +++++++++++++++++++---
1 files changed, 25 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 5be955d..70ca252 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -1509,7 +1509,24 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
CDEBUG(D_READA, "start statahead thread: sai %p, parent %pd\n",
sai, parent);
+ /*
+ * if another process started statahead thread, or deauthorized current
+ * lli_opendir_key, don't start statahead.
+ */
+ spin_lock(&lli->lli_sa_lock);
+ if (unlikely(lli->lli_sai || lli->lli_opendir_key ||
+ lli->lli_opendir_pid != current->pid)) {
+ spin_unlock(&lli->lli_sa_lock);
+
+ dput(parent);
+ iput(sai->sai_inode);
+ rc = -EAGAIN;
+ goto out;
+ }
lli->lli_sai = sai;
+ spin_unlock(&lli->lli_sa_lock);
+
+ atomic_inc(&ll_i2sbi(parent->d_inode)->ll_sa_running);
task = kthread_run(ll_statahead_thread, parent, "ll_sa_%u",
lli->lli_opendir_pid);
@@ -1518,9 +1535,12 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
rc = PTR_ERR(task);
CERROR("cannot start ll_sa thread: rc = %d\n", rc);
dput(parent);
- lli->lli_opendir_key = NULL;
+
+ spin_lock(&lli->lli_sa_lock);
thread_set_flags(thread, SVC_STOPPED);
thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
+ spin_unlock(&lli->lli_sa_lock);
+
ll_sai_put(sai);
LASSERT(!lli->lli_sai);
return -EAGAIN;
@@ -1529,7 +1549,6 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
l_wait_event(thread->t_ctl_waitq,
thread_is_running(thread) || thread_is_stopped(thread),
&lwi);
- atomic_inc(&ll_i2sbi(d_inode(parent))->ll_sa_running);
ll_sai_put(sai);
/*
@@ -1540,9 +1559,11 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
out:
kfree(sai);
+ /*
+ * once we start statahead thread failed, disable statahead so
+ * subsequent won't waste time to try it.
+ */
spin_lock(&lli->lli_sa_lock);
- lli->lli_opendir_key = NULL;
- lli->lli_opendir_pid = 0;
lli->lli_sa_enabled = 0;
spin_unlock(&lli->lli_sa_lock);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 059/124] staging: lustre: at: net AT after connect
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Boyko, Vitaly Fertman, James Simmons
From: Alexander Boyko <alexander_boyko@xyratex.com>
Once connected, the previously gathered AT statistics is not valid
anymore because may reflect other routing, etc. The connect by itself
could take a long time due to different reasons (e.g. server was not
ready) and net latency got very high (see import_select_connection())
what does not reflect the current situation.
Take into account only the current (re-)CONNECT rpc latency.
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5380
Xyratex-bug-id: MRP-1285
Reviewed-on: http://review.whamcloud.com/11155
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 4 ++--
drivers/staging/lustre/lustre/ptlrpc/import.c | 8 ++++++++
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 2 ++
3 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index bea1c16..27fe00d 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -283,8 +283,8 @@ int ptlrpc_at_get_net_latency(struct ptlrpc_request *req)
}
/* Adjust expected network latency */
-static void ptlrpc_at_adj_net_latency(struct ptlrpc_request *req,
- unsigned int service_time)
+void ptlrpc_at_adj_net_latency(struct ptlrpc_request *req,
+ unsigned int service_time)
{
unsigned int nl, oldnl;
struct imp_at *at;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 013a957..17f3fec 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -846,6 +846,14 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
imp->imp_obd->obd_self_export->exp_connect_data = *ocd;
class_export_put(exp);
+ /*
+ * The net statistics after (re-)connect is not valid anymore,
+ * because may reflect other routing, etc.
+ */
+ at_init(&imp->imp_at.iat_net_latency, 0, 0);
+ ptlrpc_at_adj_net_latency(request,
+ lustre_msg_get_service_time(request->rq_repmsg));
+
obd_import_event(imp->imp_obd, imp, IMP_EVENT_OCD);
if (aa->pcaa_initial_connect) {
diff --git a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
index a9831fa..29cfac2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
+++ b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
@@ -53,6 +53,8 @@ int ptlrpc_start_thread(struct ptlrpc_service_part *svcpt, int wait);
int ptlrpcd_start(struct ptlrpcd_ctl *pc);
/* client.c */
+void ptlrpc_at_adj_net_latency(struct ptlrpc_request *req,
+ unsigned int service_time);
struct ptlrpc_bulk_desc *ptlrpc_new_bulk(unsigned npages, unsigned max_brw,
unsigned type, unsigned portal);
int ptlrpc_request_cache_init(void);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 059/124] staging: lustre: at: net AT after connect
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Boyko, Vitaly Fertman, James Simmons
From: Alexander Boyko <alexander_boyko@xyratex.com>
Once connected, the previously gathered AT statistics is not valid
anymore because may reflect other routing, etc. The connect by itself
could take a long time due to different reasons (e.g. server was not
ready) and net latency got very high (see import_select_connection())
what does not reflect the current situation.
Take into account only the current (re-)CONNECT rpc latency.
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5380
Xyratex-bug-id: MRP-1285
Reviewed-on: http://review.whamcloud.com/11155
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 4 ++--
drivers/staging/lustre/lustre/ptlrpc/import.c | 8 ++++++++
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 2 ++
3 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index bea1c16..27fe00d 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -283,8 +283,8 @@ int ptlrpc_at_get_net_latency(struct ptlrpc_request *req)
}
/* Adjust expected network latency */
-static void ptlrpc_at_adj_net_latency(struct ptlrpc_request *req,
- unsigned int service_time)
+void ptlrpc_at_adj_net_latency(struct ptlrpc_request *req,
+ unsigned int service_time)
{
unsigned int nl, oldnl;
struct imp_at *at;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 013a957..17f3fec 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -846,6 +846,14 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
imp->imp_obd->obd_self_export->exp_connect_data = *ocd;
class_export_put(exp);
+ /*
+ * The net statistics after (re-)connect is not valid anymore,
+ * because may reflect other routing, etc.
+ */
+ at_init(&imp->imp_at.iat_net_latency, 0, 0);
+ ptlrpc_at_adj_net_latency(request,
+ lustre_msg_get_service_time(request->rq_repmsg));
+
obd_import_event(imp->imp_obd, imp, IMP_EVENT_OCD);
if (aa->pcaa_initial_connect) {
diff --git a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
index a9831fa..29cfac2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
+++ b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
@@ -53,6 +53,8 @@ int ptlrpc_start_thread(struct ptlrpc_service_part *svcpt, int wait);
int ptlrpcd_start(struct ptlrpcd_ctl *pc);
/* client.c */
+void ptlrpc_at_adj_net_latency(struct ptlrpc_request *req,
+ unsigned int service_time);
struct ptlrpc_bulk_desc *ptlrpc_new_bulk(unsigned npages, unsigned max_brw,
unsigned type, unsigned portal);
int ptlrpc_request_cache_init(void);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 060/124] staging: lustre: mdc: fix comparison between signed and unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:37 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change type of client_obd->*_mds_*size from int to __u32 and
argumanets of related create/rename/setattr functions.
Change type of op_data->op_namelen to size_t.
Change type of argument size for all mdc_*_pack() to size_t.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/11379
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 23 +++++++-------
drivers/staging/lustre/lustre/include/obd_class.h | 12 ++++----
.../staging/lustre/lustre/llite/llite_internal.h | 4 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 4 +-
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 5 ++-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 29 +++++++++---------
drivers/staging/lustre/lustre/mdc/mdc_internal.h | 30 ++++++++++---------
drivers/staging/lustre/lustre/mdc/mdc_lib.c | 27 +++++++++--------
drivers/staging/lustre/lustre/mdc/mdc_locks.c | 19 ++++++------
drivers/staging/lustre/lustre/mdc/mdc_reint.c | 10 +++---
drivers/staging/lustre/lustre/mdc/mdc_request.c | 31 ++++++++++---------
11 files changed, 101 insertions(+), 93 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 6758bf4..8a7c803 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -220,10 +220,10 @@ struct client_obd {
/* max_mds_easize is purely a performance thing so we don't have to
* call obd_size_diskmd() all the time.
*/
- int cl_default_mds_easize;
- int cl_max_mds_easize;
- int cl_default_mds_cookiesize;
- int cl_max_mds_cookiesize;
+ u32 cl_default_mds_easize;
+ u32 cl_max_mds_easize;
+ u32 cl_default_mds_cookiesize;
+ u32 cl_max_mds_cookiesize;
enum lustre_sec_part cl_sp_me;
enum lustre_sec_part cl_sp_to;
@@ -745,7 +745,7 @@ struct md_op_data {
struct lustre_handle op_handle;
s64 op_mod_time;
const char *op_name;
- int op_namelen;
+ size_t op_namelen;
__u32 op_mode;
struct lmv_stripe_md *op_mea1;
struct lmv_stripe_md *op_mea2;
@@ -976,8 +976,8 @@ struct md_ops {
int (*close)(struct obd_export *, struct md_op_data *,
struct md_open_data *, struct ptlrpc_request **);
int (*create)(struct obd_export *, struct md_op_data *,
- const void *, int, int, __u32, __u32, cfs_cap_t,
- __u64, struct ptlrpc_request **);
+ const void *, size_t, umode_t, uid_t, gid_t,
+ cfs_cap_t, __u64, struct ptlrpc_request **);
int (*done_writing)(struct obd_export *, struct md_op_data *,
struct md_open_data *);
int (*enqueue)(struct obd_export *, struct ldlm_enqueue_info *,
@@ -995,10 +995,10 @@ struct md_ops {
int (*link)(struct obd_export *, struct md_op_data *,
struct ptlrpc_request **);
int (*rename)(struct obd_export *, struct md_op_data *,
- const char *, int, const char *, int,
+ const char *, size_t, const char *, size_t,
struct ptlrpc_request **);
int (*setattr)(struct obd_export *, struct md_op_data *, void *,
- int, void *, int, struct ptlrpc_request **,
+ size_t, void *, size_t, struct ptlrpc_request **,
struct md_open_data **mod);
int (*sync)(struct obd_export *, const struct lu_fid *,
struct ptlrpc_request **);
@@ -1016,7 +1016,7 @@ struct md_ops {
u64, const char *, const char *, int, int, int,
struct ptlrpc_request **);
- int (*init_ea_size)(struct obd_export *, int, int, int, int);
+ int (*init_ea_size)(struct obd_export *, u32, u32, u32, u32);
int (*get_lustre_md)(struct obd_export *, struct ptlrpc_request *,
struct obd_export *, struct obd_export *,
@@ -1139,7 +1139,8 @@ static inline const char *lu_dev_name(const struct lu_device *lu_dev)
return lu_dev->ld_obd->obd_name;
}
-static inline bool filename_is_volatile(const char *name, int namelen, int *idx)
+static inline bool filename_is_volatile(const char *name, size_t namelen,
+ int *idx)
{
const char *start;
char *end;
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 6fc54bd..3d081cf 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1358,9 +1358,9 @@ static inline int md_close(struct obd_export *exp, struct md_op_data *op_data,
}
static inline int md_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid,
- __u32 gid, cfs_cap_t cap_effective, __u64 rdev,
- struct ptlrpc_request **request)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective,
+ __u64 rdev, struct ptlrpc_request **request)
{
int rc;
@@ -1440,8 +1440,8 @@ static inline int md_link(struct obd_export *exp, struct md_op_data *op_data,
}
static inline int md_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new,
- int newlen, struct ptlrpc_request **request)
+ const char *old, size_t oldlen, const char *new,
+ size_t newlen, struct ptlrpc_request **request)
{
int rc;
@@ -1453,7 +1453,7 @@ static inline int md_rename(struct obd_export *exp, struct md_op_data *op_data,
}
static inline int md_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request,
struct md_open_data **mod)
{
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index cd639f4..a68bea1 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -868,8 +868,8 @@ enum {
struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
struct inode *i1, struct inode *i2,
- const char *name, int namelen,
- int mode, __u32 opc, void *data);
+ const char *name, size_t namelen,
+ u32 mode, __u32 opc, void *data);
void ll_finish_md_op_data(struct md_op_data *op_data);
int ll_get_obd_name(struct inode *inode, unsigned int cmd, unsigned long arg);
char *ll_get_fsname(struct super_block *sb, char *buf, int buflen);
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index c498748..e46214f 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2324,8 +2324,8 @@ int ll_process_config(struct lustre_cfg *lcfg)
/* this function prepares md_op_data hint for passing ot down to MD stack. */
struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
struct inode *i1, struct inode *i2,
- const char *name, int namelen,
- int mode, __u32 opc, void *data)
+ const char *name, size_t namelen,
+ u32 mode, __u32 opc, void *data)
{
if (!name) {
/* Do not reuse namelen for something else. */
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 6845a5e..a96b49d 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -520,8 +520,9 @@ int lmv_intent_lock(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(fid_is_sane(&op_data->op_fid1));
CDEBUG(D_INODE, "INTENT LOCK '%s' for "DFID" '%*s' on "DFID"\n",
- LL_IT2STR(it), PFID(&op_data->op_fid2), op_data->op_namelen,
- op_data->op_name, PFID(&op_data->op_fid1));
+ LL_IT2STR(it), PFID(&op_data->op_fid2),
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1));
rc = lmv_check_connect(obd);
if (rc)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 0ffb653..0337f05 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -245,8 +245,8 @@ static int lmv_connect(const struct lu_env *env,
return rc;
}
-static int lmv_init_ea_size(struct obd_export *exp, int easize,
- int def_easize, int cookiesize, int def_cookiesize)
+static int lmv_init_ea_size(struct obd_export *exp, u32 easize, u32 def_easize,
+ u32 cookiesize, u32 def_cookiesize)
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
@@ -1660,9 +1660,9 @@ lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data,
}
static int lmv_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid,
- __u32 gid, cfs_cap_t cap_effective, __u64 rdev,
- struct ptlrpc_request **request)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective,
+ __u64 rdev, struct ptlrpc_request **request)
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
@@ -1681,8 +1681,8 @@ static int lmv_create(struct obd_export *exp, struct md_op_data *op_data,
return PTR_ERR(tgt);
CDEBUG(D_INODE, "CREATE name '%.*s' on "DFID" -> mds #%x\n",
- op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
- op_data->op_mds);
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1), op_data->op_mds);
rc = lmv_fid_alloc(NULL, exp, &op_data->op_fid2, op_data);
if (rc)
@@ -1785,8 +1785,8 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data,
return PTR_ERR(tgt);
CDEBUG(D_INODE, "GETATTR_NAME for %*s on " DFID " -> mds #%u\n",
- op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
- tgt->ltd_idx);
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1), tgt->ltd_idx);
rc = md_getattr_name(tgt->ltd_exp, op_data, preq);
if (rc != 0)
@@ -1879,7 +1879,7 @@ static int lmv_link(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(op_data->op_namelen != 0);
CDEBUG(D_INODE, "LINK "DFID":%*s to "DFID"\n",
- PFID(&op_data->op_fid2), op_data->op_namelen,
+ PFID(&op_data->op_fid2), (int)op_data->op_namelen,
op_data->op_name, PFID(&op_data->op_fid1));
op_data->op_fsuid = from_kuid(&init_user_ns, current_fsuid());
@@ -1916,7 +1916,8 @@ static int lmv_link(struct obd_export *exp, struct md_op_data *op_data,
}
static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen,
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen,
struct ptlrpc_request **request)
{
struct obd_device *obd = exp->exp_obd;
@@ -1927,9 +1928,9 @@ static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(oldlen != 0);
CDEBUG(D_INODE, "RENAME %.*s in "DFID":%d to %.*s in "DFID":%d\n",
- oldlen, old, PFID(&op_data->op_fid1),
+ (int)oldlen, old, PFID(&op_data->op_fid1),
op_data->op_mea1 ? op_data->op_mea1->lsm_md_stripe_count : 0,
- newlen, new, PFID(&op_data->op_fid2),
+ (int)newlen, new, PFID(&op_data->op_fid2),
op_data->op_mea2 ? op_data->op_mea2->lsm_md_stripe_count : 0);
rc = lmv_check_connect(obd);
@@ -2035,7 +2036,7 @@ static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
}
static int lmv_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request,
struct md_open_data **mod)
{
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_internal.h b/drivers/staging/lustre/lustre/mdc/mdc_internal.h
index c10441d..f446c1c 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_internal.h
+++ b/drivers/staging/lustre/lustre/mdc/mdc_internal.h
@@ -38,25 +38,26 @@
void lprocfs_mdc_init_vars(struct lprocfs_static_vars *lvars);
void mdc_pack_body(struct ptlrpc_request *req, const struct lu_fid *fid,
- __u64 valid, int ea_size, __u32 suppgid, int flags);
+ __u64 valid, size_t ea_size, __u32 suppgid, u32 flags);
void mdc_swap_layouts_pack(struct ptlrpc_request *req,
struct md_op_data *op_data);
-void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff, __u32 size,
+void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff, size_t size,
const struct lu_fid *fid);
-void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, int flags,
- struct md_op_data *data, int ea_size);
+void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, u32 flags,
+ struct md_op_data *data, size_t ea_size);
void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len);
+ void *ea, size_t ealen, void *ea2, size_t ea2len);
void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const void *data, int datalen, __u32 mode, __u32 uid,
- __u32 gid, cfs_cap_t capability, __u64 rdev);
+ const void *data, size_t datalen, umode_t mode, uid_t uid,
+ gid_t gid, cfs_cap_t capability, __u64 rdev);
void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- __u32 mode, __u64 rdev, __u64 flags, const void *data,
- int datalen);
+ umode_t mode, __u64 rdev, __u64 flags, const void *data,
+ size_t datalen);
void mdc_unlink_pack(struct ptlrpc_request *req, struct md_op_data *op_data);
void mdc_link_pack(struct ptlrpc_request *req, struct md_op_data *op_data);
void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen);
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen);
void mdc_close_pack(struct ptlrpc_request *req, struct md_op_data *op_data);
/* mdc/mdc_locks.c */
@@ -94,16 +95,17 @@ void mdc_commit_open(struct ptlrpc_request *req);
void mdc_replay_open(struct ptlrpc_request *req);
int mdc_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid, __u32 gid,
- cfs_cap_t capability, __u64 rdev,
+ const void *data, size_t datalen, umode_t mode, uid_t uid,
+ gid_t gid, cfs_cap_t capability, __u64 rdev,
struct ptlrpc_request **request);
int mdc_link(struct obd_export *exp, struct md_op_data *op_data,
struct ptlrpc_request **request);
int mdc_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen,
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen,
struct ptlrpc_request **request);
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request, struct md_open_data **mod);
int mdc_unlink(struct obd_export *exp, struct md_op_data *op_data,
struct ptlrpc_request **request);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_lib.c b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
index ecfe13e..82d1c4d 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_lib.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
@@ -58,7 +58,7 @@ void mdc_swap_layouts_pack(struct ptlrpc_request *req,
}
void mdc_pack_body(struct ptlrpc_request *req, const struct lu_fid *fid,
- __u64 valid, int ea_size, __u32 suppgid, int flags)
+ __u64 valid, size_t ea_size, __u32 suppgid, u32 flags)
{
struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
&RMF_MDT_BODY);
@@ -103,8 +103,8 @@ static void mdc_pack_name(struct ptlrpc_request *req,
LASSERT(cpy_len == name_len && lu_name_is_valid_2(buf, cpy_len));
}
-void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff,
- __u32 size, const struct lu_fid *fid)
+void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff, size_t size,
+ const struct lu_fid *fid)
{
struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
&RMF_MDT_BODY);
@@ -118,8 +118,8 @@ void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff,
/* packing of MDS records */
void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const void *data, int datalen, __u32 mode,
- __u32 uid, __u32 gid, cfs_cap_t cap_effective, __u64 rdev)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective, __u64 rdev)
{
struct mdt_rec_create *rec;
char *tmp;
@@ -153,7 +153,7 @@ void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
}
}
-static __u64 mds_pack_open_flags(__u64 flags, __u32 mode)
+static inline __u64 mds_pack_open_flags(__u64 flags)
{
__u64 cr_flags = (flags & (FMODE_READ | FMODE_WRITE |
MDS_OPEN_FL_INTERNAL));
@@ -182,8 +182,8 @@ static __u64 mds_pack_open_flags(__u64 flags, __u32 mode)
/* packing of MDS records */
void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- __u32 mode, __u64 rdev, __u64 flags, const void *lmm,
- int lmmlen)
+ umode_t mode, __u64 rdev, __u64 flags, const void *lmm,
+ size_t lmmlen)
{
struct mdt_rec_create *rec;
char *tmp;
@@ -201,7 +201,7 @@ void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
rec->cr_fid2 = op_data->op_fid2;
rec->cr_mode = mode;
- cr_flags = mds_pack_open_flags(flags, mode);
+ cr_flags = mds_pack_open_flags(flags);
rec->cr_rdev = rdev;
rec->cr_time = op_data->op_mod_time;
rec->cr_suppgid1 = op_data->op_suppgids[0];
@@ -307,7 +307,7 @@ static void mdc_ioepoch_pack(struct mdt_ioepoch *epoch,
}
void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len)
+ void *ea, size_t ealen, void *ea2, size_t ea2len)
{
struct mdt_rec_setattr *rec;
struct mdt_ioepoch *epoch;
@@ -388,7 +388,8 @@ void mdc_link_pack(struct ptlrpc_request *req, struct md_op_data *op_data)
}
void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen)
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen)
{
struct mdt_rec_rename *rec;
@@ -416,8 +417,8 @@ void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
mdc_pack_name(req, &RMF_SYMTGT, new, newlen);
}
-void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, int flags,
- struct md_op_data *op_data, int ea_size)
+void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, u32 flags,
+ struct md_op_data *op_data, size_t ea_size)
{
struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
&RMF_MDT_BODY);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_locks.c b/drivers/staging/lustre/lustre/mdc/mdc_locks.c
index 54de46b..f1f6c08 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_locks.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_locks.c
@@ -235,7 +235,7 @@ mdc_intent_open_pack(struct obd_export *exp, struct lookup_intent *it,
struct obd_device *obddev = class_exp2obd(exp);
struct ldlm_intent *lit;
const void *lmm = op_data->op_data;
- int lmmsize = op_data->op_data_size;
+ u32 lmmsize = op_data->op_data_size;
LIST_HEAD(cancels);
int count = 0;
int mode;
@@ -317,7 +317,8 @@ mdc_intent_getxattr_pack(struct obd_export *exp,
{
struct ptlrpc_request *req;
struct ldlm_intent *lit;
- int rc, count = 0, maxdata;
+ int rc, count = 0;
+ u32 maxdata;
LIST_HEAD(cancels);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
@@ -402,7 +403,7 @@ static struct ptlrpc_request *mdc_intent_getattr_pack(struct obd_export *exp,
OBD_MD_MEA | OBD_MD_FLACL;
struct ldlm_intent *lit;
int rc;
- int easize;
+ u32 easize;
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
&RQF_LDLM_INTENT_GETATTR);
@@ -507,7 +508,7 @@ static int mdc_finish_enqueue(struct obd_export *exp,
struct ldlm_reply *lockrep;
struct ldlm_lock *lock;
void *lvb_data = NULL;
- int lvb_len = 0;
+ u32 lvb_len = 0;
LASSERT(rc >= 0);
/* Similarly, if we're going to replay this request, we don't want to
@@ -952,8 +953,8 @@ static int mdc_finish_intent_lock(struct obd_export *exp,
}
CDEBUG(D_DENTRY,
"D_IT dentry %.*s intent: %s status %d disp %x rc %d\n",
- op_data->op_namelen, op_data->op_name, ldlm_it2str(it->it_op),
- it->it_status, it->it_disposition, rc);
+ (int)op_data->op_namelen, op_data->op_name,
+ ldlm_it2str(it->it_op), it->it_status, it->it_disposition, rc);
return rc;
}
@@ -1067,7 +1068,7 @@ int mdc_intent_lock(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(it);
CDEBUG(D_DLMTRACE, "(name: %.*s,"DFID") in obj "DFID
- ", intent: %s flags %#Lo\n", op_data->op_namelen,
+ ", intent: %s flags %#Lo\n", (int)op_data->op_namelen,
op_data->op_name, PFID(&op_data->op_fid2),
PFID(&op_data->op_fid1), ldlm_it2str(it->it_op),
it->it_flags);
@@ -1177,8 +1178,8 @@ int mdc_intent_getattr_async(struct obd_export *exp,
CDEBUG(D_DLMTRACE,
"name: %.*s in inode " DFID ", intent: %s flags %#Lo\n",
- op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
- ldlm_it2str(it->it_op), it->it_flags);
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1), ldlm_it2str(it->it_op), it->it_flags);
fid_build_reg_res_name(&op_data->op_fid1, &res_id);
req = mdc_intent_getattr_pack(exp, it, op_data);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_reint.c b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
index af5c92c..c921e47 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_reint.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
@@ -99,7 +99,7 @@ int mdc_resource_get_unused(struct obd_export *exp, const struct lu_fid *fid,
}
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request, struct md_open_data **mod)
{
LIST_HEAD(cancels);
@@ -196,9 +196,9 @@ int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
}
int mdc_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid, __u32 gid,
- cfs_cap_t cap_effective, __u64 rdev,
- struct ptlrpc_request **request)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective,
+ __u64 rdev, struct ptlrpc_request **request)
{
struct ptlrpc_request *req;
int level, rc;
@@ -389,7 +389,7 @@ int mdc_link(struct obd_export *exp, struct md_op_data *op_data,
}
int mdc_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen,
+ const char *old, size_t oldlen, const char *new, size_t newlen,
struct ptlrpc_request **request)
{
LIST_HEAD(cancels);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 157fdc0..e622b7a 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -447,7 +447,7 @@ static int mdc_get_lustre_md(struct obd_export *exp,
if (rc < 0)
goto out;
- if (rc < sizeof(*md->lsm)) {
+ if (rc < (typeof(rc))sizeof(*md->lsm)) {
CDEBUG(D_INFO,
"lsm size too small: rc < sizeof (*md->lsm) (%d < %d)\n",
rc, (int)sizeof(*md->lsm));
@@ -485,7 +485,7 @@ static int mdc_get_lustre_md(struct obd_export *exp,
if (rc < 0)
goto out;
- if (rc < sizeof(*md->lmv)) {
+ if (rc < (typeof(rc))sizeof(*md->lmv)) {
CDEBUG(D_INFO,
"size too small: rc < sizeof(*md->lmv) (%d < %d)\n",
rc, (int)sizeof(*md->lmv));
@@ -1843,7 +1843,7 @@ out:
return rc;
}
-static struct kuc_hdr *changelog_kuc_hdr(char *buf, int len, int flags)
+static struct kuc_hdr *changelog_kuc_hdr(char *buf, size_t len, u32 flags)
{
struct kuc_hdr *lh = (struct kuc_hdr *)buf;
@@ -1876,7 +1876,8 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
struct changelog_show *cs = data;
struct llog_changelog_rec *rec = (struct llog_changelog_rec *)hdr;
struct kuc_hdr *lh;
- int len, rc;
+ size_t len;
+ int rc;
if (rec->cr_hdr.lrh_type != CHANGELOG_REC) {
rc = -EINVAL;
@@ -1907,7 +1908,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
memcpy(lh + 1, &rec->cr, len - sizeof(*lh));
rc = libcfs_kkuc_msg_put(cs->cs_fp, lh);
- CDEBUG(D_HSM, "kucmsg fp %p len %d rc %d\n", cs->cs_fp, len, rc);
+ CDEBUG(D_HSM, "kucmsg fp %p len %zu rc %d\n", cs->cs_fp, len, rc);
return rc;
}
@@ -2369,7 +2370,7 @@ static void lustre_swab_hai(struct hsm_action_item *h)
static void lustre_swab_hal(struct hsm_action_list *h)
{
struct hsm_action_item *hai;
- int i;
+ u32 i;
__swab32s(&h->hal_version);
__swab32s(&h->hal_count);
@@ -2418,14 +2419,14 @@ static int mdc_ioc_hsm_ct_start(struct obd_export *exp,
* @param val KUC message (kuc_hdr + hsm_action_list)
* @param len total length of message
*/
-static int mdc_hsm_copytool_send(int len, void *val)
+static int mdc_hsm_copytool_send(size_t len, void *val)
{
struct kuc_hdr *lh = (struct kuc_hdr *)val;
struct hsm_action_list *hal = (struct hsm_action_list *)(lh + 1);
if (len < sizeof(*lh) + sizeof(*hal)) {
- CERROR("Short HSM message %d < %d\n", len,
- (int)(sizeof(*lh) + sizeof(*hal)));
+ CERROR("Short HSM message %zu < %zu\n", len,
+ sizeof(*lh) + sizeof(*hal));
return -EPROTO;
}
if (lh->kuc_magic == __swab16(KUC_MAGIC)) {
@@ -2528,18 +2529,18 @@ static int mdc_get_info(const struct lu_env *env, struct obd_export *exp,
int rc = -EINVAL;
if (KEY_IS(KEY_MAX_EASIZE)) {
- int mdsize, *max_easize;
+ u32 mdsize, *max_easize;
if (*vallen != sizeof(int))
return -EINVAL;
- mdsize = *(int *)val;
+ mdsize = *(u32 *)val;
if (mdsize > exp->exp_obd->u.cli.cl_max_mds_easize)
exp->exp_obd->u.cli.cl_max_mds_easize = mdsize;
max_easize = val;
*max_easize = exp->exp_obd->u.cli.cl_max_mds_easize;
return 0;
} else if (KEY_IS(KEY_DEFAULT_EASIZE)) {
- int *default_easize;
+ u32 *default_easize;
if (*vallen != sizeof(int))
return -EINVAL;
@@ -2556,7 +2557,7 @@ static int mdc_get_info(const struct lu_env *env, struct obd_export *exp,
*data = imp->imp_connect_data;
return 0;
} else if (KEY_IS(KEY_TGT_COUNT)) {
- *((int *)val) = 1;
+ *((u32 *)val) = 1;
return 0;
}
@@ -2784,8 +2785,8 @@ err_rpc_lock:
* a large number of stripes is possible. If a larger reply buffer is
* required it will be reallocated in the ptlrpc layer due to overflow.
*/
-static int mdc_init_ea_size(struct obd_export *exp, int easize,
- int def_easize, int cookiesize, int def_cookiesize)
+static int mdc_init_ea_size(struct obd_export *exp, u32 easize, u32 def_easize,
+ u32 cookiesize, u32 def_cookiesize)
{
struct obd_device *obd = exp->exp_obd;
struct client_obd *cli = &obd->u.cli;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 060/124] staging: lustre: mdc: fix comparison between signed and unsigned
@ 2016-09-18 20:37 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:37 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change type of client_obd->*_mds_*size from int to __u32 and
argumanets of related create/rename/setattr functions.
Change type of op_data->op_namelen to size_t.
Change type of argument size for all mdc_*_pack() to size_t.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/11379
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 23 +++++++-------
drivers/staging/lustre/lustre/include/obd_class.h | 12 ++++----
.../staging/lustre/lustre/llite/llite_internal.h | 4 +-
drivers/staging/lustre/lustre/llite/llite_lib.c | 4 +-
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 5 ++-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 29 +++++++++---------
drivers/staging/lustre/lustre/mdc/mdc_internal.h | 30 ++++++++++---------
drivers/staging/lustre/lustre/mdc/mdc_lib.c | 27 +++++++++--------
drivers/staging/lustre/lustre/mdc/mdc_locks.c | 19 ++++++------
drivers/staging/lustre/lustre/mdc/mdc_reint.c | 10 +++---
drivers/staging/lustre/lustre/mdc/mdc_request.c | 31 ++++++++++---------
11 files changed, 101 insertions(+), 93 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 6758bf4..8a7c803 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -220,10 +220,10 @@ struct client_obd {
/* max_mds_easize is purely a performance thing so we don't have to
* call obd_size_diskmd() all the time.
*/
- int cl_default_mds_easize;
- int cl_max_mds_easize;
- int cl_default_mds_cookiesize;
- int cl_max_mds_cookiesize;
+ u32 cl_default_mds_easize;
+ u32 cl_max_mds_easize;
+ u32 cl_default_mds_cookiesize;
+ u32 cl_max_mds_cookiesize;
enum lustre_sec_part cl_sp_me;
enum lustre_sec_part cl_sp_to;
@@ -745,7 +745,7 @@ struct md_op_data {
struct lustre_handle op_handle;
s64 op_mod_time;
const char *op_name;
- int op_namelen;
+ size_t op_namelen;
__u32 op_mode;
struct lmv_stripe_md *op_mea1;
struct lmv_stripe_md *op_mea2;
@@ -976,8 +976,8 @@ struct md_ops {
int (*close)(struct obd_export *, struct md_op_data *,
struct md_open_data *, struct ptlrpc_request **);
int (*create)(struct obd_export *, struct md_op_data *,
- const void *, int, int, __u32, __u32, cfs_cap_t,
- __u64, struct ptlrpc_request **);
+ const void *, size_t, umode_t, uid_t, gid_t,
+ cfs_cap_t, __u64, struct ptlrpc_request **);
int (*done_writing)(struct obd_export *, struct md_op_data *,
struct md_open_data *);
int (*enqueue)(struct obd_export *, struct ldlm_enqueue_info *,
@@ -995,10 +995,10 @@ struct md_ops {
int (*link)(struct obd_export *, struct md_op_data *,
struct ptlrpc_request **);
int (*rename)(struct obd_export *, struct md_op_data *,
- const char *, int, const char *, int,
+ const char *, size_t, const char *, size_t,
struct ptlrpc_request **);
int (*setattr)(struct obd_export *, struct md_op_data *, void *,
- int, void *, int, struct ptlrpc_request **,
+ size_t, void *, size_t, struct ptlrpc_request **,
struct md_open_data **mod);
int (*sync)(struct obd_export *, const struct lu_fid *,
struct ptlrpc_request **);
@@ -1016,7 +1016,7 @@ struct md_ops {
u64, const char *, const char *, int, int, int,
struct ptlrpc_request **);
- int (*init_ea_size)(struct obd_export *, int, int, int, int);
+ int (*init_ea_size)(struct obd_export *, u32, u32, u32, u32);
int (*get_lustre_md)(struct obd_export *, struct ptlrpc_request *,
struct obd_export *, struct obd_export *,
@@ -1139,7 +1139,8 @@ static inline const char *lu_dev_name(const struct lu_device *lu_dev)
return lu_dev->ld_obd->obd_name;
}
-static inline bool filename_is_volatile(const char *name, int namelen, int *idx)
+static inline bool filename_is_volatile(const char *name, size_t namelen,
+ int *idx)
{
const char *start;
char *end;
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 6fc54bd..3d081cf 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1358,9 +1358,9 @@ static inline int md_close(struct obd_export *exp, struct md_op_data *op_data,
}
static inline int md_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid,
- __u32 gid, cfs_cap_t cap_effective, __u64 rdev,
- struct ptlrpc_request **request)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective,
+ __u64 rdev, struct ptlrpc_request **request)
{
int rc;
@@ -1440,8 +1440,8 @@ static inline int md_link(struct obd_export *exp, struct md_op_data *op_data,
}
static inline int md_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new,
- int newlen, struct ptlrpc_request **request)
+ const char *old, size_t oldlen, const char *new,
+ size_t newlen, struct ptlrpc_request **request)
{
int rc;
@@ -1453,7 +1453,7 @@ static inline int md_rename(struct obd_export *exp, struct md_op_data *op_data,
}
static inline int md_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request,
struct md_open_data **mod)
{
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index cd639f4..a68bea1 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -868,8 +868,8 @@ enum {
struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
struct inode *i1, struct inode *i2,
- const char *name, int namelen,
- int mode, __u32 opc, void *data);
+ const char *name, size_t namelen,
+ u32 mode, __u32 opc, void *data);
void ll_finish_md_op_data(struct md_op_data *op_data);
int ll_get_obd_name(struct inode *inode, unsigned int cmd, unsigned long arg);
char *ll_get_fsname(struct super_block *sb, char *buf, int buflen);
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index c498748..e46214f 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -2324,8 +2324,8 @@ int ll_process_config(struct lustre_cfg *lcfg)
/* this function prepares md_op_data hint for passing ot down to MD stack. */
struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
struct inode *i1, struct inode *i2,
- const char *name, int namelen,
- int mode, __u32 opc, void *data)
+ const char *name, size_t namelen,
+ u32 mode, __u32 opc, void *data)
{
if (!name) {
/* Do not reuse namelen for something else. */
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 6845a5e..a96b49d 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -520,8 +520,9 @@ int lmv_intent_lock(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(fid_is_sane(&op_data->op_fid1));
CDEBUG(D_INODE, "INTENT LOCK '%s' for "DFID" '%*s' on "DFID"\n",
- LL_IT2STR(it), PFID(&op_data->op_fid2), op_data->op_namelen,
- op_data->op_name, PFID(&op_data->op_fid1));
+ LL_IT2STR(it), PFID(&op_data->op_fid2),
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1));
rc = lmv_check_connect(obd);
if (rc)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 0ffb653..0337f05 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -245,8 +245,8 @@ static int lmv_connect(const struct lu_env *env,
return rc;
}
-static int lmv_init_ea_size(struct obd_export *exp, int easize,
- int def_easize, int cookiesize, int def_cookiesize)
+static int lmv_init_ea_size(struct obd_export *exp, u32 easize, u32 def_easize,
+ u32 cookiesize, u32 def_cookiesize)
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
@@ -1660,9 +1660,9 @@ lmv_locate_mds(struct lmv_obd *lmv, struct md_op_data *op_data,
}
static int lmv_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid,
- __u32 gid, cfs_cap_t cap_effective, __u64 rdev,
- struct ptlrpc_request **request)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective,
+ __u64 rdev, struct ptlrpc_request **request)
{
struct obd_device *obd = exp->exp_obd;
struct lmv_obd *lmv = &obd->u.lmv;
@@ -1681,8 +1681,8 @@ static int lmv_create(struct obd_export *exp, struct md_op_data *op_data,
return PTR_ERR(tgt);
CDEBUG(D_INODE, "CREATE name '%.*s' on "DFID" -> mds #%x\n",
- op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
- op_data->op_mds);
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1), op_data->op_mds);
rc = lmv_fid_alloc(NULL, exp, &op_data->op_fid2, op_data);
if (rc)
@@ -1785,8 +1785,8 @@ lmv_getattr_name(struct obd_export *exp, struct md_op_data *op_data,
return PTR_ERR(tgt);
CDEBUG(D_INODE, "GETATTR_NAME for %*s on " DFID " -> mds #%u\n",
- op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
- tgt->ltd_idx);
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1), tgt->ltd_idx);
rc = md_getattr_name(tgt->ltd_exp, op_data, preq);
if (rc != 0)
@@ -1879,7 +1879,7 @@ static int lmv_link(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(op_data->op_namelen != 0);
CDEBUG(D_INODE, "LINK "DFID":%*s to "DFID"\n",
- PFID(&op_data->op_fid2), op_data->op_namelen,
+ PFID(&op_data->op_fid2), (int)op_data->op_namelen,
op_data->op_name, PFID(&op_data->op_fid1));
op_data->op_fsuid = from_kuid(&init_user_ns, current_fsuid());
@@ -1916,7 +1916,8 @@ static int lmv_link(struct obd_export *exp, struct md_op_data *op_data,
}
static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen,
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen,
struct ptlrpc_request **request)
{
struct obd_device *obd = exp->exp_obd;
@@ -1927,9 +1928,9 @@ static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(oldlen != 0);
CDEBUG(D_INODE, "RENAME %.*s in "DFID":%d to %.*s in "DFID":%d\n",
- oldlen, old, PFID(&op_data->op_fid1),
+ (int)oldlen, old, PFID(&op_data->op_fid1),
op_data->op_mea1 ? op_data->op_mea1->lsm_md_stripe_count : 0,
- newlen, new, PFID(&op_data->op_fid2),
+ (int)newlen, new, PFID(&op_data->op_fid2),
op_data->op_mea2 ? op_data->op_mea2->lsm_md_stripe_count : 0);
rc = lmv_check_connect(obd);
@@ -2035,7 +2036,7 @@ static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
}
static int lmv_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request,
struct md_open_data **mod)
{
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_internal.h b/drivers/staging/lustre/lustre/mdc/mdc_internal.h
index c10441d..f446c1c 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_internal.h
+++ b/drivers/staging/lustre/lustre/mdc/mdc_internal.h
@@ -38,25 +38,26 @@
void lprocfs_mdc_init_vars(struct lprocfs_static_vars *lvars);
void mdc_pack_body(struct ptlrpc_request *req, const struct lu_fid *fid,
- __u64 valid, int ea_size, __u32 suppgid, int flags);
+ __u64 valid, size_t ea_size, __u32 suppgid, u32 flags);
void mdc_swap_layouts_pack(struct ptlrpc_request *req,
struct md_op_data *op_data);
-void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff, __u32 size,
+void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff, size_t size,
const struct lu_fid *fid);
-void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, int flags,
- struct md_op_data *data, int ea_size);
+void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, u32 flags,
+ struct md_op_data *data, size_t ea_size);
void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len);
+ void *ea, size_t ealen, void *ea2, size_t ea2len);
void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const void *data, int datalen, __u32 mode, __u32 uid,
- __u32 gid, cfs_cap_t capability, __u64 rdev);
+ const void *data, size_t datalen, umode_t mode, uid_t uid,
+ gid_t gid, cfs_cap_t capability, __u64 rdev);
void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- __u32 mode, __u64 rdev, __u64 flags, const void *data,
- int datalen);
+ umode_t mode, __u64 rdev, __u64 flags, const void *data,
+ size_t datalen);
void mdc_unlink_pack(struct ptlrpc_request *req, struct md_op_data *op_data);
void mdc_link_pack(struct ptlrpc_request *req, struct md_op_data *op_data);
void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen);
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen);
void mdc_close_pack(struct ptlrpc_request *req, struct md_op_data *op_data);
/* mdc/mdc_locks.c */
@@ -94,16 +95,17 @@ void mdc_commit_open(struct ptlrpc_request *req);
void mdc_replay_open(struct ptlrpc_request *req);
int mdc_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid, __u32 gid,
- cfs_cap_t capability, __u64 rdev,
+ const void *data, size_t datalen, umode_t mode, uid_t uid,
+ gid_t gid, cfs_cap_t capability, __u64 rdev,
struct ptlrpc_request **request);
int mdc_link(struct obd_export *exp, struct md_op_data *op_data,
struct ptlrpc_request **request);
int mdc_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen,
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen,
struct ptlrpc_request **request);
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request, struct md_open_data **mod);
int mdc_unlink(struct obd_export *exp, struct md_op_data *op_data,
struct ptlrpc_request **request);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_lib.c b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
index ecfe13e..82d1c4d 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_lib.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
@@ -58,7 +58,7 @@ void mdc_swap_layouts_pack(struct ptlrpc_request *req,
}
void mdc_pack_body(struct ptlrpc_request *req, const struct lu_fid *fid,
- __u64 valid, int ea_size, __u32 suppgid, int flags)
+ __u64 valid, size_t ea_size, __u32 suppgid, u32 flags)
{
struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
&RMF_MDT_BODY);
@@ -103,8 +103,8 @@ static void mdc_pack_name(struct ptlrpc_request *req,
LASSERT(cpy_len == name_len && lu_name_is_valid_2(buf, cpy_len));
}
-void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff,
- __u32 size, const struct lu_fid *fid)
+void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff, size_t size,
+ const struct lu_fid *fid)
{
struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
&RMF_MDT_BODY);
@@ -118,8 +118,8 @@ void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff,
/* packing of MDS records */
void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const void *data, int datalen, __u32 mode,
- __u32 uid, __u32 gid, cfs_cap_t cap_effective, __u64 rdev)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective, __u64 rdev)
{
struct mdt_rec_create *rec;
char *tmp;
@@ -153,7 +153,7 @@ void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
}
}
-static __u64 mds_pack_open_flags(__u64 flags, __u32 mode)
+static inline __u64 mds_pack_open_flags(__u64 flags)
{
__u64 cr_flags = (flags & (FMODE_READ | FMODE_WRITE |
MDS_OPEN_FL_INTERNAL));
@@ -182,8 +182,8 @@ static __u64 mds_pack_open_flags(__u64 flags, __u32 mode)
/* packing of MDS records */
void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- __u32 mode, __u64 rdev, __u64 flags, const void *lmm,
- int lmmlen)
+ umode_t mode, __u64 rdev, __u64 flags, const void *lmm,
+ size_t lmmlen)
{
struct mdt_rec_create *rec;
char *tmp;
@@ -201,7 +201,7 @@ void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
rec->cr_fid2 = op_data->op_fid2;
rec->cr_mode = mode;
- cr_flags = mds_pack_open_flags(flags, mode);
+ cr_flags = mds_pack_open_flags(flags);
rec->cr_rdev = rdev;
rec->cr_time = op_data->op_mod_time;
rec->cr_suppgid1 = op_data->op_suppgids[0];
@@ -307,7 +307,7 @@ static void mdc_ioepoch_pack(struct mdt_ioepoch *epoch,
}
void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len)
+ void *ea, size_t ealen, void *ea2, size_t ea2len)
{
struct mdt_rec_setattr *rec;
struct mdt_ioepoch *epoch;
@@ -388,7 +388,8 @@ void mdc_link_pack(struct ptlrpc_request *req, struct md_op_data *op_data)
}
void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen)
+ const char *old, size_t oldlen,
+ const char *new, size_t newlen)
{
struct mdt_rec_rename *rec;
@@ -416,8 +417,8 @@ void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
mdc_pack_name(req, &RMF_SYMTGT, new, newlen);
}
-void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, int flags,
- struct md_op_data *op_data, int ea_size)
+void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, u32 flags,
+ struct md_op_data *op_data, size_t ea_size)
{
struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
&RMF_MDT_BODY);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_locks.c b/drivers/staging/lustre/lustre/mdc/mdc_locks.c
index 54de46b..f1f6c08 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_locks.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_locks.c
@@ -235,7 +235,7 @@ mdc_intent_open_pack(struct obd_export *exp, struct lookup_intent *it,
struct obd_device *obddev = class_exp2obd(exp);
struct ldlm_intent *lit;
const void *lmm = op_data->op_data;
- int lmmsize = op_data->op_data_size;
+ u32 lmmsize = op_data->op_data_size;
LIST_HEAD(cancels);
int count = 0;
int mode;
@@ -317,7 +317,8 @@ mdc_intent_getxattr_pack(struct obd_export *exp,
{
struct ptlrpc_request *req;
struct ldlm_intent *lit;
- int rc, count = 0, maxdata;
+ int rc, count = 0;
+ u32 maxdata;
LIST_HEAD(cancels);
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
@@ -402,7 +403,7 @@ static struct ptlrpc_request *mdc_intent_getattr_pack(struct obd_export *exp,
OBD_MD_MEA | OBD_MD_FLACL;
struct ldlm_intent *lit;
int rc;
- int easize;
+ u32 easize;
req = ptlrpc_request_alloc(class_exp2cliimp(exp),
&RQF_LDLM_INTENT_GETATTR);
@@ -507,7 +508,7 @@ static int mdc_finish_enqueue(struct obd_export *exp,
struct ldlm_reply *lockrep;
struct ldlm_lock *lock;
void *lvb_data = NULL;
- int lvb_len = 0;
+ u32 lvb_len = 0;
LASSERT(rc >= 0);
/* Similarly, if we're going to replay this request, we don't want to
@@ -952,8 +953,8 @@ static int mdc_finish_intent_lock(struct obd_export *exp,
}
CDEBUG(D_DENTRY,
"D_IT dentry %.*s intent: %s status %d disp %x rc %d\n",
- op_data->op_namelen, op_data->op_name, ldlm_it2str(it->it_op),
- it->it_status, it->it_disposition, rc);
+ (int)op_data->op_namelen, op_data->op_name,
+ ldlm_it2str(it->it_op), it->it_status, it->it_disposition, rc);
return rc;
}
@@ -1067,7 +1068,7 @@ int mdc_intent_lock(struct obd_export *exp, struct md_op_data *op_data,
LASSERT(it);
CDEBUG(D_DLMTRACE, "(name: %.*s,"DFID") in obj "DFID
- ", intent: %s flags %#Lo\n", op_data->op_namelen,
+ ", intent: %s flags %#Lo\n", (int)op_data->op_namelen,
op_data->op_name, PFID(&op_data->op_fid2),
PFID(&op_data->op_fid1), ldlm_it2str(it->it_op),
it->it_flags);
@@ -1177,8 +1178,8 @@ int mdc_intent_getattr_async(struct obd_export *exp,
CDEBUG(D_DLMTRACE,
"name: %.*s in inode " DFID ", intent: %s flags %#Lo\n",
- op_data->op_namelen, op_data->op_name, PFID(&op_data->op_fid1),
- ldlm_it2str(it->it_op), it->it_flags);
+ (int)op_data->op_namelen, op_data->op_name,
+ PFID(&op_data->op_fid1), ldlm_it2str(it->it_op), it->it_flags);
fid_build_reg_res_name(&op_data->op_fid1, &res_id);
req = mdc_intent_getattr_pack(exp, it, op_data);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_reint.c b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
index af5c92c..c921e47 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_reint.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_reint.c
@@ -99,7 +99,7 @@ int mdc_resource_get_unused(struct obd_export *exp, const struct lu_fid *fid,
}
int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
- void *ea, int ealen, void *ea2, int ea2len,
+ void *ea, size_t ealen, void *ea2, size_t ea2len,
struct ptlrpc_request **request, struct md_open_data **mod)
{
LIST_HEAD(cancels);
@@ -196,9 +196,9 @@ int mdc_setattr(struct obd_export *exp, struct md_op_data *op_data,
}
int mdc_create(struct obd_export *exp, struct md_op_data *op_data,
- const void *data, int datalen, int mode, __u32 uid, __u32 gid,
- cfs_cap_t cap_effective, __u64 rdev,
- struct ptlrpc_request **request)
+ const void *data, size_t datalen, umode_t mode,
+ uid_t uid, gid_t gid, cfs_cap_t cap_effective,
+ __u64 rdev, struct ptlrpc_request **request)
{
struct ptlrpc_request *req;
int level, rc;
@@ -389,7 +389,7 @@ int mdc_link(struct obd_export *exp, struct md_op_data *op_data,
}
int mdc_rename(struct obd_export *exp, struct md_op_data *op_data,
- const char *old, int oldlen, const char *new, int newlen,
+ const char *old, size_t oldlen, const char *new, size_t newlen,
struct ptlrpc_request **request)
{
LIST_HEAD(cancels);
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 157fdc0..e622b7a 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -447,7 +447,7 @@ static int mdc_get_lustre_md(struct obd_export *exp,
if (rc < 0)
goto out;
- if (rc < sizeof(*md->lsm)) {
+ if (rc < (typeof(rc))sizeof(*md->lsm)) {
CDEBUG(D_INFO,
"lsm size too small: rc < sizeof (*md->lsm) (%d < %d)\n",
rc, (int)sizeof(*md->lsm));
@@ -485,7 +485,7 @@ static int mdc_get_lustre_md(struct obd_export *exp,
if (rc < 0)
goto out;
- if (rc < sizeof(*md->lmv)) {
+ if (rc < (typeof(rc))sizeof(*md->lmv)) {
CDEBUG(D_INFO,
"size too small: rc < sizeof(*md->lmv) (%d < %d)\n",
rc, (int)sizeof(*md->lmv));
@@ -1843,7 +1843,7 @@ out:
return rc;
}
-static struct kuc_hdr *changelog_kuc_hdr(char *buf, int len, int flags)
+static struct kuc_hdr *changelog_kuc_hdr(char *buf, size_t len, u32 flags)
{
struct kuc_hdr *lh = (struct kuc_hdr *)buf;
@@ -1876,7 +1876,8 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
struct changelog_show *cs = data;
struct llog_changelog_rec *rec = (struct llog_changelog_rec *)hdr;
struct kuc_hdr *lh;
- int len, rc;
+ size_t len;
+ int rc;
if (rec->cr_hdr.lrh_type != CHANGELOG_REC) {
rc = -EINVAL;
@@ -1907,7 +1908,7 @@ static int changelog_kkuc_cb(const struct lu_env *env, struct llog_handle *llh,
memcpy(lh + 1, &rec->cr, len - sizeof(*lh));
rc = libcfs_kkuc_msg_put(cs->cs_fp, lh);
- CDEBUG(D_HSM, "kucmsg fp %p len %d rc %d\n", cs->cs_fp, len, rc);
+ CDEBUG(D_HSM, "kucmsg fp %p len %zu rc %d\n", cs->cs_fp, len, rc);
return rc;
}
@@ -2369,7 +2370,7 @@ static void lustre_swab_hai(struct hsm_action_item *h)
static void lustre_swab_hal(struct hsm_action_list *h)
{
struct hsm_action_item *hai;
- int i;
+ u32 i;
__swab32s(&h->hal_version);
__swab32s(&h->hal_count);
@@ -2418,14 +2419,14 @@ static int mdc_ioc_hsm_ct_start(struct obd_export *exp,
* @param val KUC message (kuc_hdr + hsm_action_list)
* @param len total length of message
*/
-static int mdc_hsm_copytool_send(int len, void *val)
+static int mdc_hsm_copytool_send(size_t len, void *val)
{
struct kuc_hdr *lh = (struct kuc_hdr *)val;
struct hsm_action_list *hal = (struct hsm_action_list *)(lh + 1);
if (len < sizeof(*lh) + sizeof(*hal)) {
- CERROR("Short HSM message %d < %d\n", len,
- (int)(sizeof(*lh) + sizeof(*hal)));
+ CERROR("Short HSM message %zu < %zu\n", len,
+ sizeof(*lh) + sizeof(*hal));
return -EPROTO;
}
if (lh->kuc_magic == __swab16(KUC_MAGIC)) {
@@ -2528,18 +2529,18 @@ static int mdc_get_info(const struct lu_env *env, struct obd_export *exp,
int rc = -EINVAL;
if (KEY_IS(KEY_MAX_EASIZE)) {
- int mdsize, *max_easize;
+ u32 mdsize, *max_easize;
if (*vallen != sizeof(int))
return -EINVAL;
- mdsize = *(int *)val;
+ mdsize = *(u32 *)val;
if (mdsize > exp->exp_obd->u.cli.cl_max_mds_easize)
exp->exp_obd->u.cli.cl_max_mds_easize = mdsize;
max_easize = val;
*max_easize = exp->exp_obd->u.cli.cl_max_mds_easize;
return 0;
} else if (KEY_IS(KEY_DEFAULT_EASIZE)) {
- int *default_easize;
+ u32 *default_easize;
if (*vallen != sizeof(int))
return -EINVAL;
@@ -2556,7 +2557,7 @@ static int mdc_get_info(const struct lu_env *env, struct obd_export *exp,
*data = imp->imp_connect_data;
return 0;
} else if (KEY_IS(KEY_TGT_COUNT)) {
- *((int *)val) = 1;
+ *((u32 *)val) = 1;
return 0;
}
@@ -2784,8 +2785,8 @@ err_rpc_lock:
* a large number of stripes is possible. If a larger reply buffer is
* required it will be reallocated in the ptlrpc layer due to overflow.
*/
-static int mdc_init_ea_size(struct obd_export *exp, int easize,
- int def_easize, int cookiesize, int def_cookiesize)
+static int mdc_init_ea_size(struct obd_export *exp, u32 easize, u32 def_easize,
+ u32 cookiesize, u32 def_cookiesize)
{
struct obd_device *obd = exp->exp_obd;
struct client_obd *cli = &obd->u.cli;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 061/124] staging: lustre: obd: cleanup struct md_op_data and uses
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Make the following changes in or around struct md_op_data:
* rename enum op_cli_flags to enum md_cli_flags.
* Change to type of the op_flags member from __u32 to enum
md_op_flags.
* Remove the used but never set member op_npages.
* Remove the set but never used member op_stripe_offset (an alias for
op_ioepoch).
* Remove the op_max_pages alias for op_valid and add a op_max_pages
member.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11734
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 23 ++++++++++-------------
drivers/staging/lustre/lustre/mdc/mdc_request.c | 3 +--
2 files changed, 11 insertions(+), 15 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 8a7c803..2d25e40 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -736,6 +736,14 @@ static inline int it_to_lock_mode(struct lookup_intent *it)
return -EINVAL;
}
+enum md_cli_flags {
+ CLI_SET_MEA = BIT(0),
+ CLI_RM_ENTRY = BIT(1),
+ CLI_HASH64 = BIT(2),
+ CLI_API32 = BIT(3),
+ CLI_MIGRATE = BIT(4),
+};
+
struct md_op_data {
struct lu_fid op_fid1; /* operation fid1 (usually parent) */
struct lu_fid op_fid2; /* operation fid2 (usually child) */
@@ -773,35 +781,24 @@ struct md_op_data {
__u64 op_offset;
/* Used by readdir */
- __u32 op_npages;
+ __u32 op_max_pages;
/* used to transfer info between the stacks of MD client
* see enum op_cli_flags
*/
- __u32 op_cli_flags;
+ enum md_cli_flags op_cli_flags;
/* File object data version for HSM release, on client */
__u64 op_data_version;
struct lustre_handle op_lease_handle;
};
-#define op_stripe_offset op_ioepoch
-#define op_max_pages op_valid
-
struct md_callback {
int (*md_blocking_ast)(struct ldlm_lock *lock,
struct ldlm_lock_desc *desc,
void *data, int flag);
};
-enum op_cli_flags {
- CLI_SET_MEA = 1 << 0,
- CLI_RM_ENTRY = 1 << 1,
- CLI_HASH64 = BIT(2),
- CLI_API32 = BIT(3),
- CLI_MIGRATE = BIT(4),
-};
-
struct md_enqueue_info;
/* metadata stat-ahead */
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 8c2f6d6..166f0c4 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1264,8 +1264,7 @@ static int mdc_read_page_remote(void *data, struct page *page0)
lu_pgs >>= LU_PAGE_SHIFT;
LASSERT(!(req->rq_bulk->bd_nob_transferred & ~LU_PAGE_MASK));
- CDEBUG(D_INODE, "read %d(%d)/%d pages\n", rd_pgs, lu_pgs,
- op_data->op_npages);
+ CDEBUG(D_INODE, "read %d(%d) pages\n", rd_pgs, lu_pgs);
mdc_adjust_dirpages(page_pool, rd_pgs, lu_pgs);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 061/124] staging: lustre: obd: cleanup struct md_op_data and uses
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Make the following changes in or around struct md_op_data:
* rename enum op_cli_flags to enum md_cli_flags.
* Change to type of the op_flags member from __u32 to enum
md_op_flags.
* Remove the used but never set member op_npages.
* Remove the set but never used member op_stripe_offset (an alias for
op_ioepoch).
* Remove the op_max_pages alias for op_valid and add a op_max_pages
member.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/11734
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 23 ++++++++++-------------
drivers/staging/lustre/lustre/mdc/mdc_request.c | 3 +--
2 files changed, 11 insertions(+), 15 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 8a7c803..2d25e40 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -736,6 +736,14 @@ static inline int it_to_lock_mode(struct lookup_intent *it)
return -EINVAL;
}
+enum md_cli_flags {
+ CLI_SET_MEA = BIT(0),
+ CLI_RM_ENTRY = BIT(1),
+ CLI_HASH64 = BIT(2),
+ CLI_API32 = BIT(3),
+ CLI_MIGRATE = BIT(4),
+};
+
struct md_op_data {
struct lu_fid op_fid1; /* operation fid1 (usually parent) */
struct lu_fid op_fid2; /* operation fid2 (usually child) */
@@ -773,35 +781,24 @@ struct md_op_data {
__u64 op_offset;
/* Used by readdir */
- __u32 op_npages;
+ __u32 op_max_pages;
/* used to transfer info between the stacks of MD client
* see enum op_cli_flags
*/
- __u32 op_cli_flags;
+ enum md_cli_flags op_cli_flags;
/* File object data version for HSM release, on client */
__u64 op_data_version;
struct lustre_handle op_lease_handle;
};
-#define op_stripe_offset op_ioepoch
-#define op_max_pages op_valid
-
struct md_callback {
int (*md_blocking_ast)(struct ldlm_lock *lock,
struct ldlm_lock_desc *desc,
void *data, int flag);
};
-enum op_cli_flags {
- CLI_SET_MEA = 1 << 0,
- CLI_RM_ENTRY = 1 << 1,
- CLI_HASH64 = BIT(2),
- CLI_API32 = BIT(3),
- CLI_MIGRATE = BIT(4),
-};
-
struct md_enqueue_info;
/* metadata stat-ahead */
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 8c2f6d6..166f0c4 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1264,8 +1264,7 @@ static int mdc_read_page_remote(void *data, struct page *page0)
lu_pgs >>= LU_PAGE_SHIFT;
LASSERT(!(req->rq_bulk->bd_nob_transferred & ~LU_PAGE_MASK));
- CDEBUG(D_INODE, "read %d(%d)/%d pages\n", rd_pgs, lu_pgs,
- op_data->op_npages);
+ CDEBUG(D_INODE, "read %d(%d) pages\n", rd_pgs, lu_pgs);
mdc_adjust_dirpages(page_pool, rd_pgs, lu_pgs);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 062/124] staging: lustre: replace direct HZ access with kernel APIs
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Jian Yu,
James Simmons
From: Jian Yu <jian.yu@intel.com>
On some customer's systems, kernel was compiled with HZ defined to
100, instead of 1000. This improves performance for HPC applications.
However, to use these systems with Lustre, customers have to re-build
Lustre for the kernel because Lustre directly uses the defined
constant HZ.
Since kernel 2.6.21, some non-HZ dependent timing APIs become non-
inline functions, which can be used in Lustre codes to replace the
direct HZ access.
These kernel APIs include:
jiffies_to_msecs()
jiffies_to_usecs()
jiffies_to_timespec()
msecs_to_jiffies()
usecs_to_jiffies()
timespec_to_jiffies()
And here are some samples of the replacement:
HZ -> msecs_to_jiffies(MSEC_PER_SEC)
n * HZ -> msecs_to_jiffies(n * MSEC_PER_SEC)
HZ / n -> msecs_to_jiffies(MSEC_PER_SEC / n)
n / HZ -> jiffies_to_msecs(n) / MSEC_PER_SEC
n / HZ * 1000 -> jiffies_to_msecs(n)
This patch replaces the direct HZ access in lustre modules.
Signed-off-by: Jian Yu <jian.yu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5443
Reviewed-on: http://review.whamcloud.com/12052
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_resource.c | 3 ++-
drivers/staging/lustre/lustre/llite/statahead.c | 3 ++-
drivers/staging/lustre/lustre/mgc/mgc_request.c | 5 +++--
drivers/staging/lustre/lustre/ptlrpc/sec.c | 7 ++++---
drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c | 9 +++++----
drivers/staging/lustre/lustre/ptlrpc/sec_gc.c | 3 ++-
6 files changed, 18 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
index 912cd68..a09c25a 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
@@ -880,7 +880,8 @@ static int __ldlm_namespace_free(struct ldlm_namespace *ns, int force)
ldlm_ns_name(ns), atomic_read(&ns->ns_bref));
force_wait:
if (force)
- lwi = LWI_TIMEOUT(obd_timeout * HZ / 4, NULL, NULL);
+ lwi = LWI_TIMEOUT(msecs_to_jiffies(obd_timeout *
+ MSEC_PER_SEC) / 4, NULL, NULL);
rc = l_wait_event(ns->ns_waitq,
atomic_read(&ns->ns_bref) == 0, &lwi);
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 70ca252..323a175 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -1154,7 +1154,8 @@ out:
*/
while (sai->sai_sent != sai->sai_replied) {
/* in case we're not woken up, timeout wait */
- lwi = LWI_TIMEOUT(HZ >> 3, NULL, NULL);
+ lwi = LWI_TIMEOUT(msecs_to_jiffies(MSEC_PER_SEC >> 3),
+ NULL, NULL);
l_wait_event(thread->t_ctl_waitq,
sai->sai_sent == sai->sai_replied, &lwi);
}
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index f3d4f7f..88d4d10 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -549,8 +549,9 @@ static int mgc_requeue_thread(void *data)
* caused the lock revocation to finish its setup, plus some
* random so everyone doesn't try to reconnect at once.
*/
- to = MGC_TIMEOUT_MIN_SECONDS * HZ;
- to += rand * HZ / 100; /* rand is centi-seconds */
+ to = msecs_to_jiffies(MGC_TIMEOUT_MIN_SECONDS * MSEC_PER_SEC);
+ /* rand is centi-seconds */
+ to += msecs_to_jiffies(rand * MSEC_PER_SEC / 100);
lwi = LWI_TIMEOUT(to, NULL, NULL);
l_wait_event(rq_waitq, rq_state & (RQ_STOP | RQ_PRECLEANUP),
&lwi);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec.c b/drivers/staging/lustre/lustre/ptlrpc/sec.c
index dbd819f..523579f 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec.c
@@ -499,7 +499,7 @@ static int sptlrpc_req_replace_dead_ctx(struct ptlrpc_request *req)
newctx, newctx->cc_flags);
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(HZ);
+ schedule_timeout(msecs_to_jiffies(MSEC_PER_SEC));
} else {
/*
* it's possible newctx == oldctx if we're switching
@@ -718,8 +718,9 @@ again:
req->rq_restart = 0;
spin_unlock(&req->rq_lock);
- lwi = LWI_TIMEOUT_INTR(timeout * HZ, ctx_refresh_timeout,
- ctx_refresh_interrupt, req);
+ lwi = LWI_TIMEOUT_INTR(msecs_to_jiffies(timeout * MSEC_PER_SEC),
+ ctx_refresh_timeout, ctx_refresh_interrupt,
+ req);
rc = l_wait_event(req->rq_reply_waitq, ctx_check_refresh(ctx), &lwi);
/*
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
index fdb32d5..6adf274 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
@@ -139,7 +139,7 @@ int sptlrpc_proc_enc_pool_seq_show(struct seq_file *m, void *v)
"cache missing: %lu\n"
"low free mark: %lu\n"
"max waitqueue depth: %u\n"
- "max wait time: %ld/%u\n",
+ "max wait time: %ld/%lu\n",
totalram_pages,
PAGES_PER_POOL,
page_pools.epp_max_pages,
@@ -158,7 +158,7 @@ int sptlrpc_proc_enc_pool_seq_show(struct seq_file *m, void *v)
page_pools.epp_st_lowfree,
page_pools.epp_st_max_wqlen,
page_pools.epp_st_max_wait,
- HZ);
+ msecs_to_jiffies(MSEC_PER_SEC));
spin_unlock(&page_pools.epp_lock);
@@ -432,12 +432,13 @@ void sptlrpc_enc_pool_fini(void)
if (page_pools.epp_st_access > 0) {
CDEBUG(D_SEC,
- "max pages %lu, grows %u, grow fails %u, shrinks %u, access %lu, missing %lu, max qlen %u, max wait %ld/%d\n",
+ "max pages %lu, grows %u, grow fails %u, shrinks %u, access %lu, missing %lu, max qlen %u, max wait %ld/%ld\n",
page_pools.epp_st_max_pages, page_pools.epp_st_grows,
page_pools.epp_st_grow_fails,
page_pools.epp_st_shrinks, page_pools.epp_st_access,
page_pools.epp_st_missings, page_pools.epp_st_max_wqlen,
- page_pools.epp_st_max_wait, HZ);
+ page_pools.epp_st_max_wait,
+ msecs_to_jiffies(MSEC_PER_SEC));
}
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
index 9b9801e..f79a889 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
@@ -182,7 +182,8 @@ again:
/* check ctx list again before sleep */
sec_process_ctx_list();
- lwi = LWI_TIMEOUT(SEC_GC_INTERVAL * HZ, NULL, NULL);
+ lwi = LWI_TIMEOUT(msecs_to_jiffies(SEC_GC_INTERVAL * MSEC_PER_SEC),
+ NULL, NULL);
l_wait_event(thread->t_ctl_waitq,
thread_is_stopping(thread) ||
thread_is_signal(thread),
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 062/124] staging: lustre: replace direct HZ access with kernel APIs
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Jian Yu,
James Simmons
From: Jian Yu <jian.yu@intel.com>
On some customer's systems, kernel was compiled with HZ defined to
100, instead of 1000. This improves performance for HPC applications.
However, to use these systems with Lustre, customers have to re-build
Lustre for the kernel because Lustre directly uses the defined
constant HZ.
Since kernel 2.6.21, some non-HZ dependent timing APIs become non-
inline functions, which can be used in Lustre codes to replace the
direct HZ access.
These kernel APIs include:
jiffies_to_msecs()
jiffies_to_usecs()
jiffies_to_timespec()
msecs_to_jiffies()
usecs_to_jiffies()
timespec_to_jiffies()
And here are some samples of the replacement:
HZ -> msecs_to_jiffies(MSEC_PER_SEC)
n * HZ -> msecs_to_jiffies(n * MSEC_PER_SEC)
HZ / n -> msecs_to_jiffies(MSEC_PER_SEC / n)
n / HZ -> jiffies_to_msecs(n) / MSEC_PER_SEC
n / HZ * 1000 -> jiffies_to_msecs(n)
This patch replaces the direct HZ access in lustre modules.
Signed-off-by: Jian Yu <jian.yu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5443
Reviewed-on: http://review.whamcloud.com/12052
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_resource.c | 3 ++-
drivers/staging/lustre/lustre/llite/statahead.c | 3 ++-
drivers/staging/lustre/lustre/mgc/mgc_request.c | 5 +++--
drivers/staging/lustre/lustre/ptlrpc/sec.c | 7 ++++---
drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c | 9 +++++----
drivers/staging/lustre/lustre/ptlrpc/sec_gc.c | 3 ++-
6 files changed, 18 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
index 912cd68..a09c25a 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c
@@ -880,7 +880,8 @@ static int __ldlm_namespace_free(struct ldlm_namespace *ns, int force)
ldlm_ns_name(ns), atomic_read(&ns->ns_bref));
force_wait:
if (force)
- lwi = LWI_TIMEOUT(obd_timeout * HZ / 4, NULL, NULL);
+ lwi = LWI_TIMEOUT(msecs_to_jiffies(obd_timeout *
+ MSEC_PER_SEC) / 4, NULL, NULL);
rc = l_wait_event(ns->ns_waitq,
atomic_read(&ns->ns_bref) == 0, &lwi);
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 70ca252..323a175 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -1154,7 +1154,8 @@ out:
*/
while (sai->sai_sent != sai->sai_replied) {
/* in case we're not woken up, timeout wait */
- lwi = LWI_TIMEOUT(HZ >> 3, NULL, NULL);
+ lwi = LWI_TIMEOUT(msecs_to_jiffies(MSEC_PER_SEC >> 3),
+ NULL, NULL);
l_wait_event(thread->t_ctl_waitq,
sai->sai_sent == sai->sai_replied, &lwi);
}
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index f3d4f7f..88d4d10 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -549,8 +549,9 @@ static int mgc_requeue_thread(void *data)
* caused the lock revocation to finish its setup, plus some
* random so everyone doesn't try to reconnect at once.
*/
- to = MGC_TIMEOUT_MIN_SECONDS * HZ;
- to += rand * HZ / 100; /* rand is centi-seconds */
+ to = msecs_to_jiffies(MGC_TIMEOUT_MIN_SECONDS * MSEC_PER_SEC);
+ /* rand is centi-seconds */
+ to += msecs_to_jiffies(rand * MSEC_PER_SEC / 100);
lwi = LWI_TIMEOUT(to, NULL, NULL);
l_wait_event(rq_waitq, rq_state & (RQ_STOP | RQ_PRECLEANUP),
&lwi);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec.c b/drivers/staging/lustre/lustre/ptlrpc/sec.c
index dbd819f..523579f 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec.c
@@ -499,7 +499,7 @@ static int sptlrpc_req_replace_dead_ctx(struct ptlrpc_request *req)
newctx, newctx->cc_flags);
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(HZ);
+ schedule_timeout(msecs_to_jiffies(MSEC_PER_SEC));
} else {
/*
* it's possible newctx == oldctx if we're switching
@@ -718,8 +718,9 @@ again:
req->rq_restart = 0;
spin_unlock(&req->rq_lock);
- lwi = LWI_TIMEOUT_INTR(timeout * HZ, ctx_refresh_timeout,
- ctx_refresh_interrupt, req);
+ lwi = LWI_TIMEOUT_INTR(msecs_to_jiffies(timeout * MSEC_PER_SEC),
+ ctx_refresh_timeout, ctx_refresh_interrupt,
+ req);
rc = l_wait_event(req->rq_reply_waitq, ctx_check_refresh(ctx), &lwi);
/*
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
index fdb32d5..6adf274 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
@@ -139,7 +139,7 @@ int sptlrpc_proc_enc_pool_seq_show(struct seq_file *m, void *v)
"cache missing: %lu\n"
"low free mark: %lu\n"
"max waitqueue depth: %u\n"
- "max wait time: %ld/%u\n",
+ "max wait time: %ld/%lu\n",
totalram_pages,
PAGES_PER_POOL,
page_pools.epp_max_pages,
@@ -158,7 +158,7 @@ int sptlrpc_proc_enc_pool_seq_show(struct seq_file *m, void *v)
page_pools.epp_st_lowfree,
page_pools.epp_st_max_wqlen,
page_pools.epp_st_max_wait,
- HZ);
+ msecs_to_jiffies(MSEC_PER_SEC));
spin_unlock(&page_pools.epp_lock);
@@ -432,12 +432,13 @@ void sptlrpc_enc_pool_fini(void)
if (page_pools.epp_st_access > 0) {
CDEBUG(D_SEC,
- "max pages %lu, grows %u, grow fails %u, shrinks %u, access %lu, missing %lu, max qlen %u, max wait %ld/%d\n",
+ "max pages %lu, grows %u, grow fails %u, shrinks %u, access %lu, missing %lu, max qlen %u, max wait %ld/%ld\n",
page_pools.epp_st_max_pages, page_pools.epp_st_grows,
page_pools.epp_st_grow_fails,
page_pools.epp_st_shrinks, page_pools.epp_st_access,
page_pools.epp_st_missings, page_pools.epp_st_max_wqlen,
- page_pools.epp_st_max_wait, HZ);
+ page_pools.epp_st_max_wait,
+ msecs_to_jiffies(MSEC_PER_SEC));
}
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
index 9b9801e..f79a889 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
@@ -182,7 +182,8 @@ again:
/* check ctx list again before sleep */
sec_process_ctx_list();
- lwi = LWI_TIMEOUT(SEC_GC_INTERVAL * HZ, NULL, NULL);
+ lwi = LWI_TIMEOUT(msecs_to_jiffies(SEC_GC_INTERVAL * MSEC_PER_SEC),
+ NULL, NULL);
l_wait_event(thread->t_ctl_waitq,
thread_is_stopping(thread) ||
thread_is_signal(thread),
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 063/124] staging: lustre: ldlm: count of pools is unsigned long
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Function ldlm_pools_count() return unsigned long but counter is int.
Use ldlm_pool_granted() everywhere.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12304
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
index 2fc319e..1147b30 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
@@ -792,7 +792,8 @@ static struct completion ldlm_pools_comp;
*/
static unsigned long ldlm_pools_count(ldlm_side_t client, gfp_t gfp_mask)
{
- int total = 0, nr_ns;
+ unsigned long total = 0;
+ int nr_ns;
struct ldlm_namespace *ns;
struct ldlm_namespace *ns_old = NULL; /* loop detection */
void *cookie;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 063/124] staging: lustre: ldlm: count of pools is unsigned long
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Function ldlm_pools_count() return unsigned long but counter is int.
Use ldlm_pool_granted() everywhere.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12304
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
index 2fc319e..1147b30 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
@@ -792,7 +792,8 @@ static struct completion ldlm_pools_comp;
*/
static unsigned long ldlm_pools_count(ldlm_side_t client, gfp_t gfp_mask)
{
- int total = 0, nr_ns;
+ unsigned long total = 0;
+ int nr_ns;
struct ldlm_namespace *ns;
struct ldlm_namespace *ns_old = NULL; /* loop detection */
void *cookie;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 064/124] staging: lustre: lu_dirent_calc_size() return type to size_t
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12383
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 15 +++------------
1 files changed, 3 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index a0ca571..445b813 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1013,12 +1013,12 @@ static inline struct lu_dirent *lu_dirent_next(struct lu_dirent *ent)
return next;
}
-static inline int lu_dirent_calc_size(int namelen, __u16 attr)
+static inline size_t lu_dirent_calc_size(size_t namelen, __u16 attr)
{
- int size;
+ size_t size;
if (attr & LUDA_TYPE) {
- const unsigned align = sizeof(struct luda_type) - 1;
+ const size_t align = sizeof(struct luda_type) - 1;
size = (sizeof(struct lu_dirent) + namelen + align) & ~align;
size += sizeof(struct luda_type);
@@ -1029,15 +1029,6 @@ static inline int lu_dirent_calc_size(int namelen, __u16 attr)
return (size + 7) & ~7;
}
-static inline int lu_dirent_size(const struct lu_dirent *ent)
-{
- if (le16_to_cpu(ent->lde_reclen) == 0) {
- return lu_dirent_calc_size(le16_to_cpu(ent->lde_namelen),
- le32_to_cpu(ent->lde_attrs));
- }
- return le16_to_cpu(ent->lde_reclen);
-}
-
#define MDS_DIR_END_OFF 0xfffffffffffffffeULL
/**
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 064/124] staging: lustre: lu_dirent_calc_size() return type to size_t
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12383
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 15 +++------------
1 files changed, 3 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index a0ca571..445b813 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1013,12 +1013,12 @@ static inline struct lu_dirent *lu_dirent_next(struct lu_dirent *ent)
return next;
}
-static inline int lu_dirent_calc_size(int namelen, __u16 attr)
+static inline size_t lu_dirent_calc_size(size_t namelen, __u16 attr)
{
- int size;
+ size_t size;
if (attr & LUDA_TYPE) {
- const unsigned align = sizeof(struct luda_type) - 1;
+ const size_t align = sizeof(struct luda_type) - 1;
size = (sizeof(struct lu_dirent) + namelen + align) & ~align;
size += sizeof(struct luda_type);
@@ -1029,15 +1029,6 @@ static inline int lu_dirent_calc_size(int namelen, __u16 attr)
return (size + 7) & ~7;
}
-static inline int lu_dirent_size(const struct lu_dirent *ent)
-{
- if (le16_to_cpu(ent->lde_reclen) == 0) {
- return lu_dirent_calc_size(le16_to_cpu(ent->lde_namelen),
- le32_to_cpu(ent->lde_attrs));
- }
- return le16_to_cpu(ent->lde_reclen);
-}
-
#define MDS_DIR_END_OFF 0xfffffffffffffffeULL
/**
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 065/124] staging: lustre: obdclass: change lu_site->ls_purge_start to unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12384
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 2 +-
drivers/staging/lustre/lustre/obdclass/lu_object.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index 7b37cae..260643e 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -602,7 +602,7 @@ struct lu_site {
/**
* index of bucket on hash table while purging
*/
- int ls_purge_start;
+ unsigned int ls_purge_start;
/**
* Top-level device for this stack.
*/
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index e284e5d..4395117 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -338,7 +338,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
struct cfs_hash_bd bd2;
struct list_head dispose;
int did_sth;
- int start;
+ unsigned int start;
int count;
int bnr;
unsigned int i;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 065/124] staging: lustre: obdclass: change lu_site->ls_purge_start to unsigned
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12384
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lu_object.h | 2 +-
drivers/staging/lustre/lustre/obdclass/lu_object.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lu_object.h b/drivers/staging/lustre/lustre/include/lu_object.h
index 7b37cae..260643e 100644
--- a/drivers/staging/lustre/lustre/include/lu_object.h
+++ b/drivers/staging/lustre/lustre/include/lu_object.h
@@ -602,7 +602,7 @@ struct lu_site {
/**
* index of bucket on hash table while purging
*/
- int ls_purge_start;
+ unsigned int ls_purge_start;
/**
* Top-level device for this stack.
*/
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index e284e5d..4395117 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -338,7 +338,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
struct cfs_hash_bd bd2;
struct list_head dispose;
int did_sth;
- int start;
+ unsigned int start;
int count;
int bnr;
unsigned int i;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 066/124] staging: lustre: lov: remove LL_IOC_RECREATE_{FID,OBJ}
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove the obsolete ioctls LL_IOC_RECREATE_FID and LL_IOC_RECREATE_OBJ
along with their handlers in llite. Remove the then unused OBD method
lov_create(). Remove OBD_FL_RECREATE_OBJS handling from osc_create().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12442
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 4 +-
drivers/staging/lustre/lustre/llite/file.c | 86 -------------------
drivers/staging/lustre/lustre/lov/lov_obd.c | 87 --------------------
drivers/staging/lustre/lustre/osc/osc_request.c | 5 -
4 files changed, 2 insertions(+), 180 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 3ef5db0..de52339 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -202,8 +202,8 @@ struct ost_id {
#define LL_IOC_LOV_SETSTRIPE _IOW('f', 154, long)
#define LL_IOC_LOV_GETSTRIPE _IOW('f', 155, long)
#define LL_IOC_LOV_SETEA _IOW('f', 156, long)
-#define LL_IOC_RECREATE_OBJ _IOW('f', 157, long)
-#define LL_IOC_RECREATE_FID _IOW('f', 157, struct lu_fid)
+/* LL_IOC_RECREATE_OBJ 157 obsolete */
+/* LL_IOC_RECREATE_FID 158 obsolete */
#define LL_IOC_GROUP_LOCK _IOW('f', 158, long)
#define LL_IOC_GROUP_UNLOCK _IOW('f', 159, long)
/* #define LL_IOC_QUOTACHECK 160 OBD_IOC_QUOTACHECK */
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index fc81551..4b10260 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1302,88 +1302,6 @@ static ssize_t ll_file_splice_read(struct file *in_file, loff_t *ppos,
return result;
}
-static int ll_lov_recreate(struct inode *inode, struct ost_id *oi, u32 ost_idx)
-{
- struct obd_export *exp = ll_i2dtexp(inode);
- struct obd_trans_info oti = { 0 };
- struct obdo *oa = NULL;
- int lsm_size;
- int rc = 0;
- struct lov_stripe_md *lsm = NULL, *lsm2;
-
- oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
- if (!oa)
- return -ENOMEM;
-
- lsm = ccc_inode_lsm_get(inode);
- if (!lsm_has_objects(lsm)) {
- rc = -ENOENT;
- goto out;
- }
-
- lsm_size = sizeof(*lsm) + (sizeof(struct lov_oinfo) *
- (lsm->lsm_stripe_count));
-
- lsm2 = libcfs_kvzalloc(lsm_size, GFP_NOFS);
- if (!lsm2) {
- rc = -ENOMEM;
- goto out;
- }
-
- oa->o_oi = *oi;
- oa->o_nlink = ost_idx;
- oa->o_flags |= OBD_FL_RECREATE_OBJS;
- oa->o_valid = OBD_MD_FLID | OBD_MD_FLFLAGS | OBD_MD_FLGROUP;
- obdo_from_inode(oa, inode, OBD_MD_FLTYPE | OBD_MD_FLATIME |
- OBD_MD_FLMTIME | OBD_MD_FLCTIME);
- obdo_set_parent_fid(oa, &ll_i2info(inode)->lli_fid);
- memcpy(lsm2, lsm, lsm_size);
- ll_inode_size_lock(inode);
- rc = obd_create(NULL, exp, oa, &lsm2, &oti);
- ll_inode_size_unlock(inode);
-
- kvfree(lsm2);
- goto out;
-out:
- ccc_inode_lsm_put(inode, lsm);
- kmem_cache_free(obdo_cachep, oa);
- return rc;
-}
-
-static int ll_lov_recreate_obj(struct inode *inode, unsigned long arg)
-{
- struct ll_recreate_obj ucreat;
- struct ost_id oi;
-
- if (!capable(CFS_CAP_SYS_ADMIN))
- return -EPERM;
-
- if (copy_from_user(&ucreat, (struct ll_recreate_obj __user *)arg,
- sizeof(ucreat)))
- return -EFAULT;
-
- ostid_set_seq_mdt0(&oi);
- ostid_set_id(&oi, ucreat.lrc_id);
- return ll_lov_recreate(inode, &oi, ucreat.lrc_ost_idx);
-}
-
-static int ll_lov_recreate_fid(struct inode *inode, unsigned long arg)
-{
- struct lu_fid fid;
- struct ost_id oi;
- u32 ost_idx;
-
- if (!capable(CFS_CAP_SYS_ADMIN))
- return -EPERM;
-
- if (copy_from_user(&fid, (struct lu_fid __user *)arg, sizeof(fid)))
- return -EFAULT;
-
- fid_to_ostid(&fid, &oi);
- ost_idx = (fid_seq(&fid) >> 16) & 0xffff;
- return ll_lov_recreate(inode, &oi, ost_idx);
-}
-
int ll_lov_setstripe_ea_info(struct inode *inode, struct dentry *dentry,
__u64 flags, struct lov_user_md *lum,
int lum_size)
@@ -2322,10 +2240,6 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
}
case LL_IOC_LOV_GETSTRIPE:
return ll_lov_getstripe(inode, arg);
- case LL_IOC_RECREATE_OBJ:
- return ll_lov_recreate_obj(inode, arg);
- case LL_IOC_RECREATE_FID:
- return ll_lov_recreate_fid(inode, arg);
case FSFILT_IOC_FIEMAP:
return ll_ioctl_fiemap(inode, arg);
case FSFILT_IOC_GETFLAGS:
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 265be08..a54cd70 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -972,92 +972,6 @@ out:
return rc;
}
-static int lov_recreate(struct obd_export *exp, struct obdo *src_oa,
- struct lov_stripe_md **ea, struct obd_trans_info *oti)
-{
- struct lov_stripe_md *obj_mdp, *lsm;
- struct lov_obd *lov = &exp->exp_obd->u.lov;
- unsigned ost_idx;
- int rc, i;
-
- LASSERT(src_oa->o_valid & OBD_MD_FLFLAGS &&
- src_oa->o_flags & OBD_FL_RECREATE_OBJS);
-
- obj_mdp = kzalloc(sizeof(*obj_mdp), GFP_NOFS);
- if (!obj_mdp)
- return -ENOMEM;
-
- ost_idx = src_oa->o_nlink;
- lsm = *ea;
- if (!lsm) {
- rc = -EINVAL;
- goto out;
- }
- if (ost_idx >= lov->desc.ld_tgt_count ||
- !lov->lov_tgts[ost_idx]) {
- rc = -EINVAL;
- goto out;
- }
-
- for (i = 0; i < lsm->lsm_stripe_count; i++) {
- struct lov_oinfo *loi = lsm->lsm_oinfo[i];
-
- if (lov_oinfo_is_dummy(loi))
- continue;
-
- if (loi->loi_ost_idx == ost_idx) {
- if (ostid_id(&loi->loi_oi) != ostid_id(&src_oa->o_oi)) {
- rc = -EINVAL;
- goto out;
- }
- break;
- }
- }
- if (i == lsm->lsm_stripe_count) {
- rc = -EINVAL;
- goto out;
- }
-
- rc = obd_create(NULL, lov->lov_tgts[ost_idx]->ltd_exp,
- src_oa, &obj_mdp, oti);
-out:
- kfree(obj_mdp);
- return rc;
-}
-
-/* the LOV expects oa->o_id to be set to the LOV object id */
-static int lov_create(const struct lu_env *env, struct obd_export *exp,
- struct obdo *src_oa, struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
-{
- struct lov_obd *lov;
- int rc = 0;
-
- LASSERT(ea);
- if (!exp)
- return -EINVAL;
-
- if ((src_oa->o_valid & OBD_MD_FLFLAGS) &&
- src_oa->o_flags == OBD_FL_DELORPHAN) {
- /* should be used with LOV anymore */
- LBUG();
- }
-
- lov = &exp->exp_obd->u.lov;
- if (!lov->desc.ld_active_tgt_count)
- return -EIO;
-
- obd_getref(exp->exp_obd);
- /* Recreate a specific object id at the given OST index */
- if ((src_oa->o_valid & OBD_MD_FLFLAGS) &&
- (src_oa->o_flags & OBD_FL_RECREATE_OBJS)) {
- rc = lov_recreate(exp, src_oa, ea, oti);
- }
-
- obd_putref(exp->exp_obd);
- return rc;
-}
-
#define ASSERT_LSM_MAGIC(lsmp) \
do { \
LASSERT((lsmp)); \
@@ -2281,7 +2195,6 @@ static struct obd_ops lov_obd_ops = {
.statfs_async = lov_statfs_async,
.packmd = lov_packmd,
.unpackmd = lov_unpackmd,
- .create = lov_create,
.destroy = lov_destroy,
.getattr_async = lov_getattr_async,
.setattr_async = lov_setattr_async,
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 7449b77..11c376e 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -696,11 +696,6 @@ static int osc_create(const struct lu_env *env, struct obd_export *exp,
LASSERT(ea);
LASSERT(oa->o_valid & OBD_MD_FLGROUP);
- if ((oa->o_valid & OBD_MD_FLFLAGS) &&
- oa->o_flags == OBD_FL_RECREATE_OBJS) {
- return osc_real_create(exp, oa, ea, oti);
- }
-
if (!fid_seq_is_mdt(ostid_seq(&oa->o_oi)))
return osc_real_create(exp, oa, ea, oti);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 066/124] staging: lustre: lov: remove LL_IOC_RECREATE_{FID, OBJ}
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove the obsolete ioctls LL_IOC_RECREATE_FID and LL_IOC_RECREATE_OBJ
along with their handlers in llite. Remove the then unused OBD method
lov_create(). Remove OBD_FL_RECREATE_OBJS handling from osc_create().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12442
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 4 +-
drivers/staging/lustre/lustre/llite/file.c | 86 -------------------
drivers/staging/lustre/lustre/lov/lov_obd.c | 87 --------------------
drivers/staging/lustre/lustre/osc/osc_request.c | 5 -
4 files changed, 2 insertions(+), 180 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index 3ef5db0..de52339 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -202,8 +202,8 @@ struct ost_id {
#define LL_IOC_LOV_SETSTRIPE _IOW('f', 154, long)
#define LL_IOC_LOV_GETSTRIPE _IOW('f', 155, long)
#define LL_IOC_LOV_SETEA _IOW('f', 156, long)
-#define LL_IOC_RECREATE_OBJ _IOW('f', 157, long)
-#define LL_IOC_RECREATE_FID _IOW('f', 157, struct lu_fid)
+/* LL_IOC_RECREATE_OBJ 157 obsolete */
+/* LL_IOC_RECREATE_FID 158 obsolete */
#define LL_IOC_GROUP_LOCK _IOW('f', 158, long)
#define LL_IOC_GROUP_UNLOCK _IOW('f', 159, long)
/* #define LL_IOC_QUOTACHECK 160 OBD_IOC_QUOTACHECK */
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index fc81551..4b10260 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1302,88 +1302,6 @@ static ssize_t ll_file_splice_read(struct file *in_file, loff_t *ppos,
return result;
}
-static int ll_lov_recreate(struct inode *inode, struct ost_id *oi, u32 ost_idx)
-{
- struct obd_export *exp = ll_i2dtexp(inode);
- struct obd_trans_info oti = { 0 };
- struct obdo *oa = NULL;
- int lsm_size;
- int rc = 0;
- struct lov_stripe_md *lsm = NULL, *lsm2;
-
- oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
- if (!oa)
- return -ENOMEM;
-
- lsm = ccc_inode_lsm_get(inode);
- if (!lsm_has_objects(lsm)) {
- rc = -ENOENT;
- goto out;
- }
-
- lsm_size = sizeof(*lsm) + (sizeof(struct lov_oinfo) *
- (lsm->lsm_stripe_count));
-
- lsm2 = libcfs_kvzalloc(lsm_size, GFP_NOFS);
- if (!lsm2) {
- rc = -ENOMEM;
- goto out;
- }
-
- oa->o_oi = *oi;
- oa->o_nlink = ost_idx;
- oa->o_flags |= OBD_FL_RECREATE_OBJS;
- oa->o_valid = OBD_MD_FLID | OBD_MD_FLFLAGS | OBD_MD_FLGROUP;
- obdo_from_inode(oa, inode, OBD_MD_FLTYPE | OBD_MD_FLATIME |
- OBD_MD_FLMTIME | OBD_MD_FLCTIME);
- obdo_set_parent_fid(oa, &ll_i2info(inode)->lli_fid);
- memcpy(lsm2, lsm, lsm_size);
- ll_inode_size_lock(inode);
- rc = obd_create(NULL, exp, oa, &lsm2, &oti);
- ll_inode_size_unlock(inode);
-
- kvfree(lsm2);
- goto out;
-out:
- ccc_inode_lsm_put(inode, lsm);
- kmem_cache_free(obdo_cachep, oa);
- return rc;
-}
-
-static int ll_lov_recreate_obj(struct inode *inode, unsigned long arg)
-{
- struct ll_recreate_obj ucreat;
- struct ost_id oi;
-
- if (!capable(CFS_CAP_SYS_ADMIN))
- return -EPERM;
-
- if (copy_from_user(&ucreat, (struct ll_recreate_obj __user *)arg,
- sizeof(ucreat)))
- return -EFAULT;
-
- ostid_set_seq_mdt0(&oi);
- ostid_set_id(&oi, ucreat.lrc_id);
- return ll_lov_recreate(inode, &oi, ucreat.lrc_ost_idx);
-}
-
-static int ll_lov_recreate_fid(struct inode *inode, unsigned long arg)
-{
- struct lu_fid fid;
- struct ost_id oi;
- u32 ost_idx;
-
- if (!capable(CFS_CAP_SYS_ADMIN))
- return -EPERM;
-
- if (copy_from_user(&fid, (struct lu_fid __user *)arg, sizeof(fid)))
- return -EFAULT;
-
- fid_to_ostid(&fid, &oi);
- ost_idx = (fid_seq(&fid) >> 16) & 0xffff;
- return ll_lov_recreate(inode, &oi, ost_idx);
-}
-
int ll_lov_setstripe_ea_info(struct inode *inode, struct dentry *dentry,
__u64 flags, struct lov_user_md *lum,
int lum_size)
@@ -2322,10 +2240,6 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
}
case LL_IOC_LOV_GETSTRIPE:
return ll_lov_getstripe(inode, arg);
- case LL_IOC_RECREATE_OBJ:
- return ll_lov_recreate_obj(inode, arg);
- case LL_IOC_RECREATE_FID:
- return ll_lov_recreate_fid(inode, arg);
case FSFILT_IOC_FIEMAP:
return ll_ioctl_fiemap(inode, arg);
case FSFILT_IOC_GETFLAGS:
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 265be08..a54cd70 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -972,92 +972,6 @@ out:
return rc;
}
-static int lov_recreate(struct obd_export *exp, struct obdo *src_oa,
- struct lov_stripe_md **ea, struct obd_trans_info *oti)
-{
- struct lov_stripe_md *obj_mdp, *lsm;
- struct lov_obd *lov = &exp->exp_obd->u.lov;
- unsigned ost_idx;
- int rc, i;
-
- LASSERT(src_oa->o_valid & OBD_MD_FLFLAGS &&
- src_oa->o_flags & OBD_FL_RECREATE_OBJS);
-
- obj_mdp = kzalloc(sizeof(*obj_mdp), GFP_NOFS);
- if (!obj_mdp)
- return -ENOMEM;
-
- ost_idx = src_oa->o_nlink;
- lsm = *ea;
- if (!lsm) {
- rc = -EINVAL;
- goto out;
- }
- if (ost_idx >= lov->desc.ld_tgt_count ||
- !lov->lov_tgts[ost_idx]) {
- rc = -EINVAL;
- goto out;
- }
-
- for (i = 0; i < lsm->lsm_stripe_count; i++) {
- struct lov_oinfo *loi = lsm->lsm_oinfo[i];
-
- if (lov_oinfo_is_dummy(loi))
- continue;
-
- if (loi->loi_ost_idx == ost_idx) {
- if (ostid_id(&loi->loi_oi) != ostid_id(&src_oa->o_oi)) {
- rc = -EINVAL;
- goto out;
- }
- break;
- }
- }
- if (i == lsm->lsm_stripe_count) {
- rc = -EINVAL;
- goto out;
- }
-
- rc = obd_create(NULL, lov->lov_tgts[ost_idx]->ltd_exp,
- src_oa, &obj_mdp, oti);
-out:
- kfree(obj_mdp);
- return rc;
-}
-
-/* the LOV expects oa->o_id to be set to the LOV object id */
-static int lov_create(const struct lu_env *env, struct obd_export *exp,
- struct obdo *src_oa, struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
-{
- struct lov_obd *lov;
- int rc = 0;
-
- LASSERT(ea);
- if (!exp)
- return -EINVAL;
-
- if ((src_oa->o_valid & OBD_MD_FLFLAGS) &&
- src_oa->o_flags == OBD_FL_DELORPHAN) {
- /* should be used with LOV anymore */
- LBUG();
- }
-
- lov = &exp->exp_obd->u.lov;
- if (!lov->desc.ld_active_tgt_count)
- return -EIO;
-
- obd_getref(exp->exp_obd);
- /* Recreate a specific object id at the given OST index */
- if ((src_oa->o_valid & OBD_MD_FLFLAGS) &&
- (src_oa->o_flags & OBD_FL_RECREATE_OBJS)) {
- rc = lov_recreate(exp, src_oa, ea, oti);
- }
-
- obd_putref(exp->exp_obd);
- return rc;
-}
-
#define ASSERT_LSM_MAGIC(lsmp) \
do { \
LASSERT((lsmp)); \
@@ -2281,7 +2195,6 @@ static struct obd_ops lov_obd_ops = {
.statfs_async = lov_statfs_async,
.packmd = lov_packmd,
.unpackmd = lov_unpackmd,
- .create = lov_create,
.destroy = lov_destroy,
.getattr_async = lov_getattr_async,
.setattr_async = lov_setattr_async,
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 7449b77..11c376e 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -696,11 +696,6 @@ static int osc_create(const struct lu_env *env, struct obd_export *exp,
LASSERT(ea);
LASSERT(oa->o_valid & OBD_MD_FLGROUP);
- if ((oa->o_valid & OBD_MD_FLFLAGS) &&
- oa->o_flags == OBD_FL_RECREATE_OBJS) {
- return osc_real_create(exp, oa, ea, oti);
- }
-
if (!fid_seq_is_mdt(ostid_seq(&oa->o_oi)))
return osc_real_create(exp, oa, ea, oti);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 067/124] staging: lustre: changelog: fix comparison between signed and unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change type of changelog_*{namelen,size}() to size_t.
Fixed string specifier for unsigned types.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12474
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 10 +++++-----
drivers/staging/lustre/lustre/llite/file.c | 4 ++--
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index de52339..e8f81a5 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -819,9 +819,9 @@ struct changelog_ext_jobid {
char cr_jobid[LUSTRE_JOBID_SIZE]; /**< zero-terminated string. */
};
-static inline unsigned int changelog_rec_offset(enum changelog_rec_flags crf)
+static inline size_t changelog_rec_offset(enum changelog_rec_flags crf)
{
- unsigned int size = sizeof(struct changelog_rec);
+ size_t size = sizeof(struct changelog_rec);
if (crf & CLF_RENAME)
size += sizeof(struct changelog_ext_rename);
@@ -832,12 +832,12 @@ static inline unsigned int changelog_rec_offset(enum changelog_rec_flags crf)
return size;
}
-static inline int changelog_rec_size(struct changelog_rec *rec)
+static inline size_t changelog_rec_size(struct changelog_rec *rec)
{
return changelog_rec_offset(rec->cr_flags);
}
-static inline unsigned int changelog_rec_varsize(struct changelog_rec *rec)
+static inline size_t changelog_rec_varsize(struct changelog_rec *rec)
{
return changelog_rec_size(rec) - sizeof(*rec) + rec->cr_namelen;
}
@@ -869,7 +869,7 @@ static inline char *changelog_rec_name(struct changelog_rec *rec)
CLF_SUPPORTED);
}
-static inline int changelog_rec_snamelen(struct changelog_rec *rec)
+static inline size_t changelog_rec_snamelen(struct changelog_rec *rec)
{
return rec->cr_namelen - strlen(changelog_rec_name(rec)) - 1;
}
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 4b10260..a562b11 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1129,7 +1129,7 @@ ll_file_io_generic(const struct lu_env *env, struct vvp_io_args *args,
struct cl_io *io;
ssize_t result;
- CDEBUG(D_VFSTRACE, "file: %s, type: %d ppos: %llu, count: %zd\n",
+ CDEBUG(D_VFSTRACE, "file: %s, type: %d ppos: %llu, count: %zu\n",
file->f_path.dentry->d_name.name, iot, *ppos, count);
restart:
@@ -1207,7 +1207,7 @@ out:
* short read/write instead of restart io.
*/
if ((result == 0 || result == -ENODATA) && io->ci_need_restart) {
- CDEBUG(D_VFSTRACE, "Restart %s on %pD from %lld, count:%zd\n",
+ CDEBUG(D_VFSTRACE, "Restart %s on %pD from %lld, count:%zu\n",
iot == CIT_READ ? "read" : "write",
file, *ppos, count);
LASSERTF(io->ci_nob == 0, "%zd\n", io->ci_nob);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 067/124] staging: lustre: changelog: fix comparison between signed and unsigned
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change type of changelog_*{namelen,size}() to size_t.
Fixed string specifier for unsigned types.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12474
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 10 +++++-----
drivers/staging/lustre/lustre/llite/file.c | 4 ++--
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index de52339..e8f81a5 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -819,9 +819,9 @@ struct changelog_ext_jobid {
char cr_jobid[LUSTRE_JOBID_SIZE]; /**< zero-terminated string. */
};
-static inline unsigned int changelog_rec_offset(enum changelog_rec_flags crf)
+static inline size_t changelog_rec_offset(enum changelog_rec_flags crf)
{
- unsigned int size = sizeof(struct changelog_rec);
+ size_t size = sizeof(struct changelog_rec);
if (crf & CLF_RENAME)
size += sizeof(struct changelog_ext_rename);
@@ -832,12 +832,12 @@ static inline unsigned int changelog_rec_offset(enum changelog_rec_flags crf)
return size;
}
-static inline int changelog_rec_size(struct changelog_rec *rec)
+static inline size_t changelog_rec_size(struct changelog_rec *rec)
{
return changelog_rec_offset(rec->cr_flags);
}
-static inline unsigned int changelog_rec_varsize(struct changelog_rec *rec)
+static inline size_t changelog_rec_varsize(struct changelog_rec *rec)
{
return changelog_rec_size(rec) - sizeof(*rec) + rec->cr_namelen;
}
@@ -869,7 +869,7 @@ static inline char *changelog_rec_name(struct changelog_rec *rec)
CLF_SUPPORTED);
}
-static inline int changelog_rec_snamelen(struct changelog_rec *rec)
+static inline size_t changelog_rec_snamelen(struct changelog_rec *rec)
{
return rec->cr_namelen - strlen(changelog_rec_name(rec)) - 1;
}
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 4b10260..a562b11 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1129,7 +1129,7 @@ ll_file_io_generic(const struct lu_env *env, struct vvp_io_args *args,
struct cl_io *io;
ssize_t result;
- CDEBUG(D_VFSTRACE, "file: %s, type: %d ppos: %llu, count: %zd\n",
+ CDEBUG(D_VFSTRACE, "file: %s, type: %d ppos: %llu, count: %zu\n",
file->f_path.dentry->d_name.name, iot, *ppos, count);
restart:
@@ -1207,7 +1207,7 @@ out:
* short read/write instead of restart io.
*/
if ((result == 0 || result == -ENODATA) && io->ci_need_restart) {
- CDEBUG(D_VFSTRACE, "Restart %s on %pD from %lld, count:%zd\n",
+ CDEBUG(D_VFSTRACE, "Restart %s on %pD from %lld, count:%zu\n",
iot == CIT_READ ? "read" : "write",
file, *ppos, count);
LASSERTF(io->ci_nob == 0, "%zd\n", io->ci_nob);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 068/124] staging: lustre: lov: remove unused {get,set}_info handlers
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In LOV and OSC remove handlers for the obsolete get and set info keys:
KEY_CAPA_KEY, KEY_CONNECT_FLAG, KEY_EVICT_BY_NID, KEY_LAST_ID,
KEY_LOCK_TO_STRIPE, KEY_MDS_CONN, KEY_NEXT_ID.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12445
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 4 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 132 ++---------------------
drivers/staging/lustre/lustre/osc/osc_request.c | 46 +--------
3 files changed, 10 insertions(+), 172 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 2d25e40..6eac63e 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -681,20 +681,16 @@ enum obd_cleanup_stage {
#define KEY_INTERMDS "inter_mds"
#define KEY_LAST_ID "last_id"
#define KEY_LAST_FID "last_fid"
-#define KEY_LOCK_TO_STRIPE "lock_to_stripe"
#define KEY_LOVDESC "lovdesc"
#define KEY_MAX_EASIZE "max_easize"
#define KEY_DEFAULT_EASIZE "default_easize"
-#define KEY_MDS_CONN "mds_conn"
#define KEY_MGSSEC "mgssec"
-#define KEY_NEXT_ID "next_id"
#define KEY_READ_ONLY "read-only"
#define KEY_REGISTER_TARGET "register_target"
#define KEY_SET_FS "set_fs"
#define KEY_TGT_COUNT "tgt_count"
/* KEY_SET_INFO in lustre_idl.h */
#define KEY_SPTLRPC_CONF "sptlrpc_conf"
-#define KEY_CONNECT_FLAG "connect_flags"
#define KEY_CACHE_SET "cache_set"
#define KEY_CACHE_LRU_SHRINK "cache_lru_shrink"
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index a54cd70..2ee4b19 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -1862,73 +1862,14 @@ static int lov_get_info(const struct lu_env *env, struct obd_export *exp,
{
struct obd_device *obddev = class_exp2obd(exp);
struct lov_obd *lov = &obddev->u.lov;
- int i, rc;
+ int rc;
if (!vallen || !val)
return -EFAULT;
obd_getref(obddev);
- if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
- struct {
- char name[16];
- struct ldlm_lock *lock;
- } *data = key;
- struct ldlm_res_id *res_id = &data->lock->l_resource->lr_name;
- struct lov_oinfo *loi;
- __u32 *stripe = val;
-
- if (*vallen < sizeof(*stripe)) {
- rc = -EFAULT;
- goto out;
- }
- *vallen = sizeof(*stripe);
-
- /* XXX This is another one of those bits that will need to
- * change if we ever actually support nested LOVs. It uses
- * the lock's export to find out which stripe it is.
- */
- /* XXX - it's assumed all the locks for deleted OSTs have
- * been cancelled. Also, the export for deleted OSTs will
- * be NULL and won't match the lock's export.
- */
- for (i = 0; i < lsm->lsm_stripe_count; i++) {
- loi = lsm->lsm_oinfo[i];
- if (lov_oinfo_is_dummy(loi))
- continue;
-
- if (!lov->lov_tgts[loi->loi_ost_idx])
- continue;
- if (lov->lov_tgts[loi->loi_ost_idx]->ltd_exp ==
- data->lock->l_conn_export &&
- ostid_res_name_eq(&loi->loi_oi, res_id)) {
- *stripe = i;
- rc = 0;
- goto out;
- }
- }
- LDLM_ERROR(data->lock, "lock on inode without such object");
- dump_lsm(D_ERROR, lsm);
- rc = -ENXIO;
- goto out;
- } else if (KEY_IS(KEY_LAST_ID)) {
- struct obd_id_info *info = val;
- __u32 size = sizeof(u64);
- struct lov_tgt_desc *tgt;
-
- LASSERT(*vallen == sizeof(struct obd_id_info));
- tgt = lov->lov_tgts[info->idx];
-
- if (!tgt || !tgt->ltd_active) {
- rc = -ESRCH;
- goto out;
- }
-
- rc = obd_get_info(env, tgt->ltd_exp, keylen, key,
- &size, info->data, NULL);
- rc = 0;
- goto out;
- } else if (KEY_IS(KEY_LOVDESC)) {
+ if (KEY_IS(KEY_LOVDESC)) {
struct lov_desc *desc_ret = val;
*desc_ret = lov->desc;
@@ -1937,22 +1878,6 @@ static int lov_get_info(const struct lu_env *env, struct obd_export *exp,
} else if (KEY_IS(KEY_FIEMAP)) {
rc = lov_fiemap(lov, keylen, key, vallen, val, lsm);
goto out;
- } else if (KEY_IS(KEY_CONNECT_FLAG)) {
- struct lov_tgt_desc *tgt;
- __u64 ost_idx = *((__u64 *)val);
-
- LASSERT(*vallen == sizeof(__u64));
- LASSERT(ost_idx < lov->desc.ld_tgt_count);
- tgt = lov->lov_tgts[ost_idx];
-
- if (!tgt || !tgt->ltd_exp) {
- rc = -ESRCH;
- goto out;
- }
-
- *((__u64 *)val) = exp_connect_flags(tgt->ltd_exp);
- rc = 0;
- goto out;
} else if (KEY_IS(KEY_TGT_COUNT)) {
*((int *)val) = lov->desc.ld_tgt_count;
rc = 0;
@@ -1975,8 +1900,7 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
u32 count;
int i, rc = 0, err;
struct lov_tgt_desc *tgt;
- unsigned int incr = 0, check_uuid = 0, do_inactive = 0, no_set = 0;
- unsigned int next_id = 0, mds_con = 0;
+ int do_inactive = 0, no_set = 0;
if (!set) {
no_set = 1;
@@ -1988,18 +1912,8 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
obd_getref(obddev);
count = lov->desc.ld_tgt_count;
- if (KEY_IS(KEY_NEXT_ID)) {
- count = vallen / sizeof(struct obd_id_info);
- vallen = sizeof(u64);
- incr = sizeof(struct obd_id_info);
- do_inactive = 1;
- next_id = 1;
- } else if (KEY_IS(KEY_CHECKSUM)) {
+ if (KEY_IS(KEY_CHECKSUM)) {
do_inactive = 1;
- } else if (KEY_IS(KEY_EVICT_BY_NID)) {
- /* use defaults: do_inactive = incr = 0; */
- } else if (KEY_IS(KEY_MDS_CONN)) {
- mds_con = 1;
} else if (KEY_IS(KEY_CACHE_SET)) {
LASSERT(!lov->lov_cache);
lov->lov_cache = val;
@@ -2007,11 +1921,9 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
cl_cache_incref(lov->lov_cache);
}
- for (i = 0; i < count; i++, val = (char *)val + incr) {
- if (next_id)
- tgt = lov->lov_tgts[((struct obd_id_info *)val)->idx];
- else
- tgt = lov->lov_tgts[i];
+ for (i = 0; i < count; i++) {
+ tgt = lov->lov_tgts[i];
+
/* OST was disconnected */
if (!tgt || !tgt->ltd_exp)
continue;
@@ -2020,34 +1932,8 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
if (!tgt->ltd_active && !do_inactive)
continue;
- if (mds_con) {
- struct mds_group_info *mgi;
-
- LASSERT(vallen == sizeof(*mgi));
- mgi = (struct mds_group_info *)val;
-
- /* Only want a specific OSC */
- if (mgi->uuid && !obd_uuid_equals(mgi->uuid,
- &tgt->ltd_uuid))
- continue;
-
- err = obd_set_info_async(env, tgt->ltd_exp,
- keylen, key, sizeof(int),
- &mgi->group, set);
- } else if (next_id) {
- err = obd_set_info_async(env, tgt->ltd_exp,
- keylen, key, vallen,
- ((struct obd_id_info *)val)->data, set);
- } else {
- /* Only want a specific OSC */
- if (check_uuid &&
- !obd_uuid_equals(val, &tgt->ltd_uuid))
- continue;
-
- err = obd_set_info_async(env, tgt->ltd_exp,
- keylen, key, vallen, val, set);
- }
-
+ err = obd_set_info_async(env, tgt->ltd_exp, keylen, key,
+ vallen, val, set);
if (!rc)
rc = err;
}
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 11c376e..1b6491a 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -2733,51 +2733,7 @@ static int osc_get_info(const struct lu_env *env, struct obd_export *exp,
if (!vallen || !val)
return -EFAULT;
- if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
- __u32 *stripe = val;
- *vallen = sizeof(*stripe);
- *stripe = 0;
- return 0;
- } else if (KEY_IS(KEY_LAST_ID)) {
- struct ptlrpc_request *req;
- u64 *reply;
- char *tmp;
- int rc;
-
- req = ptlrpc_request_alloc(class_exp2cliimp(exp),
- &RQF_OST_GET_INFO_LAST_ID);
- if (!req)
- return -ENOMEM;
-
- req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
- RCL_CLIENT, keylen);
- rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
- if (rc) {
- ptlrpc_request_free(req);
- return rc;
- }
-
- tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
- memcpy(tmp, key, keylen);
-
- req->rq_no_delay = 1;
- req->rq_no_resend = 1;
- ptlrpc_request_set_replen(req);
- rc = ptlrpc_queue_wait(req);
- if (rc)
- goto out;
-
- reply = req_capsule_server_get(&req->rq_pill, &RMF_OBD_ID);
- if (!reply) {
- rc = -EPROTO;
- goto out;
- }
-
- *((u64 *)val) = *reply;
-out:
- ptlrpc_req_finished(req);
- return rc;
- } else if (KEY_IS(KEY_FIEMAP)) {
+ if (KEY_IS(KEY_FIEMAP)) {
struct ll_fiemap_info_key *fm_key = key;
struct ldlm_res_id res_id;
ldlm_policy_data_t policy;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 068/124] staging: lustre: lov: remove unused {get, set}_info handlers
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In LOV and OSC remove handlers for the obsolete get and set info keys:
KEY_CAPA_KEY, KEY_CONNECT_FLAG, KEY_EVICT_BY_NID, KEY_LAST_ID,
KEY_LOCK_TO_STRIPE, KEY_MDS_CONN, KEY_NEXT_ID.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12445
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 4 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 132 ++---------------------
drivers/staging/lustre/lustre/osc/osc_request.c | 46 +--------
3 files changed, 10 insertions(+), 172 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 2d25e40..6eac63e 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -681,20 +681,16 @@ enum obd_cleanup_stage {
#define KEY_INTERMDS "inter_mds"
#define KEY_LAST_ID "last_id"
#define KEY_LAST_FID "last_fid"
-#define KEY_LOCK_TO_STRIPE "lock_to_stripe"
#define KEY_LOVDESC "lovdesc"
#define KEY_MAX_EASIZE "max_easize"
#define KEY_DEFAULT_EASIZE "default_easize"
-#define KEY_MDS_CONN "mds_conn"
#define KEY_MGSSEC "mgssec"
-#define KEY_NEXT_ID "next_id"
#define KEY_READ_ONLY "read-only"
#define KEY_REGISTER_TARGET "register_target"
#define KEY_SET_FS "set_fs"
#define KEY_TGT_COUNT "tgt_count"
/* KEY_SET_INFO in lustre_idl.h */
#define KEY_SPTLRPC_CONF "sptlrpc_conf"
-#define KEY_CONNECT_FLAG "connect_flags"
#define KEY_CACHE_SET "cache_set"
#define KEY_CACHE_LRU_SHRINK "cache_lru_shrink"
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index a54cd70..2ee4b19 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -1862,73 +1862,14 @@ static int lov_get_info(const struct lu_env *env, struct obd_export *exp,
{
struct obd_device *obddev = class_exp2obd(exp);
struct lov_obd *lov = &obddev->u.lov;
- int i, rc;
+ int rc;
if (!vallen || !val)
return -EFAULT;
obd_getref(obddev);
- if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
- struct {
- char name[16];
- struct ldlm_lock *lock;
- } *data = key;
- struct ldlm_res_id *res_id = &data->lock->l_resource->lr_name;
- struct lov_oinfo *loi;
- __u32 *stripe = val;
-
- if (*vallen < sizeof(*stripe)) {
- rc = -EFAULT;
- goto out;
- }
- *vallen = sizeof(*stripe);
-
- /* XXX This is another one of those bits that will need to
- * change if we ever actually support nested LOVs. It uses
- * the lock's export to find out which stripe it is.
- */
- /* XXX - it's assumed all the locks for deleted OSTs have
- * been cancelled. Also, the export for deleted OSTs will
- * be NULL and won't match the lock's export.
- */
- for (i = 0; i < lsm->lsm_stripe_count; i++) {
- loi = lsm->lsm_oinfo[i];
- if (lov_oinfo_is_dummy(loi))
- continue;
-
- if (!lov->lov_tgts[loi->loi_ost_idx])
- continue;
- if (lov->lov_tgts[loi->loi_ost_idx]->ltd_exp ==
- data->lock->l_conn_export &&
- ostid_res_name_eq(&loi->loi_oi, res_id)) {
- *stripe = i;
- rc = 0;
- goto out;
- }
- }
- LDLM_ERROR(data->lock, "lock on inode without such object");
- dump_lsm(D_ERROR, lsm);
- rc = -ENXIO;
- goto out;
- } else if (KEY_IS(KEY_LAST_ID)) {
- struct obd_id_info *info = val;
- __u32 size = sizeof(u64);
- struct lov_tgt_desc *tgt;
-
- LASSERT(*vallen == sizeof(struct obd_id_info));
- tgt = lov->lov_tgts[info->idx];
-
- if (!tgt || !tgt->ltd_active) {
- rc = -ESRCH;
- goto out;
- }
-
- rc = obd_get_info(env, tgt->ltd_exp, keylen, key,
- &size, info->data, NULL);
- rc = 0;
- goto out;
- } else if (KEY_IS(KEY_LOVDESC)) {
+ if (KEY_IS(KEY_LOVDESC)) {
struct lov_desc *desc_ret = val;
*desc_ret = lov->desc;
@@ -1937,22 +1878,6 @@ static int lov_get_info(const struct lu_env *env, struct obd_export *exp,
} else if (KEY_IS(KEY_FIEMAP)) {
rc = lov_fiemap(lov, keylen, key, vallen, val, lsm);
goto out;
- } else if (KEY_IS(KEY_CONNECT_FLAG)) {
- struct lov_tgt_desc *tgt;
- __u64 ost_idx = *((__u64 *)val);
-
- LASSERT(*vallen == sizeof(__u64));
- LASSERT(ost_idx < lov->desc.ld_tgt_count);
- tgt = lov->lov_tgts[ost_idx];
-
- if (!tgt || !tgt->ltd_exp) {
- rc = -ESRCH;
- goto out;
- }
-
- *((__u64 *)val) = exp_connect_flags(tgt->ltd_exp);
- rc = 0;
- goto out;
} else if (KEY_IS(KEY_TGT_COUNT)) {
*((int *)val) = lov->desc.ld_tgt_count;
rc = 0;
@@ -1975,8 +1900,7 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
u32 count;
int i, rc = 0, err;
struct lov_tgt_desc *tgt;
- unsigned int incr = 0, check_uuid = 0, do_inactive = 0, no_set = 0;
- unsigned int next_id = 0, mds_con = 0;
+ int do_inactive = 0, no_set = 0;
if (!set) {
no_set = 1;
@@ -1988,18 +1912,8 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
obd_getref(obddev);
count = lov->desc.ld_tgt_count;
- if (KEY_IS(KEY_NEXT_ID)) {
- count = vallen / sizeof(struct obd_id_info);
- vallen = sizeof(u64);
- incr = sizeof(struct obd_id_info);
- do_inactive = 1;
- next_id = 1;
- } else if (KEY_IS(KEY_CHECKSUM)) {
+ if (KEY_IS(KEY_CHECKSUM)) {
do_inactive = 1;
- } else if (KEY_IS(KEY_EVICT_BY_NID)) {
- /* use defaults: do_inactive = incr = 0; */
- } else if (KEY_IS(KEY_MDS_CONN)) {
- mds_con = 1;
} else if (KEY_IS(KEY_CACHE_SET)) {
LASSERT(!lov->lov_cache);
lov->lov_cache = val;
@@ -2007,11 +1921,9 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
cl_cache_incref(lov->lov_cache);
}
- for (i = 0; i < count; i++, val = (char *)val + incr) {
- if (next_id)
- tgt = lov->lov_tgts[((struct obd_id_info *)val)->idx];
- else
- tgt = lov->lov_tgts[i];
+ for (i = 0; i < count; i++) {
+ tgt = lov->lov_tgts[i];
+
/* OST was disconnected */
if (!tgt || !tgt->ltd_exp)
continue;
@@ -2020,34 +1932,8 @@ static int lov_set_info_async(const struct lu_env *env, struct obd_export *exp,
if (!tgt->ltd_active && !do_inactive)
continue;
- if (mds_con) {
- struct mds_group_info *mgi;
-
- LASSERT(vallen == sizeof(*mgi));
- mgi = (struct mds_group_info *)val;
-
- /* Only want a specific OSC */
- if (mgi->uuid && !obd_uuid_equals(mgi->uuid,
- &tgt->ltd_uuid))
- continue;
-
- err = obd_set_info_async(env, tgt->ltd_exp,
- keylen, key, sizeof(int),
- &mgi->group, set);
- } else if (next_id) {
- err = obd_set_info_async(env, tgt->ltd_exp,
- keylen, key, vallen,
- ((struct obd_id_info *)val)->data, set);
- } else {
- /* Only want a specific OSC */
- if (check_uuid &&
- !obd_uuid_equals(val, &tgt->ltd_uuid))
- continue;
-
- err = obd_set_info_async(env, tgt->ltd_exp,
- keylen, key, vallen, val, set);
- }
-
+ err = obd_set_info_async(env, tgt->ltd_exp, keylen, key,
+ vallen, val, set);
if (!rc)
rc = err;
}
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 11c376e..1b6491a 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -2733,51 +2733,7 @@ static int osc_get_info(const struct lu_env *env, struct obd_export *exp,
if (!vallen || !val)
return -EFAULT;
- if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
- __u32 *stripe = val;
- *vallen = sizeof(*stripe);
- *stripe = 0;
- return 0;
- } else if (KEY_IS(KEY_LAST_ID)) {
- struct ptlrpc_request *req;
- u64 *reply;
- char *tmp;
- int rc;
-
- req = ptlrpc_request_alloc(class_exp2cliimp(exp),
- &RQF_OST_GET_INFO_LAST_ID);
- if (!req)
- return -ENOMEM;
-
- req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
- RCL_CLIENT, keylen);
- rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
- if (rc) {
- ptlrpc_request_free(req);
- return rc;
- }
-
- tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
- memcpy(tmp, key, keylen);
-
- req->rq_no_delay = 1;
- req->rq_no_resend = 1;
- ptlrpc_request_set_replen(req);
- rc = ptlrpc_queue_wait(req);
- if (rc)
- goto out;
-
- reply = req_capsule_server_get(&req->rq_pill, &RMF_OBD_ID);
- if (!reply) {
- rc = -EPROTO;
- goto out;
- }
-
- *((u64 *)val) = *reply;
-out:
- ptlrpc_req_finished(req);
- return rc;
- } else if (KEY_IS(KEY_FIEMAP)) {
+ if (KEY_IS(KEY_FIEMAP)) {
struct ll_fiemap_info_key *fm_key = key;
struct ldlm_res_id res_id;
ldlm_policy_data_t policy;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 069/124] staging: lustre: fix messages with missing newlines
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Restore the trailing newline in the definition of OSC_DUMP_GRANT().
Remove an unnecessary CDEBUG() from ldlm_pool_recalc().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5551
Reviewed-on: http://review.whamcloud.com/11996
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 4 ----
drivers/staging/lustre/lustre/osc/osc_cache.c | 2 +-
2 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
index 1147b30..44c3abf 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
@@ -385,10 +385,6 @@ static int ldlm_pool_recalc(struct ldlm_pool *pl)
pl->pl_recalc_period;
if (recalc_interval_sec <= 0) {
/* Prevent too frequent recalculation. */
- CDEBUG(D_DLMTRACE,
- "Negative interval(%d), too short period(%lld)",
- recalc_interval_sec,
- (s64)pl->pl_recalc_period);
recalc_interval_sec = 1;
}
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 26ad360..fb18dd1 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1385,7 +1385,7 @@ static int osc_completion(const struct lu_env *env, struct osc_async_page *oap,
struct client_obd *__tmp = (cli); \
CDEBUG(lvl, "%s: grant { dirty: %ld/%ld dirty_pages: %ld/%lu " \
"dropped: %ld avail: %ld, reserved: %ld, flight: %d }" \
- "lru {in list: %ld, left: %ld, waiters: %d }" fmt, \
+ "lru {in list: %ld, left: %ld, waiters: %d }" fmt "\n", \
__tmp->cl_import->imp_obd->obd_name, \
__tmp->cl_dirty_pages, __tmp->cl_dirty_max_pages, \
atomic_long_read(&obd_dirty_pages), obd_max_dirty_pages, \
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 069/124] staging: lustre: fix messages with missing newlines
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Restore the trailing newline in the definition of OSC_DUMP_GRANT().
Remove an unnecessary CDEBUG() from ldlm_pool_recalc().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5551
Reviewed-on: http://review.whamcloud.com/11996
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 4 ----
drivers/staging/lustre/lustre/osc/osc_cache.c | 2 +-
2 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
index 1147b30..44c3abf 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
@@ -385,10 +385,6 @@ static int ldlm_pool_recalc(struct ldlm_pool *pl)
pl->pl_recalc_period;
if (recalc_interval_sec <= 0) {
/* Prevent too frequent recalculation. */
- CDEBUG(D_DLMTRACE,
- "Negative interval(%d), too short period(%lld)",
- recalc_interval_sec,
- (s64)pl->pl_recalc_period);
recalc_interval_sec = 1;
}
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 26ad360..fb18dd1 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1385,7 +1385,7 @@ static int osc_completion(const struct lu_env *env, struct osc_async_page *oap,
struct client_obd *__tmp = (cli); \
CDEBUG(lvl, "%s: grant { dirty: %ld/%ld dirty_pages: %ld/%lu " \
"dropped: %ld avail: %ld, reserved: %ld, flight: %d }" \
- "lru {in list: %ld, left: %ld, waiters: %d }" fmt, \
+ "lru {in list: %ld, left: %ld, waiters: %d }" fmt "\n", \
__tmp->cl_import->imp_obd->obd_name, \
__tmp->cl_dirty_pages, __tmp->cl_dirty_max_pages, \
atomic_long_read(&obd_dirty_pages), obd_max_dirty_pages, \
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 070/124] staging: lustre: statahead: small fixes and cleanup
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
small fixes:
* when 'unplug' is set for ll_statahead(), sa_put() shouldn't kill
the entry found, because its inflight RPC may not finish yet.
* remove 'sai_generation', add 'lli_sa_generation' because the
former one is not safe to access without lock.
* revalidate_statahead_dentry() may fail to wait for statahead
entry to become ready, in this case it should not release this
entry, because it may be used by inflight statahead RPC.
cleanups:
* rename ll_statahead_enter() to ll_statahead().
* move dentry 'lld_sa_generation' update to ll_statahead() to
simplify code and logic.
* other small cleanups.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9667
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6222
Reviewed-on: http://review.whamcloud.com/13708
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/dcache.c | 5 +-
.../staging/lustre/lustre/llite/llite_internal.h | 137 +++-----
drivers/staging/lustre/lustre/llite/namei.c | 11 +-
drivers/staging/lustre/lustre/llite/statahead.c | 353 +++++++++++---------
drivers/staging/lustre/lustre/mdc/mdc_request.c | 2 +-
5 files changed, 250 insertions(+), 258 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index 8500080..0e45d8f 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -278,14 +278,13 @@ static int ll_revalidate_dentry(struct dentry *dentry,
if (lookup_flags & (LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE))
return 1;
- if (!dentry_need_statahead(dir, dentry))
+ if (!dentry_may_statahead(dir, dentry))
return 1;
if (lookup_flags & LOOKUP_RCU)
return -ECHILD;
- do_statahead_enter(dir, &dentry, !d_inode(dentry));
- ll_statahead_mark(dir, dentry);
+ ll_statahead(dir, &dentry, !d_inode(dentry));
return 1;
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index a68bea1..bdfdff5 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -161,7 +161,7 @@ struct ll_inode_info {
/* for directory */
struct {
/* serialize normal readdir and statahead-readdir. */
- struct mutex d_readdir_mutex;
+ struct mutex lli_readdir_mutex;
/* metadata statahead */
/* since parent-child threads can share the same @file
@@ -169,44 +169,35 @@ struct ll_inode_info {
* case of parent exit before child -- it is me should
* cleanup the dir readahead.
*/
- void *d_opendir_key;
- struct ll_statahead_info *d_sai;
+ void *lli_opendir_key;
+ struct ll_statahead_info *lli_sai;
/* protect statahead stuff. */
- spinlock_t d_sa_lock;
+ spinlock_t lli_sa_lock;
/* "opendir_pid" is the token when lookup/revalidate
* -- I am the owner of dir statahead.
*/
- pid_t d_opendir_pid;
+ pid_t lli_opendir_pid;
/* stat will try to access statahead entries or start
* statahead if this flag is set, and this flag will be
* set upon dir open, and cleared when dir is closed,
* statahead hit ratio is too low, or start statahead
* thread failed.
*/
- unsigned int d_sa_enabled:1;
+ unsigned int lli_sa_enabled:1;
+ /* generation for statahead */
+ unsigned int lli_sa_generation;
/* directory stripe information */
- struct lmv_stripe_md *d_lsm_md;
+ struct lmv_stripe_md *lli_lsm_md;
/* striped directory size */
- loff_t d_stripe_size;
- /* striped directory nlink */
- __u64 d_stripe_nlink;
- } d;
-
-#define lli_readdir_mutex u.d.d_readdir_mutex
-#define lli_opendir_key u.d.d_opendir_key
-#define lli_sai u.d.d_sai
-#define lli_sa_lock u.d.d_sa_lock
-#define lli_sa_enabled u.d.d_sa_enabled
-#define lli_opendir_pid u.d.d_opendir_pid
-#define lli_lsm_md u.d.d_lsm_md
-#define lli_stripe_dir_size u.d.d_stripe_size
-#define lli_stripe_dir_nlink u.d.d_stripe_nlink
+ loff_t lli_stripe_dir_size;
+ u64 lli_stripe_dir_nlink;
+ };
/* for non-directory */
struct {
- struct mutex f_size_mutex;
- char *f_symlink_name;
- __u64 f_maxbytes;
+ struct mutex lli_size_mutex;
+ char *lli_symlink_name;
+ __u64 lli_maxbytes;
/*
* struct rw_semaphore {
* signed long count; // align d.d_def_acl
@@ -214,16 +205,16 @@ struct ll_inode_info {
* struct list_head wait_list;
* }
*/
- struct rw_semaphore f_trunc_sem;
- struct range_lock_tree f_write_tree;
+ struct rw_semaphore lli_trunc_sem;
+ struct range_lock_tree lli_write_tree;
- struct rw_semaphore f_glimpse_sem;
- unsigned long f_glimpse_time;
- struct list_head f_agl_list;
- __u64 f_agl_index;
+ struct rw_semaphore lli_glimpse_sem;
+ unsigned long lli_glimpse_time;
+ struct list_head lli_agl_list;
+ __u64 lli_agl_index;
/* for writepage() only to communicate to fsync */
- int f_async_rc;
+ int lli_async_rc;
/*
* whenever a process try to read/write the file, the
@@ -233,22 +224,9 @@ struct ll_inode_info {
* so the read/write statistics for jobid will not be
* accurate if the file is shared by different jobs.
*/
- char f_jobid[LUSTRE_JOBID_SIZE];
- } f;
-
-#define lli_size_mutex u.f.f_size_mutex
-#define lli_symlink_name u.f.f_symlink_name
-#define lli_maxbytes u.f.f_maxbytes
-#define lli_trunc_sem u.f.f_trunc_sem
-#define lli_write_tree u.f.f_write_tree
-#define lli_glimpse_sem u.f.f_glimpse_sem
-#define lli_glimpse_time u.f.f_glimpse_time
-#define lli_agl_list u.f.f_agl_list
-#define lli_agl_index u.f.f_agl_index
-#define lli_async_rc u.f.f_async_rc
-#define lli_jobid u.f.f_jobid
-
- } u;
+ char lli_jobid[LUSTRE_JOBID_SIZE];
+ };
+ };
/* XXX: For following frequent used members, although they maybe special
* used for non-directory object, it is some time-wasting to check
@@ -1095,11 +1073,10 @@ void ll_ra_stats_inc(struct inode *inode, enum ra_stat which);
/* per inode struct, for dir only */
struct ll_statahead_info {
- struct inode *sai_inode;
+ struct dentry *sai_dentry;
atomic_t sai_refcount; /* when access this struct, hold
* refcount
*/
- unsigned int sai_generation; /* generation for statahead */
unsigned int sai_max; /* max ahead of lookup */
__u64 sai_sent; /* stat requests sent count */
__u64 sai_replied; /* stat requests which received
@@ -1142,8 +1119,7 @@ struct ll_statahead_info {
atomic_t sai_cache_count; /* entry count in cache */
};
-int do_statahead_enter(struct inode *dir, struct dentry **dentry,
- int only_unplug);
+int ll_statahead(struct inode *dir, struct dentry **dentry, bool unplug);
void ll_authorize_statahead(struct inode *dir, void *key);
void ll_deauthorize_statahead(struct inode *dir, void *key);
@@ -1175,24 +1151,12 @@ static inline int ll_glimpse_size(struct inode *inode)
return rc;
}
-static inline void
-ll_statahead_mark(struct inode *dir, struct dentry *dentry)
-{
- struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_statahead_info *sai = lli->lli_sai;
- struct ll_dentry_data *ldd = ll_d2d(dentry);
-
- /* not the same process, don't mark */
- if (lli->lli_opendir_pid != current_pid())
- return;
-
- LASSERT(ldd);
- if (sai)
- ldd->lld_sa_generation = sai->sai_generation;
-}
-
+/*
+ * dentry may statahead when statahead is enabled and current process has opened
+ * parent directory, and this dentry hasn't accessed statahead cache before
+ */
static inline bool
-dentry_need_statahead(struct inode *dir, struct dentry *dentry)
+dentry_may_statahead(struct inode *dir, struct dentry *dentry)
{
struct ll_inode_info *lli;
struct ll_dentry_data *ldd;
@@ -1215,38 +1179,27 @@ dentry_need_statahead(struct inode *dir, struct dentry *dentry)
if (lli->lli_opendir_pid != current_pid())
return false;
- ldd = ll_d2d(dentry);
/*
- * When stats a dentry, the system trigger more than once "revalidate"
- * or "lookup", for "getattr", for "getxattr", and maybe for others.
- * Under patchless client mode, the operation intent is not accurate,
- * which maybe misguide the statahead thread. For example:
- * The "revalidate" call for "getattr" and "getxattr" of a dentry maybe
- * have the same operation intent -- "IT_GETATTR".
- * In fact, one dentry should has only one chance to interact with the
- * statahead thread, otherwise the statahead windows will be confused.
+ * When stating a dentry, kernel may trigger 'revalidate' or 'lookup'
+ * multiple times, eg. for 'getattr', 'getxattr' and etc.
+ * For patchless client, lookup intent is not accurate, which may
+ * misguide statahead. For example:
+ * The 'revalidate' call for 'getattr' and 'getxattr' of a dentry will
+ * have the same intent -- IT_GETATTR, while one dentry should access
+ * statahead cache once, otherwise statahead windows is messed up.
* The solution is as following:
- * Assign "lld_sa_generation" with "sai_generation" when a dentry
- * "IT_GETATTR" for the first time, and the subsequent "IT_GETATTR"
- * will bypass interacting with statahead thread for checking:
- * "lld_sa_generation == lli_sai->sai_generation"
+ * Assign 'lld_sa_generation' with 'lli_sa_generation' when a dentry
+ * IT_GETATTR for the first time, and subsequent IT_GETATTR will
+ * bypass interacting with statahead cache by checking
+ * 'lld_sa_generation == lli->lli_sa_generation'.
*/
- if (ldd && lli->lli_sai &&
- ldd->lld_sa_generation == lli->lli_sai->sai_generation)
+ ldd = ll_d2d(dentry);
+ if (ldd && ldd->lld_sa_generation == lli->lli_sa_generation)
return false;
return true;
}
-static inline int
-ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug)
-{
- if (!dentry_need_statahead(dir, *dentryp))
- return -EAGAIN;
-
- return do_statahead_enter(dir, dentryp, only_unplug);
-}
-
/* llite ioctl register support routine */
enum llioc_iter {
LLIOC_CONT = 0,
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 85f8ce7..494140a 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -522,8 +522,8 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
if (!it || it->it_op == IT_GETXATTR)
it = &lookup_it;
- if (it->it_op == IT_GETATTR) {
- rc = ll_statahead_enter(parent, &dentry, 0);
+ if (it->it_op == IT_GETATTR && dentry_may_statahead(parent, dentry)) {
+ rc = ll_statahead(parent, &dentry, 0);
if (rc == 1) {
if (dentry == save)
retval = NULL;
@@ -574,11 +574,8 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
retval = NULL;
else
retval = dentry;
- out:
- if (req)
- ptlrpc_req_finished(req);
- if (it->it_op == IT_GETATTR && (!retval || retval == dentry))
- ll_statahead_mark(parent, dentry);
+out:
+ ptlrpc_req_finished(req);
return retval;
}
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 323a175..9e76a68 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -54,12 +54,12 @@ enum se_stat {
/*
* sa_entry is not refcounted: statahead thread allocates it and do async stat,
* and in async stat callback ll_statahead_interpret() will add it into
- * sai_cb_entries, later statahead thread will call sa_handle_callback() to
+ * sai_interim_entries, later statahead thread will call sa_handle_callback() to
* instantiate entry and move it into sai_entries, and then only scanner process
* can access and free it.
*/
struct sa_entry {
- /* link into sai_cb_entries or sai_entries */
+ /* link into sai_interim_entries or sai_entries */
struct list_head se_list;
/* link into sai hash table locally */
struct list_head se_hash;
@@ -84,23 +84,20 @@ struct sa_entry {
static unsigned int sai_generation;
static DEFINE_SPINLOCK(sai_generation_lock);
-/*
- * The entry only can be released by the caller, it is necessary to hold lock.
- */
+/* sa_entry is ready to use */
static inline int sa_ready(struct sa_entry *entry)
{
smp_rmb();
return (entry->se_state != SA_ENTRY_INIT);
}
+/* hash value to put in sai_cache */
static inline int sa_hash(int val)
{
return val & LL_SA_CACHE_MASK;
}
-/*
- * Insert entry to hash SA table.
- */
+/* hash entry into sai_cache */
static inline void
sa_rehash(struct ll_statahead_info *sai, struct sa_entry *entry)
{
@@ -130,11 +127,13 @@ static inline int agl_should_run(struct ll_statahead_info *sai,
return (inode && S_ISREG(inode->i_mode) && sai->sai_agl_valid);
}
+/* statahead window is full */
static inline int sa_sent_full(struct ll_statahead_info *sai)
{
return atomic_read(&sai->sai_cache_count) >= sai->sai_max;
}
+/* got async stat replies */
static inline int sa_has_callback(struct ll_statahead_info *sai)
{
return !list_empty(&sai->sai_interim_entries);
@@ -158,7 +157,7 @@ static inline int sa_low_hit(struct ll_statahead_info *sai)
}
/*
- * If the given index is behind of statahead window more than
+ * if the given index is behind of statahead window more than
* SA_OMITTED_ENTRY_MAX, then it is old.
*/
static inline int is_omitted_entry(struct ll_statahead_info *sai, __u64 index)
@@ -167,9 +166,7 @@ static inline int is_omitted_entry(struct ll_statahead_info *sai, __u64 index)
sai->sai_index);
}
-/*
- * Insert it into sai_entries tail when init.
- */
+/* allocate sa_entry and hash it to allow scanner process to find it */
static struct sa_entry *
sa_alloc(struct dentry *parent, struct ll_statahead_info *sai, __u64 index,
const char *name, int len)
@@ -198,7 +195,7 @@ sa_alloc(struct dentry *parent, struct ll_statahead_info *sai, __u64 index,
entry->se_qstr.len = len;
entry->se_qstr.name = dname;
- lli = ll_i2info(sai->sai_inode);
+ lli = ll_i2info(sai->sai_dentry->d_inode);
spin_lock(&lli->lli_sa_lock);
INIT_LIST_HEAD(&entry->se_list);
sa_rehash(sai, entry);
@@ -246,7 +243,7 @@ sa_get(struct ll_statahead_info *sai, const struct qstr *qstr)
static inline void
sa_kill(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_dentry->d_inode);
LASSERT(!list_empty(&entry->se_hash));
LASSERT(!list_empty(&entry->se_list));
@@ -271,7 +268,7 @@ sa_put(struct ll_statahead_info *sai, struct sa_entry *entry)
struct sa_entry *tmp, *next;
if (entry && entry->se_state == SA_ENTRY_SUCC) {
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_dentry->d_inode);
sai->sai_hit++;
sai->sai_consecutive_miss = 0;
@@ -293,6 +290,7 @@ sa_put(struct ll_statahead_info *sai, struct sa_entry *entry)
break;
sa_kill(sai, tmp);
}
+
wake_up(&sai->sai_thread.t_ctl_waitq);
}
@@ -329,7 +327,7 @@ __sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
static void
sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
{
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_dentry->d_inode);
struct md_enqueue_info *minfo = entry->se_minfo;
struct ptlrpc_request *req = entry->se_req;
bool wakeup;
@@ -355,14 +353,12 @@ sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
wake_up(&sai->sai_waitq);
}
-/*
- * Insert inode into the list of sai_agls.
- */
+/* Insert inode into the list of sai_agls. */
static void ll_agl_add(struct ll_statahead_info *sai,
struct inode *inode, int index)
{
struct ll_inode_info *child = ll_i2info(inode);
- struct ll_inode_info *parent = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *parent = ll_i2info(sai->sai_dentry->d_inode);
int added = 0;
spin_lock(&child->lli_agl_lock);
@@ -387,8 +383,9 @@ static void ll_agl_add(struct ll_statahead_info *sai,
}
/* allocate sai */
-static struct ll_statahead_info *ll_sai_alloc(void)
+static struct ll_statahead_info *ll_sai_alloc(struct dentry *dentry)
{
+ struct ll_inode_info *lli = ll_i2info(dentry->d_inode);
struct ll_statahead_info *sai;
int i;
@@ -396,14 +393,9 @@ static struct ll_statahead_info *ll_sai_alloc(void)
if (!sai)
return NULL;
+ sai->sai_dentry = dget(dentry);
atomic_set(&sai->sai_refcount, 1);
- spin_lock(&sai_generation_lock);
- sai->sai_generation = ++sai_generation;
- if (unlikely(sai_generation == 0))
- sai->sai_generation = ++sai_generation;
- spin_unlock(&sai_generation_lock);
-
sai->sai_max = LL_SA_RPC_MIN;
sai->sai_index = 1;
init_waitqueue_head(&sai->sai_waitq);
@@ -420,9 +412,27 @@ static struct ll_statahead_info *ll_sai_alloc(void)
}
atomic_set(&sai->sai_cache_count, 0);
+ spin_lock(&sai_generation_lock);
+ lli->lli_sa_generation = ++sai_generation;
+ if (unlikely(!sai_generation))
+ lli->lli_sa_generation = ++sai_generation;
+ spin_unlock(&sai_generation_lock);
+
return sai;
}
+/* free sai */
+static inline void ll_sai_free(struct ll_statahead_info *sai)
+{
+ LASSERT(sai->sai_dentry);
+ dput(sai->sai_dentry);
+ kfree(sai);
+}
+
+/*
+ * take refcount of sai if sai for @dir exists, which means statahead is on for
+ * this directory.
+ */
static inline struct ll_statahead_info *ll_sai_get(struct inode *dir)
{
struct ll_inode_info *lli = ll_i2info(dir);
@@ -437,12 +447,16 @@ static inline struct ll_statahead_info *ll_sai_get(struct inode *dir)
return sai;
}
+/*
+ * put sai refcount after use, if refcount reaches zero, free sai and sa_entries
+ * attached to it.
+ */
static void ll_sai_put(struct ll_statahead_info *sai)
{
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_dentry->d_inode);
if (atomic_dec_and_lock(&sai->sai_refcount, &lli->lli_sa_lock)) {
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_dentry->d_inode);
struct sa_entry *entry, *next;
lli->lli_sai = NULL;
@@ -460,8 +474,7 @@ static void ll_sai_put(struct ll_statahead_info *sai)
LASSERT(atomic_read(&sai->sai_cache_count) == 0);
LASSERT(list_empty(&sai->sai_agls));
- iput(sai->sai_inode);
- kfree(sai);
+ ll_sai_free(sai);
atomic_dec(&sbi->ll_sa_running);
}
}
@@ -533,7 +546,7 @@ static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
static void sa_instantiate(struct ll_statahead_info *sai,
struct sa_entry *entry)
{
- struct inode *dir = sai->sai_inode;
+ struct inode *dir = sai->sai_dentry->d_inode;
struct inode *child;
struct md_enqueue_info *minfo;
struct lookup_intent *it;
@@ -609,12 +622,12 @@ out:
sa_make_ready(sai, entry, rc);
}
-/* once there are async stat replies, instantiate sa_entry */
+/* once there are async stat replies, instantiate sa_entry from replies */
static void sa_handle_callback(struct ll_statahead_info *sai)
{
struct ll_inode_info *lli;
- lli = ll_i2info(sai->sai_inode);
+ lli = ll_i2info(sai->sai_dentry->d_inode);
while (sa_has_callback(sai)) {
struct sa_entry *entry;
@@ -631,21 +644,6 @@ static void sa_handle_callback(struct ll_statahead_info *sai)
sa_instantiate(sai, entry);
}
-
- spin_lock(&lli->lli_agl_lock);
- while (!agl_list_empty(sai)) {
- struct ll_inode_info *clli;
-
- clli = list_entry(sai->sai_agls.next,
- struct ll_inode_info, lli_agl_list);
- list_del_init(&clli->lli_agl_list);
- spin_unlock(&lli->lli_agl_lock);
-
- ll_agl_trigger(&clli->lli_vfs_inode, sai);
-
- spin_lock(&lli->lli_agl_lock);
- }
- spin_unlock(&lli->lli_agl_lock);
}
/*
@@ -718,6 +716,7 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
return rc;
}
+/* finish async stat RPC arguments */
static void sa_fini_data(struct md_enqueue_info *minfo,
struct ldlm_enqueue_info *einfo)
{
@@ -775,6 +774,7 @@ static int sa_prep_data(struct inode *dir, struct inode *child,
return 0;
}
+/* async stat for file not found in dcache */
static int sa_lookup(struct inode *dir, struct sa_entry *entry)
{
struct md_enqueue_info *minfo;
@@ -786,17 +786,18 @@ static int sa_lookup(struct inode *dir, struct sa_entry *entry)
return rc;
rc = md_intent_getattr_async(ll_i2mdexp(dir), minfo, einfo);
- if (rc < 0)
+ if (rc)
sa_fini_data(minfo, einfo);
return rc;
}
/**
- * similar to ll_revalidate_it().
- * \retval 1 -- dentry valid
- * \retval 0 -- will send stat-ahead request
- * \retval others -- prepare stat-ahead request failed
+ * async stat for file found in dcache, similar to .revalidate
+ *
+ * \retval 1 dentry valid, no RPC sent
+ * \retval 0 dentry invalid, will send async stat RPC
+ * \retval negative number upon error
*/
static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
struct dentry *dentry)
@@ -831,7 +832,7 @@ static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
}
rc = md_intent_getattr_async(ll_i2mdexp(dir), minfo, einfo);
- if (rc < 0) {
+ if (rc) {
entry->se_inode = NULL;
iput(inode);
sa_fini_data(minfo, einfo);
@@ -840,6 +841,7 @@ static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
return rc;
}
+/* async stat for file with @name */
static void sa_statahead(struct dentry *parent, const char *name, int len)
{
struct inode *dir = d_inode(parent);
@@ -873,6 +875,7 @@ static void sa_statahead(struct dentry *parent, const char *name, int len)
sai->sai_index++;
}
+/* async glimpse (agl) thread main function */
static int ll_agl_thread(void *arg)
{
struct dentry *parent = arg;
@@ -946,6 +949,7 @@ static int ll_agl_thread(void *arg)
return 0;
}
+/* start agl thread */
static void ll_start_agl(struct dentry *parent, struct ll_statahead_info *sai)
{
struct ptlrpc_thread *thread = &sai->sai_agl_thread;
@@ -970,6 +974,7 @@ static void ll_start_agl(struct dentry *parent, struct ll_statahead_info *sai)
&lwi);
}
+/* statahead thread main function */
static int ll_statahead_thread(void *arg)
{
struct dentry *parent = arg;
@@ -977,7 +982,7 @@ static int ll_statahead_thread(void *arg)
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_sb_info *sbi = ll_i2sbi(dir);
struct ll_statahead_info *sai;
- struct ptlrpc_thread *thread;
+ struct ptlrpc_thread *sa_thread;
struct ptlrpc_thread *agl_thread;
struct page *page = NULL;
__u64 pos = 0;
@@ -987,9 +992,9 @@ static int ll_statahead_thread(void *arg)
struct l_wait_info lwi = { 0 };
sai = ll_sai_get(dir);
- thread = &sai->sai_thread;
+ sa_thread = &sai->sai_thread;
agl_thread = &sai->sai_agl_thread;
- thread->t_pid = current_pid();
+ sa_thread->t_pid = current_pid();
CDEBUG(D_READA, "statahead thread starting: sai %p, parent %pd\n",
sai, parent);
@@ -1007,16 +1012,16 @@ static int ll_statahead_thread(void *arg)
atomic_inc(&sbi->ll_sa_total);
spin_lock(&lli->lli_sa_lock);
- if (thread_is_init(thread))
+ if (thread_is_init(sa_thread))
/* If someone else has changed the thread state
* (e.g. already changed to SVC_STOPPING), we can't just
* blindly overwrite that setting.
*/
- thread_set_flags(thread, SVC_RUNNING);
+ thread_set_flags(sa_thread, SVC_RUNNING);
spin_unlock(&lli->lli_sa_lock);
- wake_up(&thread->t_ctl_waitq);
+ wake_up(&sa_thread->t_ctl_waitq);
- while (pos != MDS_DIR_END_OFF && thread_is_running(thread)) {
+ while (pos != MDS_DIR_END_OFF && thread_is_running(sa_thread)) {
struct lu_dirpage *dp;
struct lu_dirent *ent;
@@ -1033,7 +1038,7 @@ static int ll_statahead_thread(void *arg)
dp = page_address(page);
for (ent = lu_dirent_start(dp);
- ent && thread_is_running(thread) && !sa_low_hit(sai);
+ ent && thread_is_running(sa_thread) && !sa_low_hit(sai);
ent = lu_dirent_next(ent)) {
__u64 hash;
int namelen;
@@ -1082,15 +1087,32 @@ static int ll_statahead_thread(void *arg)
/* wait for spare statahead window */
do {
- l_wait_event(thread->t_ctl_waitq,
+ l_wait_event(sa_thread->t_ctl_waitq,
!sa_sent_full(sai) ||
sa_has_callback(sai) ||
!list_empty(&sai->sai_agls) ||
- !thread_is_running(thread),
+ !thread_is_running(sa_thread),
&lwi);
sa_handle_callback(sai);
+
+ spin_lock(&lli->lli_agl_lock);
+ while (sa_sent_full(sai) &&
+ !agl_list_empty(sai)) {
+ struct ll_inode_info *clli;
+
+ clli = list_entry(sai->sai_agls.next,
+ struct ll_inode_info, lli_agl_list);
+ list_del_init(&clli->lli_agl_list);
+ spin_unlock(&lli->lli_agl_lock);
+
+ ll_agl_trigger(&clli->lli_vfs_inode,
+ sai);
+
+ spin_lock(&lli->lli_agl_lock);
+ }
+ spin_unlock(&lli->lli_agl_lock);
} while (sa_sent_full(sai) &&
- thread_is_running(thread));
+ thread_is_running(sa_thread));
sa_statahead(parent, name, namelen);
}
@@ -1113,7 +1135,7 @@ static int ll_statahead_thread(void *arg)
if (rc < 0) {
spin_lock(&lli->lli_sa_lock);
- thread_set_flags(thread, SVC_STOPPING);
+ thread_set_flags(sa_thread, SVC_STOPPING);
lli->lli_sa_enabled = 0;
spin_unlock(&lli->lli_sa_lock);
}
@@ -1122,11 +1144,11 @@ static int ll_statahead_thread(void *arg)
* statahead is finished, but statahead entries need to be cached, wait
* for file release to stop me.
*/
- while (thread_is_running(thread)) {
- l_wait_event(thread->t_ctl_waitq,
+ while (thread_is_running(sa_thread)) {
+ l_wait_event(sa_thread->t_ctl_waitq,
sa_has_callback(sai) ||
!agl_list_empty(sai) ||
- !thread_is_running(thread),
+ !thread_is_running(sa_thread),
&lwi);
sa_handle_callback(sai);
@@ -1156,7 +1178,7 @@ out:
/* in case we're not woken up, timeout wait */
lwi = LWI_TIMEOUT(msecs_to_jiffies(MSEC_PER_SEC >> 3),
NULL, NULL);
- l_wait_event(thread->t_ctl_waitq,
+ l_wait_event(sa_thread->t_ctl_waitq,
sai->sai_sent == sai->sai_replied, &lwi);
}
@@ -1164,19 +1186,20 @@ out:
sa_handle_callback(sai);
spin_lock(&lli->lli_sa_lock);
- thread_set_flags(thread, SVC_STOPPED);
+ thread_set_flags(sa_thread, SVC_STOPPED);
spin_unlock(&lli->lli_sa_lock);
- wake_up(&sai->sai_waitq);
- wake_up(&thread->t_ctl_waitq);
- ll_sai_put(sai);
CDEBUG(D_READA, "statahead thread stopped: sai %p, parent %pd\n",
sai, parent);
- dput(parent);
+
+ wake_up(&sai->sai_waitq);
+ wake_up(&sa_thread->t_ctl_waitq);
+ ll_sai_put(sai);
+
return rc;
}
-/* authorize opened dir handle @key to statahead later */
+/* authorize opened dir handle @key to statahead */
void ll_authorize_statahead(struct inode *dir, void *key)
{
struct ll_inode_info *lli = ll_i2info(dir);
@@ -1230,7 +1253,7 @@ enum {
/**
* not first dirent, or is "."
*/
- LS_NONE_FIRST_DE = 0,
+ LS_NOT_FIRST_DE = 0,
/**
* the first non-hidden dirent
*/
@@ -1241,6 +1264,7 @@ enum {
LS_FIRST_DOT_DE
};
+/* file is first dirent under @dir */
static int is_first_dirent(struct inode *dir, struct dentry *dentry)
{
const struct qstr *target = &dentry->d_name;
@@ -1248,7 +1272,7 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
struct page *page;
__u64 pos = 0;
int dot_de;
- int rc = LS_NONE_FIRST_DE;
+ int rc = LS_NOT_FIRST_DE;
op_data = ll_prep_md_op_data(NULL, dir, dir, NULL, 0, 0,
LUSTRE_OPC_ANY, dir);
@@ -1324,7 +1348,7 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
if (target->len != namelen ||
memcmp(target->name, name, namelen) != 0)
- rc = LS_NONE_FIRST_DE;
+ rc = LS_NOT_FIRST_DE;
else if (!dot_de)
rc = LS_FIRST_DE;
else
@@ -1356,13 +1380,27 @@ out:
return rc;
}
+/**
+ * revalidate @dentryp from statahead cache
+ *
+ * \param[in] dir parent directory
+ * \param[in] sai sai structure
+ * \param[out] dentryp pointer to dentry which will be revalidated
+ * \param[in] unplug unplug statahead window only (normally for negative
+ * dentry)
+ * \retval 1 on success, dentry is saved in @dentryp
+ * \retval 0 if revalidation failed (no proper lock on client)
+ * \retval negative number upon error
+ */
static int revalidate_statahead_dentry(struct inode *dir,
struct ll_statahead_info *sai,
struct dentry **dentryp,
- int only_unplug)
+ bool unplug)
{
struct sa_entry *entry = NULL;
struct l_wait_info lwi = { 0 };
+ struct ll_dentry_data *ldd;
+ struct ll_inode_info *lli;
int rc = 0;
if ((*dentryp)->d_name.name[0] == '.') {
@@ -1392,10 +1430,15 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
+ if (unplug) {
+ rc = 1;
+ goto out_unplug;
+ }
+
entry = sa_get(sai, &(*dentryp)->d_name);
- if (!entry || only_unplug) {
- sa_put(sai, entry);
- return entry ? 1 : -EAGAIN;
+ if (!entry) {
+ rc = -EAGAIN;
+ goto out_unplug;
}
/* if statahead is busy in readdir, help it do post-work */
@@ -1406,13 +1449,15 @@ static int revalidate_statahead_dentry(struct inode *dir,
sai->sai_index_wait = entry->se_index;
lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
LWI_ON_SIGNAL_NOOP, NULL);
- rc = l_wait_event(sai->sai_waitq,
- sa_ready(entry) ||
- thread_is_stopped(&sai->sai_thread),
- &lwi);
+ rc = l_wait_event(sai->sai_waitq, sa_ready(entry), &lwi);
if (rc < 0) {
- sa_put(sai, entry);
- return -EAGAIN;
+ /*
+ * entry may not be ready, so it may be used by inflight
+ * statahead RPC, don't free it.
+ */
+ entry = NULL;
+ rc = -EAGAIN;
+ goto out_unplug;
}
}
@@ -1430,10 +1475,15 @@ static int revalidate_statahead_dentry(struct inode *dir,
alias = ll_splice_alias(inode, *dentryp);
if (IS_ERR(alias)) {
- sa_put(sai, entry);
- return PTR_ERR(alias);
+ rc = PTR_ERR(alias);
+ goto out_unplug;
}
*dentryp = alias;
+ /**
+ * statahead prepared this inode, transfer inode
+ * refcount from sa_entry to dentry
+ */
+ entry->se_inode = NULL;
} else if ((*dentryp)->d_inode != inode) {
/* revalidate, but inode is recreated */
CDEBUG(D_READA,
@@ -1445,10 +1495,7 @@ static int revalidate_statahead_dentry(struct inode *dir,
PFID(ll_inode2fid(inode)));
rc = -ESTALE;
goto out_unplug;
- } else {
- iput(inode);
}
- entry->se_inode = NULL;
if ((bits & MDS_INODELOCK_LOOKUP) &&
d_lustre_invalid(*dentryp))
@@ -1457,10 +1504,34 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
out_unplug:
+ /*
+ * statahead cached sa_entry can be used only once, and will be killed
+ * right after use, so if lookup/revalidate accessed statahead cache,
+ * set dentry ldd_sa_generation to parent lli_sa_generation, later if we
+ * stat this file again, we know we've done statahead before, see
+ * dentry_may_statahead().
+ */
+ ldd = ll_d2d(*dentryp);
+ lli = ll_i2info(dir);
+ /* ldd can be NULL if llite lookup failed. */
+ if (ldd)
+ ldd->lld_sa_generation = lli->lli_sa_generation;
sa_put(sai, entry);
return rc;
}
+/**
+ * start statahead thread
+ *
+ * \param[in] dir parent directory
+ * \param[in] dentry dentry that triggers statahead, normally the first
+ * dirent under @dir
+ * \retval -EAGAIN on success, because when this function is
+ * called, it's already in lookup call, so client should
+ * do it itself instead of waiting for statahead thread
+ * to do it asynchronously.
+ * \retval negative number upon error
+ */
static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
{
struct ll_inode_info *lli = ll_i2info(dir);
@@ -1468,60 +1539,34 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
struct l_wait_info lwi = { 0 };
struct ptlrpc_thread *thread;
struct task_struct *task;
- struct dentry *parent;
+ struct dentry *parent = dentry->d_parent;
int rc;
/* I am the "lli_opendir_pid" owner, only me can set "lli_sai". */
rc = is_first_dirent(dir, dentry);
- if (rc == LS_NONE_FIRST_DE) {
+ if (rc == LS_NOT_FIRST_DE) {
/* It is not "ls -{a}l" operation, no need statahead for it. */
- rc = -EAGAIN;
+ rc = -EFAULT;
goto out;
}
- sai = ll_sai_alloc();
+ sai = ll_sai_alloc(parent);
if (!sai) {
rc = -ENOMEM;
goto out;
}
sai->sai_ls_all = (rc == LS_FIRST_DOT_DE);
- sai->sai_inode = igrab(dir);
- if (unlikely(!sai->sai_inode)) {
- CWARN("Do not start stat ahead on dying inode "DFID"\n",
- PFID(&lli->lli_fid));
- rc = -ESTALE;
- goto out;
- }
-
- /* get parent reference count here, and put it in ll_statahead_thread */
- parent = dget(dentry->d_parent);
- if (unlikely(sai->sai_inode != d_inode(parent))) {
- struct ll_inode_info *nlli = ll_i2info(d_inode(parent));
-
- CWARN("Race condition, someone changed %pd just now: old parent "DFID", new parent "DFID"\n",
- dentry, PFID(&lli->lli_fid), PFID(&nlli->lli_fid));
- dput(parent);
- iput(sai->sai_inode);
- rc = -EAGAIN;
- goto out;
- }
-
- CDEBUG(D_READA, "start statahead thread: sai %p, parent %pd\n",
- sai, parent);
-
/*
- * if another process started statahead thread, or deauthorized current
- * lli_opendir_key, don't start statahead.
+ * if current lli_opendir_key was deauthorized, or dir re-opened by
+ * another process, don't start statahead, otherwise the newly spawned
+ * statahead thread won't be notified to quit.
*/
spin_lock(&lli->lli_sa_lock);
if (unlikely(lli->lli_sai || lli->lli_opendir_key ||
lli->lli_opendir_pid != current->pid)) {
spin_unlock(&lli->lli_sa_lock);
-
- dput(parent);
- iput(sai->sai_inode);
- rc = -EAGAIN;
+ rc = -EPERM;
goto out;
}
lli->lli_sai = sai;
@@ -1529,22 +1574,16 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
atomic_inc(&ll_i2sbi(parent->d_inode)->ll_sa_running);
+ CDEBUG(D_READA, "start statahead thread: [pid %d] [parent %pd]\n",
+ current_pid(), parent);
+
task = kthread_run(ll_statahead_thread, parent, "ll_sa_%u",
lli->lli_opendir_pid);
thread = &sai->sai_thread;
if (IS_ERR(task)) {
rc = PTR_ERR(task);
- CERROR("cannot start ll_sa thread: rc = %d\n", rc);
- dput(parent);
-
- spin_lock(&lli->lli_sa_lock);
- thread_set_flags(thread, SVC_STOPPED);
- thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
- spin_unlock(&lli->lli_sa_lock);
-
- ll_sai_put(sai);
- LASSERT(!lli->lli_sai);
- return -EAGAIN;
+ CERROR("can't start ll_sa thread, rc : %d\n", rc);
+ goto out;
}
l_wait_event(thread->t_ctl_waitq,
@@ -1559,29 +1598,35 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
return -EAGAIN;
out:
- kfree(sai);
/*
* once we start statahead thread failed, disable statahead so
- * subsequent won't waste time to try it.
+ * that subsequent stat won't waste time to try it.
*/
spin_lock(&lli->lli_sa_lock);
lli->lli_sa_enabled = 0;
+ lli->lli_sai = NULL;
spin_unlock(&lli->lli_sa_lock);
-
+ if (sai)
+ ll_sai_free(sai);
return rc;
}
/**
- * Start statahead thread if this is the first dir entry.
- * Otherwise if a thread is started already, wait it until it is ahead of me.
- * \retval 1 -- find entry with lock in cache, the caller needs to do
- * nothing.
- * \retval 0 -- find entry in cache, but without lock, the caller needs
- * refresh from MDS.
- * \retval others -- the caller need to process as non-statahead.
+ * statahead entry function, this is called when client getattr on a file, it
+ * will start statahead thread if this is the first dir entry, else revalidate
+ * dentry from statahead cache.
+ *
+ * \param[in] dir parent directory
+ * \param[out] dentryp dentry to getattr
+ * \param[in] unplug unplug statahead window only (normally for negative
+ * dentry)
+ * \retval 1 on success
+ * \retval 0 revalidation from statahead cache failed, caller needs
+ * to getattr from server directly
+ * \retval negative number on error, caller often ignores this and
+ * then getattr from server
*/
-int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
- int only_unplug)
+int ll_statahead(struct inode *dir, struct dentry **dentryp, bool unplug)
{
struct ll_statahead_info *sai;
@@ -1589,13 +1634,11 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
if (sai) {
int rc;
- rc = revalidate_statahead_dentry(dir, sai, dentryp,
- only_unplug);
+ rc = revalidate_statahead_dentry(dir, sai, dentryp, unplug);
CDEBUG(D_READA, "revalidate statahead %pd: %d.\n",
*dentryp, rc);
ll_sai_put(sai);
return rc;
}
-
return start_statahead_thread(dir, *dentryp);
}
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 166f0c4..125d882 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1367,7 +1367,7 @@ static int mdc_read_page(struct obd_export *exp, struct md_op_data *op_data,
page = mdc_page_locate(mapping, &rp_param.rp_off, &start, &end,
rp_param.rp_hash64);
if (IS_ERR(page)) {
- CERROR("%s: dir page locate: "DFID" at %llu: rc %ld\n",
+ CDEBUG(D_INFO, "%s: dir page locate: " DFID " at %llu: rc %ld\n",
exp->exp_obd->obd_name, PFID(&op_data->op_fid1),
rp_param.rp_off, PTR_ERR(page));
rc = PTR_ERR(page);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 070/124] staging: lustre: statahead: small fixes and cleanup
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Lai Siyao,
James Simmons
From: Lai Siyao <lai.siyao@intel.com>
small fixes:
* when 'unplug' is set for ll_statahead(), sa_put() shouldn't kill
the entry found, because its inflight RPC may not finish yet.
* remove 'sai_generation', add 'lli_sa_generation' because the
former one is not safe to access without lock.
* revalidate_statahead_dentry() may fail to wait for statahead
entry to become ready, in this case it should not release this
entry, because it may be used by inflight statahead RPC.
cleanups:
* rename ll_statahead_enter() to ll_statahead().
* move dentry 'lld_sa_generation' update to ll_statahead() to
simplify code and logic.
* other small cleanups.
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3270
Reviewed-on: http://review.whamcloud.com/9667
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6222
Reviewed-on: http://review.whamcloud.com/13708
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/dcache.c | 5 +-
.../staging/lustre/lustre/llite/llite_internal.h | 137 +++-----
drivers/staging/lustre/lustre/llite/namei.c | 11 +-
drivers/staging/lustre/lustre/llite/statahead.c | 353 +++++++++++---------
drivers/staging/lustre/lustre/mdc/mdc_request.c | 2 +-
5 files changed, 250 insertions(+), 258 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index 8500080..0e45d8f 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -278,14 +278,13 @@ static int ll_revalidate_dentry(struct dentry *dentry,
if (lookup_flags & (LOOKUP_PARENT | LOOKUP_OPEN | LOOKUP_CREATE))
return 1;
- if (!dentry_need_statahead(dir, dentry))
+ if (!dentry_may_statahead(dir, dentry))
return 1;
if (lookup_flags & LOOKUP_RCU)
return -ECHILD;
- do_statahead_enter(dir, &dentry, !d_inode(dentry));
- ll_statahead_mark(dir, dentry);
+ ll_statahead(dir, &dentry, !d_inode(dentry));
return 1;
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index a68bea1..bdfdff5 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -161,7 +161,7 @@ struct ll_inode_info {
/* for directory */
struct {
/* serialize normal readdir and statahead-readdir. */
- struct mutex d_readdir_mutex;
+ struct mutex lli_readdir_mutex;
/* metadata statahead */
/* since parent-child threads can share the same @file
@@ -169,44 +169,35 @@ struct ll_inode_info {
* case of parent exit before child -- it is me should
* cleanup the dir readahead.
*/
- void *d_opendir_key;
- struct ll_statahead_info *d_sai;
+ void *lli_opendir_key;
+ struct ll_statahead_info *lli_sai;
/* protect statahead stuff. */
- spinlock_t d_sa_lock;
+ spinlock_t lli_sa_lock;
/* "opendir_pid" is the token when lookup/revalidate
* -- I am the owner of dir statahead.
*/
- pid_t d_opendir_pid;
+ pid_t lli_opendir_pid;
/* stat will try to access statahead entries or start
* statahead if this flag is set, and this flag will be
* set upon dir open, and cleared when dir is closed,
* statahead hit ratio is too low, or start statahead
* thread failed.
*/
- unsigned int d_sa_enabled:1;
+ unsigned int lli_sa_enabled:1;
+ /* generation for statahead */
+ unsigned int lli_sa_generation;
/* directory stripe information */
- struct lmv_stripe_md *d_lsm_md;
+ struct lmv_stripe_md *lli_lsm_md;
/* striped directory size */
- loff_t d_stripe_size;
- /* striped directory nlink */
- __u64 d_stripe_nlink;
- } d;
-
-#define lli_readdir_mutex u.d.d_readdir_mutex
-#define lli_opendir_key u.d.d_opendir_key
-#define lli_sai u.d.d_sai
-#define lli_sa_lock u.d.d_sa_lock
-#define lli_sa_enabled u.d.d_sa_enabled
-#define lli_opendir_pid u.d.d_opendir_pid
-#define lli_lsm_md u.d.d_lsm_md
-#define lli_stripe_dir_size u.d.d_stripe_size
-#define lli_stripe_dir_nlink u.d.d_stripe_nlink
+ loff_t lli_stripe_dir_size;
+ u64 lli_stripe_dir_nlink;
+ };
/* for non-directory */
struct {
- struct mutex f_size_mutex;
- char *f_symlink_name;
- __u64 f_maxbytes;
+ struct mutex lli_size_mutex;
+ char *lli_symlink_name;
+ __u64 lli_maxbytes;
/*
* struct rw_semaphore {
* signed long count; // align d.d_def_acl
@@ -214,16 +205,16 @@ struct ll_inode_info {
* struct list_head wait_list;
* }
*/
- struct rw_semaphore f_trunc_sem;
- struct range_lock_tree f_write_tree;
+ struct rw_semaphore lli_trunc_sem;
+ struct range_lock_tree lli_write_tree;
- struct rw_semaphore f_glimpse_sem;
- unsigned long f_glimpse_time;
- struct list_head f_agl_list;
- __u64 f_agl_index;
+ struct rw_semaphore lli_glimpse_sem;
+ unsigned long lli_glimpse_time;
+ struct list_head lli_agl_list;
+ __u64 lli_agl_index;
/* for writepage() only to communicate to fsync */
- int f_async_rc;
+ int lli_async_rc;
/*
* whenever a process try to read/write the file, the
@@ -233,22 +224,9 @@ struct ll_inode_info {
* so the read/write statistics for jobid will not be
* accurate if the file is shared by different jobs.
*/
- char f_jobid[LUSTRE_JOBID_SIZE];
- } f;
-
-#define lli_size_mutex u.f.f_size_mutex
-#define lli_symlink_name u.f.f_symlink_name
-#define lli_maxbytes u.f.f_maxbytes
-#define lli_trunc_sem u.f.f_trunc_sem
-#define lli_write_tree u.f.f_write_tree
-#define lli_glimpse_sem u.f.f_glimpse_sem
-#define lli_glimpse_time u.f.f_glimpse_time
-#define lli_agl_list u.f.f_agl_list
-#define lli_agl_index u.f.f_agl_index
-#define lli_async_rc u.f.f_async_rc
-#define lli_jobid u.f.f_jobid
-
- } u;
+ char lli_jobid[LUSTRE_JOBID_SIZE];
+ };
+ };
/* XXX: For following frequent used members, although they maybe special
* used for non-directory object, it is some time-wasting to check
@@ -1095,11 +1073,10 @@ void ll_ra_stats_inc(struct inode *inode, enum ra_stat which);
/* per inode struct, for dir only */
struct ll_statahead_info {
- struct inode *sai_inode;
+ struct dentry *sai_dentry;
atomic_t sai_refcount; /* when access this struct, hold
* refcount
*/
- unsigned int sai_generation; /* generation for statahead */
unsigned int sai_max; /* max ahead of lookup */
__u64 sai_sent; /* stat requests sent count */
__u64 sai_replied; /* stat requests which received
@@ -1142,8 +1119,7 @@ struct ll_statahead_info {
atomic_t sai_cache_count; /* entry count in cache */
};
-int do_statahead_enter(struct inode *dir, struct dentry **dentry,
- int only_unplug);
+int ll_statahead(struct inode *dir, struct dentry **dentry, bool unplug);
void ll_authorize_statahead(struct inode *dir, void *key);
void ll_deauthorize_statahead(struct inode *dir, void *key);
@@ -1175,24 +1151,12 @@ static inline int ll_glimpse_size(struct inode *inode)
return rc;
}
-static inline void
-ll_statahead_mark(struct inode *dir, struct dentry *dentry)
-{
- struct ll_inode_info *lli = ll_i2info(dir);
- struct ll_statahead_info *sai = lli->lli_sai;
- struct ll_dentry_data *ldd = ll_d2d(dentry);
-
- /* not the same process, don't mark */
- if (lli->lli_opendir_pid != current_pid())
- return;
-
- LASSERT(ldd);
- if (sai)
- ldd->lld_sa_generation = sai->sai_generation;
-}
-
+/*
+ * dentry may statahead when statahead is enabled and current process has opened
+ * parent directory, and this dentry hasn't accessed statahead cache before
+ */
static inline bool
-dentry_need_statahead(struct inode *dir, struct dentry *dentry)
+dentry_may_statahead(struct inode *dir, struct dentry *dentry)
{
struct ll_inode_info *lli;
struct ll_dentry_data *ldd;
@@ -1215,38 +1179,27 @@ dentry_need_statahead(struct inode *dir, struct dentry *dentry)
if (lli->lli_opendir_pid != current_pid())
return false;
- ldd = ll_d2d(dentry);
/*
- * When stats a dentry, the system trigger more than once "revalidate"
- * or "lookup", for "getattr", for "getxattr", and maybe for others.
- * Under patchless client mode, the operation intent is not accurate,
- * which maybe misguide the statahead thread. For example:
- * The "revalidate" call for "getattr" and "getxattr" of a dentry maybe
- * have the same operation intent -- "IT_GETATTR".
- * In fact, one dentry should has only one chance to interact with the
- * statahead thread, otherwise the statahead windows will be confused.
+ * When stating a dentry, kernel may trigger 'revalidate' or 'lookup'
+ * multiple times, eg. for 'getattr', 'getxattr' and etc.
+ * For patchless client, lookup intent is not accurate, which may
+ * misguide statahead. For example:
+ * The 'revalidate' call for 'getattr' and 'getxattr' of a dentry will
+ * have the same intent -- IT_GETATTR, while one dentry should access
+ * statahead cache once, otherwise statahead windows is messed up.
* The solution is as following:
- * Assign "lld_sa_generation" with "sai_generation" when a dentry
- * "IT_GETATTR" for the first time, and the subsequent "IT_GETATTR"
- * will bypass interacting with statahead thread for checking:
- * "lld_sa_generation == lli_sai->sai_generation"
+ * Assign 'lld_sa_generation' with 'lli_sa_generation' when a dentry
+ * IT_GETATTR for the first time, and subsequent IT_GETATTR will
+ * bypass interacting with statahead cache by checking
+ * 'lld_sa_generation == lli->lli_sa_generation'.
*/
- if (ldd && lli->lli_sai &&
- ldd->lld_sa_generation == lli->lli_sai->sai_generation)
+ ldd = ll_d2d(dentry);
+ if (ldd && ldd->lld_sa_generation == lli->lli_sa_generation)
return false;
return true;
}
-static inline int
-ll_statahead_enter(struct inode *dir, struct dentry **dentryp, int only_unplug)
-{
- if (!dentry_need_statahead(dir, *dentryp))
- return -EAGAIN;
-
- return do_statahead_enter(dir, dentryp, only_unplug);
-}
-
/* llite ioctl register support routine */
enum llioc_iter {
LLIOC_CONT = 0,
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 85f8ce7..494140a 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -522,8 +522,8 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
if (!it || it->it_op == IT_GETXATTR)
it = &lookup_it;
- if (it->it_op == IT_GETATTR) {
- rc = ll_statahead_enter(parent, &dentry, 0);
+ if (it->it_op == IT_GETATTR && dentry_may_statahead(parent, dentry)) {
+ rc = ll_statahead(parent, &dentry, 0);
if (rc == 1) {
if (dentry == save)
retval = NULL;
@@ -574,11 +574,8 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
retval = NULL;
else
retval = dentry;
- out:
- if (req)
- ptlrpc_req_finished(req);
- if (it->it_op == IT_GETATTR && (!retval || retval == dentry))
- ll_statahead_mark(parent, dentry);
+out:
+ ptlrpc_req_finished(req);
return retval;
}
diff --git a/drivers/staging/lustre/lustre/llite/statahead.c b/drivers/staging/lustre/lustre/llite/statahead.c
index 323a175..9e76a68 100644
--- a/drivers/staging/lustre/lustre/llite/statahead.c
+++ b/drivers/staging/lustre/lustre/llite/statahead.c
@@ -54,12 +54,12 @@ enum se_stat {
/*
* sa_entry is not refcounted: statahead thread allocates it and do async stat,
* and in async stat callback ll_statahead_interpret() will add it into
- * sai_cb_entries, later statahead thread will call sa_handle_callback() to
+ * sai_interim_entries, later statahead thread will call sa_handle_callback() to
* instantiate entry and move it into sai_entries, and then only scanner process
* can access and free it.
*/
struct sa_entry {
- /* link into sai_cb_entries or sai_entries */
+ /* link into sai_interim_entries or sai_entries */
struct list_head se_list;
/* link into sai hash table locally */
struct list_head se_hash;
@@ -84,23 +84,20 @@ struct sa_entry {
static unsigned int sai_generation;
static DEFINE_SPINLOCK(sai_generation_lock);
-/*
- * The entry only can be released by the caller, it is necessary to hold lock.
- */
+/* sa_entry is ready to use */
static inline int sa_ready(struct sa_entry *entry)
{
smp_rmb();
return (entry->se_state != SA_ENTRY_INIT);
}
+/* hash value to put in sai_cache */
static inline int sa_hash(int val)
{
return val & LL_SA_CACHE_MASK;
}
-/*
- * Insert entry to hash SA table.
- */
+/* hash entry into sai_cache */
static inline void
sa_rehash(struct ll_statahead_info *sai, struct sa_entry *entry)
{
@@ -130,11 +127,13 @@ static inline int agl_should_run(struct ll_statahead_info *sai,
return (inode && S_ISREG(inode->i_mode) && sai->sai_agl_valid);
}
+/* statahead window is full */
static inline int sa_sent_full(struct ll_statahead_info *sai)
{
return atomic_read(&sai->sai_cache_count) >= sai->sai_max;
}
+/* got async stat replies */
static inline int sa_has_callback(struct ll_statahead_info *sai)
{
return !list_empty(&sai->sai_interim_entries);
@@ -158,7 +157,7 @@ static inline int sa_low_hit(struct ll_statahead_info *sai)
}
/*
- * If the given index is behind of statahead window more than
+ * if the given index is behind of statahead window more than
* SA_OMITTED_ENTRY_MAX, then it is old.
*/
static inline int is_omitted_entry(struct ll_statahead_info *sai, __u64 index)
@@ -167,9 +166,7 @@ static inline int is_omitted_entry(struct ll_statahead_info *sai, __u64 index)
sai->sai_index);
}
-/*
- * Insert it into sai_entries tail when init.
- */
+/* allocate sa_entry and hash it to allow scanner process to find it */
static struct sa_entry *
sa_alloc(struct dentry *parent, struct ll_statahead_info *sai, __u64 index,
const char *name, int len)
@@ -198,7 +195,7 @@ sa_alloc(struct dentry *parent, struct ll_statahead_info *sai, __u64 index,
entry->se_qstr.len = len;
entry->se_qstr.name = dname;
- lli = ll_i2info(sai->sai_inode);
+ lli = ll_i2info(sai->sai_dentry->d_inode);
spin_lock(&lli->lli_sa_lock);
INIT_LIST_HEAD(&entry->se_list);
sa_rehash(sai, entry);
@@ -246,7 +243,7 @@ sa_get(struct ll_statahead_info *sai, const struct qstr *qstr)
static inline void
sa_kill(struct ll_statahead_info *sai, struct sa_entry *entry)
{
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_dentry->d_inode);
LASSERT(!list_empty(&entry->se_hash));
LASSERT(!list_empty(&entry->se_list));
@@ -271,7 +268,7 @@ sa_put(struct ll_statahead_info *sai, struct sa_entry *entry)
struct sa_entry *tmp, *next;
if (entry && entry->se_state == SA_ENTRY_SUCC) {
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_dentry->d_inode);
sai->sai_hit++;
sai->sai_consecutive_miss = 0;
@@ -293,6 +290,7 @@ sa_put(struct ll_statahead_info *sai, struct sa_entry *entry)
break;
sa_kill(sai, tmp);
}
+
wake_up(&sai->sai_thread.t_ctl_waitq);
}
@@ -329,7 +327,7 @@ __sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
static void
sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
{
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_dentry->d_inode);
struct md_enqueue_info *minfo = entry->se_minfo;
struct ptlrpc_request *req = entry->se_req;
bool wakeup;
@@ -355,14 +353,12 @@ sa_make_ready(struct ll_statahead_info *sai, struct sa_entry *entry, int ret)
wake_up(&sai->sai_waitq);
}
-/*
- * Insert inode into the list of sai_agls.
- */
+/* Insert inode into the list of sai_agls. */
static void ll_agl_add(struct ll_statahead_info *sai,
struct inode *inode, int index)
{
struct ll_inode_info *child = ll_i2info(inode);
- struct ll_inode_info *parent = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *parent = ll_i2info(sai->sai_dentry->d_inode);
int added = 0;
spin_lock(&child->lli_agl_lock);
@@ -387,8 +383,9 @@ static void ll_agl_add(struct ll_statahead_info *sai,
}
/* allocate sai */
-static struct ll_statahead_info *ll_sai_alloc(void)
+static struct ll_statahead_info *ll_sai_alloc(struct dentry *dentry)
{
+ struct ll_inode_info *lli = ll_i2info(dentry->d_inode);
struct ll_statahead_info *sai;
int i;
@@ -396,14 +393,9 @@ static struct ll_statahead_info *ll_sai_alloc(void)
if (!sai)
return NULL;
+ sai->sai_dentry = dget(dentry);
atomic_set(&sai->sai_refcount, 1);
- spin_lock(&sai_generation_lock);
- sai->sai_generation = ++sai_generation;
- if (unlikely(sai_generation == 0))
- sai->sai_generation = ++sai_generation;
- spin_unlock(&sai_generation_lock);
-
sai->sai_max = LL_SA_RPC_MIN;
sai->sai_index = 1;
init_waitqueue_head(&sai->sai_waitq);
@@ -420,9 +412,27 @@ static struct ll_statahead_info *ll_sai_alloc(void)
}
atomic_set(&sai->sai_cache_count, 0);
+ spin_lock(&sai_generation_lock);
+ lli->lli_sa_generation = ++sai_generation;
+ if (unlikely(!sai_generation))
+ lli->lli_sa_generation = ++sai_generation;
+ spin_unlock(&sai_generation_lock);
+
return sai;
}
+/* free sai */
+static inline void ll_sai_free(struct ll_statahead_info *sai)
+{
+ LASSERT(sai->sai_dentry);
+ dput(sai->sai_dentry);
+ kfree(sai);
+}
+
+/*
+ * take refcount of sai if sai for @dir exists, which means statahead is on for
+ * this directory.
+ */
static inline struct ll_statahead_info *ll_sai_get(struct inode *dir)
{
struct ll_inode_info *lli = ll_i2info(dir);
@@ -437,12 +447,16 @@ static inline struct ll_statahead_info *ll_sai_get(struct inode *dir)
return sai;
}
+/*
+ * put sai refcount after use, if refcount reaches zero, free sai and sa_entries
+ * attached to it.
+ */
static void ll_sai_put(struct ll_statahead_info *sai)
{
- struct ll_inode_info *lli = ll_i2info(sai->sai_inode);
+ struct ll_inode_info *lli = ll_i2info(sai->sai_dentry->d_inode);
if (atomic_dec_and_lock(&sai->sai_refcount, &lli->lli_sa_lock)) {
- struct ll_sb_info *sbi = ll_i2sbi(sai->sai_inode);
+ struct ll_sb_info *sbi = ll_i2sbi(sai->sai_dentry->d_inode);
struct sa_entry *entry, *next;
lli->lli_sai = NULL;
@@ -460,8 +474,7 @@ static void ll_sai_put(struct ll_statahead_info *sai)
LASSERT(atomic_read(&sai->sai_cache_count) == 0);
LASSERT(list_empty(&sai->sai_agls));
- iput(sai->sai_inode);
- kfree(sai);
+ ll_sai_free(sai);
atomic_dec(&sbi->ll_sa_running);
}
}
@@ -533,7 +546,7 @@ static void ll_agl_trigger(struct inode *inode, struct ll_statahead_info *sai)
static void sa_instantiate(struct ll_statahead_info *sai,
struct sa_entry *entry)
{
- struct inode *dir = sai->sai_inode;
+ struct inode *dir = sai->sai_dentry->d_inode;
struct inode *child;
struct md_enqueue_info *minfo;
struct lookup_intent *it;
@@ -609,12 +622,12 @@ out:
sa_make_ready(sai, entry, rc);
}
-/* once there are async stat replies, instantiate sa_entry */
+/* once there are async stat replies, instantiate sa_entry from replies */
static void sa_handle_callback(struct ll_statahead_info *sai)
{
struct ll_inode_info *lli;
- lli = ll_i2info(sai->sai_inode);
+ lli = ll_i2info(sai->sai_dentry->d_inode);
while (sa_has_callback(sai)) {
struct sa_entry *entry;
@@ -631,21 +644,6 @@ static void sa_handle_callback(struct ll_statahead_info *sai)
sa_instantiate(sai, entry);
}
-
- spin_lock(&lli->lli_agl_lock);
- while (!agl_list_empty(sai)) {
- struct ll_inode_info *clli;
-
- clli = list_entry(sai->sai_agls.next,
- struct ll_inode_info, lli_agl_list);
- list_del_init(&clli->lli_agl_list);
- spin_unlock(&lli->lli_agl_lock);
-
- ll_agl_trigger(&clli->lli_vfs_inode, sai);
-
- spin_lock(&lli->lli_agl_lock);
- }
- spin_unlock(&lli->lli_agl_lock);
}
/*
@@ -718,6 +716,7 @@ static int ll_statahead_interpret(struct ptlrpc_request *req,
return rc;
}
+/* finish async stat RPC arguments */
static void sa_fini_data(struct md_enqueue_info *minfo,
struct ldlm_enqueue_info *einfo)
{
@@ -775,6 +774,7 @@ static int sa_prep_data(struct inode *dir, struct inode *child,
return 0;
}
+/* async stat for file not found in dcache */
static int sa_lookup(struct inode *dir, struct sa_entry *entry)
{
struct md_enqueue_info *minfo;
@@ -786,17 +786,18 @@ static int sa_lookup(struct inode *dir, struct sa_entry *entry)
return rc;
rc = md_intent_getattr_async(ll_i2mdexp(dir), minfo, einfo);
- if (rc < 0)
+ if (rc)
sa_fini_data(minfo, einfo);
return rc;
}
/**
- * similar to ll_revalidate_it().
- * \retval 1 -- dentry valid
- * \retval 0 -- will send stat-ahead request
- * \retval others -- prepare stat-ahead request failed
+ * async stat for file found in dcache, similar to .revalidate
+ *
+ * \retval 1 dentry valid, no RPC sent
+ * \retval 0 dentry invalid, will send async stat RPC
+ * \retval negative number upon error
*/
static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
struct dentry *dentry)
@@ -831,7 +832,7 @@ static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
}
rc = md_intent_getattr_async(ll_i2mdexp(dir), minfo, einfo);
- if (rc < 0) {
+ if (rc) {
entry->se_inode = NULL;
iput(inode);
sa_fini_data(minfo, einfo);
@@ -840,6 +841,7 @@ static int sa_revalidate(struct inode *dir, struct sa_entry *entry,
return rc;
}
+/* async stat for file with @name */
static void sa_statahead(struct dentry *parent, const char *name, int len)
{
struct inode *dir = d_inode(parent);
@@ -873,6 +875,7 @@ static void sa_statahead(struct dentry *parent, const char *name, int len)
sai->sai_index++;
}
+/* async glimpse (agl) thread main function */
static int ll_agl_thread(void *arg)
{
struct dentry *parent = arg;
@@ -946,6 +949,7 @@ static int ll_agl_thread(void *arg)
return 0;
}
+/* start agl thread */
static void ll_start_agl(struct dentry *parent, struct ll_statahead_info *sai)
{
struct ptlrpc_thread *thread = &sai->sai_agl_thread;
@@ -970,6 +974,7 @@ static void ll_start_agl(struct dentry *parent, struct ll_statahead_info *sai)
&lwi);
}
+/* statahead thread main function */
static int ll_statahead_thread(void *arg)
{
struct dentry *parent = arg;
@@ -977,7 +982,7 @@ static int ll_statahead_thread(void *arg)
struct ll_inode_info *lli = ll_i2info(dir);
struct ll_sb_info *sbi = ll_i2sbi(dir);
struct ll_statahead_info *sai;
- struct ptlrpc_thread *thread;
+ struct ptlrpc_thread *sa_thread;
struct ptlrpc_thread *agl_thread;
struct page *page = NULL;
__u64 pos = 0;
@@ -987,9 +992,9 @@ static int ll_statahead_thread(void *arg)
struct l_wait_info lwi = { 0 };
sai = ll_sai_get(dir);
- thread = &sai->sai_thread;
+ sa_thread = &sai->sai_thread;
agl_thread = &sai->sai_agl_thread;
- thread->t_pid = current_pid();
+ sa_thread->t_pid = current_pid();
CDEBUG(D_READA, "statahead thread starting: sai %p, parent %pd\n",
sai, parent);
@@ -1007,16 +1012,16 @@ static int ll_statahead_thread(void *arg)
atomic_inc(&sbi->ll_sa_total);
spin_lock(&lli->lli_sa_lock);
- if (thread_is_init(thread))
+ if (thread_is_init(sa_thread))
/* If someone else has changed the thread state
* (e.g. already changed to SVC_STOPPING), we can't just
* blindly overwrite that setting.
*/
- thread_set_flags(thread, SVC_RUNNING);
+ thread_set_flags(sa_thread, SVC_RUNNING);
spin_unlock(&lli->lli_sa_lock);
- wake_up(&thread->t_ctl_waitq);
+ wake_up(&sa_thread->t_ctl_waitq);
- while (pos != MDS_DIR_END_OFF && thread_is_running(thread)) {
+ while (pos != MDS_DIR_END_OFF && thread_is_running(sa_thread)) {
struct lu_dirpage *dp;
struct lu_dirent *ent;
@@ -1033,7 +1038,7 @@ static int ll_statahead_thread(void *arg)
dp = page_address(page);
for (ent = lu_dirent_start(dp);
- ent && thread_is_running(thread) && !sa_low_hit(sai);
+ ent && thread_is_running(sa_thread) && !sa_low_hit(sai);
ent = lu_dirent_next(ent)) {
__u64 hash;
int namelen;
@@ -1082,15 +1087,32 @@ static int ll_statahead_thread(void *arg)
/* wait for spare statahead window */
do {
- l_wait_event(thread->t_ctl_waitq,
+ l_wait_event(sa_thread->t_ctl_waitq,
!sa_sent_full(sai) ||
sa_has_callback(sai) ||
!list_empty(&sai->sai_agls) ||
- !thread_is_running(thread),
+ !thread_is_running(sa_thread),
&lwi);
sa_handle_callback(sai);
+
+ spin_lock(&lli->lli_agl_lock);
+ while (sa_sent_full(sai) &&
+ !agl_list_empty(sai)) {
+ struct ll_inode_info *clli;
+
+ clli = list_entry(sai->sai_agls.next,
+ struct ll_inode_info, lli_agl_list);
+ list_del_init(&clli->lli_agl_list);
+ spin_unlock(&lli->lli_agl_lock);
+
+ ll_agl_trigger(&clli->lli_vfs_inode,
+ sai);
+
+ spin_lock(&lli->lli_agl_lock);
+ }
+ spin_unlock(&lli->lli_agl_lock);
} while (sa_sent_full(sai) &&
- thread_is_running(thread));
+ thread_is_running(sa_thread));
sa_statahead(parent, name, namelen);
}
@@ -1113,7 +1135,7 @@ static int ll_statahead_thread(void *arg)
if (rc < 0) {
spin_lock(&lli->lli_sa_lock);
- thread_set_flags(thread, SVC_STOPPING);
+ thread_set_flags(sa_thread, SVC_STOPPING);
lli->lli_sa_enabled = 0;
spin_unlock(&lli->lli_sa_lock);
}
@@ -1122,11 +1144,11 @@ static int ll_statahead_thread(void *arg)
* statahead is finished, but statahead entries need to be cached, wait
* for file release to stop me.
*/
- while (thread_is_running(thread)) {
- l_wait_event(thread->t_ctl_waitq,
+ while (thread_is_running(sa_thread)) {
+ l_wait_event(sa_thread->t_ctl_waitq,
sa_has_callback(sai) ||
!agl_list_empty(sai) ||
- !thread_is_running(thread),
+ !thread_is_running(sa_thread),
&lwi);
sa_handle_callback(sai);
@@ -1156,7 +1178,7 @@ out:
/* in case we're not woken up, timeout wait */
lwi = LWI_TIMEOUT(msecs_to_jiffies(MSEC_PER_SEC >> 3),
NULL, NULL);
- l_wait_event(thread->t_ctl_waitq,
+ l_wait_event(sa_thread->t_ctl_waitq,
sai->sai_sent == sai->sai_replied, &lwi);
}
@@ -1164,19 +1186,20 @@ out:
sa_handle_callback(sai);
spin_lock(&lli->lli_sa_lock);
- thread_set_flags(thread, SVC_STOPPED);
+ thread_set_flags(sa_thread, SVC_STOPPED);
spin_unlock(&lli->lli_sa_lock);
- wake_up(&sai->sai_waitq);
- wake_up(&thread->t_ctl_waitq);
- ll_sai_put(sai);
CDEBUG(D_READA, "statahead thread stopped: sai %p, parent %pd\n",
sai, parent);
- dput(parent);
+
+ wake_up(&sai->sai_waitq);
+ wake_up(&sa_thread->t_ctl_waitq);
+ ll_sai_put(sai);
+
return rc;
}
-/* authorize opened dir handle @key to statahead later */
+/* authorize opened dir handle @key to statahead */
void ll_authorize_statahead(struct inode *dir, void *key)
{
struct ll_inode_info *lli = ll_i2info(dir);
@@ -1230,7 +1253,7 @@ enum {
/**
* not first dirent, or is "."
*/
- LS_NONE_FIRST_DE = 0,
+ LS_NOT_FIRST_DE = 0,
/**
* the first non-hidden dirent
*/
@@ -1241,6 +1264,7 @@ enum {
LS_FIRST_DOT_DE
};
+/* file is first dirent under @dir */
static int is_first_dirent(struct inode *dir, struct dentry *dentry)
{
const struct qstr *target = &dentry->d_name;
@@ -1248,7 +1272,7 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
struct page *page;
__u64 pos = 0;
int dot_de;
- int rc = LS_NONE_FIRST_DE;
+ int rc = LS_NOT_FIRST_DE;
op_data = ll_prep_md_op_data(NULL, dir, dir, NULL, 0, 0,
LUSTRE_OPC_ANY, dir);
@@ -1324,7 +1348,7 @@ static int is_first_dirent(struct inode *dir, struct dentry *dentry)
if (target->len != namelen ||
memcmp(target->name, name, namelen) != 0)
- rc = LS_NONE_FIRST_DE;
+ rc = LS_NOT_FIRST_DE;
else if (!dot_de)
rc = LS_FIRST_DE;
else
@@ -1356,13 +1380,27 @@ out:
return rc;
}
+/**
+ * revalidate @dentryp from statahead cache
+ *
+ * \param[in] dir parent directory
+ * \param[in] sai sai structure
+ * \param[out] dentryp pointer to dentry which will be revalidated
+ * \param[in] unplug unplug statahead window only (normally for negative
+ * dentry)
+ * \retval 1 on success, dentry is saved in @dentryp
+ * \retval 0 if revalidation failed (no proper lock on client)
+ * \retval negative number upon error
+ */
static int revalidate_statahead_dentry(struct inode *dir,
struct ll_statahead_info *sai,
struct dentry **dentryp,
- int only_unplug)
+ bool unplug)
{
struct sa_entry *entry = NULL;
struct l_wait_info lwi = { 0 };
+ struct ll_dentry_data *ldd;
+ struct ll_inode_info *lli;
int rc = 0;
if ((*dentryp)->d_name.name[0] == '.') {
@@ -1392,10 +1430,15 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
+ if (unplug) {
+ rc = 1;
+ goto out_unplug;
+ }
+
entry = sa_get(sai, &(*dentryp)->d_name);
- if (!entry || only_unplug) {
- sa_put(sai, entry);
- return entry ? 1 : -EAGAIN;
+ if (!entry) {
+ rc = -EAGAIN;
+ goto out_unplug;
}
/* if statahead is busy in readdir, help it do post-work */
@@ -1406,13 +1449,15 @@ static int revalidate_statahead_dentry(struct inode *dir,
sai->sai_index_wait = entry->se_index;
lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(30), NULL,
LWI_ON_SIGNAL_NOOP, NULL);
- rc = l_wait_event(sai->sai_waitq,
- sa_ready(entry) ||
- thread_is_stopped(&sai->sai_thread),
- &lwi);
+ rc = l_wait_event(sai->sai_waitq, sa_ready(entry), &lwi);
if (rc < 0) {
- sa_put(sai, entry);
- return -EAGAIN;
+ /*
+ * entry may not be ready, so it may be used by inflight
+ * statahead RPC, don't free it.
+ */
+ entry = NULL;
+ rc = -EAGAIN;
+ goto out_unplug;
}
}
@@ -1430,10 +1475,15 @@ static int revalidate_statahead_dentry(struct inode *dir,
alias = ll_splice_alias(inode, *dentryp);
if (IS_ERR(alias)) {
- sa_put(sai, entry);
- return PTR_ERR(alias);
+ rc = PTR_ERR(alias);
+ goto out_unplug;
}
*dentryp = alias;
+ /**
+ * statahead prepared this inode, transfer inode
+ * refcount from sa_entry to dentry
+ */
+ entry->se_inode = NULL;
} else if ((*dentryp)->d_inode != inode) {
/* revalidate, but inode is recreated */
CDEBUG(D_READA,
@@ -1445,10 +1495,7 @@ static int revalidate_statahead_dentry(struct inode *dir,
PFID(ll_inode2fid(inode)));
rc = -ESTALE;
goto out_unplug;
- } else {
- iput(inode);
}
- entry->se_inode = NULL;
if ((bits & MDS_INODELOCK_LOOKUP) &&
d_lustre_invalid(*dentryp))
@@ -1457,10 +1504,34 @@ static int revalidate_statahead_dentry(struct inode *dir,
}
}
out_unplug:
+ /*
+ * statahead cached sa_entry can be used only once, and will be killed
+ * right after use, so if lookup/revalidate accessed statahead cache,
+ * set dentry ldd_sa_generation to parent lli_sa_generation, later if we
+ * stat this file again, we know we've done statahead before, see
+ * dentry_may_statahead().
+ */
+ ldd = ll_d2d(*dentryp);
+ lli = ll_i2info(dir);
+ /* ldd can be NULL if llite lookup failed. */
+ if (ldd)
+ ldd->lld_sa_generation = lli->lli_sa_generation;
sa_put(sai, entry);
return rc;
}
+/**
+ * start statahead thread
+ *
+ * \param[in] dir parent directory
+ * \param[in] dentry dentry that triggers statahead, normally the first
+ * dirent under @dir
+ * \retval -EAGAIN on success, because when this function is
+ * called, it's already in lookup call, so client should
+ * do it itself instead of waiting for statahead thread
+ * to do it asynchronously.
+ * \retval negative number upon error
+ */
static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
{
struct ll_inode_info *lli = ll_i2info(dir);
@@ -1468,60 +1539,34 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
struct l_wait_info lwi = { 0 };
struct ptlrpc_thread *thread;
struct task_struct *task;
- struct dentry *parent;
+ struct dentry *parent = dentry->d_parent;
int rc;
/* I am the "lli_opendir_pid" owner, only me can set "lli_sai". */
rc = is_first_dirent(dir, dentry);
- if (rc == LS_NONE_FIRST_DE) {
+ if (rc == LS_NOT_FIRST_DE) {
/* It is not "ls -{a}l" operation, no need statahead for it. */
- rc = -EAGAIN;
+ rc = -EFAULT;
goto out;
}
- sai = ll_sai_alloc();
+ sai = ll_sai_alloc(parent);
if (!sai) {
rc = -ENOMEM;
goto out;
}
sai->sai_ls_all = (rc == LS_FIRST_DOT_DE);
- sai->sai_inode = igrab(dir);
- if (unlikely(!sai->sai_inode)) {
- CWARN("Do not start stat ahead on dying inode "DFID"\n",
- PFID(&lli->lli_fid));
- rc = -ESTALE;
- goto out;
- }
-
- /* get parent reference count here, and put it in ll_statahead_thread */
- parent = dget(dentry->d_parent);
- if (unlikely(sai->sai_inode != d_inode(parent))) {
- struct ll_inode_info *nlli = ll_i2info(d_inode(parent));
-
- CWARN("Race condition, someone changed %pd just now: old parent "DFID", new parent "DFID"\n",
- dentry, PFID(&lli->lli_fid), PFID(&nlli->lli_fid));
- dput(parent);
- iput(sai->sai_inode);
- rc = -EAGAIN;
- goto out;
- }
-
- CDEBUG(D_READA, "start statahead thread: sai %p, parent %pd\n",
- sai, parent);
-
/*
- * if another process started statahead thread, or deauthorized current
- * lli_opendir_key, don't start statahead.
+ * if current lli_opendir_key was deauthorized, or dir re-opened by
+ * another process, don't start statahead, otherwise the newly spawned
+ * statahead thread won't be notified to quit.
*/
spin_lock(&lli->lli_sa_lock);
if (unlikely(lli->lli_sai || lli->lli_opendir_key ||
lli->lli_opendir_pid != current->pid)) {
spin_unlock(&lli->lli_sa_lock);
-
- dput(parent);
- iput(sai->sai_inode);
- rc = -EAGAIN;
+ rc = -EPERM;
goto out;
}
lli->lli_sai = sai;
@@ -1529,22 +1574,16 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
atomic_inc(&ll_i2sbi(parent->d_inode)->ll_sa_running);
+ CDEBUG(D_READA, "start statahead thread: [pid %d] [parent %pd]\n",
+ current_pid(), parent);
+
task = kthread_run(ll_statahead_thread, parent, "ll_sa_%u",
lli->lli_opendir_pid);
thread = &sai->sai_thread;
if (IS_ERR(task)) {
rc = PTR_ERR(task);
- CERROR("cannot start ll_sa thread: rc = %d\n", rc);
- dput(parent);
-
- spin_lock(&lli->lli_sa_lock);
- thread_set_flags(thread, SVC_STOPPED);
- thread_set_flags(&sai->sai_agl_thread, SVC_STOPPED);
- spin_unlock(&lli->lli_sa_lock);
-
- ll_sai_put(sai);
- LASSERT(!lli->lli_sai);
- return -EAGAIN;
+ CERROR("can't start ll_sa thread, rc : %d\n", rc);
+ goto out;
}
l_wait_event(thread->t_ctl_waitq,
@@ -1559,29 +1598,35 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry)
return -EAGAIN;
out:
- kfree(sai);
/*
* once we start statahead thread failed, disable statahead so
- * subsequent won't waste time to try it.
+ * that subsequent stat won't waste time to try it.
*/
spin_lock(&lli->lli_sa_lock);
lli->lli_sa_enabled = 0;
+ lli->lli_sai = NULL;
spin_unlock(&lli->lli_sa_lock);
-
+ if (sai)
+ ll_sai_free(sai);
return rc;
}
/**
- * Start statahead thread if this is the first dir entry.
- * Otherwise if a thread is started already, wait it until it is ahead of me.
- * \retval 1 -- find entry with lock in cache, the caller needs to do
- * nothing.
- * \retval 0 -- find entry in cache, but without lock, the caller needs
- * refresh from MDS.
- * \retval others -- the caller need to process as non-statahead.
+ * statahead entry function, this is called when client getattr on a file, it
+ * will start statahead thread if this is the first dir entry, else revalidate
+ * dentry from statahead cache.
+ *
+ * \param[in] dir parent directory
+ * \param[out] dentryp dentry to getattr
+ * \param[in] unplug unplug statahead window only (normally for negative
+ * dentry)
+ * \retval 1 on success
+ * \retval 0 revalidation from statahead cache failed, caller needs
+ * to getattr from server directly
+ * \retval negative number on error, caller often ignores this and
+ * then getattr from server
*/
-int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
- int only_unplug)
+int ll_statahead(struct inode *dir, struct dentry **dentryp, bool unplug)
{
struct ll_statahead_info *sai;
@@ -1589,13 +1634,11 @@ int do_statahead_enter(struct inode *dir, struct dentry **dentryp,
if (sai) {
int rc;
- rc = revalidate_statahead_dentry(dir, sai, dentryp,
- only_unplug);
+ rc = revalidate_statahead_dentry(dir, sai, dentryp, unplug);
CDEBUG(D_READA, "revalidate statahead %pd: %d.\n",
*dentryp, rc);
ll_sai_put(sai);
return rc;
}
-
return start_statahead_thread(dir, *dentryp);
}
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 166f0c4..125d882 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -1367,7 +1367,7 @@ static int mdc_read_page(struct obd_export *exp, struct md_op_data *op_data,
page = mdc_page_locate(mapping, &rp_param.rp_off, &start, &end,
rp_param.rp_hash64);
if (IS_ERR(page)) {
- CERROR("%s: dir page locate: "DFID" at %llu: rc %ld\n",
+ CDEBUG(D_INFO, "%s: dir page locate: " DFID " at %llu: rc %ld\n",
exp->exp_obd->obd_name, PFID(&op_data->op_fid1),
rp_param.rp_off, PTR_ERR(page));
rc = PTR_ERR(page);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* Re: [PATCH 070/124] staging: lustre: statahead: small fixes and cleanup
2016-09-18 20:38 ` [lustre-devel] " James Simmons
@ 2016-09-19 7:51 ` Greg Kroah-Hartman
-1 siblings, 0 replies; 263+ messages in thread
From: Greg Kroah-Hartman @ 2016-09-19 7:51 UTC (permalink / raw)
To: James Simmons
Cc: devel, Andreas Dilger, Oleg Drokin, Lai Siyao,
Linux Kernel Mailing List, Lustre Development List
On Sun, Sep 18, 2016 at 04:38:09PM -0400, James Simmons wrote:
> From: Lai Siyao <lai.siyao@intel.com>
>
> small fixes:
> * when 'unplug' is set for ll_statahead(), sa_put() shouldn't kill
> the entry found, because its inflight RPC may not finish yet.
> * remove 'sai_generation', add 'lli_sa_generation' because the
> former one is not safe to access without lock.
> * revalidate_statahead_dentry() may fail to wait for statahead
> entry to become ready, in this case it should not release this
> entry, because it may be used by inflight statahead RPC.
>
> cleanups:
> * rename ll_statahead_enter() to ll_statahead().
> * move dentry 'lld_sa_generation' update to ll_statahead() to
> simplify code and logic.
> * other small cleanups.
That's a lot of different things to be doing all at once in a single
patch. I know this came from your tree, but in the future, please push
back and ask for broken up patches, as that is required for "upstream"
here. I'll take this for now :(
greg k-h
^ permalink raw reply [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 070/124] staging: lustre: statahead: small fixes and cleanup
@ 2016-09-19 7:51 ` Greg Kroah-Hartman
0 siblings, 0 replies; 263+ messages in thread
From: Greg Kroah-Hartman @ 2016-09-19 7:51 UTC (permalink / raw)
To: James Simmons
Cc: devel, Andreas Dilger, Oleg Drokin, Lai Siyao,
Linux Kernel Mailing List, Lustre Development List
On Sun, Sep 18, 2016 at 04:38:09PM -0400, James Simmons wrote:
> From: Lai Siyao <lai.siyao@intel.com>
>
> small fixes:
> * when 'unplug' is set for ll_statahead(), sa_put() shouldn't kill
> the entry found, because its inflight RPC may not finish yet.
> * remove 'sai_generation', add 'lli_sa_generation' because the
> former one is not safe to access without lock.
> * revalidate_statahead_dentry() may fail to wait for statahead
> entry to become ready, in this case it should not release this
> entry, because it may be used by inflight statahead RPC.
>
> cleanups:
> * rename ll_statahead_enter() to ll_statahead().
> * move dentry 'lld_sa_generation' update to ll_statahead() to
> simplify code and logic.
> * other small cleanups.
That's a lot of different things to be doing all at once in a single
patch. I know this came from your tree, but in the future, please push
back and ask for broken up patches, as that is required for "upstream"
here. I'll take this for now :(
greg k-h
^ permalink raw reply [flat|nested] 263+ messages in thread
* [PATCH 071/124] staging: lustre: obd: remove unused obd methods
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove no longer used osc_packmd() and osc_getstripe().
Several ioctls cases that are no longer used are removed.
Remove no longer used adjust_kms() infrastructure.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2785
Reviewed-on: http://review.whamcloud.com/8547
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 3 -
drivers/staging/lustre/lustre/include/obd_class.h | 13 --
drivers/staging/lustre/lustre/lov/lov_internal.h | 2 -
drivers/staging/lustre/lustre/lov/lov_merge.c | 39 ------
drivers/staging/lustre/lustre/lov/lov_obd.c | 1 -
drivers/staging/lustre/lustre/mgc/mgc_request.c | 2 -
drivers/staging/lustre/lustre/osc/osc_request.c | 147 ---------------------
7 files changed, 0 insertions(+), 207 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 6eac63e..04a9b97 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -866,7 +866,6 @@ struct obd_ops {
int (*unpackmd)(struct obd_export *exp,
struct lov_stripe_md **mem_tgt,
struct lov_mds_md *disk_src, int disk_len);
- int (*preallocate)(struct lustre_handle *, u32 *req, u64 *ids);
int (*create)(const struct lu_env *env, struct obd_export *exp,
struct obdo *oa, struct lov_stripe_md **ea,
struct obd_trans_info *oti);
@@ -882,8 +881,6 @@ struct obd_ops {
struct obd_info *oinfo);
int (*getattr_async)(struct obd_export *exp, struct obd_info *oinfo,
struct ptlrpc_request_set *set);
- int (*adjust_kms)(struct obd_export *exp, struct lov_stripe_md *lsm,
- u64 size, int shrink);
int (*preprw)(const struct lu_env *env, int cmd,
struct obd_export *exp, struct obdo *oa, int objcount,
struct obd_ioobj *obj, struct niobuf_remote *remote,
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 3d081cf..3245952 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1152,19 +1152,6 @@ static inline int obd_commitrw(const struct lu_env *env, int cmd,
return rc;
}
-static inline int obd_adjust_kms(struct obd_export *exp,
- struct lov_stripe_md *lsm, u64 size,
- int shrink)
-{
- int rc;
-
- EXP_CHECK_DT_OP(exp, adjust_kms);
- EXP_COUNTER_INCREMENT(exp, adjust_kms);
-
- rc = OBP(exp->exp_obd, adjust_kms)(exp, lsm, size, shrink);
- return rc;
-}
-
static inline int obd_iocontrol(unsigned int cmd, struct obd_export *exp,
int len, void *karg, void __user *uarg)
{
diff --git a/drivers/staging/lustre/lustre/lov/lov_internal.h b/drivers/staging/lustre/lustre/lov/lov_internal.h
index 12bd511..5ddaeda 100644
--- a/drivers/staging/lustre/lustre/lov/lov_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_internal.h
@@ -134,8 +134,6 @@ static inline void lov_put_reqset(struct lov_request_set *set)
/* lov_merge.c */
void lov_merge_attrs(struct obdo *tgt, struct obdo *src, u64 valid,
struct lov_stripe_md *lsm, int stripeno, int *set);
-int lov_adjust_kms(struct obd_export *exp, struct lov_stripe_md *lsm,
- u64 size, int shrink);
int lov_merge_lvb_kms(struct lov_stripe_md *lsm,
struct ost_lvb *lvb, __u64 *kms_place);
diff --git a/drivers/staging/lustre/lustre/lov/lov_merge.c b/drivers/staging/lustre/lustre/lov/lov_merge.c
index b9c9086..674af10 100644
--- a/drivers/staging/lustre/lustre/lov/lov_merge.c
+++ b/drivers/staging/lustre/lustre/lov/lov_merge.c
@@ -105,45 +105,6 @@ int lov_merge_lvb_kms(struct lov_stripe_md *lsm,
return rc;
}
-/* Must be called under the lov_stripe_lock() */
-int lov_adjust_kms(struct obd_export *exp, struct lov_stripe_md *lsm,
- u64 size, int shrink)
-{
- struct lov_oinfo *loi;
- int stripe = 0;
- __u64 kms;
-
- assert_spin_locked(&lsm->lsm_lock);
- LASSERT(lsm->lsm_lock_owner == current_pid());
-
- if (shrink) {
- for (; stripe < lsm->lsm_stripe_count; stripe++) {
- struct lov_oinfo *loi = lsm->lsm_oinfo[stripe];
-
- kms = lov_size_to_stripe(lsm, size, stripe);
- CDEBUG(D_INODE,
- "stripe %d KMS %sing %llu->%llu\n",
- stripe, kms > loi->loi_kms ? "increase":"shrink",
- loi->loi_kms, kms);
- loi->loi_lvb.lvb_size = kms;
- loi_kms_set(loi, loi->loi_lvb.lvb_size);
- }
- return 0;
- }
-
- if (size > 0)
- stripe = lov_stripe_number(lsm, size - 1);
- kms = lov_size_to_stripe(lsm, size, stripe);
- loi = lsm->lsm_oinfo[stripe];
-
- CDEBUG(D_INODE, "stripe %d KMS %sincreasing %llu->%llu\n",
- stripe, kms > loi->loi_kms ? "" : "not ", loi->loi_kms, kms);
- if (kms > loi->loi_kms)
- loi_kms_set(loi, kms);
-
- return 0;
-}
-
void lov_merge_attrs(struct obdo *tgt, struct obdo *src, u64 valid,
struct lov_stripe_md *lsm, int stripeno, int *set)
{
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 2ee4b19..f433515 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -2084,7 +2084,6 @@ static struct obd_ops lov_obd_ops = {
.destroy = lov_destroy,
.getattr_async = lov_getattr_async,
.setattr_async = lov_setattr_async,
- .adjust_kms = lov_adjust_kms,
.iocontrol = lov_iocontrol,
.get_info = lov_get_info,
.set_info_async = lov_set_info_async,
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index 88d4d10..2191ca1 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -1739,8 +1739,6 @@ static struct obd_ops mgc_obd_ops = {
.del_conn = client_import_del_conn,
.connect = client_connect_import,
.disconnect = client_disconnect_export,
- /* .enqueue = mgc_enqueue, */
- /* .iocontrol = mgc_iocontrol, */
.set_info_async = mgc_set_info_async,
.get_info = mgc_get_info,
.import_event = mgc_import_event,
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 1b6491a..71f6abe 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -103,36 +103,6 @@ static void osc_release_ppga(struct brw_page **ppga, u32 count);
static int brw_interpret(const struct lu_env *env,
struct ptlrpc_request *req, void *data, int rc);
-/* Pack OSC object metadata for disk storage (LE byte order). */
-static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
- struct lov_stripe_md *lsm)
-{
- int lmm_size;
-
- lmm_size = sizeof(**lmmp);
- if (!lmmp)
- return lmm_size;
-
- if (*lmmp && !lsm) {
- kfree(*lmmp);
- *lmmp = NULL;
- return 0;
- } else if (unlikely(lsm && ostid_id(&lsm->lsm_oi) == 0)) {
- return -EBADF;
- }
-
- if (!*lmmp) {
- *lmmp = kzalloc(lmm_size, GFP_NOFS);
- if (!*lmmp)
- return -ENOMEM;
- }
-
- if (lsm)
- ostid_cpu_to_le(&lsm->lsm_oi, &(*lmmp)->lmm_oi);
-
- return lmm_size;
-}
-
/* Unpack OSC object metadata from disk storage (LE byte order). */
static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
struct lov_mds_md *lmm, int lmm_bytes)
@@ -2570,71 +2540,6 @@ static int osc_statfs(const struct lu_env *env, struct obd_export *exp,
return rc;
}
-/* Retrieve object striping information.
- *
- * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
- * the maximum number of OST indices which will fit in the user buffer.
- * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
- */
-static int osc_getstripe(struct lov_stripe_md *lsm,
- struct lov_user_md __user *lump)
-{
- /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
- struct lov_user_md_v3 lum, *lumk;
- struct lov_user_ost_data_v1 *lmm_objects;
- int rc = 0, lum_size;
-
- if (!lsm)
- return -ENODATA;
-
- /* we only need the header part from user space to get lmm_magic and
- * lmm_stripe_count, (the header part is common to v1 and v3)
- */
- lum_size = sizeof(struct lov_user_md_v1);
- if (copy_from_user(&lum, lump, lum_size))
- return -EFAULT;
-
- if ((lum.lmm_magic != LOV_USER_MAGIC_V1) &&
- (lum.lmm_magic != LOV_USER_MAGIC_V3))
- return -EINVAL;
-
- /* lov_user_md_vX and lov_mds_md_vX must have the same size */
- LASSERT(sizeof(struct lov_user_md_v1) == sizeof(struct lov_mds_md_v1));
- LASSERT(sizeof(struct lov_user_md_v3) == sizeof(struct lov_mds_md_v3));
- LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lumk->lmm_objects[0]));
-
- /* we can use lov_mds_md_size() to compute lum_size
- * because lov_user_md_vX and lov_mds_md_vX have the same size
- */
- if (lum.lmm_stripe_count > 0) {
- lum_size = lov_mds_md_size(lum.lmm_stripe_count, lum.lmm_magic);
- lumk = kzalloc(lum_size, GFP_NOFS);
- if (!lumk)
- return -ENOMEM;
-
- if (lum.lmm_magic == LOV_USER_MAGIC_V1)
- lmm_objects =
- &(((struct lov_user_md_v1 *)lumk)->lmm_objects[0]);
- else
- lmm_objects = &lumk->lmm_objects[0];
- lmm_objects->l_ost_oi = lsm->lsm_oi;
- } else {
- lum_size = lov_mds_md_size(0, lum.lmm_magic);
- lumk = &lum;
- }
-
- lumk->lmm_oi = lsm->lsm_oi;
- lumk->lmm_stripe_count = 1;
-
- if (copy_to_user(lump, lumk, lum_size))
- rc = -EFAULT;
-
- if (lumk != &lum)
- kfree(lumk);
-
- return rc;
-}
-
static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
void *karg, void __user *uarg)
{
@@ -2648,57 +2553,6 @@ static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
return -EINVAL;
}
switch (cmd) {
- case OBD_IOC_LOV_GET_CONFIG: {
- char *buf;
- struct lov_desc *desc;
- struct obd_uuid uuid;
-
- buf = NULL;
- len = 0;
- if (obd_ioctl_getdata(&buf, &len, uarg)) {
- err = -EINVAL;
- goto out;
- }
-
- data = (struct obd_ioctl_data *)buf;
-
- if (sizeof(*desc) > data->ioc_inllen1) {
- obd_ioctl_freedata(buf, len);
- err = -EINVAL;
- goto out;
- }
-
- if (data->ioc_inllen2 < sizeof(uuid)) {
- obd_ioctl_freedata(buf, len);
- err = -EINVAL;
- goto out;
- }
-
- desc = (struct lov_desc *)data->ioc_inlbuf1;
- desc->ld_tgt_count = 1;
- desc->ld_active_tgt_count = 1;
- desc->ld_default_stripe_count = 1;
- desc->ld_default_stripe_size = 0;
- desc->ld_default_stripe_offset = 0;
- desc->ld_pattern = 0;
- memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
-
- memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
-
- err = copy_to_user(uarg, buf, len);
- if (err)
- err = -EFAULT;
- obd_ioctl_freedata(buf, len);
- goto out;
- }
- case LL_IOC_LOV_SETSTRIPE:
- err = obd_alloc_memmd(exp, karg);
- if (err > 0)
- err = 0;
- goto out;
- case LL_IOC_LOV_GETSTRIPE:
- err = osc_getstripe(karg, uarg);
- goto out;
case OBD_IOC_CLIENT_RECOVER:
err = ptlrpc_recover_import(obd->u.cli.cl_import,
data->ioc_inlbuf1, 0);
@@ -3287,7 +3141,6 @@ static struct obd_ops osc_obd_ops = {
.disconnect = osc_disconnect,
.statfs = osc_statfs,
.statfs_async = osc_statfs_async,
- .packmd = osc_packmd,
.unpackmd = osc_unpackmd,
.create = osc_create,
.destroy = osc_destroy,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 071/124] staging: lustre: obd: remove unused obd methods
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove no longer used osc_packmd() and osc_getstripe().
Several ioctls cases that are no longer used are removed.
Remove no longer used adjust_kms() infrastructure.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2785
Reviewed-on: http://review.whamcloud.com/8547
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 3 -
drivers/staging/lustre/lustre/include/obd_class.h | 13 --
drivers/staging/lustre/lustre/lov/lov_internal.h | 2 -
drivers/staging/lustre/lustre/lov/lov_merge.c | 39 ------
drivers/staging/lustre/lustre/lov/lov_obd.c | 1 -
drivers/staging/lustre/lustre/mgc/mgc_request.c | 2 -
drivers/staging/lustre/lustre/osc/osc_request.c | 147 ---------------------
7 files changed, 0 insertions(+), 207 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 6eac63e..04a9b97 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -866,7 +866,6 @@ struct obd_ops {
int (*unpackmd)(struct obd_export *exp,
struct lov_stripe_md **mem_tgt,
struct lov_mds_md *disk_src, int disk_len);
- int (*preallocate)(struct lustre_handle *, u32 *req, u64 *ids);
int (*create)(const struct lu_env *env, struct obd_export *exp,
struct obdo *oa, struct lov_stripe_md **ea,
struct obd_trans_info *oti);
@@ -882,8 +881,6 @@ struct obd_ops {
struct obd_info *oinfo);
int (*getattr_async)(struct obd_export *exp, struct obd_info *oinfo,
struct ptlrpc_request_set *set);
- int (*adjust_kms)(struct obd_export *exp, struct lov_stripe_md *lsm,
- u64 size, int shrink);
int (*preprw)(const struct lu_env *env, int cmd,
struct obd_export *exp, struct obdo *oa, int objcount,
struct obd_ioobj *obj, struct niobuf_remote *remote,
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 3d081cf..3245952 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1152,19 +1152,6 @@ static inline int obd_commitrw(const struct lu_env *env, int cmd,
return rc;
}
-static inline int obd_adjust_kms(struct obd_export *exp,
- struct lov_stripe_md *lsm, u64 size,
- int shrink)
-{
- int rc;
-
- EXP_CHECK_DT_OP(exp, adjust_kms);
- EXP_COUNTER_INCREMENT(exp, adjust_kms);
-
- rc = OBP(exp->exp_obd, adjust_kms)(exp, lsm, size, shrink);
- return rc;
-}
-
static inline int obd_iocontrol(unsigned int cmd, struct obd_export *exp,
int len, void *karg, void __user *uarg)
{
diff --git a/drivers/staging/lustre/lustre/lov/lov_internal.h b/drivers/staging/lustre/lustre/lov/lov_internal.h
index 12bd511..5ddaeda 100644
--- a/drivers/staging/lustre/lustre/lov/lov_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_internal.h
@@ -134,8 +134,6 @@ static inline void lov_put_reqset(struct lov_request_set *set)
/* lov_merge.c */
void lov_merge_attrs(struct obdo *tgt, struct obdo *src, u64 valid,
struct lov_stripe_md *lsm, int stripeno, int *set);
-int lov_adjust_kms(struct obd_export *exp, struct lov_stripe_md *lsm,
- u64 size, int shrink);
int lov_merge_lvb_kms(struct lov_stripe_md *lsm,
struct ost_lvb *lvb, __u64 *kms_place);
diff --git a/drivers/staging/lustre/lustre/lov/lov_merge.c b/drivers/staging/lustre/lustre/lov/lov_merge.c
index b9c9086..674af10 100644
--- a/drivers/staging/lustre/lustre/lov/lov_merge.c
+++ b/drivers/staging/lustre/lustre/lov/lov_merge.c
@@ -105,45 +105,6 @@ int lov_merge_lvb_kms(struct lov_stripe_md *lsm,
return rc;
}
-/* Must be called under the lov_stripe_lock() */
-int lov_adjust_kms(struct obd_export *exp, struct lov_stripe_md *lsm,
- u64 size, int shrink)
-{
- struct lov_oinfo *loi;
- int stripe = 0;
- __u64 kms;
-
- assert_spin_locked(&lsm->lsm_lock);
- LASSERT(lsm->lsm_lock_owner == current_pid());
-
- if (shrink) {
- for (; stripe < lsm->lsm_stripe_count; stripe++) {
- struct lov_oinfo *loi = lsm->lsm_oinfo[stripe];
-
- kms = lov_size_to_stripe(lsm, size, stripe);
- CDEBUG(D_INODE,
- "stripe %d KMS %sing %llu->%llu\n",
- stripe, kms > loi->loi_kms ? "increase":"shrink",
- loi->loi_kms, kms);
- loi->loi_lvb.lvb_size = kms;
- loi_kms_set(loi, loi->loi_lvb.lvb_size);
- }
- return 0;
- }
-
- if (size > 0)
- stripe = lov_stripe_number(lsm, size - 1);
- kms = lov_size_to_stripe(lsm, size, stripe);
- loi = lsm->lsm_oinfo[stripe];
-
- CDEBUG(D_INODE, "stripe %d KMS %sincreasing %llu->%llu\n",
- stripe, kms > loi->loi_kms ? "" : "not ", loi->loi_kms, kms);
- if (kms > loi->loi_kms)
- loi_kms_set(loi, kms);
-
- return 0;
-}
-
void lov_merge_attrs(struct obdo *tgt, struct obdo *src, u64 valid,
struct lov_stripe_md *lsm, int stripeno, int *set)
{
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 2ee4b19..f433515 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -2084,7 +2084,6 @@ static struct obd_ops lov_obd_ops = {
.destroy = lov_destroy,
.getattr_async = lov_getattr_async,
.setattr_async = lov_setattr_async,
- .adjust_kms = lov_adjust_kms,
.iocontrol = lov_iocontrol,
.get_info = lov_get_info,
.set_info_async = lov_set_info_async,
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index 88d4d10..2191ca1 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -1739,8 +1739,6 @@ static struct obd_ops mgc_obd_ops = {
.del_conn = client_import_del_conn,
.connect = client_connect_import,
.disconnect = client_disconnect_export,
- /* .enqueue = mgc_enqueue, */
- /* .iocontrol = mgc_iocontrol, */
.set_info_async = mgc_set_info_async,
.get_info = mgc_get_info,
.import_event = mgc_import_event,
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 1b6491a..71f6abe 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -103,36 +103,6 @@ static void osc_release_ppga(struct brw_page **ppga, u32 count);
static int brw_interpret(const struct lu_env *env,
struct ptlrpc_request *req, void *data, int rc);
-/* Pack OSC object metadata for disk storage (LE byte order). */
-static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
- struct lov_stripe_md *lsm)
-{
- int lmm_size;
-
- lmm_size = sizeof(**lmmp);
- if (!lmmp)
- return lmm_size;
-
- if (*lmmp && !lsm) {
- kfree(*lmmp);
- *lmmp = NULL;
- return 0;
- } else if (unlikely(lsm && ostid_id(&lsm->lsm_oi) == 0)) {
- return -EBADF;
- }
-
- if (!*lmmp) {
- *lmmp = kzalloc(lmm_size, GFP_NOFS);
- if (!*lmmp)
- return -ENOMEM;
- }
-
- if (lsm)
- ostid_cpu_to_le(&lsm->lsm_oi, &(*lmmp)->lmm_oi);
-
- return lmm_size;
-}
-
/* Unpack OSC object metadata from disk storage (LE byte order). */
static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
struct lov_mds_md *lmm, int lmm_bytes)
@@ -2570,71 +2540,6 @@ static int osc_statfs(const struct lu_env *env, struct obd_export *exp,
return rc;
}
-/* Retrieve object striping information.
- *
- * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
- * the maximum number of OST indices which will fit in the user buffer.
- * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
- */
-static int osc_getstripe(struct lov_stripe_md *lsm,
- struct lov_user_md __user *lump)
-{
- /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
- struct lov_user_md_v3 lum, *lumk;
- struct lov_user_ost_data_v1 *lmm_objects;
- int rc = 0, lum_size;
-
- if (!lsm)
- return -ENODATA;
-
- /* we only need the header part from user space to get lmm_magic and
- * lmm_stripe_count, (the header part is common to v1 and v3)
- */
- lum_size = sizeof(struct lov_user_md_v1);
- if (copy_from_user(&lum, lump, lum_size))
- return -EFAULT;
-
- if ((lum.lmm_magic != LOV_USER_MAGIC_V1) &&
- (lum.lmm_magic != LOV_USER_MAGIC_V3))
- return -EINVAL;
-
- /* lov_user_md_vX and lov_mds_md_vX must have the same size */
- LASSERT(sizeof(struct lov_user_md_v1) == sizeof(struct lov_mds_md_v1));
- LASSERT(sizeof(struct lov_user_md_v3) == sizeof(struct lov_mds_md_v3));
- LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lumk->lmm_objects[0]));
-
- /* we can use lov_mds_md_size() to compute lum_size
- * because lov_user_md_vX and lov_mds_md_vX have the same size
- */
- if (lum.lmm_stripe_count > 0) {
- lum_size = lov_mds_md_size(lum.lmm_stripe_count, lum.lmm_magic);
- lumk = kzalloc(lum_size, GFP_NOFS);
- if (!lumk)
- return -ENOMEM;
-
- if (lum.lmm_magic == LOV_USER_MAGIC_V1)
- lmm_objects =
- &(((struct lov_user_md_v1 *)lumk)->lmm_objects[0]);
- else
- lmm_objects = &lumk->lmm_objects[0];
- lmm_objects->l_ost_oi = lsm->lsm_oi;
- } else {
- lum_size = lov_mds_md_size(0, lum.lmm_magic);
- lumk = &lum;
- }
-
- lumk->lmm_oi = lsm->lsm_oi;
- lumk->lmm_stripe_count = 1;
-
- if (copy_to_user(lump, lumk, lum_size))
- rc = -EFAULT;
-
- if (lumk != &lum)
- kfree(lumk);
-
- return rc;
-}
-
static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
void *karg, void __user *uarg)
{
@@ -2648,57 +2553,6 @@ static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
return -EINVAL;
}
switch (cmd) {
- case OBD_IOC_LOV_GET_CONFIG: {
- char *buf;
- struct lov_desc *desc;
- struct obd_uuid uuid;
-
- buf = NULL;
- len = 0;
- if (obd_ioctl_getdata(&buf, &len, uarg)) {
- err = -EINVAL;
- goto out;
- }
-
- data = (struct obd_ioctl_data *)buf;
-
- if (sizeof(*desc) > data->ioc_inllen1) {
- obd_ioctl_freedata(buf, len);
- err = -EINVAL;
- goto out;
- }
-
- if (data->ioc_inllen2 < sizeof(uuid)) {
- obd_ioctl_freedata(buf, len);
- err = -EINVAL;
- goto out;
- }
-
- desc = (struct lov_desc *)data->ioc_inlbuf1;
- desc->ld_tgt_count = 1;
- desc->ld_active_tgt_count = 1;
- desc->ld_default_stripe_count = 1;
- desc->ld_default_stripe_size = 0;
- desc->ld_default_stripe_offset = 0;
- desc->ld_pattern = 0;
- memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
-
- memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
-
- err = copy_to_user(uarg, buf, len);
- if (err)
- err = -EFAULT;
- obd_ioctl_freedata(buf, len);
- goto out;
- }
- case LL_IOC_LOV_SETSTRIPE:
- err = obd_alloc_memmd(exp, karg);
- if (err > 0)
- err = 0;
- goto out;
- case LL_IOC_LOV_GETSTRIPE:
- err = osc_getstripe(karg, uarg);
- goto out;
case OBD_IOC_CLIENT_RECOVER:
err = ptlrpc_recover_import(obd->u.cli.cl_import,
data->ioc_inlbuf1, 0);
@@ -3287,7 +3141,6 @@ static struct obd_ops osc_obd_ops = {
.disconnect = osc_disconnect,
.statfs = osc_statfs,
.statfs_async = osc_statfs_async,
- .packmd = osc_packmd,
.unpackmd = osc_unpackmd,
.create = osc_create,
.destroy = osc_destroy,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 072/124] staging: lustre: echo: replace lov_stripe_md with lov_oinfo
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In echo_client replace uses of struct lov_stripe_md with struct
lov_oinfo (since the instances of the former really only contained a
single instance of the latter). Remove the then unneccessary functions
echo_alloc_memmd(), echo_free_memmd(), osc_unpackmd(), and
obd_alloc_memmd(). Remove the struct lov_stripe_md * parameter from
obd_create().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5418
Reviewed-on: http://review.whamcloud.com/12447
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 3 +-
drivers/staging/lustre/lustre/include/obd_class.h | 14 +--
.../staging/lustre/lustre/obdecho/echo_client.c | 143 +++++--------------
drivers/staging/lustre/lustre/osc/osc_request.c | 44 +------
4 files changed, 45 insertions(+), 159 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 04a9b97..ca63891 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -867,8 +867,7 @@ struct obd_ops {
struct lov_stripe_md **mem_tgt,
struct lov_mds_md *disk_src, int disk_len);
int (*create)(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md **ea,
- struct obd_trans_info *oti);
+ struct obdo *oa, struct obd_trans_info *oti);
int (*destroy)(const struct lu_env *env, struct obd_export *exp,
struct obdo *oa, struct lov_stripe_md *ea,
struct obd_trans_info *oti, struct obd_export *md_exp);
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 3245952..2bf2826 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -677,15 +677,6 @@ static inline int obd_unpackmd(struct obd_export *exp,
return rc;
}
-/* helper functions */
-static inline int obd_alloc_memmd(struct obd_export *exp,
- struct lov_stripe_md **mem_tgt)
-{
- LASSERT(mem_tgt);
- LASSERT(!*mem_tgt);
- return obd_unpackmd(exp, mem_tgt, NULL, 0);
-}
-
static inline int obd_free_memmd(struct obd_export *exp,
struct lov_stripe_md **mem_tgt)
{
@@ -699,15 +690,14 @@ static inline int obd_free_memmd(struct obd_export *exp,
}
static inline int obd_create(const struct lu_env *env, struct obd_export *exp,
- struct obdo *obdo, struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
+ struct obdo *obdo, struct obd_trans_info *oti)
{
int rc;
EXP_CHECK_DT_OP(exp, create);
EXP_COUNTER_INCREMENT(exp, create);
- rc = OBP(exp->exp_obd, create)(env, exp, obdo, ea, oti);
+ rc = OBP(exp->exp_obd, create)(env, exp, obdo, oti);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index cddc2d2..e487fda 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -65,14 +65,14 @@ struct echo_object {
struct echo_device *eo_dev;
struct list_head eo_obj_chain;
- struct lov_stripe_md *eo_lsm;
+ struct lov_oinfo *eo_oinfo;
atomic_t eo_npages;
int eo_deleted;
};
struct echo_object_conf {
struct cl_object_conf eoc_cl;
- struct lov_stripe_md **eoc_md;
+ struct lov_oinfo **eoc_oinfo;
};
struct echo_page {
@@ -153,9 +153,6 @@ struct echo_object_conf *cl2echo_conf(const struct cl_object_conf *c)
}
/** @} echo_helpers */
-
-static struct echo_object *cl_echo_object_find(struct echo_device *d,
- struct lov_stripe_md **lsm);
static int cl_echo_object_put(struct echo_object *eco);
static int cl_echo_object_brw(struct echo_object *eco, int rw, u64 offset,
struct page **pages, int npages, int async);
@@ -414,10 +411,13 @@ static int echo_object_init(const struct lu_env *env, struct lu_object *obj,
cconf = lu2cl_conf(conf);
econf = cl2echo_conf(cconf);
- LASSERT(econf->eoc_md);
- eco->eo_lsm = *econf->eoc_md;
- /* clear the lsm pointer so that it won't get freed. */
- *econf->eoc_md = NULL;
+ LASSERT(econf->eoc_oinfo);
+ /*
+ * Transfer the oinfo pointer to eco that it won't be
+ * freed.
+ */
+ eco->eo_oinfo = *econf->eoc_oinfo;
+ *econf->eoc_oinfo = NULL;
eco->eo_dev = ed;
atomic_set(&eco->eo_npages, 0);
@@ -430,52 +430,6 @@ static int echo_object_init(const struct lu_env *env, struct lu_object *obj,
return 0;
}
-/* taken from osc_unpackmd() */
-static int echo_alloc_memmd(struct echo_device *ed,
- struct lov_stripe_md **lsmp)
-{
- int lsm_size;
-
- /* If export is lov/osc then use their obd method */
- if (ed->ed_next)
- return obd_alloc_memmd(ed->ed_ec->ec_exp, lsmp);
- /* OFD has no unpackmd method, do everything here */
- lsm_size = lov_stripe_md_size(1);
-
- LASSERT(!*lsmp);
- *lsmp = kzalloc(lsm_size, GFP_NOFS);
- if (!*lsmp)
- return -ENOMEM;
-
- (*lsmp)->lsm_oinfo[0] = kzalloc(sizeof(struct lov_oinfo), GFP_NOFS);
- if (!(*lsmp)->lsm_oinfo[0]) {
- kfree(*lsmp);
- return -ENOMEM;
- }
-
- loi_init((*lsmp)->lsm_oinfo[0]);
- (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
- ostid_set_seq_echo(&(*lsmp)->lsm_oi);
-
- return lsm_size;
-}
-
-static int echo_free_memmd(struct echo_device *ed, struct lov_stripe_md **lsmp)
-{
- int lsm_size;
-
- /* If export is lov/osc then use their obd method */
- if (ed->ed_next)
- return obd_free_memmd(ed->ed_ec->ec_exp, lsmp);
- /* OFD has no unpackmd method, do everything here */
- lsm_size = lov_stripe_md_size(1);
-
- kfree((*lsmp)->lsm_oinfo[0]);
- kfree(*lsmp);
- *lsmp = NULL;
- return 0;
-}
-
static void echo_object_free(const struct lu_env *env, struct lu_object *obj)
{
struct echo_object *eco = cl2echo_obj(lu2cl(obj));
@@ -490,8 +444,7 @@ static void echo_object_free(const struct lu_env *env, struct lu_object *obj)
lu_object_fini(obj);
lu_object_header_fini(obj->lo_header);
- if (eco->eo_lsm)
- echo_free_memmd(eco->eo_dev, &eco->eo_lsm);
+ kfree(eco->eo_oinfo);
kmem_cache_free(echo_object_kmem, eco);
}
@@ -865,25 +818,21 @@ static struct lu_device_type echo_device_type = {
*/
/* Interfaces to echo client obd device */
-static struct echo_object *cl_echo_object_find(struct echo_device *d,
- struct lov_stripe_md **lsmp)
+static struct echo_object *
+cl_echo_object_find(struct echo_device *d, const struct ost_id *oi)
{
struct lu_env *env;
struct echo_thread_info *info;
struct echo_object_conf *conf;
- struct lov_stripe_md *lsm;
+ struct lov_oinfo *oinfo = NULL;
struct echo_object *eco;
struct cl_object *obj;
struct lu_fid *fid;
int refcheck;
int rc;
- LASSERT(lsmp);
- lsm = *lsmp;
- LASSERT(lsm);
- LASSERTF(ostid_id(&lsm->lsm_oi) != 0, DOSTID"\n", POSTID(&lsm->lsm_oi));
- LASSERTF(ostid_seq(&lsm->lsm_oi) == FID_SEQ_ECHO, DOSTID"\n",
- POSTID(&lsm->lsm_oi));
+ LASSERTF(ostid_id(oi), DOSTID "\n", POSTID(oi));
+ LASSERTF(ostid_seq(oi) == FID_SEQ_ECHO, DOSTID "\n", POSTID(oi));
/* Never return an object if the obd is to be freed. */
if (echo_dev2cl(d)->cd_lu_dev.ld_obd->obd_stopping)
@@ -896,16 +845,24 @@ static struct echo_object *cl_echo_object_find(struct echo_device *d,
info = echo_env_info(env);
conf = &info->eti_conf;
if (d->ed_next) {
- struct lov_oinfo *oinfo = lsm->lsm_oinfo[0];
+ oinfo = kzalloc(sizeof(*oinfo), GFP_NOFS);
+ if (!oinfo) {
+ eco = ERR_PTR(-ENOMEM);
+ goto out;
+ }
- LASSERT(oinfo);
- oinfo->loi_oi = lsm->lsm_oi;
+ oinfo->loi_oi = *oi;
conf->eoc_cl.u.coc_oinfo = oinfo;
}
- conf->eoc_md = lsmp;
+
+ /*
+ * If echo_object_init() is successful then ownership of oinfo
+ * is transferred to the object.
+ */
+ conf->eoc_oinfo = &oinfo;
fid = &info->eti_fid;
- rc = ostid_to_fid(fid, &lsm->lsm_oi, 0);
+ rc = ostid_to_fid(fid, (struct ost_id *)oi, 0);
if (rc != 0) {
eco = ERR_PTR(rc);
goto out;
@@ -928,6 +885,7 @@ static struct echo_object *cl_echo_object_find(struct echo_device *d,
}
out:
+ kfree(oinfo);
cl_env_put(env, &refcheck);
return eco;
}
@@ -1146,7 +1104,6 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
{
struct echo_object *eco;
struct echo_client_obd *ec = ed->ed_ec;
- struct lov_stripe_md *lsm = NULL;
int rc;
int created = 0;
@@ -1157,30 +1114,19 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
return -EINVAL;
}
- rc = echo_alloc_memmd(ed, &lsm);
- if (rc < 0) {
- CERROR("Cannot allocate md: rc = %d\n", rc);
- goto failed;
- }
-
- /* setup object ID here */
- lsm->lsm_oi = oa->o_oi;
+ if (!ostid_id(&oa->o_oi))
+ ostid_set_id(&oa->o_oi, ++last_object_id);
- if (ostid_id(&lsm->lsm_oi) == 0)
- ostid_set_id(&lsm->lsm_oi, ++last_object_id);
-
- rc = obd_create(env, ec->ec_exp, oa, &lsm, oti);
+ rc = obd_create(env, ec->ec_exp, oa, oti);
if (rc != 0) {
CERROR("Cannot create objects: rc = %d\n", rc);
goto failed;
}
created = 1;
- /* See what object ID we were given */
- oa->o_oi = lsm->lsm_oi;
oa->o_valid |= OBD_MD_FLID;
- eco = cl_echo_object_find(ed, &lsm);
+ eco = cl_echo_object_find(ed, &oa->o_oi);
if (IS_ERR(eco)) {
rc = PTR_ERR(eco);
goto failed;
@@ -1191,9 +1137,7 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
failed:
if (created && rc)
- obd_destroy(env, ec->ec_exp, oa, lsm, oti, NULL);
- if (lsm)
- echo_free_memmd(ed, &lsm);
+ obd_destroy(env, ec->ec_exp, oa, NULL, oti, NULL);
if (rc)
CERROR("create object failed with: rc = %d\n", rc);
return rc;
@@ -1202,32 +1146,21 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
static int echo_get_object(struct echo_object **ecop, struct echo_device *ed,
struct obdo *oa)
{
- struct lov_stripe_md *lsm = NULL;
struct echo_object *eco;
int rc;
- if ((oa->o_valid & OBD_MD_FLID) == 0 || ostid_id(&oa->o_oi) == 0) {
- /* disallow use of object id 0 */
- CERROR("No valid oid\n");
+ if (!(oa->o_valid & OBD_MD_FLID) || !(oa->o_valid & OBD_MD_FLGROUP) ||
+ !ostid_id(&oa->o_oi)) {
+ CERROR("invalid oid " DOSTID "\n", POSTID(&oa->o_oi));
return -EINVAL;
}
- rc = echo_alloc_memmd(ed, &lsm);
- if (rc < 0)
- return rc;
-
- lsm->lsm_oi = oa->o_oi;
- if (!(oa->o_valid & OBD_MD_FLGROUP))
- ostid_set_seq_echo(&lsm->lsm_oi);
-
rc = 0;
- eco = cl_echo_object_find(ed, &lsm);
+ eco = cl_echo_object_find(ed, &oa->o_oi);
if (!IS_ERR(eco))
*ecop = eco;
else
rc = PTR_ERR(eco);
- if (lsm)
- echo_free_memmd(ed, &lsm);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 71f6abe..aefd776 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -398,24 +398,16 @@ static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
oinfo->oi_cb_up, oinfo, rqset);
}
-static int osc_real_create(struct obd_export *exp, struct obdo *oa,
- struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
+static int osc_create(const struct lu_env *env, struct obd_export *exp,
+ struct obdo *oa, struct obd_trans_info *oti)
{
struct ptlrpc_request *req;
struct ost_body *body;
- struct lov_stripe_md *lsm;
int rc;
LASSERT(oa);
- LASSERT(ea);
-
- lsm = *ea;
- if (!lsm) {
- rc = obd_alloc_memmd(exp, &lsm);
- if (rc < 0)
- return rc;
- }
+ LASSERT(oa->o_valid & OBD_MD_FLGROUP);
+ LASSERT(fid_seq_is_echo(ostid_seq(&oa->o_oi)));
req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
if (!req) {
@@ -461,13 +453,6 @@ static int osc_real_create(struct obd_export *exp, struct obdo *oa,
oa->o_blksize = cli_brw_size(exp->exp_obd);
oa->o_valid |= OBD_MD_FLBLKSZ;
- /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
- * have valid lsm_oinfo data structs, so don't go touching that.
- * This needs to be fixed in a big way.
- */
- lsm->lsm_oi = oa->o_oi;
- *ea = lsm;
-
if (oti && oa->o_valid & OBD_MD_FLCOOKIE) {
if (!oti->oti_logcookies)
oti->oti_logcookies = &oti->oti_onecookie;
@@ -479,8 +464,6 @@ static int osc_real_create(struct obd_export *exp, struct obdo *oa,
out_req:
ptlrpc_req_finished(req);
out:
- if (rc && !*ea)
- obd_free_memmd(exp, &lsm);
return rc;
}
@@ -656,25 +639,6 @@ static int osc_can_send_destroy(struct client_obd *cli)
return 0;
}
-static int osc_create(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
-{
- int rc = 0;
-
- LASSERT(oa);
- LASSERT(ea);
- LASSERT(oa->o_valid & OBD_MD_FLGROUP);
-
- if (!fid_seq_is_mdt(ostid_seq(&oa->o_oi)))
- return osc_real_create(exp, oa, ea, oti);
-
- /* we should not get here anymore */
- LBUG();
-
- return rc;
-}
-
/* Destroy requests can be async always on the client, and we don't even really
* care about the return code since the client cannot do anything at all about
* a destroy failure.
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 072/124] staging: lustre: echo: replace lov_stripe_md with lov_oinfo
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In echo_client replace uses of struct lov_stripe_md with struct
lov_oinfo (since the instances of the former really only contained a
single instance of the latter). Remove the then unneccessary functions
echo_alloc_memmd(), echo_free_memmd(), osc_unpackmd(), and
obd_alloc_memmd(). Remove the struct lov_stripe_md * parameter from
obd_create().
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5418
Reviewed-on: http://review.whamcloud.com/12447
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 3 +-
drivers/staging/lustre/lustre/include/obd_class.h | 14 +--
.../staging/lustre/lustre/obdecho/echo_client.c | 143 +++++--------------
drivers/staging/lustre/lustre/osc/osc_request.c | 44 +------
4 files changed, 45 insertions(+), 159 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 04a9b97..ca63891 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -867,8 +867,7 @@ struct obd_ops {
struct lov_stripe_md **mem_tgt,
struct lov_mds_md *disk_src, int disk_len);
int (*create)(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md **ea,
- struct obd_trans_info *oti);
+ struct obdo *oa, struct obd_trans_info *oti);
int (*destroy)(const struct lu_env *env, struct obd_export *exp,
struct obdo *oa, struct lov_stripe_md *ea,
struct obd_trans_info *oti, struct obd_export *md_exp);
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 3245952..2bf2826 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -677,15 +677,6 @@ static inline int obd_unpackmd(struct obd_export *exp,
return rc;
}
-/* helper functions */
-static inline int obd_alloc_memmd(struct obd_export *exp,
- struct lov_stripe_md **mem_tgt)
-{
- LASSERT(mem_tgt);
- LASSERT(!*mem_tgt);
- return obd_unpackmd(exp, mem_tgt, NULL, 0);
-}
-
static inline int obd_free_memmd(struct obd_export *exp,
struct lov_stripe_md **mem_tgt)
{
@@ -699,15 +690,14 @@ static inline int obd_free_memmd(struct obd_export *exp,
}
static inline int obd_create(const struct lu_env *env, struct obd_export *exp,
- struct obdo *obdo, struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
+ struct obdo *obdo, struct obd_trans_info *oti)
{
int rc;
EXP_CHECK_DT_OP(exp, create);
EXP_COUNTER_INCREMENT(exp, create);
- rc = OBP(exp->exp_obd, create)(env, exp, obdo, ea, oti);
+ rc = OBP(exp->exp_obd, create)(env, exp, obdo, oti);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index cddc2d2..e487fda 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -65,14 +65,14 @@ struct echo_object {
struct echo_device *eo_dev;
struct list_head eo_obj_chain;
- struct lov_stripe_md *eo_lsm;
+ struct lov_oinfo *eo_oinfo;
atomic_t eo_npages;
int eo_deleted;
};
struct echo_object_conf {
struct cl_object_conf eoc_cl;
- struct lov_stripe_md **eoc_md;
+ struct lov_oinfo **eoc_oinfo;
};
struct echo_page {
@@ -153,9 +153,6 @@ struct echo_object_conf *cl2echo_conf(const struct cl_object_conf *c)
}
/** @} echo_helpers */
-
-static struct echo_object *cl_echo_object_find(struct echo_device *d,
- struct lov_stripe_md **lsm);
static int cl_echo_object_put(struct echo_object *eco);
static int cl_echo_object_brw(struct echo_object *eco, int rw, u64 offset,
struct page **pages, int npages, int async);
@@ -414,10 +411,13 @@ static int echo_object_init(const struct lu_env *env, struct lu_object *obj,
cconf = lu2cl_conf(conf);
econf = cl2echo_conf(cconf);
- LASSERT(econf->eoc_md);
- eco->eo_lsm = *econf->eoc_md;
- /* clear the lsm pointer so that it won't get freed. */
- *econf->eoc_md = NULL;
+ LASSERT(econf->eoc_oinfo);
+ /*
+ * Transfer the oinfo pointer to eco that it won't be
+ * freed.
+ */
+ eco->eo_oinfo = *econf->eoc_oinfo;
+ *econf->eoc_oinfo = NULL;
eco->eo_dev = ed;
atomic_set(&eco->eo_npages, 0);
@@ -430,52 +430,6 @@ static int echo_object_init(const struct lu_env *env, struct lu_object *obj,
return 0;
}
-/* taken from osc_unpackmd() */
-static int echo_alloc_memmd(struct echo_device *ed,
- struct lov_stripe_md **lsmp)
-{
- int lsm_size;
-
- /* If export is lov/osc then use their obd method */
- if (ed->ed_next)
- return obd_alloc_memmd(ed->ed_ec->ec_exp, lsmp);
- /* OFD has no unpackmd method, do everything here */
- lsm_size = lov_stripe_md_size(1);
-
- LASSERT(!*lsmp);
- *lsmp = kzalloc(lsm_size, GFP_NOFS);
- if (!*lsmp)
- return -ENOMEM;
-
- (*lsmp)->lsm_oinfo[0] = kzalloc(sizeof(struct lov_oinfo), GFP_NOFS);
- if (!(*lsmp)->lsm_oinfo[0]) {
- kfree(*lsmp);
- return -ENOMEM;
- }
-
- loi_init((*lsmp)->lsm_oinfo[0]);
- (*lsmp)->lsm_maxbytes = LUSTRE_EXT3_STRIPE_MAXBYTES;
- ostid_set_seq_echo(&(*lsmp)->lsm_oi);
-
- return lsm_size;
-}
-
-static int echo_free_memmd(struct echo_device *ed, struct lov_stripe_md **lsmp)
-{
- int lsm_size;
-
- /* If export is lov/osc then use their obd method */
- if (ed->ed_next)
- return obd_free_memmd(ed->ed_ec->ec_exp, lsmp);
- /* OFD has no unpackmd method, do everything here */
- lsm_size = lov_stripe_md_size(1);
-
- kfree((*lsmp)->lsm_oinfo[0]);
- kfree(*lsmp);
- *lsmp = NULL;
- return 0;
-}
-
static void echo_object_free(const struct lu_env *env, struct lu_object *obj)
{
struct echo_object *eco = cl2echo_obj(lu2cl(obj));
@@ -490,8 +444,7 @@ static void echo_object_free(const struct lu_env *env, struct lu_object *obj)
lu_object_fini(obj);
lu_object_header_fini(obj->lo_header);
- if (eco->eo_lsm)
- echo_free_memmd(eco->eo_dev, &eco->eo_lsm);
+ kfree(eco->eo_oinfo);
kmem_cache_free(echo_object_kmem, eco);
}
@@ -865,25 +818,21 @@ static struct lu_device_type echo_device_type = {
*/
/* Interfaces to echo client obd device */
-static struct echo_object *cl_echo_object_find(struct echo_device *d,
- struct lov_stripe_md **lsmp)
+static struct echo_object *
+cl_echo_object_find(struct echo_device *d, const struct ost_id *oi)
{
struct lu_env *env;
struct echo_thread_info *info;
struct echo_object_conf *conf;
- struct lov_stripe_md *lsm;
+ struct lov_oinfo *oinfo = NULL;
struct echo_object *eco;
struct cl_object *obj;
struct lu_fid *fid;
int refcheck;
int rc;
- LASSERT(lsmp);
- lsm = *lsmp;
- LASSERT(lsm);
- LASSERTF(ostid_id(&lsm->lsm_oi) != 0, DOSTID"\n", POSTID(&lsm->lsm_oi));
- LASSERTF(ostid_seq(&lsm->lsm_oi) == FID_SEQ_ECHO, DOSTID"\n",
- POSTID(&lsm->lsm_oi));
+ LASSERTF(ostid_id(oi), DOSTID "\n", POSTID(oi));
+ LASSERTF(ostid_seq(oi) == FID_SEQ_ECHO, DOSTID "\n", POSTID(oi));
/* Never return an object if the obd is to be freed. */
if (echo_dev2cl(d)->cd_lu_dev.ld_obd->obd_stopping)
@@ -896,16 +845,24 @@ static struct echo_object *cl_echo_object_find(struct echo_device *d,
info = echo_env_info(env);
conf = &info->eti_conf;
if (d->ed_next) {
- struct lov_oinfo *oinfo = lsm->lsm_oinfo[0];
+ oinfo = kzalloc(sizeof(*oinfo), GFP_NOFS);
+ if (!oinfo) {
+ eco = ERR_PTR(-ENOMEM);
+ goto out;
+ }
- LASSERT(oinfo);
- oinfo->loi_oi = lsm->lsm_oi;
+ oinfo->loi_oi = *oi;
conf->eoc_cl.u.coc_oinfo = oinfo;
}
- conf->eoc_md = lsmp;
+
+ /*
+ * If echo_object_init() is successful then ownership of oinfo
+ * is transferred to the object.
+ */
+ conf->eoc_oinfo = &oinfo;
fid = &info->eti_fid;
- rc = ostid_to_fid(fid, &lsm->lsm_oi, 0);
+ rc = ostid_to_fid(fid, (struct ost_id *)oi, 0);
if (rc != 0) {
eco = ERR_PTR(rc);
goto out;
@@ -928,6 +885,7 @@ static struct echo_object *cl_echo_object_find(struct echo_device *d,
}
out:
+ kfree(oinfo);
cl_env_put(env, &refcheck);
return eco;
}
@@ -1146,7 +1104,6 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
{
struct echo_object *eco;
struct echo_client_obd *ec = ed->ed_ec;
- struct lov_stripe_md *lsm = NULL;
int rc;
int created = 0;
@@ -1157,30 +1114,19 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
return -EINVAL;
}
- rc = echo_alloc_memmd(ed, &lsm);
- if (rc < 0) {
- CERROR("Cannot allocate md: rc = %d\n", rc);
- goto failed;
- }
-
- /* setup object ID here */
- lsm->lsm_oi = oa->o_oi;
+ if (!ostid_id(&oa->o_oi))
+ ostid_set_id(&oa->o_oi, ++last_object_id);
- if (ostid_id(&lsm->lsm_oi) == 0)
- ostid_set_id(&lsm->lsm_oi, ++last_object_id);
-
- rc = obd_create(env, ec->ec_exp, oa, &lsm, oti);
+ rc = obd_create(env, ec->ec_exp, oa, oti);
if (rc != 0) {
CERROR("Cannot create objects: rc = %d\n", rc);
goto failed;
}
created = 1;
- /* See what object ID we were given */
- oa->o_oi = lsm->lsm_oi;
oa->o_valid |= OBD_MD_FLID;
- eco = cl_echo_object_find(ed, &lsm);
+ eco = cl_echo_object_find(ed, &oa->o_oi);
if (IS_ERR(eco)) {
rc = PTR_ERR(eco);
goto failed;
@@ -1191,9 +1137,7 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
failed:
if (created && rc)
- obd_destroy(env, ec->ec_exp, oa, lsm, oti, NULL);
- if (lsm)
- echo_free_memmd(ed, &lsm);
+ obd_destroy(env, ec->ec_exp, oa, NULL, oti, NULL);
if (rc)
CERROR("create object failed with: rc = %d\n", rc);
return rc;
@@ -1202,32 +1146,21 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
static int echo_get_object(struct echo_object **ecop, struct echo_device *ed,
struct obdo *oa)
{
- struct lov_stripe_md *lsm = NULL;
struct echo_object *eco;
int rc;
- if ((oa->o_valid & OBD_MD_FLID) == 0 || ostid_id(&oa->o_oi) == 0) {
- /* disallow use of object id 0 */
- CERROR("No valid oid\n");
+ if (!(oa->o_valid & OBD_MD_FLID) || !(oa->o_valid & OBD_MD_FLGROUP) ||
+ !ostid_id(&oa->o_oi)) {
+ CERROR("invalid oid " DOSTID "\n", POSTID(&oa->o_oi));
return -EINVAL;
}
- rc = echo_alloc_memmd(ed, &lsm);
- if (rc < 0)
- return rc;
-
- lsm->lsm_oi = oa->o_oi;
- if (!(oa->o_valid & OBD_MD_FLGROUP))
- ostid_set_seq_echo(&lsm->lsm_oi);
-
rc = 0;
- eco = cl_echo_object_find(ed, &lsm);
+ eco = cl_echo_object_find(ed, &oa->o_oi);
if (!IS_ERR(eco))
*ecop = eco;
else
rc = PTR_ERR(eco);
- if (lsm)
- echo_free_memmd(ed, &lsm);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 71f6abe..aefd776 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -398,24 +398,16 @@ static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
oinfo->oi_cb_up, oinfo, rqset);
}
-static int osc_real_create(struct obd_export *exp, struct obdo *oa,
- struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
+static int osc_create(const struct lu_env *env, struct obd_export *exp,
+ struct obdo *oa, struct obd_trans_info *oti)
{
struct ptlrpc_request *req;
struct ost_body *body;
- struct lov_stripe_md *lsm;
int rc;
LASSERT(oa);
- LASSERT(ea);
-
- lsm = *ea;
- if (!lsm) {
- rc = obd_alloc_memmd(exp, &lsm);
- if (rc < 0)
- return rc;
- }
+ LASSERT(oa->o_valid & OBD_MD_FLGROUP);
+ LASSERT(fid_seq_is_echo(ostid_seq(&oa->o_oi)));
req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
if (!req) {
@@ -461,13 +453,6 @@ static int osc_real_create(struct obd_export *exp, struct obdo *oa,
oa->o_blksize = cli_brw_size(exp->exp_obd);
oa->o_valid |= OBD_MD_FLBLKSZ;
- /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
- * have valid lsm_oinfo data structs, so don't go touching that.
- * This needs to be fixed in a big way.
- */
- lsm->lsm_oi = oa->o_oi;
- *ea = lsm;
-
if (oti && oa->o_valid & OBD_MD_FLCOOKIE) {
if (!oti->oti_logcookies)
oti->oti_logcookies = &oti->oti_onecookie;
@@ -479,8 +464,6 @@ static int osc_real_create(struct obd_export *exp, struct obdo *oa,
out_req:
ptlrpc_req_finished(req);
out:
- if (rc && !*ea)
- obd_free_memmd(exp, &lsm);
return rc;
}
@@ -656,25 +639,6 @@ static int osc_can_send_destroy(struct client_obd *cli)
return 0;
}
-static int osc_create(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md **ea,
- struct obd_trans_info *oti)
-{
- int rc = 0;
-
- LASSERT(oa);
- LASSERT(ea);
- LASSERT(oa->o_valid & OBD_MD_FLGROUP);
-
- if (!fid_seq_is_mdt(ostid_seq(&oa->o_oi)))
- return osc_real_create(exp, oa, ea, oti);
-
- /* we should not get here anymore */
- LBUG();
-
- return rc;
-}
-
/* Destroy requests can be async always on the client, and we don't even really
* care about the return code since the client cannot do anything at all about
* a destroy failure.
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 073/124] staging: lustre: llite: remove ll_objects_destroy()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove ll_objects_destroy(). This function is not needed for
interoperability with servers of version 2.4 or higher.
Remove the then unused function lov_destroy() and its supporting
functions. Remove the lsm_destroy method of struct lsm_operations.
Remove the unused struct lov_stripe_md, MD export, and capa parameters
from obd_destroy() and its implementations.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12618
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 5 +-
drivers/staging/lustre/lustre/include/obd_class.h | 6 +-
drivers/staging/lustre/lustre/llite/file.c | 6 --
.../staging/lustre/lustre/llite/llite_internal.h | 2 -
drivers/staging/lustre/lustre/llite/namei.c | 73 ------------------
drivers/staging/lustre/lustre/lov/lov_ea.c | 8 --
drivers/staging/lustre/lustre/lov/lov_internal.h | 5 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 54 --------------
drivers/staging/lustre/lustre/lov/lov_request.c | 78 --------------------
.../staging/lustre/lustre/obdecho/echo_client.c | 5 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 3 +-
11 files changed, 6 insertions(+), 239 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index ca63891..cd7f552 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -869,8 +869,7 @@ struct obd_ops {
int (*create)(const struct lu_env *env, struct obd_export *exp,
struct obdo *oa, struct obd_trans_info *oti);
int (*destroy)(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md *ea,
- struct obd_trans_info *oti, struct obd_export *md_exp);
+ struct obdo *oa, struct obd_trans_info *oti);
int (*setattr)(const struct lu_env *, struct obd_export *exp,
struct obd_info *oinfo, struct obd_trans_info *oti);
int (*setattr_async)(struct obd_export *exp, struct obd_info *oinfo,
@@ -1058,8 +1057,6 @@ struct md_ops {
struct lsm_operations {
void (*lsm_free)(struct lov_stripe_md *);
- int (*lsm_destroy)(struct lov_stripe_md *, struct obdo *oa,
- struct obd_export *md_exp);
void (*lsm_stripe_by_index)(struct lov_stripe_md *, int *, u64 *,
u64 *);
void (*lsm_stripe_by_offset)(struct lov_stripe_md *, int *, u64 *,
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 2bf2826..9836aed 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -702,16 +702,14 @@ static inline int obd_create(const struct lu_env *env, struct obd_export *exp,
}
static inline int obd_destroy(const struct lu_env *env, struct obd_export *exp,
- struct obdo *obdo, struct lov_stripe_md *ea,
- struct obd_trans_info *oti,
- struct obd_export *md_exp)
+ struct obdo *obdo, struct obd_trans_info *oti)
{
int rc;
EXP_CHECK_DT_OP(exp, destroy);
EXP_COUNTER_INCREMENT(exp, destroy);
- rc = OBP(exp->exp_obd, destroy)(env, exp, obdo, ea, oti, md_exp);
+ rc = OBP(exp->exp_obd, destroy)(env, exp, obdo, oti);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index a562b11..0b3b387 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -189,12 +189,6 @@ static int ll_close_inode_openhandle(struct obd_export *md_exp,
spin_unlock(&lli->lli_lock);
}
- if (rc == 0) {
- rc = ll_objects_destroy(req, inode);
- if (rc)
- CERROR("inode %lu ll_objects destroy: rc = %d\n",
- inode->i_ino, rc);
- }
if (rc == 0 && op_data->op_bias & MDS_HSM_RELEASE) {
struct mdt_body *body;
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index bdfdff5..0dc47f2 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -721,8 +721,6 @@ void ll_release_page(struct inode *inode, struct page *page, bool remove);
/* llite/namei.c */
extern const struct inode_operations ll_special_inode_operations;
-int ll_objects_destroy(struct ptlrpc_request *request,
- struct inode *dir);
struct inode *ll_iget(struct super_block *sb, ino_t hash,
struct lustre_md *lic);
int ll_test_inode_by_fid(struct inode *inode, void *opaque);
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 494140a..1dd5df5 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -890,77 +890,6 @@ static int ll_create_nd(struct inode *dir, struct dentry *dentry,
return rc;
}
-int ll_objects_destroy(struct ptlrpc_request *request, struct inode *dir)
-{
- struct mdt_body *body;
- struct lov_mds_md *eadata;
- struct lov_stripe_md *lsm = NULL;
- struct obd_trans_info oti = { 0 };
- struct obdo *oa;
- int rc;
-
- /* req is swabbed so this is safe */
- body = req_capsule_server_get(&request->rq_pill, &RMF_MDT_BODY);
- if (!(body->mbo_valid & OBD_MD_FLEASIZE))
- return 0;
-
- if (body->mbo_eadatasize == 0) {
- CERROR("OBD_MD_FLEASIZE set but eadatasize zero\n");
- rc = -EPROTO;
- goto out;
- }
-
- /* The MDS sent back the EA because we unlinked the last reference
- * to this file. Use this EA to unlink the objects on the OST.
- * It's opaque so we don't swab here; we leave it to obd_unpackmd() to
- * check it is complete and sensible.
- */
- eadata = req_capsule_server_sized_get(&request->rq_pill, &RMF_MDT_MD,
- body->mbo_eadatasize);
- LASSERT(eadata);
-
- rc = obd_unpackmd(ll_i2dtexp(dir), &lsm, eadata, body->mbo_eadatasize);
- if (rc < 0) {
- CERROR("obd_unpackmd: %d\n", rc);
- goto out;
- }
- LASSERT(rc >= sizeof(*lsm));
-
- oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
- if (!oa) {
- rc = -ENOMEM;
- goto out_free_memmd;
- }
-
- oa->o_oi = lsm->lsm_oi;
- oa->o_mode = body->mbo_mode & S_IFMT;
- oa->o_valid = OBD_MD_FLID | OBD_MD_FLTYPE | OBD_MD_FLGROUP;
-
- if (body->mbo_valid & OBD_MD_FLCOOKIE) {
- oa->o_valid |= OBD_MD_FLCOOKIE;
- oti.oti_logcookies =
- req_capsule_server_sized_get(&request->rq_pill,
- &RMF_LOGCOOKIES,
- sizeof(struct llog_cookie) *
- lsm->lsm_stripe_count);
- if (!oti.oti_logcookies) {
- oa->o_valid &= ~OBD_MD_FLCOOKIE;
- body->mbo_valid &= ~OBD_MD_FLCOOKIE;
- }
- }
-
- rc = obd_destroy(NULL, ll_i2dtexp(dir), oa, lsm, &oti,
- ll_i2mdexp(dir));
- if (rc)
- CERROR("obd destroy objid "DOSTID" error %d\n",
- POSTID(&lsm->lsm_oi), rc);
-out_free_memmd:
- obd_free_memmd(ll_i2dtexp(dir), &lsm);
- kmem_cache_free(obdo_cachep, oa);
-out:
- return rc;
-}
-
/* ll_unlink() doesn't update the inode with the new link count.
* Instead, ll_ddelete() and ll_d_iput() will update it based upon if there
* is any lock existing. They will recycle dentries and inodes based upon locks
@@ -994,7 +923,6 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild)
ll_update_times(request, dir);
ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_UNLINK, 1);
- rc = ll_objects_destroy(request, dir);
out:
ptlrpc_req_finished(request);
return rc;
@@ -1130,7 +1058,6 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild,
ll_update_times(request, src);
ll_update_times(request, tgt);
ll_stats_ops_tally(sbi, LPROC_LL_RENAME, 1);
- err = ll_objects_destroy(request, src);
}
ptlrpc_req_finished(request);
diff --git a/drivers/staging/lustre/lustre/lov/lov_ea.c b/drivers/staging/lustre/lustre/lov/lov_ea.c
index 9f0b5b7..214c561 100644
--- a/drivers/staging/lustre/lustre/lov/lov_ea.c
+++ b/drivers/staging/lustre/lustre/lov/lov_ea.c
@@ -147,12 +147,6 @@ lsm_stripe_by_offset_plain(struct lov_stripe_md *lsm, int *stripeno,
*swidth = (u64)lsm->lsm_stripe_size * lsm->lsm_stripe_count;
}
-static int lsm_destroy_plain(struct lov_stripe_md *lsm, struct obdo *oa,
- struct obd_export *md_exp)
-{
- return 0;
-}
-
/* Find minimum stripe maxbytes value. For inactive or
* reconnecting targets use LUSTRE_EXT3_STRIPE_MAXBYTES.
*/
@@ -246,7 +240,6 @@ static int lsm_unpackmd_v1(struct lov_obd *lov, struct lov_stripe_md *lsm,
const struct lsm_operations lsm_v1_ops = {
.lsm_free = lsm_free_plain,
- .lsm_destroy = lsm_destroy_plain,
.lsm_stripe_by_index = lsm_stripe_by_index_plain,
.lsm_stripe_by_offset = lsm_stripe_by_offset_plain,
.lsm_lmm_verify = lsm_lmm_verify_v1,
@@ -336,7 +329,6 @@ static int lsm_unpackmd_v3(struct lov_obd *lov, struct lov_stripe_md *lsm,
const struct lsm_operations lsm_v3_ops = {
.lsm_free = lsm_free_plain,
- .lsm_destroy = lsm_destroy_plain,
.lsm_stripe_by_index = lsm_stripe_by_index_plain,
.lsm_stripe_by_offset = lsm_stripe_by_offset_plain,
.lsm_lmm_verify = lsm_lmm_verify_v3,
diff --git a/drivers/staging/lustre/lustre/lov/lov_internal.h b/drivers/staging/lustre/lustre/lov/lov_internal.h
index 5ddaeda..0e5fc2e 100644
--- a/drivers/staging/lustre/lustre/lov/lov_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_internal.h
@@ -155,11 +155,6 @@ int lov_update_common_set(struct lov_request_set *set,
int lov_prep_getattr_set(struct obd_export *exp, struct obd_info *oinfo,
struct lov_request_set **reqset);
int lov_fini_getattr_set(struct lov_request_set *set);
-int lov_prep_destroy_set(struct obd_export *exp, struct obd_info *oinfo,
- struct obdo *src_oa, struct lov_stripe_md *lsm,
- struct obd_trans_info *oti,
- struct lov_request_set **reqset);
-int lov_fini_destroy_set(struct lov_request_set *set);
int lov_prep_setattr_set(struct obd_export *exp, struct obd_info *oinfo,
struct obd_trans_info *oti,
struct lov_request_set **reqset);
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index f433515..9b12159 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -980,59 +980,6 @@ do { \
"%p->lsm_magic=%x\n", (lsmp), (lsmp)->lsm_magic); \
} while (0)
-static int lov_destroy(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md *lsm,
- struct obd_trans_info *oti, struct obd_export *md_exp)
-{
- struct lov_request_set *set;
- struct obd_info oinfo;
- struct lov_request *req;
- struct lov_obd *lov;
- int rc = 0, err = 0;
-
- ASSERT_LSM_MAGIC(lsm);
-
- if (!exp || !exp->exp_obd)
- return -ENODEV;
-
- if (oa->o_valid & OBD_MD_FLCOOKIE) {
- LASSERT(oti);
- LASSERT(oti->oti_logcookies);
- }
-
- lov = &exp->exp_obd->u.lov;
- obd_getref(exp->exp_obd);
- rc = lov_prep_destroy_set(exp, &oinfo, oa, lsm, oti, &set);
- if (rc)
- goto out;
-
- list_for_each_entry(req, &set->set_list, rq_link) {
- if (oa->o_valid & OBD_MD_FLCOOKIE)
- oti->oti_logcookies = set->set_cookies + req->rq_stripe;
-
- err = obd_destroy(env, lov->lov_tgts[req->rq_idx]->ltd_exp,
- req->rq_oi.oi_oa, NULL, oti, NULL);
- err = lov_update_common_set(set, req, err);
- if (err) {
- CERROR("%s: destroying objid "DOSTID" subobj "
- DOSTID" on OST idx %d: rc = %d\n",
- exp->exp_obd->obd_name, POSTID(&oa->o_oi),
- POSTID(&req->rq_oi.oi_oa->o_oi),
- req->rq_idx, err);
- if (!rc)
- rc = err;
- }
- }
-
- if (rc == 0)
- rc = lsm_op_find(lsm->lsm_magic)->lsm_destroy(lsm, oa, md_exp);
-
- err = lov_fini_destroy_set(set);
-out:
- obd_putref(exp->exp_obd);
- return rc ? rc : err;
-}
-
static int lov_getattr_interpret(struct ptlrpc_request_set *rqset,
void *data, int rc)
{
@@ -2081,7 +2028,6 @@ static struct obd_ops lov_obd_ops = {
.statfs_async = lov_statfs_async,
.packmd = lov_packmd,
.unpackmd = lov_unpackmd,
- .destroy = lov_destroy,
.getattr_async = lov_getattr_async,
.setattr_async = lov_setattr_async,
.iocontrol = lov_iocontrol,
diff --git a/drivers/staging/lustre/lustre/lov/lov_request.c b/drivers/staging/lustre/lustre/lov/lov_request.c
index 4099b51..09dcaf4 100644
--- a/drivers/staging/lustre/lustre/lov/lov_request.c
+++ b/drivers/staging/lustre/lustre/lov/lov_request.c
@@ -325,84 +325,6 @@ out_set:
return rc;
}
-int lov_fini_destroy_set(struct lov_request_set *set)
-{
- if (!set)
- return 0;
- LASSERT(set->set_exp);
- if (atomic_read(&set->set_completes)) {
- /* FIXME update qos data here */
- }
-
- lov_put_reqset(set);
-
- return 0;
-}
-
-int lov_prep_destroy_set(struct obd_export *exp, struct obd_info *oinfo,
- struct obdo *src_oa, struct lov_stripe_md *lsm,
- struct obd_trans_info *oti,
- struct lov_request_set **reqset)
-{
- struct lov_request_set *set;
- struct lov_obd *lov = &exp->exp_obd->u.lov;
- int rc = 0, i;
-
- set = kzalloc(sizeof(*set), GFP_NOFS);
- if (!set)
- return -ENOMEM;
- lov_init_set(set);
-
- set->set_exp = exp;
- set->set_oi = oinfo;
- set->set_oi->oi_md = lsm;
- set->set_oi->oi_oa = src_oa;
- if (oti && src_oa->o_valid & OBD_MD_FLCOOKIE)
- set->set_cookies = oti->oti_logcookies;
-
- for (i = 0; i < lsm->lsm_stripe_count; i++) {
- struct lov_oinfo *loi;
- struct lov_request *req;
-
- loi = lsm->lsm_oinfo[i];
- if (lov_oinfo_is_dummy(loi))
- continue;
-
- if (!lov_check_and_wait_active(lov, loi->loi_ost_idx)) {
- CDEBUG(D_HA, "lov idx %d inactive\n", loi->loi_ost_idx);
- continue;
- }
-
- req = kzalloc(sizeof(*req), GFP_NOFS);
- if (!req) {
- rc = -ENOMEM;
- goto out_set;
- }
-
- req->rq_stripe = i;
- req->rq_idx = loi->loi_ost_idx;
-
- req->rq_oi.oi_oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
- if (!req->rq_oi.oi_oa) {
- kfree(req);
- rc = -ENOMEM;
- goto out_set;
- }
- memcpy(req->rq_oi.oi_oa, src_oa, sizeof(*req->rq_oi.oi_oa));
- req->rq_oi.oi_oa->o_oi = loi->loi_oi;
- lov_set_add_req(req, set);
- }
- if (!set->set_count) {
- rc = -EIO;
- goto out_set;
- }
- *reqset = set;
- return rc;
-out_set:
- lov_fini_destroy_set(set);
- return rc;
-}
-
int lov_fini_setattr_set(struct lov_request_set *set)
{
int rc = 0;
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index e487fda..91d28b5 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1137,7 +1137,7 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
failed:
if (created && rc)
- obd_destroy(env, ec->ec_exp, oa, NULL, oti, NULL);
+ obd_destroy(env, ec->ec_exp, oa, oti);
if (rc)
CERROR("create object failed with: rc = %d\n", rc);
return rc;
@@ -1546,8 +1546,7 @@ echo_client_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
rc = echo_get_object(&eco, ed, oa);
if (rc == 0) {
- rc = obd_destroy(env, ec->ec_exp, oa, NULL,
- &dummy_oti, NULL);
+ rc = obd_destroy(env, ec->ec_exp, oa, &dummy_oti);
if (rc == 0)
eco->eo_deleted = 1;
echo_put_object(eco);
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index aefd776..59ce7a8 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -651,8 +651,7 @@ static int osc_can_send_destroy(struct client_obd *cli)
* cookies to the MDS after committing destroy transactions.
*/
static int osc_destroy(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md *ea,
- struct obd_trans_info *oti, struct obd_export *md_export)
+ struct obdo *oa, struct obd_trans_info *oti)
{
struct client_obd *cli = &exp->exp_obd->u.cli;
struct ptlrpc_request *req;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 073/124] staging: lustre: llite: remove ll_objects_destroy()
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Remove ll_objects_destroy(). This function is not needed for
interoperability with servers of version 2.4 or higher.
Remove the then unused function lov_destroy() and its supporting
functions. Remove the lsm_destroy method of struct lsm_operations.
Remove the unused struct lov_stripe_md, MD export, and capa parameters
from obd_destroy() and its implementations.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12618
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 5 +-
drivers/staging/lustre/lustre/include/obd_class.h | 6 +-
drivers/staging/lustre/lustre/llite/file.c | 6 --
.../staging/lustre/lustre/llite/llite_internal.h | 2 -
drivers/staging/lustre/lustre/llite/namei.c | 73 ------------------
drivers/staging/lustre/lustre/lov/lov_ea.c | 8 --
drivers/staging/lustre/lustre/lov/lov_internal.h | 5 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 54 --------------
drivers/staging/lustre/lustre/lov/lov_request.c | 78 --------------------
.../staging/lustre/lustre/obdecho/echo_client.c | 5 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 3 +-
11 files changed, 6 insertions(+), 239 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index ca63891..cd7f552 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -869,8 +869,7 @@ struct obd_ops {
int (*create)(const struct lu_env *env, struct obd_export *exp,
struct obdo *oa, struct obd_trans_info *oti);
int (*destroy)(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md *ea,
- struct obd_trans_info *oti, struct obd_export *md_exp);
+ struct obdo *oa, struct obd_trans_info *oti);
int (*setattr)(const struct lu_env *, struct obd_export *exp,
struct obd_info *oinfo, struct obd_trans_info *oti);
int (*setattr_async)(struct obd_export *exp, struct obd_info *oinfo,
@@ -1058,8 +1057,6 @@ struct md_ops {
struct lsm_operations {
void (*lsm_free)(struct lov_stripe_md *);
- int (*lsm_destroy)(struct lov_stripe_md *, struct obdo *oa,
- struct obd_export *md_exp);
void (*lsm_stripe_by_index)(struct lov_stripe_md *, int *, u64 *,
u64 *);
void (*lsm_stripe_by_offset)(struct lov_stripe_md *, int *, u64 *,
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 2bf2826..9836aed 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -702,16 +702,14 @@ static inline int obd_create(const struct lu_env *env, struct obd_export *exp,
}
static inline int obd_destroy(const struct lu_env *env, struct obd_export *exp,
- struct obdo *obdo, struct lov_stripe_md *ea,
- struct obd_trans_info *oti,
- struct obd_export *md_exp)
+ struct obdo *obdo, struct obd_trans_info *oti)
{
int rc;
EXP_CHECK_DT_OP(exp, destroy);
EXP_COUNTER_INCREMENT(exp, destroy);
- rc = OBP(exp->exp_obd, destroy)(env, exp, obdo, ea, oti, md_exp);
+ rc = OBP(exp->exp_obd, destroy)(env, exp, obdo, oti);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index a562b11..0b3b387 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -189,12 +189,6 @@ static int ll_close_inode_openhandle(struct obd_export *md_exp,
spin_unlock(&lli->lli_lock);
}
- if (rc == 0) {
- rc = ll_objects_destroy(req, inode);
- if (rc)
- CERROR("inode %lu ll_objects destroy: rc = %d\n",
- inode->i_ino, rc);
- }
if (rc == 0 && op_data->op_bias & MDS_HSM_RELEASE) {
struct mdt_body *body;
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index bdfdff5..0dc47f2 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -721,8 +721,6 @@ void ll_release_page(struct inode *inode, struct page *page, bool remove);
/* llite/namei.c */
extern const struct inode_operations ll_special_inode_operations;
-int ll_objects_destroy(struct ptlrpc_request *request,
- struct inode *dir);
struct inode *ll_iget(struct super_block *sb, ino_t hash,
struct lustre_md *lic);
int ll_test_inode_by_fid(struct inode *inode, void *opaque);
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 494140a..1dd5df5 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -890,77 +890,6 @@ static int ll_create_nd(struct inode *dir, struct dentry *dentry,
return rc;
}
-int ll_objects_destroy(struct ptlrpc_request *request, struct inode *dir)
-{
- struct mdt_body *body;
- struct lov_mds_md *eadata;
- struct lov_stripe_md *lsm = NULL;
- struct obd_trans_info oti = { 0 };
- struct obdo *oa;
- int rc;
-
- /* req is swabbed so this is safe */
- body = req_capsule_server_get(&request->rq_pill, &RMF_MDT_BODY);
- if (!(body->mbo_valid & OBD_MD_FLEASIZE))
- return 0;
-
- if (body->mbo_eadatasize == 0) {
- CERROR("OBD_MD_FLEASIZE set but eadatasize zero\n");
- rc = -EPROTO;
- goto out;
- }
-
- /* The MDS sent back the EA because we unlinked the last reference
- * to this file. Use this EA to unlink the objects on the OST.
- * It's opaque so we don't swab here; we leave it to obd_unpackmd() to
- * check it is complete and sensible.
- */
- eadata = req_capsule_server_sized_get(&request->rq_pill, &RMF_MDT_MD,
- body->mbo_eadatasize);
- LASSERT(eadata);
-
- rc = obd_unpackmd(ll_i2dtexp(dir), &lsm, eadata, body->mbo_eadatasize);
- if (rc < 0) {
- CERROR("obd_unpackmd: %d\n", rc);
- goto out;
- }
- LASSERT(rc >= sizeof(*lsm));
-
- oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
- if (!oa) {
- rc = -ENOMEM;
- goto out_free_memmd;
- }
-
- oa->o_oi = lsm->lsm_oi;
- oa->o_mode = body->mbo_mode & S_IFMT;
- oa->o_valid = OBD_MD_FLID | OBD_MD_FLTYPE | OBD_MD_FLGROUP;
-
- if (body->mbo_valid & OBD_MD_FLCOOKIE) {
- oa->o_valid |= OBD_MD_FLCOOKIE;
- oti.oti_logcookies =
- req_capsule_server_sized_get(&request->rq_pill,
- &RMF_LOGCOOKIES,
- sizeof(struct llog_cookie) *
- lsm->lsm_stripe_count);
- if (!oti.oti_logcookies) {
- oa->o_valid &= ~OBD_MD_FLCOOKIE;
- body->mbo_valid &= ~OBD_MD_FLCOOKIE;
- }
- }
-
- rc = obd_destroy(NULL, ll_i2dtexp(dir), oa, lsm, &oti,
- ll_i2mdexp(dir));
- if (rc)
- CERROR("obd destroy objid "DOSTID" error %d\n",
- POSTID(&lsm->lsm_oi), rc);
-out_free_memmd:
- obd_free_memmd(ll_i2dtexp(dir), &lsm);
- kmem_cache_free(obdo_cachep, oa);
-out:
- return rc;
-}
-
/* ll_unlink() doesn't update the inode with the new link count.
* Instead, ll_ddelete() and ll_d_iput() will update it based upon if there
* is any lock existing. They will recycle dentries and inodes based upon locks
@@ -994,7 +923,6 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild)
ll_update_times(request, dir);
ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_UNLINK, 1);
- rc = ll_objects_destroy(request, dir);
out:
ptlrpc_req_finished(request);
return rc;
@@ -1130,7 +1058,6 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild,
ll_update_times(request, src);
ll_update_times(request, tgt);
ll_stats_ops_tally(sbi, LPROC_LL_RENAME, 1);
- err = ll_objects_destroy(request, src);
}
ptlrpc_req_finished(request);
diff --git a/drivers/staging/lustre/lustre/lov/lov_ea.c b/drivers/staging/lustre/lustre/lov/lov_ea.c
index 9f0b5b7..214c561 100644
--- a/drivers/staging/lustre/lustre/lov/lov_ea.c
+++ b/drivers/staging/lustre/lustre/lov/lov_ea.c
@@ -147,12 +147,6 @@ lsm_stripe_by_offset_plain(struct lov_stripe_md *lsm, int *stripeno,
*swidth = (u64)lsm->lsm_stripe_size * lsm->lsm_stripe_count;
}
-static int lsm_destroy_plain(struct lov_stripe_md *lsm, struct obdo *oa,
- struct obd_export *md_exp)
-{
- return 0;
-}
-
/* Find minimum stripe maxbytes value. For inactive or
* reconnecting targets use LUSTRE_EXT3_STRIPE_MAXBYTES.
*/
@@ -246,7 +240,6 @@ static int lsm_unpackmd_v1(struct lov_obd *lov, struct lov_stripe_md *lsm,
const struct lsm_operations lsm_v1_ops = {
.lsm_free = lsm_free_plain,
- .lsm_destroy = lsm_destroy_plain,
.lsm_stripe_by_index = lsm_stripe_by_index_plain,
.lsm_stripe_by_offset = lsm_stripe_by_offset_plain,
.lsm_lmm_verify = lsm_lmm_verify_v1,
@@ -336,7 +329,6 @@ static int lsm_unpackmd_v3(struct lov_obd *lov, struct lov_stripe_md *lsm,
const struct lsm_operations lsm_v3_ops = {
.lsm_free = lsm_free_plain,
- .lsm_destroy = lsm_destroy_plain,
.lsm_stripe_by_index = lsm_stripe_by_index_plain,
.lsm_stripe_by_offset = lsm_stripe_by_offset_plain,
.lsm_lmm_verify = lsm_lmm_verify_v3,
diff --git a/drivers/staging/lustre/lustre/lov/lov_internal.h b/drivers/staging/lustre/lustre/lov/lov_internal.h
index 5ddaeda..0e5fc2e 100644
--- a/drivers/staging/lustre/lustre/lov/lov_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_internal.h
@@ -155,11 +155,6 @@ int lov_update_common_set(struct lov_request_set *set,
int lov_prep_getattr_set(struct obd_export *exp, struct obd_info *oinfo,
struct lov_request_set **reqset);
int lov_fini_getattr_set(struct lov_request_set *set);
-int lov_prep_destroy_set(struct obd_export *exp, struct obd_info *oinfo,
- struct obdo *src_oa, struct lov_stripe_md *lsm,
- struct obd_trans_info *oti,
- struct lov_request_set **reqset);
-int lov_fini_destroy_set(struct lov_request_set *set);
int lov_prep_setattr_set(struct obd_export *exp, struct obd_info *oinfo,
struct obd_trans_info *oti,
struct lov_request_set **reqset);
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index f433515..9b12159 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -980,59 +980,6 @@ do { \
"%p->lsm_magic=%x\n", (lsmp), (lsmp)->lsm_magic); \
} while (0)
-static int lov_destroy(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md *lsm,
- struct obd_trans_info *oti, struct obd_export *md_exp)
-{
- struct lov_request_set *set;
- struct obd_info oinfo;
- struct lov_request *req;
- struct lov_obd *lov;
- int rc = 0, err = 0;
-
- ASSERT_LSM_MAGIC(lsm);
-
- if (!exp || !exp->exp_obd)
- return -ENODEV;
-
- if (oa->o_valid & OBD_MD_FLCOOKIE) {
- LASSERT(oti);
- LASSERT(oti->oti_logcookies);
- }
-
- lov = &exp->exp_obd->u.lov;
- obd_getref(exp->exp_obd);
- rc = lov_prep_destroy_set(exp, &oinfo, oa, lsm, oti, &set);
- if (rc)
- goto out;
-
- list_for_each_entry(req, &set->set_list, rq_link) {
- if (oa->o_valid & OBD_MD_FLCOOKIE)
- oti->oti_logcookies = set->set_cookies + req->rq_stripe;
-
- err = obd_destroy(env, lov->lov_tgts[req->rq_idx]->ltd_exp,
- req->rq_oi.oi_oa, NULL, oti, NULL);
- err = lov_update_common_set(set, req, err);
- if (err) {
- CERROR("%s: destroying objid "DOSTID" subobj "
- DOSTID" on OST idx %d: rc = %d\n",
- exp->exp_obd->obd_name, POSTID(&oa->o_oi),
- POSTID(&req->rq_oi.oi_oa->o_oi),
- req->rq_idx, err);
- if (!rc)
- rc = err;
- }
- }
-
- if (rc == 0)
- rc = lsm_op_find(lsm->lsm_magic)->lsm_destroy(lsm, oa, md_exp);
-
- err = lov_fini_destroy_set(set);
-out:
- obd_putref(exp->exp_obd);
- return rc ? rc : err;
-}
-
static int lov_getattr_interpret(struct ptlrpc_request_set *rqset,
void *data, int rc)
{
@@ -2081,7 +2028,6 @@ static struct obd_ops lov_obd_ops = {
.statfs_async = lov_statfs_async,
.packmd = lov_packmd,
.unpackmd = lov_unpackmd,
- .destroy = lov_destroy,
.getattr_async = lov_getattr_async,
.setattr_async = lov_setattr_async,
.iocontrol = lov_iocontrol,
diff --git a/drivers/staging/lustre/lustre/lov/lov_request.c b/drivers/staging/lustre/lustre/lov/lov_request.c
index 4099b51..09dcaf4 100644
--- a/drivers/staging/lustre/lustre/lov/lov_request.c
+++ b/drivers/staging/lustre/lustre/lov/lov_request.c
@@ -325,84 +325,6 @@ out_set:
return rc;
}
-int lov_fini_destroy_set(struct lov_request_set *set)
-{
- if (!set)
- return 0;
- LASSERT(set->set_exp);
- if (atomic_read(&set->set_completes)) {
- /* FIXME update qos data here */
- }
-
- lov_put_reqset(set);
-
- return 0;
-}
-
-int lov_prep_destroy_set(struct obd_export *exp, struct obd_info *oinfo,
- struct obdo *src_oa, struct lov_stripe_md *lsm,
- struct obd_trans_info *oti,
- struct lov_request_set **reqset)
-{
- struct lov_request_set *set;
- struct lov_obd *lov = &exp->exp_obd->u.lov;
- int rc = 0, i;
-
- set = kzalloc(sizeof(*set), GFP_NOFS);
- if (!set)
- return -ENOMEM;
- lov_init_set(set);
-
- set->set_exp = exp;
- set->set_oi = oinfo;
- set->set_oi->oi_md = lsm;
- set->set_oi->oi_oa = src_oa;
- if (oti && src_oa->o_valid & OBD_MD_FLCOOKIE)
- set->set_cookies = oti->oti_logcookies;
-
- for (i = 0; i < lsm->lsm_stripe_count; i++) {
- struct lov_oinfo *loi;
- struct lov_request *req;
-
- loi = lsm->lsm_oinfo[i];
- if (lov_oinfo_is_dummy(loi))
- continue;
-
- if (!lov_check_and_wait_active(lov, loi->loi_ost_idx)) {
- CDEBUG(D_HA, "lov idx %d inactive\n", loi->loi_ost_idx);
- continue;
- }
-
- req = kzalloc(sizeof(*req), GFP_NOFS);
- if (!req) {
- rc = -ENOMEM;
- goto out_set;
- }
-
- req->rq_stripe = i;
- req->rq_idx = loi->loi_ost_idx;
-
- req->rq_oi.oi_oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
- if (!req->rq_oi.oi_oa) {
- kfree(req);
- rc = -ENOMEM;
- goto out_set;
- }
- memcpy(req->rq_oi.oi_oa, src_oa, sizeof(*req->rq_oi.oi_oa));
- req->rq_oi.oi_oa->o_oi = loi->loi_oi;
- lov_set_add_req(req, set);
- }
- if (!set->set_count) {
- rc = -EIO;
- goto out_set;
- }
- *reqset = set;
- return rc;
-out_set:
- lov_fini_destroy_set(set);
- return rc;
-}
-
int lov_fini_setattr_set(struct lov_request_set *set)
{
int rc = 0;
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index e487fda..91d28b5 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1137,7 +1137,7 @@ static int echo_create_object(const struct lu_env *env, struct echo_device *ed,
failed:
if (created && rc)
- obd_destroy(env, ec->ec_exp, oa, NULL, oti, NULL);
+ obd_destroy(env, ec->ec_exp, oa, oti);
if (rc)
CERROR("create object failed with: rc = %d\n", rc);
return rc;
@@ -1546,8 +1546,7 @@ echo_client_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
rc = echo_get_object(&eco, ed, oa);
if (rc == 0) {
- rc = obd_destroy(env, ec->ec_exp, oa, NULL,
- &dummy_oti, NULL);
+ rc = obd_destroy(env, ec->ec_exp, oa, &dummy_oti);
if (rc == 0)
eco->eo_deleted = 1;
echo_put_object(eco);
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index aefd776..59ce7a8 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -651,8 +651,7 @@ static int osc_can_send_destroy(struct client_obd *cli)
* cookies to the MDS after committing destroy transactions.
*/
static int osc_destroy(const struct lu_env *env, struct obd_export *exp,
- struct obdo *oa, struct lov_stripe_md *ea,
- struct obd_trans_info *oti, struct obd_export *md_export)
+ struct obdo *oa, struct obd_trans_info *oti)
{
struct client_obd *cli = &exp->exp_obd->u.cli;
struct ptlrpc_request *req;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 074/124] staging: lustre: changelog: Proper record remapping
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Fixed changelog_remap_rec() to correctly remap records emitted
with jobid_var=disabled, i.e. delivered by new servers but with
no jobid field.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5862
Reviewed-on: http://review.whamcloud.com/12574
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Robert Read <robert.read@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 66 +++++++++++++-------
1 files changed, 44 insertions(+), 22 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index e8f81a5..ba679ad 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -881,41 +881,63 @@ static inline char *changelog_rec_sname(struct changelog_rec *rec)
return cr_name + strlen(cr_name) + 1;
}
-/*
+/**
* Remap a record to the desired format as specified by the crf flags.
* The record must be big enough to contain the final remapped version.
+ * Superfluous extension fields are removed and missing ones are added
+ * and zeroed. The flags of the record are updated accordingly.
+ *
+ * The jobid and rename extensions can be added to a record, to match the
+ * format an application expects, typically. In this case, the newly added
+ * fields will be zeroed.
+ * The Jobid field can be removed, to guarantee compatibility with older
+ * clients that don't expect this field in the records they process.
+ *
+ * The following assumptions are being made:
+ * - CLF_RENAME will not be removed
+ * - CLF_JOBID will not be added without CLF_RENAME being added too
+ *
+ * @param[in,out] rec The record to remap.
+ * @param[in] crf_wanted Flags describing the desired extensions.
*/
static inline void changelog_remap_rec(struct changelog_rec *rec,
- enum changelog_rec_flags crf)
+ enum changelog_rec_flags crf_wanted)
{
- size_t var_size;
- char *var_part;
+ char *jid_mov, *rnm_mov;
- crf &= CLF_SUPPORTED;
+ crf_wanted &= CLF_SUPPORTED;
- if ((rec->cr_flags & CLF_SUPPORTED) == crf)
+ if ((rec->cr_flags & CLF_SUPPORTED) == crf_wanted)
return;
- if ((crf & CLF_JOBID) && rec->cr_flags & CLF_JOBID) {
- var_part = (char *)changelog_rec_jobid(rec);
- var_size = rec->cr_namelen + sizeof(struct changelog_ext_jobid);
- } else {
- var_part = changelog_rec_name(rec);
- var_size = rec->cr_namelen;
- }
+ /* First move the variable-length name field */
+ memmove((char *)rec + changelog_rec_offset(crf_wanted),
+ changelog_rec_name(rec), rec->cr_namelen);
+
+ /* Locations of jobid and rename extensions in the remapped record */
+ jid_mov = (char *)rec +
+ changelog_rec_offset(crf_wanted & ~CLF_JOBID);
+ rnm_mov = (char *)rec +
+ changelog_rec_offset(crf_wanted & ~(CLF_JOBID | CLF_RENAME));
+
+ /* Move the extension fields to the desired positions */
+ if ((crf_wanted & CLF_JOBID) && (rec->cr_flags & CLF_JOBID))
+ memmove(jid_mov, changelog_rec_jobid(rec),
+ sizeof(struct changelog_ext_jobid));
- memmove((char *)rec + changelog_rec_offset(crf & ~CLF_JOBID), var_part,
- var_size);
+ if ((crf_wanted & CLF_RENAME) && (rec->cr_flags & CLF_RENAME))
+ memmove(rnm_mov, changelog_rec_rename(rec),
+ sizeof(struct changelog_ext_rename));
- if ((crf & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
- memset(changelog_rec_rename(rec), 0,
- sizeof(struct changelog_ext_rename));
+ /* Clear newly added fields */
+ if ((crf_wanted & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
+ memset(jid_mov, 0, sizeof(struct changelog_ext_jobid));
- if ((crf & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
- memset(changelog_rec_jobid(rec), 0,
- sizeof(struct changelog_ext_jobid));
+ if ((crf_wanted & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
+ memset(rnm_mov, 0, sizeof(struct changelog_ext_rename));
- rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf;
+ /* Update the record's flags accordingly */
+ rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf_wanted;
}
struct ioc_changelog {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 074/124] staging: lustre: changelog: Proper record remapping
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Henri Doreau,
James Simmons
From: Henri Doreau <henri.doreau@cea.fr>
Fixed changelog_remap_rec() to correctly remap records emitted
with jobid_var=disabled, i.e. delivered by new servers but with
no jobid field.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5862
Reviewed-on: http://review.whamcloud.com/12574
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Robert Read <robert.read@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_user.h | 66 +++++++++++++-------
1 files changed, 44 insertions(+), 22 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index e8f81a5..ba679ad 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -881,41 +881,63 @@ static inline char *changelog_rec_sname(struct changelog_rec *rec)
return cr_name + strlen(cr_name) + 1;
}
-/*
+/**
* Remap a record to the desired format as specified by the crf flags.
* The record must be big enough to contain the final remapped version.
+ * Superfluous extension fields are removed and missing ones are added
+ * and zeroed. The flags of the record are updated accordingly.
+ *
+ * The jobid and rename extensions can be added to a record, to match the
+ * format an application expects, typically. In this case, the newly added
+ * fields will be zeroed.
+ * The Jobid field can be removed, to guarantee compatibility with older
+ * clients that don't expect this field in the records they process.
+ *
+ * The following assumptions are being made:
+ * - CLF_RENAME will not be removed
+ * - CLF_JOBID will not be added without CLF_RENAME being added too
+ *
+ * @param[in,out] rec The record to remap.
+ * @param[in] crf_wanted Flags describing the desired extensions.
*/
static inline void changelog_remap_rec(struct changelog_rec *rec,
- enum changelog_rec_flags crf)
+ enum changelog_rec_flags crf_wanted)
{
- size_t var_size;
- char *var_part;
+ char *jid_mov, *rnm_mov;
- crf &= CLF_SUPPORTED;
+ crf_wanted &= CLF_SUPPORTED;
- if ((rec->cr_flags & CLF_SUPPORTED) == crf)
+ if ((rec->cr_flags & CLF_SUPPORTED) == crf_wanted)
return;
- if ((crf & CLF_JOBID) && rec->cr_flags & CLF_JOBID) {
- var_part = (char *)changelog_rec_jobid(rec);
- var_size = rec->cr_namelen + sizeof(struct changelog_ext_jobid);
- } else {
- var_part = changelog_rec_name(rec);
- var_size = rec->cr_namelen;
- }
+ /* First move the variable-length name field */
+ memmove((char *)rec + changelog_rec_offset(crf_wanted),
+ changelog_rec_name(rec), rec->cr_namelen);
+
+ /* Locations of jobid and rename extensions in the remapped record */
+ jid_mov = (char *)rec +
+ changelog_rec_offset(crf_wanted & ~CLF_JOBID);
+ rnm_mov = (char *)rec +
+ changelog_rec_offset(crf_wanted & ~(CLF_JOBID | CLF_RENAME));
+
+ /* Move the extension fields to the desired positions */
+ if ((crf_wanted & CLF_JOBID) && (rec->cr_flags & CLF_JOBID))
+ memmove(jid_mov, changelog_rec_jobid(rec),
+ sizeof(struct changelog_ext_jobid));
- memmove((char *)rec + changelog_rec_offset(crf & ~CLF_JOBID), var_part,
- var_size);
+ if ((crf_wanted & CLF_RENAME) && (rec->cr_flags & CLF_RENAME))
+ memmove(rnm_mov, changelog_rec_rename(rec),
+ sizeof(struct changelog_ext_rename));
- if ((crf & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
- memset(changelog_rec_rename(rec), 0,
- sizeof(struct changelog_ext_rename));
+ /* Clear newly added fields */
+ if ((crf_wanted & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
+ memset(jid_mov, 0, sizeof(struct changelog_ext_jobid));
- if ((crf & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
- memset(changelog_rec_jobid(rec), 0,
- sizeof(struct changelog_ext_jobid));
+ if ((crf_wanted & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
+ memset(rnm_mov, 0, sizeof(struct changelog_ext_rename));
- rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf;
+ /* Update the record's flags accordingly */
+ rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf_wanted;
}
struct ioc_changelog {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 075/124] staging: lustre: recovery: don't replay closed open
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Niu Yawei,
James Simmons
From: Niu Yawei <yawei.niu@intel.com>
To avoid scanning the replay open list every time in the
ptlrpc_free_committed(), the fix of LU-2613 (4322e0f9) changed
the ptlrpc_free_committed() to skip the open list unless the
import generation is changed. That introduced a race which could
make a closed open being replayed:
1. Application calls ll_close_inode_openhandle()-> mdc_close(),
to close file, rq_replay is cleared, but the open request is
still on the imp_committed_list;
2. Before the md_clear_open_replay_data() is called for close,
client start replay, and that closed open will be replayed
mistakenly;
3. Open replay interpret callback (mdc_replay_open) could race
with the mdc_clear_open_replay_data() at the end;
This patch fix the ptlrpc_free_committed() to make sure the
open list is scanned on recovery to prevent the closed open request
from being replayed.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5507
Reviewed-on: http://review.whamcloud.com/12667
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 6 +++++-
1 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index 27fe00d..a29ccaa 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -2455,7 +2455,8 @@ void ptlrpc_free_committed(struct obd_import *imp)
imp->imp_obd->obd_name, imp->imp_peer_committed_transno,
imp->imp_generation);
- if (imp->imp_generation != imp->imp_last_generation_checked)
+ if (imp->imp_generation != imp->imp_last_generation_checked ||
+ !imp->imp_last_transno_checked)
skip_committed_list = false;
imp->imp_last_transno_checked = imp->imp_peer_committed_transno;
@@ -2503,6 +2504,9 @@ free_req:
if (req->rq_import_generation < imp->imp_generation) {
DEBUG_REQ(D_RPCTRACE, req, "free stale open request");
ptlrpc_free_request(req);
+ } else if (!req->rq_replay) {
+ DEBUG_REQ(D_RPCTRACE, req, "free closed open request");
+ ptlrpc_free_request(req);
}
}
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 075/124] staging: lustre: recovery: don't replay closed open
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Niu Yawei,
James Simmons
From: Niu Yawei <yawei.niu@intel.com>
To avoid scanning the replay open list every time in the
ptlrpc_free_committed(), the fix of LU-2613 (4322e0f9) changed
the ptlrpc_free_committed() to skip the open list unless the
import generation is changed. That introduced a race which could
make a closed open being replayed:
1. Application calls ll_close_inode_openhandle()-> mdc_close(),
to close file, rq_replay is cleared, but the open request is
still on the imp_committed_list;
2. Before the md_clear_open_replay_data() is called for close,
client start replay, and that closed open will be replayed
mistakenly;
3. Open replay interpret callback (mdc_replay_open) could race
with the mdc_clear_open_replay_data() at the end;
This patch fix the ptlrpc_free_committed() to make sure the
open list is scanned on recovery to prevent the closed open request
from being replayed.
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5507
Reviewed-on: http://review.whamcloud.com/12667
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 6 +++++-
1 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index 27fe00d..a29ccaa 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -2455,7 +2455,8 @@ void ptlrpc_free_committed(struct obd_import *imp)
imp->imp_obd->obd_name, imp->imp_peer_committed_transno,
imp->imp_generation);
- if (imp->imp_generation != imp->imp_last_generation_checked)
+ if (imp->imp_generation != imp->imp_last_generation_checked ||
+ !imp->imp_last_transno_checked)
skip_committed_list = false;
imp->imp_last_transno_checked = imp->imp_peer_committed_transno;
@@ -2503,6 +2504,9 @@ free_req:
if (req->rq_import_generation < imp->imp_generation) {
DEBUG_REQ(D_RPCTRACE, req, "free stale open request");
ptlrpc_free_request(req);
+ } else if (!req->rq_replay) {
+ DEBUG_REQ(D_RPCTRACE, req, "free closed open request");
+ ptlrpc_free_request(req);
}
}
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 076/124] staging: lustre: ldlm: revert the changes for lock canceling policy
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
The changes for LRU lock policy was introduced by commit bfae5a4e,
where I was trying to revise the policy to pick locks for canceling.
However, this caused two problems as mentioned in LU-5727. The first
problem is that the lock can only be picked for canceling only if
the number of LRU locks is over preset LRU number AND it's aged; the
second problem is that mdc_cancel_weight() tends to not cancel OPEN
locks, therefore open locks can be kept forever and finally exhausts
memory on the MDT side.
The commit 7b2d26b0 ("revert changes to ldlm_cancel_aged_policy") fixed
the first problem. This patch will revert the rest of changes related
to LRU policy revise.
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5727
Reviewed-on: http://review.whamcloud.com/12733
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 6 ------
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index 46842dd..582de52 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -1198,9 +1198,6 @@ static enum ldlm_policy_res ldlm_cancel_lrur_policy(struct ldlm_namespace *ns,
if (slv == 0 || lv < slv)
return LDLM_POLICY_KEEP_LOCK;
- if (ns->ns_cancel && ns->ns_cancel(lock) == 0)
- return LDLM_POLICY_KEEP_LOCK;
-
return LDLM_POLICY_CANCEL_LOCK;
}
@@ -1244,9 +1241,6 @@ static enum ldlm_policy_res ldlm_cancel_aged_policy(struct ldlm_namespace *ns,
cfs_time_add(lock->l_last_used, ns->ns_max_age)))
return LDLM_POLICY_KEEP_LOCK;
- if (ns->ns_cancel && ns->ns_cancel(lock) == 0)
- return LDLM_POLICY_KEEP_LOCK;
-
return LDLM_POLICY_CANCEL_LOCK;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 076/124] staging: lustre: ldlm: revert the changes for lock canceling policy
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
The changes for LRU lock policy was introduced by commit bfae5a4e,
where I was trying to revise the policy to pick locks for canceling.
However, this caused two problems as mentioned in LU-5727. The first
problem is that the lock can only be picked for canceling only if
the number of LRU locks is over preset LRU number AND it's aged; the
second problem is that mdc_cancel_weight() tends to not cancel OPEN
locks, therefore open locks can be kept forever and finally exhausts
memory on the MDT side.
The commit 7b2d26b0 ("revert changes to ldlm_cancel_aged_policy") fixed
the first problem. This patch will revert the rest of changes related
to LRU policy revise.
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5727
Reviewed-on: http://review.whamcloud.com/12733
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 6 ------
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index 46842dd..582de52 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -1198,9 +1198,6 @@ static enum ldlm_policy_res ldlm_cancel_lrur_policy(struct ldlm_namespace *ns,
if (slv == 0 || lv < slv)
return LDLM_POLICY_KEEP_LOCK;
- if (ns->ns_cancel && ns->ns_cancel(lock) == 0)
- return LDLM_POLICY_KEEP_LOCK;
-
return LDLM_POLICY_CANCEL_LOCK;
}
@@ -1244,9 +1241,6 @@ static enum ldlm_policy_res ldlm_cancel_aged_policy(struct ldlm_namespace *ns,
cfs_time_add(lock->l_last_used, ns->ns_max_age)))
return LDLM_POLICY_KEEP_LOCK;
- if (ns->ns_cancel && ns->ns_cancel(lock) == 0)
- return LDLM_POLICY_KEEP_LOCK;
-
return LDLM_POLICY_CANCEL_LOCK;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 077/124] staging: lustre: ptlrpc: quiet errors on initial connection
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Andreas Dilger, Bobi Jam, James Simmons
From: Andreas Dilger <andreas.dilger@intel.com>
It may be that a client or MDS is trying to connect to a target (OST
or peer MDT) before that target is finished setup. Rather than
spamming the console logs during initial connection, only print a
console error message if there are repeated failures trying to
connect to the target, which may indicate an error on that node.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3456
Reviewed-on: http://review.whamcloud.com/10057
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 52 +++++++++++---------
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 2 +-
2 files changed, 30 insertions(+), 24 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index a29ccaa..f3914cc 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -1075,36 +1075,42 @@ static int ptlrpc_import_delay_req(struct obd_import *imp,
}
/**
- * Decide if the error message regarding provided request \a req
- * should be printed to the console or not.
- * Makes it's decision on request status and other properties.
- * Returns 1 to print error on the system console or 0 if not.
+ * Decide if the error message should be printed to the console or not.
+ * Makes its decision based on request type, status, and failure frequency.
+ *
+ * \param[in] req request that failed and may need a console message
+ *
+ * \retval false if no message should be printed
+ * \retval true if console message should be printed
*/
-static int ptlrpc_console_allow(struct ptlrpc_request *req)
+static bool ptlrpc_console_allow(struct ptlrpc_request *req)
{
__u32 opc;
- int err;
LASSERT(req->rq_reqmsg);
opc = lustre_msg_get_opc(req->rq_reqmsg);
- /*
- * Suppress particular reconnect errors which are to be expected. No
- * errors are suppressed for the initial connection on an import
- */
- if ((lustre_handle_is_used(&req->rq_import->imp_remote_handle)) &&
- (opc == OST_CONNECT || opc == MDS_CONNECT || opc == MGS_CONNECT)) {
+ /* Suppress particular reconnect errors which are to be expected. */
+ if (opc == OST_CONNECT || opc == MDS_CONNECT || opc == MGS_CONNECT) {
+ int err;
+
/* Suppress timed out reconnect requests */
- if (req->rq_timedout)
- return 0;
+ if (lustre_handle_is_used(&req->rq_import->imp_remote_handle) ||
+ req->rq_timedout)
+ return false;
- /* Suppress unavailable/again reconnect requests */
+ /*
+ * Suppress most unavailable/again reconnect requests, but
+ * print occasionally so it is clear client is trying to
+ * connect to a server where no target is running.
+ */
err = lustre_msg_get_status(req->rq_repmsg);
- if (err == -ENODEV || err == -EAGAIN)
- return 0;
+ if ((err == -ENODEV || err == -EAGAIN) &&
+ req->rq_import->imp_conn_cnt % 30 != 20)
+ return false;
}
- return 1;
+ return true;
}
/**
@@ -1118,14 +1124,14 @@ static int ptlrpc_check_status(struct ptlrpc_request *req)
err = lustre_msg_get_status(req->rq_repmsg);
if (lustre_msg_get_type(req->rq_repmsg) == PTL_RPC_MSG_ERR) {
struct obd_import *imp = req->rq_import;
+ lnet_nid_t nid = imp->imp_connection->c_peer.nid;
__u32 opc = lustre_msg_get_opc(req->rq_reqmsg);
if (ptlrpc_console_allow(req))
- LCONSOLE_ERROR_MSG(0x011, "%s: Communicating with %s, operation %s failed with %d.\n",
+ LCONSOLE_ERROR_MSG(0x011, "%s: operation %s to node %s failed: rc = %d\n",
imp->imp_obd->obd_name,
- libcfs_nid2str(
- imp->imp_connection->c_peer.nid),
- ll_opcode2str(opc), err);
+ ll_opcode2str(opc),
+ libcfs_nid2str(nid), err);
return err < 0 ? err : -EINVAL;
}
@@ -1282,7 +1288,7 @@ static int after_reply(struct ptlrpc_request *req)
* some reason. Try to reconnect, and if that fails, punt to
* the upcall.
*/
- if (ll_rpc_recoverable_error(rc)) {
+ if (ptlrpc_recoverable_error(rc)) {
if (req->rq_send_state != LUSTRE_IMP_FULL ||
imp->imp_obd->obd_no_recov || imp->imp_dlm_fake) {
return rc;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
index 29cfac2..b420aa8 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
+++ b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
@@ -270,7 +270,7 @@ void sptlrpc_conf_fini(void);
int sptlrpc_init(void);
void sptlrpc_fini(void);
-static inline int ll_rpc_recoverable_error(int rc)
+static inline bool ptlrpc_recoverable_error(int rc)
{
return (rc == -ENOTCONN || rc == -ENODEV);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 077/124] staging: lustre: ptlrpc: quiet errors on initial connection
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Andreas Dilger <andreas.dilger@intel.com>
It may be that a client or MDS is trying to connect to a target (OST
or peer MDT) before that target is finished setup. Rather than
spamming the console logs during initial connection, only print a
console error message if there are repeated failures trying to
connect to the target, which may indicate an error on that node.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3456
Reviewed-on: http://review.whamcloud.com/10057
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 52 +++++++++++---------
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 2 +-
2 files changed, 30 insertions(+), 24 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index a29ccaa..f3914cc 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -1075,36 +1075,42 @@ static int ptlrpc_import_delay_req(struct obd_import *imp,
}
/**
- * Decide if the error message regarding provided request \a req
- * should be printed to the console or not.
- * Makes it's decision on request status and other properties.
- * Returns 1 to print error on the system console or 0 if not.
+ * Decide if the error message should be printed to the console or not.
+ * Makes its decision based on request type, status, and failure frequency.
+ *
+ * \param[in] req request that failed and may need a console message
+ *
+ * \retval false if no message should be printed
+ * \retval true if console message should be printed
*/
-static int ptlrpc_console_allow(struct ptlrpc_request *req)
+static bool ptlrpc_console_allow(struct ptlrpc_request *req)
{
__u32 opc;
- int err;
LASSERT(req->rq_reqmsg);
opc = lustre_msg_get_opc(req->rq_reqmsg);
- /*
- * Suppress particular reconnect errors which are to be expected. No
- * errors are suppressed for the initial connection on an import
- */
- if ((lustre_handle_is_used(&req->rq_import->imp_remote_handle)) &&
- (opc == OST_CONNECT || opc == MDS_CONNECT || opc == MGS_CONNECT)) {
+ /* Suppress particular reconnect errors which are to be expected. */
+ if (opc == OST_CONNECT || opc == MDS_CONNECT || opc == MGS_CONNECT) {
+ int err;
+
/* Suppress timed out reconnect requests */
- if (req->rq_timedout)
- return 0;
+ if (lustre_handle_is_used(&req->rq_import->imp_remote_handle) ||
+ req->rq_timedout)
+ return false;
- /* Suppress unavailable/again reconnect requests */
+ /*
+ * Suppress most unavailable/again reconnect requests, but
+ * print occasionally so it is clear client is trying to
+ * connect to a server where no target is running.
+ */
err = lustre_msg_get_status(req->rq_repmsg);
- if (err == -ENODEV || err == -EAGAIN)
- return 0;
+ if ((err == -ENODEV || err == -EAGAIN) &&
+ req->rq_import->imp_conn_cnt % 30 != 20)
+ return false;
}
- return 1;
+ return true;
}
/**
@@ -1118,14 +1124,14 @@ static int ptlrpc_check_status(struct ptlrpc_request *req)
err = lustre_msg_get_status(req->rq_repmsg);
if (lustre_msg_get_type(req->rq_repmsg) == PTL_RPC_MSG_ERR) {
struct obd_import *imp = req->rq_import;
+ lnet_nid_t nid = imp->imp_connection->c_peer.nid;
__u32 opc = lustre_msg_get_opc(req->rq_reqmsg);
if (ptlrpc_console_allow(req))
- LCONSOLE_ERROR_MSG(0x011, "%s: Communicating with %s, operation %s failed with %d.\n",
+ LCONSOLE_ERROR_MSG(0x011, "%s: operation %s to node %s failed: rc = %d\n",
imp->imp_obd->obd_name,
- libcfs_nid2str(
- imp->imp_connection->c_peer.nid),
- ll_opcode2str(opc), err);
+ ll_opcode2str(opc),
+ libcfs_nid2str(nid), err);
return err < 0 ? err : -EINVAL;
}
@@ -1282,7 +1288,7 @@ static int after_reply(struct ptlrpc_request *req)
* some reason. Try to reconnect, and if that fails, punt to
* the upcall.
*/
- if (ll_rpc_recoverable_error(rc)) {
+ if (ptlrpc_recoverable_error(rc)) {
if (req->rq_send_state != LUSTRE_IMP_FULL ||
imp->imp_obd->obd_no_recov || imp->imp_dlm_fake) {
return rc;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
index 29cfac2..b420aa8 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
+++ b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
@@ -270,7 +270,7 @@ void sptlrpc_conf_fini(void);
int sptlrpc_init(void);
void sptlrpc_fini(void);
-static inline int ll_rpc_recoverable_error(int rc)
+static inline bool ptlrpc_recoverable_error(int rc)
{
return (rc == -ENOTCONN || rc == -ENODEV);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 078/124] staging: lustre: llog: prevent out-of-bound index
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
llog_process_thread() can be called from llog_cat_process_cb with an
index already out of bound, leading to the following crash:
LustreError: 3773:0:(llog.c:310:llog_process_thread())
ASSERTION(index <= last_index + 1 ) failed:
LustreError: 3773:0:(llog.c:310:llog_process_thread()) LBUG
#0 [ffff8801144bf900] machine_kexec at ffffffff81038f3b
#1 [ffff8801144bf960] crash_kexec at ffffffff810c5d82
#2 [ffff8801144bfa30] panic at ffffffff8152798a
#3 [ffff8801144bfab0] lbug_with_loc at ffffffffa02f8eeb [libcfs]
#4 [ffff8801144bfad0] llog_process_thread at ffffffffa0413fff [obdclass]
#5 [ffff8801144bfb80] llog_process_or_fork at ffffffffa041585f [obdclass]
#6 [ffff8801144bfbd0] llog_cat_process_cb at ffffffffa0418612 [obdclass]
#7 [ffff8801144bfc30] llog_process_thread at ffffffffa0413c22 [obdclass]
#8 [ffff8801144bfce0] llog_process_or_fork at ffffffffa041585f [obdclass]
#9 [ffff8801144bfd30] llog_cat_process_or_fork at ffffffffa0416b9d [obdclass]
If index is too big, simply return success.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5635
Reviewed-on: http://review.whamcloud.com/12161
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/llog.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/llog.c b/drivers/staging/lustre/lustre/obdclass/llog.c
index 119372c..43797f1 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog.c
@@ -235,6 +235,10 @@ static int llog_process_thread(void *arg)
else
last_index = LLOG_BITMAP_BYTES * 8 - 1;
+ /* Record is not in this buffer. */
+ if (index > last_index)
+ goto out;
+
while (rc == 0) {
struct llog_rec_hdr *rec;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 078/124] staging: lustre: llog: prevent out-of-bound index
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
llog_process_thread() can be called from llog_cat_process_cb with an
index already out of bound, leading to the following crash:
LustreError: 3773:0:(llog.c:310:llog_process_thread())
ASSERTION(index <= last_index + 1 ) failed:
LustreError: 3773:0:(llog.c:310:llog_process_thread()) LBUG
#0 [ffff8801144bf900] machine_kexec at ffffffff81038f3b
#1 [ffff8801144bf960] crash_kexec at ffffffff810c5d82
#2 [ffff8801144bfa30] panic at ffffffff8152798a
#3 [ffff8801144bfab0] lbug_with_loc at ffffffffa02f8eeb [libcfs]
#4 [ffff8801144bfad0] llog_process_thread at ffffffffa0413fff [obdclass]
#5 [ffff8801144bfb80] llog_process_or_fork at ffffffffa041585f [obdclass]
#6 [ffff8801144bfbd0] llog_cat_process_cb at ffffffffa0418612 [obdclass]
#7 [ffff8801144bfc30] llog_process_thread at ffffffffa0413c22 [obdclass]
#8 [ffff8801144bfce0] llog_process_or_fork at ffffffffa041585f [obdclass]
#9 [ffff8801144bfd30] llog_cat_process_or_fork at ffffffffa0416b9d [obdclass]
If index is too big, simply return success.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5635
Reviewed-on: http://review.whamcloud.com/12161
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/llog.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/llog.c b/drivers/staging/lustre/lustre/obdclass/llog.c
index 119372c..43797f1 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog.c
@@ -235,6 +235,10 @@ static int llog_process_thread(void *arg)
else
last_index = LLOG_BITMAP_BYTES * 8 - 1;
+ /* Record is not in this buffer. */
+ if (index > last_index)
+ goto out;
+
while (rc == 0) {
struct llog_rec_hdr *rec;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 079/124] staging: lustre: mgc: add nid iteration
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Boyko, James Simmons
From: Alexander Boyko <alexander.boyko@seagate.com>
mgc_apply_recover_logs use only first nid from entry,
this could be the problem for a cluster with several network
address for a one node.
Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5950
Xyratex-bug-id: MRP-2255
Reviewed-on: http://review.whamcloud.com/12829
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mgc/mgc_request.c | 16 +++++++++++-----
1 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index 2191ca1..23374ca 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -1159,7 +1159,7 @@ static int mgc_apply_recover_logs(struct obd_device *mgc,
while (datalen > 0) {
int entry_len = sizeof(*entry);
- int is_ost;
+ int is_ost, i;
struct obd_device *obd;
char *obdname;
char *cname;
@@ -1265,11 +1265,17 @@ static int mgc_apply_recover_logs(struct obd_device *mgc,
continue;
}
- /* TODO: iterate all nids to find one */
+ /* iterate all nids to find one */
/* find uuid by nid */
- rc = client_import_find_conn(obd->u.cli.cl_import,
- entry->u.nids[0],
- (struct obd_uuid *)uuid);
+ rc = -ENOENT;
+ for (i = 0; i < entry->mne_nid_count; i++) {
+ rc = client_import_find_conn(obd->u.cli.cl_import,
+ entry->u.nids[0],
+ (struct obd_uuid *)uuid);
+ if (!rc)
+ break;
+ }
+
up_read(&obd->u.cli.cl_sem);
if (rc < 0) {
CERROR("mgc: cannot find uuid by nid %s\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 079/124] staging: lustre: mgc: add nid iteration
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Boyko, James Simmons
From: Alexander Boyko <alexander.boyko@seagate.com>
mgc_apply_recover_logs use only first nid from entry,
this could be the problem for a cluster with several network
address for a one node.
Signed-off-by: Alexander Boyko <alexander.boyko@seagate.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5950
Xyratex-bug-id: MRP-2255
Reviewed-on: http://review.whamcloud.com/12829
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Ann Koehler <amk@cray.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mgc/mgc_request.c | 16 +++++++++++-----
1 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mgc/mgc_request.c b/drivers/staging/lustre/lustre/mgc/mgc_request.c
index 2191ca1..23374ca 100644
--- a/drivers/staging/lustre/lustre/mgc/mgc_request.c
+++ b/drivers/staging/lustre/lustre/mgc/mgc_request.c
@@ -1159,7 +1159,7 @@ static int mgc_apply_recover_logs(struct obd_device *mgc,
while (datalen > 0) {
int entry_len = sizeof(*entry);
- int is_ost;
+ int is_ost, i;
struct obd_device *obd;
char *obdname;
char *cname;
@@ -1265,11 +1265,17 @@ static int mgc_apply_recover_logs(struct obd_device *mgc,
continue;
}
- /* TODO: iterate all nids to find one */
+ /* iterate all nids to find one */
/* find uuid by nid */
- rc = client_import_find_conn(obd->u.cli.cl_import,
- entry->u.nids[0],
- (struct obd_uuid *)uuid);
+ rc = -ENOENT;
+ for (i = 0; i < entry->mne_nid_count; i++) {
+ rc = client_import_find_conn(obd->u.cli.cl_import,
+ entry->u.nids[0],
+ (struct obd_uuid *)uuid);
+ if (!rc)
+ break;
+ }
+
up_read(&obd->u.cli.cl_sem);
if (rc < 0) {
CERROR("mgc: cannot find uuid by nid %s\n",
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 080/124] staging: lustre: llite: fix dup flags names
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bob Glossman,
James Simmons
From: Bob Glossman <bob.glossman@intel.com>
The name 'xattr' is used for two different ll_flags bits.
Change the names to be distinct and different, reflecting
the names of the bits used in LL_SBI_xbitnamex #defines.
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5586
Reviewed-on: http://review.whamcloud.com/12892
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/llite/llite_internal.h | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 0dc47f2..05f2aa3 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -407,7 +407,7 @@ enum stats_track_type {
"nolck", \
"checksum", \
"flock", \
- "xattr", \
+ "user_xattr", \
"acl", \
"???", \
"???", \
@@ -423,7 +423,7 @@ enum stats_track_type {
"verbose", \
"layout", \
"user_fid2path",\
- "xattr", \
+ "xattr_cache", \
"norootsquash", \
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 080/124] staging: lustre: llite: fix dup flags names
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bob Glossman,
James Simmons
From: Bob Glossman <bob.glossman@intel.com>
The name 'xattr' is used for two different ll_flags bits.
Change the names to be distinct and different, reflecting
the names of the bits used in LL_SBI_xbitnamex #defines.
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5586
Reviewed-on: http://review.whamcloud.com/12892
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/llite/llite_internal.h | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 0dc47f2..05f2aa3 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -407,7 +407,7 @@ enum stats_track_type {
"nolck", \
"checksum", \
"flock", \
- "xattr", \
+ "user_xattr", \
"acl", \
"???", \
"???", \
@@ -423,7 +423,7 @@ enum stats_track_type {
"verbose", \
"layout", \
"user_fid2path",\
- "xattr", \
+ "xattr_cache", \
"norootsquash", \
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 081/124] staging: lustre: obdclass: lu_htable_order() return type to long
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12385
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/lu_object.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 4395117..7339c8b 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -352,7 +352,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
* the dispose list, removing them from LRU and hash table.
*/
start = s->ls_purge_start;
- bnr = (nr == ~0) ? -1 : nr / CFS_HASH_NBKT(s->ls_obj_hash) + 1;
+ bnr = (nr == ~0) ? -1 : nr / (int)CFS_HASH_NBKT(s->ls_obj_hash) + 1;
again:
/*
* It doesn't make any sense to make purge threads parallel, that can
@@ -869,11 +869,11 @@ EXPORT_SYMBOL(lu_site_print);
/**
* Return desired hash table order.
*/
-static unsigned int lu_htable_order(struct lu_device *top)
+static unsigned long lu_htable_order(struct lu_device *top)
{
unsigned long bits_max = LU_SITE_BITS_MAX;
unsigned long cache_size;
- unsigned int bits;
+ unsigned long bits;
/*
* Calculate hash table size, assuming that we want reasonable
@@ -990,7 +990,7 @@ int lu_site_init(struct lu_site *s, struct lu_device *top)
memset(s, 0, sizeof(*s));
mutex_init(&s->ls_purge_mutex);
- snprintf(name, 16, "lu_site_%s", top->ld_type->ldt_name);
+ snprintf(name, sizeof(name), "lu_site_%s", top->ld_type->ldt_name);
for (bits = lu_htable_order(top); bits >= LU_SITE_BITS_MIN; bits--) {
s->ls_obj_hash = cfs_hash_create(name, bits, bits,
bits - LU_SITE_BKT_BITS,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 081/124] staging: lustre: obdclass: lu_htable_order() return type to long
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12385
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/lu_object.c | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 4395117..7339c8b 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -352,7 +352,7 @@ int lu_site_purge(const struct lu_env *env, struct lu_site *s, int nr)
* the dispose list, removing them from LRU and hash table.
*/
start = s->ls_purge_start;
- bnr = (nr == ~0) ? -1 : nr / CFS_HASH_NBKT(s->ls_obj_hash) + 1;
+ bnr = (nr == ~0) ? -1 : nr / (int)CFS_HASH_NBKT(s->ls_obj_hash) + 1;
again:
/*
* It doesn't make any sense to make purge threads parallel, that can
@@ -869,11 +869,11 @@ EXPORT_SYMBOL(lu_site_print);
/**
* Return desired hash table order.
*/
-static unsigned int lu_htable_order(struct lu_device *top)
+static unsigned long lu_htable_order(struct lu_device *top)
{
unsigned long bits_max = LU_SITE_BITS_MAX;
unsigned long cache_size;
- unsigned int bits;
+ unsigned long bits;
/*
* Calculate hash table size, assuming that we want reasonable
@@ -990,7 +990,7 @@ int lu_site_init(struct lu_site *s, struct lu_device *top)
memset(s, 0, sizeof(*s));
mutex_init(&s->ls_purge_mutex);
- snprintf(name, 16, "lu_site_%s", top->ld_type->ldt_name);
+ snprintf(name, sizeof(name), "lu_site_%s", top->ld_type->ldt_name);
for (bits = lu_htable_order(top); bits >= LU_SITE_BITS_MIN; bits--) {
s->ls_obj_hash = cfs_hash_create(name, bits, bits,
bits - LU_SITE_BKT_BITS,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 082/124] staging: lustre: mdc: Proper accessing struct lov_user_md
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Yoshifumi Uemura, James Simmons
From: Yoshifumi Uemura <kogexe@gmail.com>
In mdc_setattr_pack() access the members of struct lov_user_md by
little endian byte order.
Signed-off-by: Yoshifumi Uemura <kogexe@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5889
Reviewed-on: http://review.whamcloud.com/12683
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mdc/mdc_lib.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_lib.c b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
index 82d1c4d..aac7e04 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_lib.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
@@ -328,7 +328,7 @@ void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
lum = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
if (!ea) { /* Remove LOV EA */
- lum->lmm_magic = LOV_USER_MAGIC_V1;
+ lum->lmm_magic = cpu_to_le32(LOV_USER_MAGIC_V1);
lum->lmm_stripe_size = 0;
lum->lmm_stripe_count = 0;
lum->lmm_stripe_offset = (typeof(lum->lmm_stripe_offset))(-1);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 082/124] staging: lustre: mdc: Proper accessing struct lov_user_md
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Yoshifumi Uemura, James Simmons
From: Yoshifumi Uemura <kogexe@gmail.com>
In mdc_setattr_pack() access the members of struct lov_user_md by
little endian byte order.
Signed-off-by: Yoshifumi Uemura <kogexe@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5889
Reviewed-on: http://review.whamcloud.com/12683
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Yang Sheng <yang.sheng@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/mdc/mdc_lib.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_lib.c b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
index 82d1c4d..aac7e04 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_lib.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_lib.c
@@ -328,7 +328,7 @@ void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
lum = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
if (!ea) { /* Remove LOV EA */
- lum->lmm_magic = LOV_USER_MAGIC_V1;
+ lum->lmm_magic = cpu_to_le32(LOV_USER_MAGIC_V1);
lum->lmm_stripe_size = 0;
lum->lmm_stripe_count = 0;
lum->lmm_stripe_offset = (typeof(lum->lmm_stripe_offset))(-1);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 083/124] staging: lustre: ldlm: evict clients returning errors on ASTs
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexey Lyashkov, James Simmons
From: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
To test proper behavior of clients returning errors on ASTs
we can induce a failure with setting OBD_FAIL_LDLM_BL_CALLBACK_NET.
Handle the new additonal case of cfs_fail_err being set as well
so that the cfs_fail_err can be sent back in a reply.
Signed-off-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5581
Xyratex-bug-id: MRP-2041
Reviewed-on: http://review.whamcloud.com/11752
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c
index b91b26d..ff2c28e 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c
@@ -559,8 +559,11 @@ static int ldlm_callback_handler(struct ptlrpc_request *req)
switch (lustre_msg_get_opc(req->rq_reqmsg)) {
case LDLM_BL_CALLBACK:
- if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET)) {
+ if (cfs_fail_err)
+ ldlm_callback_reply(req, -(int)cfs_fail_err);
return 0;
+ }
break;
case LDLM_CP_CALLBACK:
if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_CP_CALLBACK_NET))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 083/124] staging: lustre: ldlm: evict clients returning errors on ASTs
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexey Lyashkov, James Simmons
From: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
To test proper behavior of clients returning errors on ASTs
we can induce a failure with setting OBD_FAIL_LDLM_BL_CALLBACK_NET.
Handle the new additonal case of cfs_fail_err being set as well
so that the cfs_fail_err can be sent back in a reply.
Signed-off-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5581
Xyratex-bug-id: MRP-2041
Reviewed-on: http://review.whamcloud.com/11752
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c
index b91b26d..ff2c28e 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lockd.c
@@ -559,8 +559,11 @@ static int ldlm_callback_handler(struct ptlrpc_request *req)
switch (lustre_msg_get_opc(req->rq_reqmsg)) {
case LDLM_BL_CALLBACK:
- if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET))
+ if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_BL_CALLBACK_NET)) {
+ if (cfs_fail_err)
+ ldlm_callback_reply(req, -(int)cfs_fail_err);
return 0;
+ }
break;
case LDLM_CP_CALLBACK:
if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_CP_CALLBACK_NET))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 084/124] staging: lustre: fiemap: set FIEMAP_EXTENT_LAST correctly
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
When we've collected enough extents as user requested, we'd check one
further to decide whether we've reached the last extent of the file.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5933
Reviewed-on: http://review.whamcloud.com/12781
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lov/lov_obd.c | 12 ++++++++++--
1 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 9b12159..9ab3033 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -1548,6 +1548,8 @@ static int lov_fiemap(struct lov_obd *lov, __u32 keylen, void *key,
u64 fm_start, fm_end, fm_length, fm_end_offset;
u64 curr_loc;
int current_extent = 0, rc = 0, i;
+ /* Whether have we collected enough extents */
+ bool enough = false;
int ost_eof = 0; /* EOF for object */
int ost_done = 0; /* done with required mapping for this OST? */
int last_stripe;
@@ -1682,7 +1684,7 @@ static int lov_fiemap(struct lov_obd *lov, __u32 keylen, void *key,
lun_start += len_mapped_single_call;
fm_local->fm_length = req_fm_len - len_mapped_single_call;
req_fm_len = fm_local->fm_length;
- fm_local->fm_extent_count = count_local;
+ fm_local->fm_extent_count = enough ? 1 : count_local;
fm_local->fm_mapped_extents = 0;
fm_local->fm_flags = fiemap->fm_flags;
@@ -1730,6 +1732,12 @@ inactive_tgt:
goto finish;
}
break;
+ } else if (enough) {
+ /*
+ * We've collected enough extents and there are
+ * more extents after it.
+ */
+ goto finish;
}
/* If we just need num of extents then go to next device */
@@ -1769,7 +1777,7 @@ inactive_tgt:
/* Ran out of available extents? */
if (current_extent >= fiemap->fm_extent_count)
- goto finish;
+ enough = true;
} while (ost_done == 0 && ost_eof == 0);
if (cur_stripe_wrap == last_stripe)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 084/124] staging: lustre: fiemap: set FIEMAP_EXTENT_LAST correctly
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
When we've collected enough extents as user requested, we'd check one
further to decide whether we've reached the last extent of the file.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5933
Reviewed-on: http://review.whamcloud.com/12781
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lov/lov_obd.c | 12 ++++++++++--
1 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 9b12159..9ab3033 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -1548,6 +1548,8 @@ static int lov_fiemap(struct lov_obd *lov, __u32 keylen, void *key,
u64 fm_start, fm_end, fm_length, fm_end_offset;
u64 curr_loc;
int current_extent = 0, rc = 0, i;
+ /* Whether have we collected enough extents */
+ bool enough = false;
int ost_eof = 0; /* EOF for object */
int ost_done = 0; /* done with required mapping for this OST? */
int last_stripe;
@@ -1682,7 +1684,7 @@ static int lov_fiemap(struct lov_obd *lov, __u32 keylen, void *key,
lun_start += len_mapped_single_call;
fm_local->fm_length = req_fm_len - len_mapped_single_call;
req_fm_len = fm_local->fm_length;
- fm_local->fm_extent_count = count_local;
+ fm_local->fm_extent_count = enough ? 1 : count_local;
fm_local->fm_mapped_extents = 0;
fm_local->fm_flags = fiemap->fm_flags;
@@ -1730,6 +1732,12 @@ inactive_tgt:
goto finish;
}
break;
+ } else if (enough) {
+ /*
+ * We've collected enough extents and there are
+ * more extents after it.
+ */
+ goto finish;
}
/* If we just need num of extents then go to next device */
@@ -1769,7 +1777,7 @@ inactive_tgt:
/* Ran out of available extents? */
if (current_extent >= fiemap->fm_extent_count)
- goto finish;
+ enough = true;
} while (ost_done == 0 && ost_eof == 0);
if (cur_stripe_wrap == last_stripe)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 085/124] staging: lustre: obdclass: change loop indexes to unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Cleanup warnings about comparison between signed and unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12387
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/cl_io.c | 6 +++---
drivers/staging/lustre/lustre/obdclass/cl_object.c | 10 +++++-----
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_io.c b/drivers/staging/lustre/lustre/obdclass/cl_io.c
index 6643e78..7894cf9 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_io.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_io.c
@@ -1211,7 +1211,7 @@ void cl_req_page_add(const struct lu_env *env,
{
struct cl_object *obj;
struct cl_req_obj *rqo;
- int i;
+ unsigned int i;
LASSERT(list_empty(&page->cp_flight));
LASSERT(!page->cp_req);
@@ -1258,7 +1258,7 @@ EXPORT_SYMBOL(cl_req_page_done);
*/
int cl_req_prep(const struct lu_env *env, struct cl_req *req)
{
- int i;
+ unsigned int i;
int result;
const struct cl_req_slice *slice;
@@ -1291,7 +1291,7 @@ void cl_req_attr_set(const struct lu_env *env, struct cl_req *req,
{
const struct cl_req_slice *slice;
struct cl_page *page;
- int i;
+ unsigned int i;
LASSERT(!list_empty(&req->crq_pages));
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_object.c b/drivers/staging/lustre/lustre/obdclass/cl_object.c
index 91a5806..bd85455 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_object.c
@@ -377,7 +377,7 @@ static void cl_env_percpu_refill(void);
*/
int cl_site_init(struct cl_site *s, struct cl_device *d)
{
- int i;
+ size_t i;
int result;
result = lu_site_init(&s->cs_lu, &d->cd_lu_dev);
@@ -411,7 +411,7 @@ static struct cache_stats cl_env_stats = {
*/
int cl_site_stats_print(const struct cl_site *site, struct seq_file *m)
{
- int i;
+ size_t i;
static const char *pstate[] = {
[CPS_CACHED] = "c",
[CPS_OWNED] = "o",
@@ -1126,7 +1126,7 @@ static void *cl_key_init(const struct lu_context *ctx,
info = cl0_key_init(ctx, key);
if (!IS_ERR(info)) {
- int i;
+ size_t i;
for (i = 0; i < ARRAY_SIZE(info->clt_counters); ++i)
lu_ref_init(&info->clt_counters[i].ctc_locks_locked);
@@ -1138,7 +1138,7 @@ static void cl_key_fini(const struct lu_context *ctx,
struct lu_context_key *key, void *data)
{
struct cl_thread_info *info;
- int i;
+ size_t i;
info = data;
for (i = 0; i < ARRAY_SIZE(info->clt_counters); ++i)
@@ -1150,7 +1150,7 @@ static void cl_key_exit(const struct lu_context *ctx,
struct lu_context_key *key, void *data)
{
struct cl_thread_info *info = data;
- int i;
+ size_t i;
for (i = 0; i < ARRAY_SIZE(info->clt_counters); ++i) {
LASSERT(info->clt_counters[i].ctc_nr_held == 0);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 085/124] staging: lustre: obdclass: change loop indexes to unsigned
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Cleanup warnings about comparison between signed and unsigned.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12387
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/cl_io.c | 6 +++---
drivers/staging/lustre/lustre/obdclass/cl_object.c | 10 +++++-----
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_io.c b/drivers/staging/lustre/lustre/obdclass/cl_io.c
index 6643e78..7894cf9 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_io.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_io.c
@@ -1211,7 +1211,7 @@ void cl_req_page_add(const struct lu_env *env,
{
struct cl_object *obj;
struct cl_req_obj *rqo;
- int i;
+ unsigned int i;
LASSERT(list_empty(&page->cp_flight));
LASSERT(!page->cp_req);
@@ -1258,7 +1258,7 @@ EXPORT_SYMBOL(cl_req_page_done);
*/
int cl_req_prep(const struct lu_env *env, struct cl_req *req)
{
- int i;
+ unsigned int i;
int result;
const struct cl_req_slice *slice;
@@ -1291,7 +1291,7 @@ void cl_req_attr_set(const struct lu_env *env, struct cl_req *req,
{
const struct cl_req_slice *slice;
struct cl_page *page;
- int i;
+ unsigned int i;
LASSERT(!list_empty(&req->crq_pages));
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_object.c b/drivers/staging/lustre/lustre/obdclass/cl_object.c
index 91a5806..bd85455 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_object.c
@@ -377,7 +377,7 @@ static void cl_env_percpu_refill(void);
*/
int cl_site_init(struct cl_site *s, struct cl_device *d)
{
- int i;
+ size_t i;
int result;
result = lu_site_init(&s->cs_lu, &d->cd_lu_dev);
@@ -411,7 +411,7 @@ static struct cache_stats cl_env_stats = {
*/
int cl_site_stats_print(const struct cl_site *site, struct seq_file *m)
{
- int i;
+ size_t i;
static const char *pstate[] = {
[CPS_CACHED] = "c",
[CPS_OWNED] = "o",
@@ -1126,7 +1126,7 @@ static void *cl_key_init(const struct lu_context *ctx,
info = cl0_key_init(ctx, key);
if (!IS_ERR(info)) {
- int i;
+ size_t i;
for (i = 0; i < ARRAY_SIZE(info->clt_counters); ++i)
lu_ref_init(&info->clt_counters[i].ctc_locks_locked);
@@ -1138,7 +1138,7 @@ static void cl_key_fini(const struct lu_context *ctx,
struct lu_context_key *key, void *data)
{
struct cl_thread_info *info;
- int i;
+ size_t i;
info = data;
for (i = 0; i < ARRAY_SIZE(info->clt_counters); ++i)
@@ -1150,7 +1150,7 @@ static void cl_key_exit(const struct lu_context *ctx,
struct lu_context_key *key, void *data)
{
struct cl_thread_info *info = data;
- int i;
+ size_t i;
for (i = 0; i < ARRAY_SIZE(info->clt_counters); ++i) {
LASSERT(info->clt_counters[i].ctc_nr_held == 0);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 086/124] staging: lustre: obdclass: eliminate NULL error return
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bob Glossman,
James Simmons
From: Bob Glossman <bob.glossman@intel.com>
Always return an ERR_PTR() on errors, never return a NULL,
in lu_object_find_slice(). Also clean up callers who
no longer need special case handling of NULL returns.
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5858
Reviewed-on: http://review.whamcloud.com/12554
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/lu_object.c | 14 ++++++++------
1 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 7339c8b..3f1fb28 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -767,13 +767,15 @@ struct lu_object *lu_object_find_slice(const struct lu_env *env,
struct lu_object *obj;
top = lu_object_find(env, dev, f, conf);
- if (!IS_ERR(top)) {
- obj = lu_object_locate(top->lo_header, dev->ld_type);
- if (!obj)
- lu_object_put(env, top);
- } else {
- obj = top;
+ if (IS_ERR(top))
+ return top;
+
+ obj = lu_object_locate(top->lo_header, dev->ld_type);
+ if (unlikely(!obj)) {
+ lu_object_put(env, top);
+ obj = ERR_PTR(-ENOENT);
}
+
return obj;
}
EXPORT_SYMBOL(lu_object_find_slice);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 086/124] staging: lustre: obdclass: eliminate NULL error return
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bob Glossman,
James Simmons
From: Bob Glossman <bob.glossman@intel.com>
Always return an ERR_PTR() on errors, never return a NULL,
in lu_object_find_slice(). Also clean up callers who
no longer need special case handling of NULL returns.
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5858
Reviewed-on: http://review.whamcloud.com/12554
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/lu_object.c | 14 ++++++++------
1 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 7339c8b..3f1fb28 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -767,13 +767,15 @@ struct lu_object *lu_object_find_slice(const struct lu_env *env,
struct lu_object *obj;
top = lu_object_find(env, dev, f, conf);
- if (!IS_ERR(top)) {
- obj = lu_object_locate(top->lo_header, dev->ld_type);
- if (!obj)
- lu_object_put(env, top);
- } else {
- obj = top;
+ if (IS_ERR(top))
+ return top;
+
+ obj = lu_object_locate(top->lo_header, dev->ld_type);
+ if (unlikely(!obj)) {
+ lu_object_put(env, top);
+ obj = ERR_PTR(-ENOENT);
}
+
return obj;
}
EXPORT_SYMBOL(lu_object_find_slice);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 087/124] staging: lustre: ptlrpc: Suppress error message when imp_sec is freed
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Amir Shehata,
James Simmons
From: Amir Shehata <amir.shehata@intel.com>
There is a race condition on client reconnect when the import
is being destroyed. Some outstanding client bound requests
are being processed when the imp_sec has alread been freed.
Ensure to suppress the error message in import_sec_validate_get()
in that case
Signed-off-by: Amir Shehata <amir.shehata@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3353
Reviewed-on: http://review.whamcloud.com/10200
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/sec.c | 19 +++++++++++++++++--
1 files changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec.c b/drivers/staging/lustre/lustre/ptlrpc/sec.c
index 523579f..5d3995d 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec.c
@@ -311,6 +311,19 @@ static int import_sec_check_expire(struct obd_import *imp)
return sptlrpc_import_sec_adapt(imp, NULL, NULL);
}
+/**
+ * Get and validate the client side ptlrpc security facilities from
+ * \a imp. There is a race condition on client reconnect when the import is
+ * being destroyed while there are outstanding client bound requests. In
+ * this case do not output any error messages if import secuity is not
+ * found.
+ *
+ * \param[in] imp obd import associated with client
+ * \param[out] sec client side ptlrpc security
+ *
+ * \retval 0 if security retrieved successfully
+ * \retval -ve errno if there was a problem
+ */
static int import_sec_validate_get(struct obd_import *imp,
struct ptlrpc_sec **sec)
{
@@ -323,9 +336,11 @@ static int import_sec_validate_get(struct obd_import *imp,
}
*sec = sptlrpc_import_sec_ref(imp);
+ /* Only output an error when the import is still active */
if (!*sec) {
- CERROR("import %p (%s) with no sec\n",
- imp, ptlrpc_import_state_name(imp->imp_state));
+ if (list_empty(&imp->imp_zombie_chain))
+ CERROR("import %p (%s) with no sec\n",
+ imp, ptlrpc_import_state_name(imp->imp_state));
return -EACCES;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 087/124] staging: lustre: ptlrpc: Suppress error message when imp_sec is freed
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Amir Shehata,
James Simmons
From: Amir Shehata <amir.shehata@intel.com>
There is a race condition on client reconnect when the import
is being destroyed. Some outstanding client bound requests
are being processed when the imp_sec has alread been freed.
Ensure to suppress the error message in import_sec_validate_get()
in that case
Signed-off-by: Amir Shehata <amir.shehata@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3353
Reviewed-on: http://review.whamcloud.com/10200
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/sec.c | 19 +++++++++++++++++--
1 files changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec.c b/drivers/staging/lustre/lustre/ptlrpc/sec.c
index 523579f..5d3995d 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec.c
@@ -311,6 +311,19 @@ static int import_sec_check_expire(struct obd_import *imp)
return sptlrpc_import_sec_adapt(imp, NULL, NULL);
}
+/**
+ * Get and validate the client side ptlrpc security facilities from
+ * \a imp. There is a race condition on client reconnect when the import is
+ * being destroyed while there are outstanding client bound requests. In
+ * this case do not output any error messages if import secuity is not
+ * found.
+ *
+ * \param[in] imp obd import associated with client
+ * \param[out] sec client side ptlrpc security
+ *
+ * \retval 0 if security retrieved successfully
+ * \retval -ve errno if there was a problem
+ */
static int import_sec_validate_get(struct obd_import *imp,
struct ptlrpc_sec **sec)
{
@@ -323,9 +336,11 @@ static int import_sec_validate_get(struct obd_import *imp,
}
*sec = sptlrpc_import_sec_ref(imp);
+ /* Only output an error when the import is still active */
if (!*sec) {
- CERROR("import %p (%s) with no sec\n",
- imp, ptlrpc_import_state_name(imp->imp_state));
+ if (list_empty(&imp->imp_zombie_chain))
+ CERROR("import %p (%s) with no sec\n",
+ imp, ptlrpc_import_state_name(imp->imp_state));
return -EACCES;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 088/124] staging: lustre: ldlm: Recalculate interval in ldlm_pool_recalc()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Nathaniel Clark, James Simmons
From: Nathaniel Clark <nathaniel.l.clark@intel.com>
Instead of rechecking a static value, recalculate to see if pool stats
need to be updated.
Add newline so message will print instead of warning about missing
newline.
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4536
Reviewed-on: http://review.whamcloud.com/12547
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 38 ++++++++++++++----------
1 files changed, 22 insertions(+), 16 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
index 44c3abf..9a1136e 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
@@ -357,33 +357,39 @@ static int ldlm_pool_recalc(struct ldlm_pool *pl)
int count;
recalc_interval_sec = ktime_get_seconds() - pl->pl_recalc_time;
- if (recalc_interval_sec <= 0)
- goto recalc;
-
- spin_lock(&pl->pl_lock);
if (recalc_interval_sec > 0) {
- /*
- * Update pool statistics every 1s.
- */
- ldlm_pool_recalc_stats(pl);
-
- /*
- * Zero out all rates and speed for the last period.
- */
- atomic_set(&pl->pl_grant_rate, 0);
- atomic_set(&pl->pl_cancel_rate, 0);
+ spin_lock(&pl->pl_lock);
+ recalc_interval_sec = ktime_get_seconds() - pl->pl_recalc_time;
+
+ if (recalc_interval_sec > 0) {
+ /*
+ * Update pool statistics every 1s.
+ */
+ ldlm_pool_recalc_stats(pl);
+
+ /*
+ * Zero out all rates and speed for the last period.
+ */
+ atomic_set(&pl->pl_grant_rate, 0);
+ atomic_set(&pl->pl_cancel_rate, 0);
+ }
+ spin_unlock(&pl->pl_lock);
}
- spin_unlock(&pl->pl_lock);
- recalc:
if (pl->pl_ops->po_recalc) {
count = pl->pl_ops->po_recalc(pl);
lprocfs_counter_add(pl->pl_stats, LDLM_POOL_RECALC_STAT,
count);
}
+
recalc_interval_sec = pl->pl_recalc_time - ktime_get_seconds() +
pl->pl_recalc_period;
if (recalc_interval_sec <= 0) {
+ /* DEBUG: should be re-removed after LU-4536 is fixed */
+ CDEBUG(D_DLMTRACE, "%s: Negative interval(%ld), too short period(%ld)\n",
+ pl->pl_name, (long)recalc_interval_sec,
+ (long)pl->pl_recalc_period);
+
/* Prevent too frequent recalculation. */
recalc_interval_sec = 1;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 088/124] staging: lustre: ldlm: Recalculate interval in ldlm_pool_recalc()
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Nathaniel Clark, James Simmons
From: Nathaniel Clark <nathaniel.l.clark@intel.com>
Instead of rechecking a static value, recalculate to see if pool stats
need to be updated.
Add newline so message will print instead of warning about missing
newline.
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4536
Reviewed-on: http://review.whamcloud.com/12547
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ldlm/ldlm_pool.c | 38 ++++++++++++++----------
1 files changed, 22 insertions(+), 16 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
index 44c3abf..9a1136e 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_pool.c
@@ -357,33 +357,39 @@ static int ldlm_pool_recalc(struct ldlm_pool *pl)
int count;
recalc_interval_sec = ktime_get_seconds() - pl->pl_recalc_time;
- if (recalc_interval_sec <= 0)
- goto recalc;
-
- spin_lock(&pl->pl_lock);
if (recalc_interval_sec > 0) {
- /*
- * Update pool statistics every 1s.
- */
- ldlm_pool_recalc_stats(pl);
-
- /*
- * Zero out all rates and speed for the last period.
- */
- atomic_set(&pl->pl_grant_rate, 0);
- atomic_set(&pl->pl_cancel_rate, 0);
+ spin_lock(&pl->pl_lock);
+ recalc_interval_sec = ktime_get_seconds() - pl->pl_recalc_time;
+
+ if (recalc_interval_sec > 0) {
+ /*
+ * Update pool statistics every 1s.
+ */
+ ldlm_pool_recalc_stats(pl);
+
+ /*
+ * Zero out all rates and speed for the last period.
+ */
+ atomic_set(&pl->pl_grant_rate, 0);
+ atomic_set(&pl->pl_cancel_rate, 0);
+ }
+ spin_unlock(&pl->pl_lock);
}
- spin_unlock(&pl->pl_lock);
- recalc:
if (pl->pl_ops->po_recalc) {
count = pl->pl_ops->po_recalc(pl);
lprocfs_counter_add(pl->pl_stats, LDLM_POOL_RECALC_STAT,
count);
}
+
recalc_interval_sec = pl->pl_recalc_time - ktime_get_seconds() +
pl->pl_recalc_period;
if (recalc_interval_sec <= 0) {
+ /* DEBUG: should be re-removed after LU-4536 is fixed */
+ CDEBUG(D_DLMTRACE, "%s: Negative interval(%ld), too short period(%ld)\n",
+ pl->pl_name, (long)recalc_interval_sec,
+ (long)pl->pl_recalc_period);
+
/* Prevent too frequent recalculation. */
recalc_interval_sec = 1;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 089/124] staging: lustre: obd: change brw_page->count to unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Pages count is unsigned. So, change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12378
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/osc/osc_cache.c | 10 ++++++----
.../staging/lustre/lustre/osc/osc_cl_internal.h | 2 +-
3 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index cd7f552..d1c4ccc 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -176,7 +176,7 @@ struct obd_type {
struct brw_page {
u64 off;
struct page *pg;
- int count;
+ unsigned int count;
u32 flag;
};
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index fb18dd1..31deab0 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1099,7 +1099,7 @@ static int osc_extent_make_ready(const struct lu_env *env,
struct osc_async_page *oap;
struct osc_async_page *last = NULL;
struct osc_object *obj = ext->oe_obj;
- int page_count = 0;
+ unsigned int page_count = 0;
int rc;
/* we're going to grab page lock, so object lock must not be taken. */
@@ -1140,9 +1140,11 @@ static int osc_extent_make_ready(const struct lu_env *env,
* the size of file.
*/
if (!(last->oap_async_flags & ASYNC_COUNT_STABLE)) {
- last->oap_count = osc_refresh_count(env, last, OBD_BRW_WRITE);
- LASSERT(last->oap_count > 0);
- LASSERT(last->oap_page_off + last->oap_count <= PAGE_SIZE);
+ int last_oap_count = osc_refresh_count(env, last, OBD_BRW_WRITE);
+
+ LASSERT(last_oap_count > 0);
+ LASSERT(last->oap_page_off + last_oap_count <= PAGE_SIZE);
+ last->oap_count = last_oap_count;
spin_lock(&last->oap_lock);
last->oap_async_flags |= ASYNC_COUNT_STABLE;
spin_unlock(&last->oap_lock);
diff --git a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
index d41680b..3068113 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
+++ b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
@@ -64,7 +64,7 @@ struct osc_io {
/** true if this io is lockless. */
unsigned int oi_lockless;
/** how many LRU pages are reserved for this IO */
- int oi_lru_reserved;
+ unsigned long oi_lru_reserved;
/** active extents, we know how many bytes is going to be written,
* so having an active extent will prevent it from being fragmented
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 089/124] staging: lustre: obd: change brw_page->count to unsigned
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Pages count is unsigned. So, change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12378
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/osc/osc_cache.c | 10 ++++++----
.../staging/lustre/lustre/osc/osc_cl_internal.h | 2 +-
3 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index cd7f552..d1c4ccc 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -176,7 +176,7 @@ struct obd_type {
struct brw_page {
u64 off;
struct page *pg;
- int count;
+ unsigned int count;
u32 flag;
};
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index fb18dd1..31deab0 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1099,7 +1099,7 @@ static int osc_extent_make_ready(const struct lu_env *env,
struct osc_async_page *oap;
struct osc_async_page *last = NULL;
struct osc_object *obj = ext->oe_obj;
- int page_count = 0;
+ unsigned int page_count = 0;
int rc;
/* we're going to grab page lock, so object lock must not be taken. */
@@ -1140,9 +1140,11 @@ static int osc_extent_make_ready(const struct lu_env *env,
* the size of file.
*/
if (!(last->oap_async_flags & ASYNC_COUNT_STABLE)) {
- last->oap_count = osc_refresh_count(env, last, OBD_BRW_WRITE);
- LASSERT(last->oap_count > 0);
- LASSERT(last->oap_page_off + last->oap_count <= PAGE_SIZE);
+ int last_oap_count = osc_refresh_count(env, last, OBD_BRW_WRITE);
+
+ LASSERT(last_oap_count > 0);
+ LASSERT(last->oap_page_off + last_oap_count <= PAGE_SIZE);
+ last->oap_count = last_oap_count;
spin_lock(&last->oap_lock);
last->oap_async_flags |= ASYNC_COUNT_STABLE;
spin_unlock(&last->oap_lock);
diff --git a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
index d41680b..3068113 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
+++ b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
@@ -64,7 +64,7 @@ struct osc_io {
/** true if this io is lockless. */
unsigned int oi_lockless;
/** how many LRU pages are reserved for this IO */
- int oi_lru_reserved;
+ unsigned long oi_lru_reserved;
/** active extents, we know how many bytes is going to be written,
* so having an active extent will prevent it from being fragmented
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 090/124] staging: lustre: obdclass: change cl_fault_io->ft_nob to size_t
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12380
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 4 ++--
drivers/staging/lustre/lustre/llite/rw26.c | 4 ++--
drivers/staging/lustre/lustre/obdclass/cl_page.c | 4 ++--
.../staging/lustre/lustre/obdecho/echo_client.c | 2 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 4 ++--
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 5af1153..f360ec9 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -1766,7 +1766,7 @@ struct cl_io {
/** page index within file. */
pgoff_t ft_index;
/** bytes valid byte on a faulted page. */
- int ft_nob;
+ size_t ft_nob;
/** writable page? for nopage() only */
int ft_writable;
/** page of an executable? */
@@ -2297,7 +2297,7 @@ int cl_page_is_under_lock(const struct lu_env *env, struct cl_io *io,
struct cl_page *page, pgoff_t *max_index);
loff_t cl_offset(const struct cl_object *obj, pgoff_t idx);
pgoff_t cl_index(const struct cl_object *obj, loff_t offset);
-int cl_page_size(const struct cl_object *obj);
+size_t cl_page_size(const struct cl_object *obj);
int cl_pages_prune(const struct lu_env *env, struct cl_object *obj);
void cl_lock_print(const struct lu_env *env, void *cookie,
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index f35eb2e..df0318e 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -213,10 +213,10 @@ ssize_t ll_direct_rw_pages(const struct lu_env *env, struct cl_io *io,
int i;
ssize_t rc = 0;
loff_t file_offset = pv->ldp_start_offset;
- long size = pv->ldp_size;
+ size_t size = pv->ldp_size;
int page_count = pv->ldp_nr;
struct page **pages = pv->ldp_pages;
- long page_size = cl_page_size(obj);
+ size_t page_size = cl_page_size(obj);
bool do_io;
int io_pages = 0;
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index 5c89dbd..18784a8 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -1032,9 +1032,9 @@ pgoff_t cl_index(const struct cl_object *obj, loff_t offset)
}
EXPORT_SYMBOL(cl_index);
-int cl_page_size(const struct cl_object *obj)
+size_t cl_page_size(const struct cl_object *obj)
{
- return 1 << PAGE_SHIFT;
+ return 1UL << PAGE_SHIFT;
}
EXPORT_SYMBOL(cl_page_size);
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index 91d28b5..9df117b 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1010,7 +1010,7 @@ static int cl_echo_object_brw(struct echo_object *eco, int rw, u64 offset,
struct cl_io *io;
struct cl_page *clp;
struct lustre_handle lh = { 0 };
- int page_size = cl_page_size(obj);
+ size_t page_size = cl_page_size(obj);
int refcheck;
int rc;
int i;
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 7698054..96beba4 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -200,7 +200,7 @@ static int osc_io_submit(const struct lu_env *env,
* Expand stripe KMS if necessary.
*/
static void osc_page_touch_at(const struct lu_env *env,
- struct cl_object *obj, pgoff_t idx, unsigned to)
+ struct cl_object *obj, pgoff_t idx, size_t to)
{
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
@@ -369,7 +369,7 @@ static int osc_io_fault_start(const struct lu_env *env,
io = ios->cis_io;
fio = &io->u.ci_fault;
- CDEBUG(D_INFO, "%lu %d %d\n",
+ CDEBUG(D_INFO, "%lu %d %zu\n",
fio->ft_index, fio->ft_writable, fio->ft_nob);
/*
* If mapping is writeable, adjust kms to cover this page,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 090/124] staging: lustre: obdclass: change cl_fault_io->ft_nob to size_t
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12380
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 4 ++--
drivers/staging/lustre/lustre/llite/rw26.c | 4 ++--
drivers/staging/lustre/lustre/obdclass/cl_page.c | 4 ++--
.../staging/lustre/lustre/obdecho/echo_client.c | 2 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 4 ++--
5 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 5af1153..f360ec9 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -1766,7 +1766,7 @@ struct cl_io {
/** page index within file. */
pgoff_t ft_index;
/** bytes valid byte on a faulted page. */
- int ft_nob;
+ size_t ft_nob;
/** writable page? for nopage() only */
int ft_writable;
/** page of an executable? */
@@ -2297,7 +2297,7 @@ int cl_page_is_under_lock(const struct lu_env *env, struct cl_io *io,
struct cl_page *page, pgoff_t *max_index);
loff_t cl_offset(const struct cl_object *obj, pgoff_t idx);
pgoff_t cl_index(const struct cl_object *obj, loff_t offset);
-int cl_page_size(const struct cl_object *obj);
+size_t cl_page_size(const struct cl_object *obj);
int cl_pages_prune(const struct lu_env *env, struct cl_object *obj);
void cl_lock_print(const struct lu_env *env, void *cookie,
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index f35eb2e..df0318e 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -213,10 +213,10 @@ ssize_t ll_direct_rw_pages(const struct lu_env *env, struct cl_io *io,
int i;
ssize_t rc = 0;
loff_t file_offset = pv->ldp_start_offset;
- long size = pv->ldp_size;
+ size_t size = pv->ldp_size;
int page_count = pv->ldp_nr;
struct page **pages = pv->ldp_pages;
- long page_size = cl_page_size(obj);
+ size_t page_size = cl_page_size(obj);
bool do_io;
int io_pages = 0;
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index 5c89dbd..18784a8 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -1032,9 +1032,9 @@ pgoff_t cl_index(const struct cl_object *obj, loff_t offset)
}
EXPORT_SYMBOL(cl_index);
-int cl_page_size(const struct cl_object *obj)
+size_t cl_page_size(const struct cl_object *obj)
{
- return 1 << PAGE_SHIFT;
+ return 1UL << PAGE_SHIFT;
}
EXPORT_SYMBOL(cl_page_size);
diff --git a/drivers/staging/lustre/lustre/obdecho/echo_client.c b/drivers/staging/lustre/lustre/obdecho/echo_client.c
index 91d28b5..9df117b 100644
--- a/drivers/staging/lustre/lustre/obdecho/echo_client.c
+++ b/drivers/staging/lustre/lustre/obdecho/echo_client.c
@@ -1010,7 +1010,7 @@ static int cl_echo_object_brw(struct echo_object *eco, int rw, u64 offset,
struct cl_io *io;
struct cl_page *clp;
struct lustre_handle lh = { 0 };
- int page_size = cl_page_size(obj);
+ size_t page_size = cl_page_size(obj);
int refcheck;
int rc;
int i;
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 7698054..96beba4 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -200,7 +200,7 @@ static int osc_io_submit(const struct lu_env *env,
* Expand stripe KMS if necessary.
*/
static void osc_page_touch_at(const struct lu_env *env,
- struct cl_object *obj, pgoff_t idx, unsigned to)
+ struct cl_object *obj, pgoff_t idx, size_t to)
{
struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
struct cl_attr *attr = &osc_env_info(env)->oti_attr;
@@ -369,7 +369,7 @@ static int osc_io_fault_start(const struct lu_env *env,
io = ios->cis_io;
fio = &io->u.ci_fault;
- CDEBUG(D_INFO, "%lu %d %d\n",
+ CDEBUG(D_INFO, "%lu %d %zu\n",
fio->ft_index, fio->ft_writable, fio->ft_nob);
/*
* If mapping is writeable, adjust kms to cover this page,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 091/124] staging: lustre: clio: add coo_getstripe interface
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
Use cl_object_operations::coo_getstripe() to handle
LL_IOC_LOV_GETSTRIPE ops.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5823
Reviewed-on: http://review.whamcloud.com/12452
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 7 +++
drivers/staging/lustre/lustre/llite/file.c | 38 +++++++++---------
.../staging/lustre/lustre/lov/lov_cl_internal.h | 6 +++
drivers/staging/lustre/lustre/lov/lov_internal.h | 2 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 3 -
drivers/staging/lustre/lustre/lov/lov_object.c | 35 +++++++++++++----
drivers/staging/lustre/lustre/lov/lov_pack.c | 40 +++++++++++++-------
drivers/staging/lustre/lustre/obdclass/cl_object.c | 21 ++++++++++
8 files changed, 105 insertions(+), 47 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index f360ec9..c1bcff8 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -395,6 +395,11 @@ struct cl_object_operations {
* mainly pages and locks.
*/
int (*coo_prune)(const struct lu_env *env, struct cl_object *obj);
+ /**
+ * Object getstripe method.
+ */
+ int (*coo_getstripe)(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *lum);
};
/**
@@ -2174,6 +2179,8 @@ int cl_conf_set(const struct lu_env *env, struct cl_object *obj,
const struct cl_object_conf *conf);
int cl_object_prune(const struct lu_env *env, struct cl_object *obj);
void cl_object_kill(const struct lu_env *env, struct cl_object *obj);
+int cl_object_getstripe(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *lum);
/**
* Returns true, iff \a o0 and \a o1 are slices of the same object.
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 0b3b387..2cb0b65 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1449,6 +1449,22 @@ static int ll_lov_setea(struct inode *inode, struct file *file,
return rc;
}
+static int ll_file_getstripe(struct inode *inode,
+ struct lov_user_md __user *lum)
+{
+ struct lu_env *env;
+ int refcheck;
+ int rc;
+
+ env = cl_env_get(&refcheck);
+ if (IS_ERR(env))
+ return PTR_ERR(env);
+
+ rc = cl_object_getstripe(env, ll_i2info(inode)->lli_clob, lum);
+ cl_env_put(env, &refcheck);
+ return rc;
+}
+
static int ll_lov_setstripe(struct inode *inode, struct file *file,
unsigned long arg)
{
@@ -1466,35 +1482,18 @@ static int ll_lov_setstripe(struct inode *inode, struct file *file,
lum_size);
cl_lov_delay_create_clear(&file->f_flags);
if (rc == 0) {
- struct lov_stripe_md *lsm;
__u32 gen;
put_user(0, &lum->lmm_stripe_count);
ll_layout_refresh(inode, &gen);
- lsm = ccc_inode_lsm_get(inode);
- rc = obd_iocontrol(LL_IOC_LOV_GETSTRIPE, ll_i2dtexp(inode),
- 0, lsm, lum);
- ccc_inode_lsm_put(inode, lsm);
+ rc = ll_file_getstripe(inode, (struct lov_user_md __user *)arg);
}
kfree(klum);
return rc;
}
-static int ll_lov_getstripe(struct inode *inode, unsigned long arg)
-{
- struct lov_stripe_md *lsm;
- int rc = -ENODATA;
-
- lsm = ccc_inode_lsm_get(inode);
- if (lsm)
- rc = obd_iocontrol(LL_IOC_LOV_GETSTRIPE, ll_i2dtexp(inode), 0,
- lsm, (void __user *)arg);
- ccc_inode_lsm_put(inode, lsm);
- return rc;
-}
-
static int
ll_get_grouplock(struct inode *inode, struct file *file, unsigned long arg)
{
@@ -2233,7 +2232,8 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
return rc;
}
case LL_IOC_LOV_GETSTRIPE:
- return ll_lov_getstripe(inode, arg);
+ return ll_file_getstripe(inode,
+ (struct lov_user_md __user *)arg);
case FSFILT_IOC_FIEMAP:
return ll_ioctl_fiemap(inode, arg);
case FSFILT_IOC_GETFLAGS:
diff --git a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
index 43d1a3f..ec40cf2 100644
--- a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
@@ -556,6 +556,8 @@ struct lov_lock_link *lov_lock_link_find(const struct lu_env *env,
struct lovsub_lock *sub);
struct lov_io_sub *lov_page_subio(const struct lu_env *env, struct lov_io *lio,
const struct cl_page_slice *slice);
+
+struct lov_stripe_md *lov_lsm_addref(struct lov_object *lov);
int lov_page_stripe(const struct cl_page *page);
#define lov_foreach_target(lov, var) \
@@ -747,6 +749,10 @@ static inline struct lov_layout_raid0 *lov_r0(struct lov_object *lov)
return &lov->u.raid0;
}
+/* lov_pack.c */
+int lov_getstripe(struct lov_object *obj, struct lov_stripe_md *lsm,
+ struct lov_user_md __user *lump);
+
/** @} lov */
#endif
diff --git a/drivers/staging/lustre/lustre/lov/lov_internal.h b/drivers/staging/lustre/lustre/lov/lov_internal.h
index 0e5fc2e..07e5ede 100644
--- a/drivers/staging/lustre/lustre/lov/lov_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_internal.h
@@ -190,8 +190,6 @@ int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmm,
struct lov_stripe_md *lsm);
int lov_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
struct lov_mds_md *lmm, int lmm_bytes);
-int lov_getstripe(struct obd_export *exp,
- struct lov_stripe_md *lsm, struct lov_user_md __user *lump);
int lov_alloc_memmd(struct lov_stripe_md **lsmp, __u16 stripe_count,
int pattern, int magic);
int lov_free_memmd(struct lov_stripe_md **lsmp);
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 9ab3033..b23016f 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -1299,9 +1299,6 @@ static int lov_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
obd_ioctl_freedata(buf, len);
break;
}
- case LL_IOC_LOV_GETSTRIPE:
- rc = lov_getstripe(exp, karg, uarg);
- break;
case OBD_IOC_QUOTACTL: {
struct if_quotactl *qctl = karg;
struct lov_tgt_desc *tgt = NULL;
diff --git a/drivers/staging/lustre/lustre/lov/lov_object.c b/drivers/staging/lustre/lustre/lov/lov_object.c
index 4f766e0..01f95a9 100644
--- a/drivers/staging/lustre/lustre/lov/lov_object.c
+++ b/drivers/staging/lustre/lustre/lov/lov_object.c
@@ -75,6 +75,13 @@ struct lov_layout_operations {
static int lov_layout_wait(const struct lu_env *env, struct lov_object *lov);
+void lov_lsm_put(struct cl_object *unused, struct lov_stripe_md *lsm)
+{
+ if (lsm)
+ lov_free_memmd(&lsm);
+}
+EXPORT_SYMBOL(lov_lsm_put);
+
/*****************************************************************************
*
* Lov object layout operations.
@@ -904,13 +911,30 @@ int lov_lock_init(const struct lu_env *env, struct cl_object *obj,
io);
}
+static int lov_object_getstripe(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *lum)
+{
+ struct lov_object *lov = cl2lov(obj);
+ struct lov_stripe_md *lsm;
+ int rc = 0;
+
+ lsm = lov_lsm_addref(lov);
+ if (!lsm)
+ return -ENODATA;
+
+ rc = lov_getstripe(cl2lov(obj), lsm, lum);
+ lov_lsm_put(obj, lsm);
+ return rc;
+}
+
static const struct cl_object_operations lov_ops = {
.coo_page_init = lov_page_init,
.coo_lock_init = lov_lock_init,
.coo_io_init = lov_io_init,
.coo_attr_get = lov_attr_get,
.coo_attr_set = lov_attr_set,
- .coo_conf_set = lov_conf_set
+ .coo_conf_set = lov_conf_set,
+ .coo_getstripe = lov_object_getstripe
};
static const struct lu_object_operations lov_lu_obj_ops = {
@@ -947,7 +971,7 @@ struct lu_object *lov_object_alloc(const struct lu_env *env,
return obj;
}
-static struct lov_stripe_md *lov_lsm_addref(struct lov_object *lov)
+struct lov_stripe_md *lov_lsm_addref(struct lov_object *lov)
{
struct lov_stripe_md *lsm = NULL;
@@ -978,13 +1002,6 @@ struct lov_stripe_md *lov_lsm_get(struct cl_object *clobj)
}
EXPORT_SYMBOL(lov_lsm_get);
-void lov_lsm_put(struct cl_object *unused, struct lov_stripe_md *lsm)
-{
- if (lsm)
- lov_free_memmd(&lsm);
-}
-EXPORT_SYMBOL(lov_lsm_put);
-
int lov_read_and_clear_async_rc(struct cl_object *clob)
{
struct lu_object *luobj;
diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
index 595cf16..be6e985 100644
--- a/drivers/staging/lustre/lustre/lov/lov_pack.c
+++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
@@ -45,6 +45,7 @@
#include "../include/lustre/lustre_user.h"
#include "lov_internal.h"
+#include "lov_cl_internal.h"
void lov_dump_lmm_common(int level, void *lmmp)
{
@@ -104,11 +105,9 @@ void lov_dump_lmm_v3(int level, struct lov_mds_md_v3 *lmm)
* LOVs properly. For now lov_mds_md_size() just assumes one u64
* per stripe.
*/
-int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
- struct lov_stripe_md *lsm)
+int lov_obd_packmd(struct lov_obd *lov, struct lov_mds_md **lmmp,
+ struct lov_stripe_md *lsm)
{
- struct obd_device *obd = class_exp2obd(exp);
- struct lov_obd *lov = &obd->u.lov;
struct lov_mds_md_v1 *lmmv1;
struct lov_mds_md_v3 *lmmv3;
__u16 stripe_count;
@@ -220,6 +219,15 @@ int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
return lmm_size;
}
+int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
+ struct lov_stripe_md *lsm)
+{
+ struct obd_device *obd = class_exp2obd(exp);
+ struct lov_obd *lov = &obd->u.lov;
+
+ return lov_obd_packmd(lov, lmmp, lsm);
+}
+
/* Find the max stripecount we should use */
__u16 lov_get_stripecnt(struct lov_obd *lov, __u32 magic, __u16 stripe_count)
{
@@ -367,16 +375,17 @@ int lov_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
* the maximum number of OST indices which will fit in the user buffer.
* lmm_magic must be LOV_USER_MAGIC.
*/
-int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
+int lov_getstripe(struct lov_object *obj, struct lov_stripe_md *lsm,
struct lov_user_md __user *lump)
{
/*
* XXX huge struct allocated on stack.
*/
/* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
+ struct lov_obd *lov;
struct lov_user_md_v3 lum;
struct lov_mds_md *lmmk = NULL;
- int rc, lmm_size;
+ int rc, lmmk_size, lmm_size;
int lum_size;
mm_segment_t seg;
@@ -396,13 +405,13 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
lum_size = sizeof(struct lov_user_md_v1);
if (copy_from_user(&lum, lump, lum_size)) {
rc = -EFAULT;
- goto out_set;
+ goto out;
}
if (lum.lmm_magic != LOV_USER_MAGIC_V1 &&
lum.lmm_magic != LOV_USER_MAGIC_V3 &&
lum.lmm_magic != LOV_USER_MAGIC_SPECIFIC) {
rc = -EINVAL;
- goto out_set;
+ goto out;
}
if (lum.lmm_stripe_count &&
@@ -411,11 +420,13 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
lum.lmm_stripe_count = lsm->lsm_stripe_count;
rc = copy_to_user(lump, &lum, lum_size);
rc = -EOVERFLOW;
- goto out_set;
+ goto out;
}
- rc = lov_packmd(exp, &lmmk, lsm);
+ lov = lu2lov_dev(obj->lo_cl.co_lu.lo_dev)->ld_lov;
+ rc = lov_obd_packmd(lov, &lmmk, lsm);
if (rc < 0)
- goto out_set;
+ goto out;
+ lmmk_size = rc;
lmm_size = rc;
rc = 0;
@@ -451,7 +462,7 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
lmm_size = lum_size;
} else if (lum.lmm_stripe_count < lmmk->lmm_stripe_count) {
rc = -EOVERFLOW;
- goto out_set;
+ goto out_free;
}
/*
* Have a difference between lov_mds_md & lov_user_md.
@@ -464,8 +475,9 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
if (copy_to_user(lump, lmmk, lmm_size))
rc = -EFAULT;
- obd_free_diskmd(exp, &lmmk);
-out_set:
+out_free:
+ kfree(lmmk);
+out:
set_fs(seg);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_object.c b/drivers/staging/lustre/lustre/obdclass/cl_object.c
index bd85455..ef5b1aa 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_object.c
@@ -321,6 +321,27 @@ int cl_object_prune(const struct lu_env *env, struct cl_object *obj)
EXPORT_SYMBOL(cl_object_prune);
/**
+ * Get stripe information of this object.
+ */
+int cl_object_getstripe(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *uarg)
+{
+ struct lu_object_header *top;
+ int result = 0;
+
+ top = obj->co_lu.lo_header;
+ list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
+ if (obj->co_ops->coo_getstripe) {
+ result = obj->co_ops->coo_getstripe(env, obj, uarg);
+ if (result)
+ break;
+ }
+ }
+ return result;
+}
+EXPORT_SYMBOL(cl_object_getstripe);
+
+/**
* Helper function removing all object locks, and marking object for
* deletion. All object pages must have been deleted at this point.
*
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 091/124] staging: lustre: clio: add coo_getstripe interface
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
Use cl_object_operations::coo_getstripe() to handle
LL_IOC_LOV_GETSTRIPE ops.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5823
Reviewed-on: http://review.whamcloud.com/12452
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 7 +++
drivers/staging/lustre/lustre/llite/file.c | 38 +++++++++---------
.../staging/lustre/lustre/lov/lov_cl_internal.h | 6 +++
drivers/staging/lustre/lustre/lov/lov_internal.h | 2 -
drivers/staging/lustre/lustre/lov/lov_obd.c | 3 -
drivers/staging/lustre/lustre/lov/lov_object.c | 35 +++++++++++++----
drivers/staging/lustre/lustre/lov/lov_pack.c | 40 +++++++++++++-------
drivers/staging/lustre/lustre/obdclass/cl_object.c | 21 ++++++++++
8 files changed, 105 insertions(+), 47 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index f360ec9..c1bcff8 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -395,6 +395,11 @@ struct cl_object_operations {
* mainly pages and locks.
*/
int (*coo_prune)(const struct lu_env *env, struct cl_object *obj);
+ /**
+ * Object getstripe method.
+ */
+ int (*coo_getstripe)(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *lum);
};
/**
@@ -2174,6 +2179,8 @@ int cl_conf_set(const struct lu_env *env, struct cl_object *obj,
const struct cl_object_conf *conf);
int cl_object_prune(const struct lu_env *env, struct cl_object *obj);
void cl_object_kill(const struct lu_env *env, struct cl_object *obj);
+int cl_object_getstripe(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *lum);
/**
* Returns true, iff \a o0 and \a o1 are slices of the same object.
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 0b3b387..2cb0b65 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1449,6 +1449,22 @@ static int ll_lov_setea(struct inode *inode, struct file *file,
return rc;
}
+static int ll_file_getstripe(struct inode *inode,
+ struct lov_user_md __user *lum)
+{
+ struct lu_env *env;
+ int refcheck;
+ int rc;
+
+ env = cl_env_get(&refcheck);
+ if (IS_ERR(env))
+ return PTR_ERR(env);
+
+ rc = cl_object_getstripe(env, ll_i2info(inode)->lli_clob, lum);
+ cl_env_put(env, &refcheck);
+ return rc;
+}
+
static int ll_lov_setstripe(struct inode *inode, struct file *file,
unsigned long arg)
{
@@ -1466,35 +1482,18 @@ static int ll_lov_setstripe(struct inode *inode, struct file *file,
lum_size);
cl_lov_delay_create_clear(&file->f_flags);
if (rc == 0) {
- struct lov_stripe_md *lsm;
__u32 gen;
put_user(0, &lum->lmm_stripe_count);
ll_layout_refresh(inode, &gen);
- lsm = ccc_inode_lsm_get(inode);
- rc = obd_iocontrol(LL_IOC_LOV_GETSTRIPE, ll_i2dtexp(inode),
- 0, lsm, lum);
- ccc_inode_lsm_put(inode, lsm);
+ rc = ll_file_getstripe(inode, (struct lov_user_md __user *)arg);
}
kfree(klum);
return rc;
}
-static int ll_lov_getstripe(struct inode *inode, unsigned long arg)
-{
- struct lov_stripe_md *lsm;
- int rc = -ENODATA;
-
- lsm = ccc_inode_lsm_get(inode);
- if (lsm)
- rc = obd_iocontrol(LL_IOC_LOV_GETSTRIPE, ll_i2dtexp(inode), 0,
- lsm, (void __user *)arg);
- ccc_inode_lsm_put(inode, lsm);
- return rc;
-}
-
static int
ll_get_grouplock(struct inode *inode, struct file *file, unsigned long arg)
{
@@ -2233,7 +2232,8 @@ ll_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
return rc;
}
case LL_IOC_LOV_GETSTRIPE:
- return ll_lov_getstripe(inode, arg);
+ return ll_file_getstripe(inode,
+ (struct lov_user_md __user *)arg);
case FSFILT_IOC_FIEMAP:
return ll_ioctl_fiemap(inode, arg);
case FSFILT_IOC_GETFLAGS:
diff --git a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
index 43d1a3f..ec40cf2 100644
--- a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
@@ -556,6 +556,8 @@ struct lov_lock_link *lov_lock_link_find(const struct lu_env *env,
struct lovsub_lock *sub);
struct lov_io_sub *lov_page_subio(const struct lu_env *env, struct lov_io *lio,
const struct cl_page_slice *slice);
+
+struct lov_stripe_md *lov_lsm_addref(struct lov_object *lov);
int lov_page_stripe(const struct cl_page *page);
#define lov_foreach_target(lov, var) \
@@ -747,6 +749,10 @@ static inline struct lov_layout_raid0 *lov_r0(struct lov_object *lov)
return &lov->u.raid0;
}
+/* lov_pack.c */
+int lov_getstripe(struct lov_object *obj, struct lov_stripe_md *lsm,
+ struct lov_user_md __user *lump);
+
/** @} lov */
#endif
diff --git a/drivers/staging/lustre/lustre/lov/lov_internal.h b/drivers/staging/lustre/lustre/lov/lov_internal.h
index 0e5fc2e..07e5ede 100644
--- a/drivers/staging/lustre/lustre/lov/lov_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_internal.h
@@ -190,8 +190,6 @@ int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmm,
struct lov_stripe_md *lsm);
int lov_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
struct lov_mds_md *lmm, int lmm_bytes);
-int lov_getstripe(struct obd_export *exp,
- struct lov_stripe_md *lsm, struct lov_user_md __user *lump);
int lov_alloc_memmd(struct lov_stripe_md **lsmp, __u16 stripe_count,
int pattern, int magic);
int lov_free_memmd(struct lov_stripe_md **lsmp);
diff --git a/drivers/staging/lustre/lustre/lov/lov_obd.c b/drivers/staging/lustre/lustre/lov/lov_obd.c
index 9ab3033..b23016f 100644
--- a/drivers/staging/lustre/lustre/lov/lov_obd.c
+++ b/drivers/staging/lustre/lustre/lov/lov_obd.c
@@ -1299,9 +1299,6 @@ static int lov_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
obd_ioctl_freedata(buf, len);
break;
}
- case LL_IOC_LOV_GETSTRIPE:
- rc = lov_getstripe(exp, karg, uarg);
- break;
case OBD_IOC_QUOTACTL: {
struct if_quotactl *qctl = karg;
struct lov_tgt_desc *tgt = NULL;
diff --git a/drivers/staging/lustre/lustre/lov/lov_object.c b/drivers/staging/lustre/lustre/lov/lov_object.c
index 4f766e0..01f95a9 100644
--- a/drivers/staging/lustre/lustre/lov/lov_object.c
+++ b/drivers/staging/lustre/lustre/lov/lov_object.c
@@ -75,6 +75,13 @@ struct lov_layout_operations {
static int lov_layout_wait(const struct lu_env *env, struct lov_object *lov);
+void lov_lsm_put(struct cl_object *unused, struct lov_stripe_md *lsm)
+{
+ if (lsm)
+ lov_free_memmd(&lsm);
+}
+EXPORT_SYMBOL(lov_lsm_put);
+
/*****************************************************************************
*
* Lov object layout operations.
@@ -904,13 +911,30 @@ int lov_lock_init(const struct lu_env *env, struct cl_object *obj,
io);
}
+static int lov_object_getstripe(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *lum)
+{
+ struct lov_object *lov = cl2lov(obj);
+ struct lov_stripe_md *lsm;
+ int rc = 0;
+
+ lsm = lov_lsm_addref(lov);
+ if (!lsm)
+ return -ENODATA;
+
+ rc = lov_getstripe(cl2lov(obj), lsm, lum);
+ lov_lsm_put(obj, lsm);
+ return rc;
+}
+
static const struct cl_object_operations lov_ops = {
.coo_page_init = lov_page_init,
.coo_lock_init = lov_lock_init,
.coo_io_init = lov_io_init,
.coo_attr_get = lov_attr_get,
.coo_attr_set = lov_attr_set,
- .coo_conf_set = lov_conf_set
+ .coo_conf_set = lov_conf_set,
+ .coo_getstripe = lov_object_getstripe
};
static const struct lu_object_operations lov_lu_obj_ops = {
@@ -947,7 +971,7 @@ struct lu_object *lov_object_alloc(const struct lu_env *env,
return obj;
}
-static struct lov_stripe_md *lov_lsm_addref(struct lov_object *lov)
+struct lov_stripe_md *lov_lsm_addref(struct lov_object *lov)
{
struct lov_stripe_md *lsm = NULL;
@@ -978,13 +1002,6 @@ struct lov_stripe_md *lov_lsm_get(struct cl_object *clobj)
}
EXPORT_SYMBOL(lov_lsm_get);
-void lov_lsm_put(struct cl_object *unused, struct lov_stripe_md *lsm)
-{
- if (lsm)
- lov_free_memmd(&lsm);
-}
-EXPORT_SYMBOL(lov_lsm_put);
-
int lov_read_and_clear_async_rc(struct cl_object *clob)
{
struct lu_object *luobj;
diff --git a/drivers/staging/lustre/lustre/lov/lov_pack.c b/drivers/staging/lustre/lustre/lov/lov_pack.c
index 595cf16..be6e985 100644
--- a/drivers/staging/lustre/lustre/lov/lov_pack.c
+++ b/drivers/staging/lustre/lustre/lov/lov_pack.c
@@ -45,6 +45,7 @@
#include "../include/lustre/lustre_user.h"
#include "lov_internal.h"
+#include "lov_cl_internal.h"
void lov_dump_lmm_common(int level, void *lmmp)
{
@@ -104,11 +105,9 @@ void lov_dump_lmm_v3(int level, struct lov_mds_md_v3 *lmm)
* LOVs properly. For now lov_mds_md_size() just assumes one u64
* per stripe.
*/
-int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
- struct lov_stripe_md *lsm)
+int lov_obd_packmd(struct lov_obd *lov, struct lov_mds_md **lmmp,
+ struct lov_stripe_md *lsm)
{
- struct obd_device *obd = class_exp2obd(exp);
- struct lov_obd *lov = &obd->u.lov;
struct lov_mds_md_v1 *lmmv1;
struct lov_mds_md_v3 *lmmv3;
__u16 stripe_count;
@@ -220,6 +219,15 @@ int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
return lmm_size;
}
+int lov_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
+ struct lov_stripe_md *lsm)
+{
+ struct obd_device *obd = class_exp2obd(exp);
+ struct lov_obd *lov = &obd->u.lov;
+
+ return lov_obd_packmd(lov, lmmp, lsm);
+}
+
/* Find the max stripecount we should use */
__u16 lov_get_stripecnt(struct lov_obd *lov, __u32 magic, __u16 stripe_count)
{
@@ -367,16 +375,17 @@ int lov_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
* the maximum number of OST indices which will fit in the user buffer.
* lmm_magic must be LOV_USER_MAGIC.
*/
-int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
+int lov_getstripe(struct lov_object *obj, struct lov_stripe_md *lsm,
struct lov_user_md __user *lump)
{
/*
* XXX huge struct allocated on stack.
*/
/* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
+ struct lov_obd *lov;
struct lov_user_md_v3 lum;
struct lov_mds_md *lmmk = NULL;
- int rc, lmm_size;
+ int rc, lmmk_size, lmm_size;
int lum_size;
mm_segment_t seg;
@@ -396,13 +405,13 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
lum_size = sizeof(struct lov_user_md_v1);
if (copy_from_user(&lum, lump, lum_size)) {
rc = -EFAULT;
- goto out_set;
+ goto out;
}
if (lum.lmm_magic != LOV_USER_MAGIC_V1 &&
lum.lmm_magic != LOV_USER_MAGIC_V3 &&
lum.lmm_magic != LOV_USER_MAGIC_SPECIFIC) {
rc = -EINVAL;
- goto out_set;
+ goto out;
}
if (lum.lmm_stripe_count &&
@@ -411,11 +420,13 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
lum.lmm_stripe_count = lsm->lsm_stripe_count;
rc = copy_to_user(lump, &lum, lum_size);
rc = -EOVERFLOW;
- goto out_set;
+ goto out;
}
- rc = lov_packmd(exp, &lmmk, lsm);
+ lov = lu2lov_dev(obj->lo_cl.co_lu.lo_dev)->ld_lov;
+ rc = lov_obd_packmd(lov, &lmmk, lsm);
if (rc < 0)
- goto out_set;
+ goto out;
+ lmmk_size = rc;
lmm_size = rc;
rc = 0;
@@ -451,7 +462,7 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
lmm_size = lum_size;
} else if (lum.lmm_stripe_count < lmmk->lmm_stripe_count) {
rc = -EOVERFLOW;
- goto out_set;
+ goto out_free;
}
/*
* Have a difference between lov_mds_md & lov_user_md.
@@ -464,8 +475,9 @@ int lov_getstripe(struct obd_export *exp, struct lov_stripe_md *lsm,
if (copy_to_user(lump, lmmk, lmm_size))
rc = -EFAULT;
- obd_free_diskmd(exp, &lmmk);
-out_set:
+out_free:
+ kfree(lmmk);
+out:
set_fs(seg);
return rc;
}
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_object.c b/drivers/staging/lustre/lustre/obdclass/cl_object.c
index bd85455..ef5b1aa 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_object.c
@@ -321,6 +321,27 @@ int cl_object_prune(const struct lu_env *env, struct cl_object *obj)
EXPORT_SYMBOL(cl_object_prune);
/**
+ * Get stripe information of this object.
+ */
+int cl_object_getstripe(const struct lu_env *env, struct cl_object *obj,
+ struct lov_user_md __user *uarg)
+{
+ struct lu_object_header *top;
+ int result = 0;
+
+ top = obj->co_lu.lo_header;
+ list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
+ if (obj->co_ops->coo_getstripe) {
+ result = obj->co_ops->coo_getstripe(env, obj, uarg);
+ if (result)
+ break;
+ }
+ }
+ return result;
+}
+EXPORT_SYMBOL(cl_object_getstripe);
+
+/**
* Helper function removing all object locks, and marking object for
* deletion. All object pages must have been deleted at this point.
*
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 092/124] staging: lustre: ptlrpc: fix comparison between signed and unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change return type and size argiments of lustre_msg_hdr_size(),
lustre_msg_buf{len,count}() and req_capsule_*_size() to __u32.
Change type of req_format->rf_idx and req_format->rf_fields.nr
to size_t. Also return zero for incorrect message magic instead
of -EINVAL. This will be more robust because of few of them after
LASSERTF(0, "...") and will not be returned. In the rest places
it return zero size instead of huge number after implicit
unsigned conversion.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12475
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_net.h | 34 ++++----
.../lustre/lustre/include/lustre_req_layout.h | 20 +++---
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 4 +-
drivers/staging/lustre/lustre/ptlrpc/layout.c | 83 ++++++++++----------
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 62 +++++++--------
5 files changed, 101 insertions(+), 102 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 531c628..21e2267 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -2616,9 +2616,9 @@ int ptlrpc_reconnect_import(struct obd_import *imp);
* @{
*/
int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
- int index);
+ u32 index);
void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
- int index);
+ u32 index);
int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len);
int ptlrpc_unpack_req_msg(struct ptlrpc_request *req, int len);
@@ -2637,27 +2637,27 @@ int lustre_shrink_msg(struct lustre_msg *msg, int segment,
unsigned int newlen, int move_data);
void lustre_free_reply_state(struct ptlrpc_reply_state *rs);
int __lustre_unpack_msg(struct lustre_msg *m, int len);
-int lustre_msg_hdr_size(__u32 magic, int count);
-int lustre_msg_size(__u32 magic, int count, __u32 *lengths);
-int lustre_msg_size_v2(int count, __u32 *lengths);
-int lustre_packed_msg_size(struct lustre_msg *msg);
-int lustre_msg_early_size(void);
-void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size);
-void *lustre_msg_buf(struct lustre_msg *m, int n, int minlen);
-int lustre_msg_buflen(struct lustre_msg *m, int n);
-int lustre_msg_bufcount(struct lustre_msg *m);
-char *lustre_msg_string(struct lustre_msg *m, int n, int max_len);
+u32 lustre_msg_hdr_size(__u32 magic, u32 count);
+u32 lustre_msg_size(__u32 magic, int count, __u32 *lengths);
+u32 lustre_msg_size_v2(int count, __u32 *lengths);
+u32 lustre_packed_msg_size(struct lustre_msg *msg);
+u32 lustre_msg_early_size(void);
+void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, u32 n, u32 min_size);
+void *lustre_msg_buf(struct lustre_msg *m, u32 n, u32 minlen);
+u32 lustre_msg_buflen(struct lustre_msg *m, u32 n);
+u32 lustre_msg_bufcount(struct lustre_msg *m);
+char *lustre_msg_string(struct lustre_msg *m, u32 n, u32 max_len);
__u32 lustre_msghdr_get_flags(struct lustre_msg *msg);
void lustre_msghdr_set_flags(struct lustre_msg *msg, __u32 flags);
__u32 lustre_msg_get_flags(struct lustre_msg *msg);
-void lustre_msg_add_flags(struct lustre_msg *msg, int flags);
-void lustre_msg_set_flags(struct lustre_msg *msg, int flags);
-void lustre_msg_clear_flags(struct lustre_msg *msg, int flags);
+void lustre_msg_add_flags(struct lustre_msg *msg, u32 flags);
+void lustre_msg_set_flags(struct lustre_msg *msg, u32 flags);
+void lustre_msg_clear_flags(struct lustre_msg *msg, u32 flags);
__u32 lustre_msg_get_op_flags(struct lustre_msg *msg);
-void lustre_msg_add_op_flags(struct lustre_msg *msg, int flags);
+void lustre_msg_add_op_flags(struct lustre_msg *msg, u32 flags);
struct lustre_handle *lustre_msg_get_handle(struct lustre_msg *msg);
__u32 lustre_msg_get_type(struct lustre_msg *msg);
-void lustre_msg_add_version(struct lustre_msg *msg, int version);
+void lustre_msg_add_version(struct lustre_msg *msg, u32 version);
__u32 lustre_msg_get_opc(struct lustre_msg *msg);
__u64 lustre_msg_get_last_committed(struct lustre_msg *msg);
__u64 *lustre_msg_get_versions(struct lustre_msg *msg);
diff --git a/drivers/staging/lustre/lustre/include/lustre_req_layout.h b/drivers/staging/lustre/lustre/include/lustre_req_layout.h
index ca0e683..a13558e 100644
--- a/drivers/staging/lustre/lustre/include/lustre_req_layout.h
+++ b/drivers/staging/lustre/lustre/include/lustre_req_layout.h
@@ -76,7 +76,8 @@ void req_capsule_init(struct req_capsule *pill, struct ptlrpc_request *req,
void req_capsule_fini(struct req_capsule *pill);
void req_capsule_set(struct req_capsule *pill, const struct req_format *fmt);
-int req_capsule_filled_sizes(struct req_capsule *pill, enum req_location loc);
+size_t req_capsule_filled_sizes(struct req_capsule *pill,
+ enum req_location loc);
int req_capsule_server_pack(struct req_capsule *pill);
void *req_capsule_client_get(struct req_capsule *pill,
@@ -86,27 +87,27 @@ void *req_capsule_client_swab_get(struct req_capsule *pill,
void *swabber);
void *req_capsule_client_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len);
+ u32 len);
void *req_capsule_server_get(struct req_capsule *pill,
const struct req_msg_field *field);
void *req_capsule_server_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len);
+ u32 len);
void *req_capsule_server_swab_get(struct req_capsule *pill,
const struct req_msg_field *field,
void *swabber);
void *req_capsule_server_sized_swab_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len, void *swabber);
+ u32 len, void *swabber);
void req_capsule_set_size(struct req_capsule *pill,
const struct req_msg_field *field,
- enum req_location loc, int size);
-int req_capsule_get_size(const struct req_capsule *pill,
+ enum req_location loc, u32 size);
+u32 req_capsule_get_size(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc);
-int req_capsule_msg_size(struct req_capsule *pill, enum req_location loc);
-int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
+u32 req_capsule_msg_size(struct req_capsule *pill, enum req_location loc);
+u32 req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
enum req_location loc);
void req_capsule_extend(struct req_capsule *pill, const struct req_format *fmt);
@@ -115,8 +116,7 @@ int req_capsule_has_field(const struct req_capsule *pill,
enum req_location loc);
void req_capsule_shrink(struct req_capsule *pill,
const struct req_msg_field *field,
- unsigned int newlen,
- enum req_location loc);
+ u32 newlen, enum req_location loc);
int req_layout_init(void);
void req_layout_fini(void);
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index 582de52..35ba6f1 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -561,7 +561,7 @@ static inline int ldlm_capsule_handles_avail(struct req_capsule *pill,
enum req_location loc,
int off)
{
- int size = req_capsule_msg_size(pill, loc);
+ u32 size = req_capsule_msg_size(pill, loc);
return ldlm_req_handles_avail(size, off);
}
@@ -570,7 +570,7 @@ static inline int ldlm_format_handles_avail(struct obd_import *imp,
const struct req_format *fmt,
enum req_location loc, int off)
{
- int size = req_capsule_fmt_size(imp->imp_msg_magic, fmt, loc);
+ u32 size = req_capsule_fmt_size(imp->imp_msg_magic, fmt, loc);
return ldlm_req_handles_avail(size, off);
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/layout.c b/drivers/staging/lustre/lustre/ptlrpc/layout.c
index 101ac87..839ef3e 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/layout.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/layout.c
@@ -1113,9 +1113,9 @@ EXPORT_SYMBOL(RMF_SWAP_LAYOUTS);
struct req_format {
const char *rf_name;
- int rf_idx;
+ size_t rf_idx;
struct {
- int nr;
+ size_t nr;
const struct req_msg_field **d;
} rf_fields[RCL_NR];
};
@@ -1555,9 +1555,9 @@ EXPORT_SYMBOL(RQF_OST_GET_INFO_FIEMAP);
*/
int req_layout_init(void)
{
- int i;
- int j;
- int k;
+ size_t i;
+ size_t j;
+ size_t k;
struct req_format *rf = NULL;
for (i = 0; i < ARRAY_SIZE(req_formats); ++i) {
@@ -1599,7 +1599,7 @@ EXPORT_SYMBOL(req_layout_fini);
*/
static void req_capsule_init_area(struct req_capsule *pill)
{
- int i;
+ size_t i;
for (i = 0; i < ARRAY_SIZE(pill->rc_area[RCL_CLIENT]); i++) {
pill->rc_area[RCL_CLIENT][i] = -1;
@@ -1650,8 +1650,7 @@ EXPORT_SYMBOL(req_capsule_fini);
static int __req_format_is_sane(const struct req_format *fmt)
{
- return
- 0 <= fmt->rf_idx && fmt->rf_idx < ARRAY_SIZE(req_formats) &&
+ return fmt->rf_idx < ARRAY_SIZE(req_formats) &&
req_formats[fmt->rf_idx] == fmt;
}
@@ -1685,11 +1684,11 @@ EXPORT_SYMBOL(req_capsule_set);
* variable-sized fields. The field sizes come from the declared \a rmf_size
* field of a \a pill's \a rc_fmt's RMF's.
*/
-int req_capsule_filled_sizes(struct req_capsule *pill,
- enum req_location loc)
+size_t req_capsule_filled_sizes(struct req_capsule *pill,
+ enum req_location loc)
{
const struct req_format *fmt = pill->rc_fmt;
- int i;
+ size_t i;
for (i = 0; i < fmt->rf_fields[loc].nr; ++i) {
if (pill->rc_area[loc][i] == -1) {
@@ -1744,11 +1743,11 @@ EXPORT_SYMBOL(req_capsule_server_pack);
* Returns the PTLRPC request or reply (\a loc) buffer offset of a \a pill
* corresponding to the given RMF (\a field).
*/
-static int __req_capsule_offset(const struct req_capsule *pill,
+static u32 __req_capsule_offset(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc)
{
- int offset;
+ u32 offset;
offset = field->rmf_offset[pill->rc_fmt->rf_idx][loc];
LASSERTF(offset > 0, "%s:%s, off=%d, loc=%d\n", pill->rc_fmt->rf_name,
@@ -1852,10 +1851,10 @@ static void *__req_capsule_get(struct req_capsule *pill,
const struct req_format *fmt;
struct lustre_msg *msg;
void *value;
- int len;
- int offset;
+ u32 len;
+ u32 offset;
- void *(*getter)(struct lustre_msg *m, int n, int minlen);
+ void *(*getter)(struct lustre_msg *m, u32 n, u32 minlen);
static const char *rcl_names[RCL_NR] = {
[RCL_CLIENT] = "client",
@@ -1882,20 +1881,20 @@ static void *__req_capsule_get(struct req_capsule *pill,
*/
len = lustre_msg_buflen(msg, offset);
if ((len % field->rmf_size) != 0) {
- CERROR("%s: array field size mismatch %d modulo %d != 0 (%d)\n",
+ CERROR("%s: array field size mismatch %d modulo %u != 0 (%d)\n",
field->rmf_name, len, field->rmf_size, loc);
return NULL;
}
} else if (pill->rc_area[loc][offset] != -1) {
len = pill->rc_area[loc][offset];
} else {
- len = max(field->rmf_size, 0);
+ len = max_t(typeof(field->rmf_size), field->rmf_size, 0);
}
value = getter(msg, offset, len);
if (!value) {
DEBUG_REQ(D_ERROR, pill->rc_req,
- "Wrong buffer for field `%s' (%d of %d) in format `%s': %d vs. %d (%s)\n",
+ "Wrong buffer for field `%s' (%u of %u) in format `%s': %u vs. %u (%s)\n",
field->rmf_name, offset, lustre_msg_bufcount(msg),
fmt->rf_name, lustre_msg_buflen(msg, offset), len,
rcl_names[loc]);
@@ -1941,7 +1940,7 @@ EXPORT_SYMBOL(req_capsule_client_swab_get);
*/
void *req_capsule_client_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len)
+ u32 len)
{
req_capsule_set_size(pill, field, RCL_CLIENT, len);
return __req_capsule_get(pill, field, RCL_CLIENT, NULL, 0);
@@ -1982,7 +1981,7 @@ EXPORT_SYMBOL(req_capsule_server_swab_get);
*/
void *req_capsule_server_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len)
+ u32 len)
{
req_capsule_set_size(pill, field, RCL_SERVER, len);
return __req_capsule_get(pill, field, RCL_SERVER, NULL, 0);
@@ -1991,7 +1990,7 @@ EXPORT_SYMBOL(req_capsule_server_sized_get);
void *req_capsule_server_sized_swab_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len, void *swabber)
+ u32 len, void *swabber)
{
req_capsule_set_size(pill, field, RCL_SERVER, len);
return __req_capsule_get(pill, field, RCL_SERVER, swabber, 0);
@@ -2007,23 +2006,25 @@ EXPORT_SYMBOL(req_capsule_server_sized_swab_get);
*/
void req_capsule_set_size(struct req_capsule *pill,
const struct req_msg_field *field,
- enum req_location loc, int size)
+ enum req_location loc, u32 size)
{
LASSERT(loc == RCL_SERVER || loc == RCL_CLIENT);
- if ((size != field->rmf_size) &&
+ if ((size != (u32)field->rmf_size) &&
(field->rmf_size != -1) &&
!(field->rmf_flags & RMF_F_NO_SIZE_CHECK) &&
(size > 0)) {
+ u32 rmf_size = (u32)field->rmf_size;
+
if ((field->rmf_flags & RMF_F_STRUCT_ARRAY) &&
- (size % field->rmf_size != 0)) {
- CERROR("%s: array field size mismatch %d %% %d != 0 (%d)\n",
- field->rmf_name, size, field->rmf_size, loc);
+ (size % rmf_size != 0)) {
+ CERROR("%s: array field size mismatch %u %% %u != 0 (%d)\n",
+ field->rmf_name, size, rmf_size, loc);
LBUG();
} else if (!(field->rmf_flags & RMF_F_STRUCT_ARRAY) &&
- size < field->rmf_size) {
- CERROR("%s: field size mismatch %d != %d (%d)\n",
- field->rmf_name, size, field->rmf_size, loc);
+ size < rmf_size) {
+ CERROR("%s: field size mismatch %u != %u (%d)\n",
+ field->rmf_name, size, rmf_size, loc);
LBUG();
}
}
@@ -2040,7 +2041,7 @@ EXPORT_SYMBOL(req_capsule_set_size);
* actually sets the size in pill.rc_area[loc][offset], but this function
* returns the message buflen[offset], maybe we should use another name.
*/
-int req_capsule_get_size(const struct req_capsule *pill,
+u32 req_capsule_get_size(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc)
{
@@ -2058,7 +2059,7 @@ EXPORT_SYMBOL(req_capsule_get_size);
*
* See also req_capsule_set_size().
*/
-int req_capsule_msg_size(struct req_capsule *pill, enum req_location loc)
+u32 req_capsule_msg_size(struct req_capsule *pill, enum req_location loc)
{
return lustre_msg_size(pill->rc_req->rq_import->imp_msg_magic,
pill->rc_fmt->rf_fields[loc].nr,
@@ -2073,10 +2074,11 @@ int req_capsule_msg_size(struct req_capsule *pill, enum req_location loc)
* This function should not be used for formats which contain variable size
* fields.
*/
-int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
+u32 req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
enum req_location loc)
{
- int size, i = 0;
+ size_t i = 0;
+ u32 size;
/*
* This function should probably LASSERT() that fmt has no fields with
@@ -2086,7 +2088,7 @@ int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
* we do.
*/
size = lustre_msg_hdr_size(magic, fmt->rf_fields[loc].nr);
- if (size < 0)
+ if (!size)
return size;
for (; i < fmt->rf_fields[loc].nr; ++i)
@@ -2118,7 +2120,7 @@ int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
void req_capsule_extend(struct req_capsule *pill, const struct req_format *fmt)
{
int i;
- int j;
+ size_t j;
const struct req_format *old;
@@ -2176,7 +2178,7 @@ static int req_capsule_field_present(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc)
{
- int offset;
+ u32 offset;
LASSERT(loc == RCL_SERVER || loc == RCL_CLIENT);
LASSERT(req_capsule_has_field(pill, field, loc));
@@ -2193,12 +2195,11 @@ static int req_capsule_field_present(const struct req_capsule *pill,
*/
void req_capsule_shrink(struct req_capsule *pill,
const struct req_msg_field *field,
- unsigned int newlen,
- enum req_location loc)
+ u32 newlen, enum req_location loc)
{
const struct req_format *fmt;
struct lustre_msg *msg;
- int len;
+ u32 len;
int offset;
fmt = pill->rc_fmt;
@@ -2211,7 +2212,7 @@ void req_capsule_shrink(struct req_capsule *pill,
msg = __req_msg(pill, loc);
len = lustre_msg_buflen(msg, offset);
- LASSERTF(newlen <= len, "%s:%s, oldlen=%d, newlen=%d\n",
+ LASSERTF(newlen <= len, "%s:%s, oldlen=%u, newlen=%u\n",
fmt->rf_name, field->rmf_name, len, newlen);
if (loc == RCL_CLIENT)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 87027c5..a1ea0c2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -50,26 +50,26 @@
#include "ptlrpc_internal.h"
-static inline int lustre_msg_hdr_size_v2(int count)
+static inline u32 lustre_msg_hdr_size_v2(u32 count)
{
return cfs_size_round(offsetof(struct lustre_msg_v2,
lm_buflens[count]));
}
-int lustre_msg_hdr_size(__u32 magic, int count)
+u32 lustre_msg_hdr_size(__u32 magic, u32 count)
{
switch (magic) {
case LUSTRE_MSG_MAGIC_V2:
return lustre_msg_hdr_size_v2(count);
default:
LASSERTF(0, "incorrect message magic: %08x\n", magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_hdr_size);
void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
- int index)
+ u32 index)
{
if (inout)
lustre_set_req_swabbed(req, index);
@@ -79,7 +79,7 @@ void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
EXPORT_SYMBOL(ptlrpc_buf_set_swabbed);
int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
- int index)
+ u32 index)
{
if (inout)
return (ptlrpc_req_need_swab(req) &&
@@ -91,9 +91,9 @@ int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
EXPORT_SYMBOL(ptlrpc_buf_need_swab);
/* early reply size */
-int lustre_msg_early_size(void)
+u32 lustre_msg_early_size(void)
{
- static int size;
+ static u32 size;
if (!size) {
/* Always reply old ptlrpc_body_v2 to keep interoperability
@@ -111,9 +111,9 @@ int lustre_msg_early_size(void)
}
EXPORT_SYMBOL(lustre_msg_early_size);
-int lustre_msg_size_v2(int count, __u32 *lengths)
+u32 lustre_msg_size_v2(int count, __u32 *lengths)
{
- int size;
+ u32 size;
int i;
size = lustre_msg_hdr_size_v2(count);
@@ -131,7 +131,7 @@ EXPORT_SYMBOL(lustre_msg_size_v2);
* target then the first buffer will be stripped because the ptlrpc
* data is part of the lustre_msg_v1 header. b=14043
*/
-int lustre_msg_size(__u32 magic, int count, __u32 *lens)
+u32 lustre_msg_size(__u32 magic, int count, __u32 *lens)
{
__u32 size[] = { sizeof(struct ptlrpc_body) };
@@ -148,7 +148,7 @@ int lustre_msg_size(__u32 magic, int count, __u32 *lens)
return lustre_msg_size_v2(count, lens);
default:
LASSERTF(0, "incorrect message magic: %08x\n", magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_size);
@@ -156,7 +156,7 @@ EXPORT_SYMBOL(lustre_msg_size);
/* This is used to determine the size of a buffer that was already packed
* and will correctly handle the different message formats.
*/
-int lustre_packed_msg_size(struct lustre_msg *msg)
+u32 lustre_packed_msg_size(struct lustre_msg *msg)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
@@ -378,11 +378,9 @@ int lustre_pack_reply(struct ptlrpc_request *req, int count, __u32 *lens,
}
EXPORT_SYMBOL(lustre_pack_reply);
-void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size)
+void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, u32 n, u32 min_size)
{
- int i, offset, buflen, bufcount;
-
- LASSERT(n >= 0);
+ u32 i, offset, buflen, bufcount;
bufcount = m->lm_bufcount;
if (unlikely(n >= bufcount)) {
@@ -406,7 +404,7 @@ void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size)
return (char *)m + offset;
}
-void *lustre_msg_buf(struct lustre_msg *m, int n, int min_size)
+void *lustre_msg_buf(struct lustre_msg *m, u32 n, u32 min_size)
{
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
@@ -419,7 +417,7 @@ void *lustre_msg_buf(struct lustre_msg *m, int n, int min_size)
}
EXPORT_SYMBOL(lustre_msg_buf);
-static int lustre_shrink_msg_v2(struct lustre_msg_v2 *msg, int segment,
+static int lustre_shrink_msg_v2(struct lustre_msg_v2 *msg, u32 segment,
unsigned int newlen, int move_data)
{
char *tail = NULL, *newpos;
@@ -647,7 +645,7 @@ int lustre_unpack_rep_ptlrpc_body(struct ptlrpc_request *req, int offset)
}
}
-static inline int lustre_msg_buflen_v2(struct lustre_msg_v2 *m, int n)
+static inline u32 lustre_msg_buflen_v2(struct lustre_msg_v2 *m, u32 n)
{
if (n >= m->lm_bufcount)
return 0;
@@ -662,14 +660,14 @@ static inline int lustre_msg_buflen_v2(struct lustre_msg_v2 *m, int n)
*
* returns zero for non-existent message indices
*/
-int lustre_msg_buflen(struct lustre_msg *m, int n)
+u32 lustre_msg_buflen(struct lustre_msg *m, u32 n)
{
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
return lustre_msg_buflen_v2(m, n);
default:
CERROR("incorrect message magic: %08x\n", m->lm_magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_buflen);
@@ -677,23 +675,23 @@ EXPORT_SYMBOL(lustre_msg_buflen);
/* NB return the bufcount for lustre_msg_v2 format, so if message is packed
* in V1 format, the result is one bigger. (add struct ptlrpc_body).
*/
-int lustre_msg_bufcount(struct lustre_msg *m)
+u32 lustre_msg_bufcount(struct lustre_msg *m)
{
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
return m->lm_bufcount;
default:
CERROR("incorrect message magic: %08x\n", m->lm_magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_bufcount);
-char *lustre_msg_string(struct lustre_msg *m, int index, int max_len)
+char *lustre_msg_string(struct lustre_msg *m, u32 index, u32 max_len)
{
/* max_len == 0 means the string should fill the buffer */
char *str;
- int slen, blen;
+ u32 slen, blen;
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
@@ -734,8 +732,8 @@ char *lustre_msg_string(struct lustre_msg *m, int index, int max_len)
EXPORT_SYMBOL(lustre_msg_string);
/* Wrap up the normal fixed length cases */
-static inline void *__lustre_swab_buf(struct lustre_msg *msg, int index,
- int min_size, void *swabber)
+static inline void *__lustre_swab_buf(struct lustre_msg *msg, u32 index,
+ u32 min_size, void *swabber)
{
void *ptr = NULL;
@@ -804,7 +802,7 @@ __u32 lustre_msg_get_flags(struct lustre_msg *msg)
}
EXPORT_SYMBOL(lustre_msg_get_flags);
-void lustre_msg_add_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_add_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -820,7 +818,7 @@ void lustre_msg_add_flags(struct lustre_msg *msg, int flags)
}
EXPORT_SYMBOL(lustre_msg_add_flags);
-void lustre_msg_set_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_set_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -836,7 +834,7 @@ void lustre_msg_set_flags(struct lustre_msg *msg, int flags)
}
EXPORT_SYMBOL(lustre_msg_set_flags);
-void lustre_msg_clear_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_clear_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -870,7 +868,7 @@ __u32 lustre_msg_get_op_flags(struct lustre_msg *msg)
}
EXPORT_SYMBOL(lustre_msg_get_op_flags);
-void lustre_msg_add_op_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_add_op_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -924,7 +922,7 @@ __u32 lustre_msg_get_type(struct lustre_msg *msg)
}
EXPORT_SYMBOL(lustre_msg_get_type);
-void lustre_msg_add_version(struct lustre_msg *msg, int version)
+void lustre_msg_add_version(struct lustre_msg *msg, u32 version)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 092/124] staging: lustre: ptlrpc: fix comparison between signed and unsigned
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change return type and size argiments of lustre_msg_hdr_size(),
lustre_msg_buf{len,count}() and req_capsule_*_size() to __u32.
Change type of req_format->rf_idx and req_format->rf_fields.nr
to size_t. Also return zero for incorrect message magic instead
of -EINVAL. This will be more robust because of few of them after
LASSERTF(0, "...") and will not be returned. In the rest places
it return zero size instead of huge number after implicit
unsigned conversion.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12475
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_net.h | 34 ++++----
.../lustre/lustre/include/lustre_req_layout.h | 20 +++---
drivers/staging/lustre/lustre/ldlm/ldlm_request.c | 4 +-
drivers/staging/lustre/lustre/ptlrpc/layout.c | 83 ++++++++++----------
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 62 +++++++--------
5 files changed, 101 insertions(+), 102 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 531c628..21e2267 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -2616,9 +2616,9 @@ int ptlrpc_reconnect_import(struct obd_import *imp);
* @{
*/
int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
- int index);
+ u32 index);
void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
- int index);
+ u32 index);
int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len);
int ptlrpc_unpack_req_msg(struct ptlrpc_request *req, int len);
@@ -2637,27 +2637,27 @@ int lustre_shrink_msg(struct lustre_msg *msg, int segment,
unsigned int newlen, int move_data);
void lustre_free_reply_state(struct ptlrpc_reply_state *rs);
int __lustre_unpack_msg(struct lustre_msg *m, int len);
-int lustre_msg_hdr_size(__u32 magic, int count);
-int lustre_msg_size(__u32 magic, int count, __u32 *lengths);
-int lustre_msg_size_v2(int count, __u32 *lengths);
-int lustre_packed_msg_size(struct lustre_msg *msg);
-int lustre_msg_early_size(void);
-void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size);
-void *lustre_msg_buf(struct lustre_msg *m, int n, int minlen);
-int lustre_msg_buflen(struct lustre_msg *m, int n);
-int lustre_msg_bufcount(struct lustre_msg *m);
-char *lustre_msg_string(struct lustre_msg *m, int n, int max_len);
+u32 lustre_msg_hdr_size(__u32 magic, u32 count);
+u32 lustre_msg_size(__u32 magic, int count, __u32 *lengths);
+u32 lustre_msg_size_v2(int count, __u32 *lengths);
+u32 lustre_packed_msg_size(struct lustre_msg *msg);
+u32 lustre_msg_early_size(void);
+void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, u32 n, u32 min_size);
+void *lustre_msg_buf(struct lustre_msg *m, u32 n, u32 minlen);
+u32 lustre_msg_buflen(struct lustre_msg *m, u32 n);
+u32 lustre_msg_bufcount(struct lustre_msg *m);
+char *lustre_msg_string(struct lustre_msg *m, u32 n, u32 max_len);
__u32 lustre_msghdr_get_flags(struct lustre_msg *msg);
void lustre_msghdr_set_flags(struct lustre_msg *msg, __u32 flags);
__u32 lustre_msg_get_flags(struct lustre_msg *msg);
-void lustre_msg_add_flags(struct lustre_msg *msg, int flags);
-void lustre_msg_set_flags(struct lustre_msg *msg, int flags);
-void lustre_msg_clear_flags(struct lustre_msg *msg, int flags);
+void lustre_msg_add_flags(struct lustre_msg *msg, u32 flags);
+void lustre_msg_set_flags(struct lustre_msg *msg, u32 flags);
+void lustre_msg_clear_flags(struct lustre_msg *msg, u32 flags);
__u32 lustre_msg_get_op_flags(struct lustre_msg *msg);
-void lustre_msg_add_op_flags(struct lustre_msg *msg, int flags);
+void lustre_msg_add_op_flags(struct lustre_msg *msg, u32 flags);
struct lustre_handle *lustre_msg_get_handle(struct lustre_msg *msg);
__u32 lustre_msg_get_type(struct lustre_msg *msg);
-void lustre_msg_add_version(struct lustre_msg *msg, int version);
+void lustre_msg_add_version(struct lustre_msg *msg, u32 version);
__u32 lustre_msg_get_opc(struct lustre_msg *msg);
__u64 lustre_msg_get_last_committed(struct lustre_msg *msg);
__u64 *lustre_msg_get_versions(struct lustre_msg *msg);
diff --git a/drivers/staging/lustre/lustre/include/lustre_req_layout.h b/drivers/staging/lustre/lustre/include/lustre_req_layout.h
index ca0e683..a13558e 100644
--- a/drivers/staging/lustre/lustre/include/lustre_req_layout.h
+++ b/drivers/staging/lustre/lustre/include/lustre_req_layout.h
@@ -76,7 +76,8 @@ void req_capsule_init(struct req_capsule *pill, struct ptlrpc_request *req,
void req_capsule_fini(struct req_capsule *pill);
void req_capsule_set(struct req_capsule *pill, const struct req_format *fmt);
-int req_capsule_filled_sizes(struct req_capsule *pill, enum req_location loc);
+size_t req_capsule_filled_sizes(struct req_capsule *pill,
+ enum req_location loc);
int req_capsule_server_pack(struct req_capsule *pill);
void *req_capsule_client_get(struct req_capsule *pill,
@@ -86,27 +87,27 @@ void *req_capsule_client_swab_get(struct req_capsule *pill,
void *swabber);
void *req_capsule_client_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len);
+ u32 len);
void *req_capsule_server_get(struct req_capsule *pill,
const struct req_msg_field *field);
void *req_capsule_server_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len);
+ u32 len);
void *req_capsule_server_swab_get(struct req_capsule *pill,
const struct req_msg_field *field,
void *swabber);
void *req_capsule_server_sized_swab_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len, void *swabber);
+ u32 len, void *swabber);
void req_capsule_set_size(struct req_capsule *pill,
const struct req_msg_field *field,
- enum req_location loc, int size);
-int req_capsule_get_size(const struct req_capsule *pill,
+ enum req_location loc, u32 size);
+u32 req_capsule_get_size(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc);
-int req_capsule_msg_size(struct req_capsule *pill, enum req_location loc);
-int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
+u32 req_capsule_msg_size(struct req_capsule *pill, enum req_location loc);
+u32 req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
enum req_location loc);
void req_capsule_extend(struct req_capsule *pill, const struct req_format *fmt);
@@ -115,8 +116,7 @@ int req_capsule_has_field(const struct req_capsule *pill,
enum req_location loc);
void req_capsule_shrink(struct req_capsule *pill,
const struct req_msg_field *field,
- unsigned int newlen,
- enum req_location loc);
+ u32 newlen, enum req_location loc);
int req_layout_init(void);
void req_layout_fini(void);
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
index 582de52..35ba6f1 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_request.c
@@ -561,7 +561,7 @@ static inline int ldlm_capsule_handles_avail(struct req_capsule *pill,
enum req_location loc,
int off)
{
- int size = req_capsule_msg_size(pill, loc);
+ u32 size = req_capsule_msg_size(pill, loc);
return ldlm_req_handles_avail(size, off);
}
@@ -570,7 +570,7 @@ static inline int ldlm_format_handles_avail(struct obd_import *imp,
const struct req_format *fmt,
enum req_location loc, int off)
{
- int size = req_capsule_fmt_size(imp->imp_msg_magic, fmt, loc);
+ u32 size = req_capsule_fmt_size(imp->imp_msg_magic, fmt, loc);
return ldlm_req_handles_avail(size, off);
}
diff --git a/drivers/staging/lustre/lustre/ptlrpc/layout.c b/drivers/staging/lustre/lustre/ptlrpc/layout.c
index 101ac87..839ef3e 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/layout.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/layout.c
@@ -1113,9 +1113,9 @@ EXPORT_SYMBOL(RMF_SWAP_LAYOUTS);
struct req_format {
const char *rf_name;
- int rf_idx;
+ size_t rf_idx;
struct {
- int nr;
+ size_t nr;
const struct req_msg_field **d;
} rf_fields[RCL_NR];
};
@@ -1555,9 +1555,9 @@ EXPORT_SYMBOL(RQF_OST_GET_INFO_FIEMAP);
*/
int req_layout_init(void)
{
- int i;
- int j;
- int k;
+ size_t i;
+ size_t j;
+ size_t k;
struct req_format *rf = NULL;
for (i = 0; i < ARRAY_SIZE(req_formats); ++i) {
@@ -1599,7 +1599,7 @@ EXPORT_SYMBOL(req_layout_fini);
*/
static void req_capsule_init_area(struct req_capsule *pill)
{
- int i;
+ size_t i;
for (i = 0; i < ARRAY_SIZE(pill->rc_area[RCL_CLIENT]); i++) {
pill->rc_area[RCL_CLIENT][i] = -1;
@@ -1650,8 +1650,7 @@ EXPORT_SYMBOL(req_capsule_fini);
static int __req_format_is_sane(const struct req_format *fmt)
{
- return
- 0 <= fmt->rf_idx && fmt->rf_idx < ARRAY_SIZE(req_formats) &&
+ return fmt->rf_idx < ARRAY_SIZE(req_formats) &&
req_formats[fmt->rf_idx] == fmt;
}
@@ -1685,11 +1684,11 @@ EXPORT_SYMBOL(req_capsule_set);
* variable-sized fields. The field sizes come from the declared \a rmf_size
* field of a \a pill's \a rc_fmt's RMF's.
*/
-int req_capsule_filled_sizes(struct req_capsule *pill,
- enum req_location loc)
+size_t req_capsule_filled_sizes(struct req_capsule *pill,
+ enum req_location loc)
{
const struct req_format *fmt = pill->rc_fmt;
- int i;
+ size_t i;
for (i = 0; i < fmt->rf_fields[loc].nr; ++i) {
if (pill->rc_area[loc][i] == -1) {
@@ -1744,11 +1743,11 @@ EXPORT_SYMBOL(req_capsule_server_pack);
* Returns the PTLRPC request or reply (\a loc) buffer offset of a \a pill
* corresponding to the given RMF (\a field).
*/
-static int __req_capsule_offset(const struct req_capsule *pill,
+static u32 __req_capsule_offset(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc)
{
- int offset;
+ u32 offset;
offset = field->rmf_offset[pill->rc_fmt->rf_idx][loc];
LASSERTF(offset > 0, "%s:%s, off=%d, loc=%d\n", pill->rc_fmt->rf_name,
@@ -1852,10 +1851,10 @@ static void *__req_capsule_get(struct req_capsule *pill,
const struct req_format *fmt;
struct lustre_msg *msg;
void *value;
- int len;
- int offset;
+ u32 len;
+ u32 offset;
- void *(*getter)(struct lustre_msg *m, int n, int minlen);
+ void *(*getter)(struct lustre_msg *m, u32 n, u32 minlen);
static const char *rcl_names[RCL_NR] = {
[RCL_CLIENT] = "client",
@@ -1882,20 +1881,20 @@ static void *__req_capsule_get(struct req_capsule *pill,
*/
len = lustre_msg_buflen(msg, offset);
if ((len % field->rmf_size) != 0) {
- CERROR("%s: array field size mismatch %d modulo %d != 0 (%d)\n",
+ CERROR("%s: array field size mismatch %d modulo %u != 0 (%d)\n",
field->rmf_name, len, field->rmf_size, loc);
return NULL;
}
} else if (pill->rc_area[loc][offset] != -1) {
len = pill->rc_area[loc][offset];
} else {
- len = max(field->rmf_size, 0);
+ len = max_t(typeof(field->rmf_size), field->rmf_size, 0);
}
value = getter(msg, offset, len);
if (!value) {
DEBUG_REQ(D_ERROR, pill->rc_req,
- "Wrong buffer for field `%s' (%d of %d) in format `%s': %d vs. %d (%s)\n",
+ "Wrong buffer for field `%s' (%u of %u) in format `%s': %u vs. %u (%s)\n",
field->rmf_name, offset, lustre_msg_bufcount(msg),
fmt->rf_name, lustre_msg_buflen(msg, offset), len,
rcl_names[loc]);
@@ -1941,7 +1940,7 @@ EXPORT_SYMBOL(req_capsule_client_swab_get);
*/
void *req_capsule_client_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len)
+ u32 len)
{
req_capsule_set_size(pill, field, RCL_CLIENT, len);
return __req_capsule_get(pill, field, RCL_CLIENT, NULL, 0);
@@ -1982,7 +1981,7 @@ EXPORT_SYMBOL(req_capsule_server_swab_get);
*/
void *req_capsule_server_sized_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len)
+ u32 len)
{
req_capsule_set_size(pill, field, RCL_SERVER, len);
return __req_capsule_get(pill, field, RCL_SERVER, NULL, 0);
@@ -1991,7 +1990,7 @@ EXPORT_SYMBOL(req_capsule_server_sized_get);
void *req_capsule_server_sized_swab_get(struct req_capsule *pill,
const struct req_msg_field *field,
- int len, void *swabber)
+ u32 len, void *swabber)
{
req_capsule_set_size(pill, field, RCL_SERVER, len);
return __req_capsule_get(pill, field, RCL_SERVER, swabber, 0);
@@ -2007,23 +2006,25 @@ EXPORT_SYMBOL(req_capsule_server_sized_swab_get);
*/
void req_capsule_set_size(struct req_capsule *pill,
const struct req_msg_field *field,
- enum req_location loc, int size)
+ enum req_location loc, u32 size)
{
LASSERT(loc == RCL_SERVER || loc == RCL_CLIENT);
- if ((size != field->rmf_size) &&
+ if ((size != (u32)field->rmf_size) &&
(field->rmf_size != -1) &&
!(field->rmf_flags & RMF_F_NO_SIZE_CHECK) &&
(size > 0)) {
+ u32 rmf_size = (u32)field->rmf_size;
+
if ((field->rmf_flags & RMF_F_STRUCT_ARRAY) &&
- (size % field->rmf_size != 0)) {
- CERROR("%s: array field size mismatch %d %% %d != 0 (%d)\n",
- field->rmf_name, size, field->rmf_size, loc);
+ (size % rmf_size != 0)) {
+ CERROR("%s: array field size mismatch %u %% %u != 0 (%d)\n",
+ field->rmf_name, size, rmf_size, loc);
LBUG();
} else if (!(field->rmf_flags & RMF_F_STRUCT_ARRAY) &&
- size < field->rmf_size) {
- CERROR("%s: field size mismatch %d != %d (%d)\n",
- field->rmf_name, size, field->rmf_size, loc);
+ size < rmf_size) {
+ CERROR("%s: field size mismatch %u != %u (%d)\n",
+ field->rmf_name, size, rmf_size, loc);
LBUG();
}
}
@@ -2040,7 +2041,7 @@ EXPORT_SYMBOL(req_capsule_set_size);
* actually sets the size in pill.rc_area[loc][offset], but this function
* returns the message buflen[offset], maybe we should use another name.
*/
-int req_capsule_get_size(const struct req_capsule *pill,
+u32 req_capsule_get_size(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc)
{
@@ -2058,7 +2059,7 @@ EXPORT_SYMBOL(req_capsule_get_size);
*
* See also req_capsule_set_size().
*/
-int req_capsule_msg_size(struct req_capsule *pill, enum req_location loc)
+u32 req_capsule_msg_size(struct req_capsule *pill, enum req_location loc)
{
return lustre_msg_size(pill->rc_req->rq_import->imp_msg_magic,
pill->rc_fmt->rf_fields[loc].nr,
@@ -2073,10 +2074,11 @@ int req_capsule_msg_size(struct req_capsule *pill, enum req_location loc)
* This function should not be used for formats which contain variable size
* fields.
*/
-int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
+u32 req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
enum req_location loc)
{
- int size, i = 0;
+ size_t i = 0;
+ u32 size;
/*
* This function should probably LASSERT() that fmt has no fields with
@@ -2086,7 +2088,7 @@ int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
* we do.
*/
size = lustre_msg_hdr_size(magic, fmt->rf_fields[loc].nr);
- if (size < 0)
+ if (!size)
return size;
for (; i < fmt->rf_fields[loc].nr; ++i)
@@ -2118,7 +2120,7 @@ int req_capsule_fmt_size(__u32 magic, const struct req_format *fmt,
void req_capsule_extend(struct req_capsule *pill, const struct req_format *fmt)
{
int i;
- int j;
+ size_t j;
const struct req_format *old;
@@ -2176,7 +2178,7 @@ static int req_capsule_field_present(const struct req_capsule *pill,
const struct req_msg_field *field,
enum req_location loc)
{
- int offset;
+ u32 offset;
LASSERT(loc == RCL_SERVER || loc == RCL_CLIENT);
LASSERT(req_capsule_has_field(pill, field, loc));
@@ -2193,12 +2195,11 @@ static int req_capsule_field_present(const struct req_capsule *pill,
*/
void req_capsule_shrink(struct req_capsule *pill,
const struct req_msg_field *field,
- unsigned int newlen,
- enum req_location loc)
+ u32 newlen, enum req_location loc)
{
const struct req_format *fmt;
struct lustre_msg *msg;
- int len;
+ u32 len;
int offset;
fmt = pill->rc_fmt;
@@ -2211,7 +2212,7 @@ void req_capsule_shrink(struct req_capsule *pill,
msg = __req_msg(pill, loc);
len = lustre_msg_buflen(msg, offset);
- LASSERTF(newlen <= len, "%s:%s, oldlen=%d, newlen=%d\n",
+ LASSERTF(newlen <= len, "%s:%s, oldlen=%u, newlen=%u\n",
fmt->rf_name, field->rmf_name, len, newlen);
if (loc == RCL_CLIENT)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index 87027c5..a1ea0c2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -50,26 +50,26 @@
#include "ptlrpc_internal.h"
-static inline int lustre_msg_hdr_size_v2(int count)
+static inline u32 lustre_msg_hdr_size_v2(u32 count)
{
return cfs_size_round(offsetof(struct lustre_msg_v2,
lm_buflens[count]));
}
-int lustre_msg_hdr_size(__u32 magic, int count)
+u32 lustre_msg_hdr_size(__u32 magic, u32 count)
{
switch (magic) {
case LUSTRE_MSG_MAGIC_V2:
return lustre_msg_hdr_size_v2(count);
default:
LASSERTF(0, "incorrect message magic: %08x\n", magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_hdr_size);
void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
- int index)
+ u32 index)
{
if (inout)
lustre_set_req_swabbed(req, index);
@@ -79,7 +79,7 @@ void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
EXPORT_SYMBOL(ptlrpc_buf_set_swabbed);
int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
- int index)
+ u32 index)
{
if (inout)
return (ptlrpc_req_need_swab(req) &&
@@ -91,9 +91,9 @@ int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
EXPORT_SYMBOL(ptlrpc_buf_need_swab);
/* early reply size */
-int lustre_msg_early_size(void)
+u32 lustre_msg_early_size(void)
{
- static int size;
+ static u32 size;
if (!size) {
/* Always reply old ptlrpc_body_v2 to keep interoperability
@@ -111,9 +111,9 @@ int lustre_msg_early_size(void)
}
EXPORT_SYMBOL(lustre_msg_early_size);
-int lustre_msg_size_v2(int count, __u32 *lengths)
+u32 lustre_msg_size_v2(int count, __u32 *lengths)
{
- int size;
+ u32 size;
int i;
size = lustre_msg_hdr_size_v2(count);
@@ -131,7 +131,7 @@ EXPORT_SYMBOL(lustre_msg_size_v2);
* target then the first buffer will be stripped because the ptlrpc
* data is part of the lustre_msg_v1 header. b=14043
*/
-int lustre_msg_size(__u32 magic, int count, __u32 *lens)
+u32 lustre_msg_size(__u32 magic, int count, __u32 *lens)
{
__u32 size[] = { sizeof(struct ptlrpc_body) };
@@ -148,7 +148,7 @@ int lustre_msg_size(__u32 magic, int count, __u32 *lens)
return lustre_msg_size_v2(count, lens);
default:
LASSERTF(0, "incorrect message magic: %08x\n", magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_size);
@@ -156,7 +156,7 @@ EXPORT_SYMBOL(lustre_msg_size);
/* This is used to determine the size of a buffer that was already packed
* and will correctly handle the different message formats.
*/
-int lustre_packed_msg_size(struct lustre_msg *msg)
+u32 lustre_packed_msg_size(struct lustre_msg *msg)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
@@ -378,11 +378,9 @@ int lustre_pack_reply(struct ptlrpc_request *req, int count, __u32 *lens,
}
EXPORT_SYMBOL(lustre_pack_reply);
-void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size)
+void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, u32 n, u32 min_size)
{
- int i, offset, buflen, bufcount;
-
- LASSERT(n >= 0);
+ u32 i, offset, buflen, bufcount;
bufcount = m->lm_bufcount;
if (unlikely(n >= bufcount)) {
@@ -406,7 +404,7 @@ void *lustre_msg_buf_v2(struct lustre_msg_v2 *m, int n, int min_size)
return (char *)m + offset;
}
-void *lustre_msg_buf(struct lustre_msg *m, int n, int min_size)
+void *lustre_msg_buf(struct lustre_msg *m, u32 n, u32 min_size)
{
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
@@ -419,7 +417,7 @@ void *lustre_msg_buf(struct lustre_msg *m, int n, int min_size)
}
EXPORT_SYMBOL(lustre_msg_buf);
-static int lustre_shrink_msg_v2(struct lustre_msg_v2 *msg, int segment,
+static int lustre_shrink_msg_v2(struct lustre_msg_v2 *msg, u32 segment,
unsigned int newlen, int move_data)
{
char *tail = NULL, *newpos;
@@ -647,7 +645,7 @@ int lustre_unpack_rep_ptlrpc_body(struct ptlrpc_request *req, int offset)
}
}
-static inline int lustre_msg_buflen_v2(struct lustre_msg_v2 *m, int n)
+static inline u32 lustre_msg_buflen_v2(struct lustre_msg_v2 *m, u32 n)
{
if (n >= m->lm_bufcount)
return 0;
@@ -662,14 +660,14 @@ static inline int lustre_msg_buflen_v2(struct lustre_msg_v2 *m, int n)
*
* returns zero for non-existent message indices
*/
-int lustre_msg_buflen(struct lustre_msg *m, int n)
+u32 lustre_msg_buflen(struct lustre_msg *m, u32 n)
{
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
return lustre_msg_buflen_v2(m, n);
default:
CERROR("incorrect message magic: %08x\n", m->lm_magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_buflen);
@@ -677,23 +675,23 @@ EXPORT_SYMBOL(lustre_msg_buflen);
/* NB return the bufcount for lustre_msg_v2 format, so if message is packed
* in V1 format, the result is one bigger. (add struct ptlrpc_body).
*/
-int lustre_msg_bufcount(struct lustre_msg *m)
+u32 lustre_msg_bufcount(struct lustre_msg *m)
{
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
return m->lm_bufcount;
default:
CERROR("incorrect message magic: %08x\n", m->lm_magic);
- return -EINVAL;
+ return 0;
}
}
EXPORT_SYMBOL(lustre_msg_bufcount);
-char *lustre_msg_string(struct lustre_msg *m, int index, int max_len)
+char *lustre_msg_string(struct lustre_msg *m, u32 index, u32 max_len)
{
/* max_len == 0 means the string should fill the buffer */
char *str;
- int slen, blen;
+ u32 slen, blen;
switch (m->lm_magic) {
case LUSTRE_MSG_MAGIC_V2:
@@ -734,8 +732,8 @@ char *lustre_msg_string(struct lustre_msg *m, int index, int max_len)
EXPORT_SYMBOL(lustre_msg_string);
/* Wrap up the normal fixed length cases */
-static inline void *__lustre_swab_buf(struct lustre_msg *msg, int index,
- int min_size, void *swabber)
+static inline void *__lustre_swab_buf(struct lustre_msg *msg, u32 index,
+ u32 min_size, void *swabber)
{
void *ptr = NULL;
@@ -804,7 +802,7 @@ __u32 lustre_msg_get_flags(struct lustre_msg *msg)
}
EXPORT_SYMBOL(lustre_msg_get_flags);
-void lustre_msg_add_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_add_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -820,7 +818,7 @@ void lustre_msg_add_flags(struct lustre_msg *msg, int flags)
}
EXPORT_SYMBOL(lustre_msg_add_flags);
-void lustre_msg_set_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_set_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -836,7 +834,7 @@ void lustre_msg_set_flags(struct lustre_msg *msg, int flags)
}
EXPORT_SYMBOL(lustre_msg_set_flags);
-void lustre_msg_clear_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_clear_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -870,7 +868,7 @@ __u32 lustre_msg_get_op_flags(struct lustre_msg *msg)
}
EXPORT_SYMBOL(lustre_msg_get_op_flags);
-void lustre_msg_add_op_flags(struct lustre_msg *msg, int flags)
+void lustre_msg_add_op_flags(struct lustre_msg *msg, u32 flags)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
@@ -924,7 +922,7 @@ __u32 lustre_msg_get_type(struct lustre_msg *msg)
}
EXPORT_SYMBOL(lustre_msg_get_type);
-void lustre_msg_add_version(struct lustre_msg *msg, int version)
+void lustre_msg_add_version(struct lustre_msg *msg, u32 version)
{
switch (msg->lm_magic) {
case LUSTRE_MSG_MAGIC_V2: {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 093/124] staging: lustre: ldlm: move LDLM_GID_ANY to lustre_dlm.h
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
lustre_idl.h only includes wire data; lustre_dlm.h is the
right place for LDLM_GID_ANY.
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6028
Reviewed-on: http://review.whamcloud.com/13074
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 2 --
drivers/staging/lustre/lustre/include/lustre_dlm.h | 5 +++++
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 445b813..acb2806 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -2716,8 +2716,6 @@ struct ldlm_extent {
__u64 gid;
};
-#define LDLM_GID_ANY ((__u64)-1)
-
static inline int ldlm_extent_overlap(const struct ldlm_extent *ex1,
const struct ldlm_extent *ex2)
{
diff --git a/drivers/staging/lustre/lustre/include/lustre_dlm.h b/drivers/staging/lustre/lustre/include/lustre_dlm.h
index 1ec4231..d035344 100644
--- a/drivers/staging/lustre/lustre/include/lustre_dlm.h
+++ b/drivers/staging/lustre/lustre/include/lustre_dlm.h
@@ -573,6 +573,11 @@ enum lvb_type {
};
/**
+ * LDLM_GID_ANY is used to match any group id in ldlm_lock_match().
+ */
+#define LDLM_GID_ANY ((__u64)-1)
+
+/**
* LDLM lock structure
*
* Represents a single LDLM lock and its state in memory. Each lock is
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 093/124] staging: lustre: ldlm: move LDLM_GID_ANY to lustre_dlm.h
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Jinshan Xiong, James Simmons
From: Jinshan Xiong <jinshan.xiong@intel.com>
lustre_idl.h only includes wire data; lustre_dlm.h is the
right place for LDLM_GID_ANY.
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6028
Reviewed-on: http://review.whamcloud.com/13074
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 2 --
drivers/staging/lustre/lustre/include/lustre_dlm.h | 5 +++++
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index 445b813..acb2806 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -2716,8 +2716,6 @@ struct ldlm_extent {
__u64 gid;
};
-#define LDLM_GID_ANY ((__u64)-1)
-
static inline int ldlm_extent_overlap(const struct ldlm_extent *ex1,
const struct ldlm_extent *ex2)
{
diff --git a/drivers/staging/lustre/lustre/include/lustre_dlm.h b/drivers/staging/lustre/lustre/include/lustre_dlm.h
index 1ec4231..d035344 100644
--- a/drivers/staging/lustre/lustre/include/lustre_dlm.h
+++ b/drivers/staging/lustre/lustre/include/lustre_dlm.h
@@ -573,6 +573,11 @@ enum lvb_type {
};
/**
+ * LDLM_GID_ANY is used to match any group id in ldlm_lock_match().
+ */
+#define LDLM_GID_ANY ((__u64)-1)
+
+/**
* LDLM lock structure
*
* Represents a single LDLM lock and its state in memory. Each lock is
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 094/124] staging: lustre: lov: flatten struct lov_stripe_md
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Flatten out the lsm_wire struct from the middle of struct
lov_stripe_md and remove the member name macros.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12581
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 30 +++++--------------
.../staging/lustre/lustre/lov/lov_cl_internal.h | 4 +-
2 files changed, 10 insertions(+), 24 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index d1c4ccc..8e426b8 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -88,31 +88,17 @@ struct lov_stripe_md {
/* maximum possible file size, might change as OSTs status changes,
* e.g. disconnected, deactivated
*/
- __u64 lsm_maxbytes;
- struct {
- /* Public members. */
- struct ost_id lw_object_oi; /* lov object id/seq */
-
- /* LOV-private members start here -- only for use in lov/. */
- __u32 lw_magic;
- __u32 lw_stripe_size; /* size of the stripe */
- __u32 lw_pattern; /* striping pattern (RAID0, RAID1) */
- __u16 lw_stripe_count; /* number of objects being striped over */
- __u16 lw_layout_gen; /* generation of the layout */
- char lw_pool_name[LOV_MAXPOOLNAME + 1]; /* pool name */
- } lsm_wire;
-
+ __u64 lsm_maxbytes;
+ struct ost_id lsm_oi;
+ __u32 lsm_magic;
+ __u32 lsm_stripe_size;
+ __u32 lsm_pattern; /* striping pattern (RAID0, RAID1) */
+ __u16 lsm_stripe_count;
+ __u16 lsm_layout_gen;
+ char lsm_pool_name[LOV_MAXPOOLNAME + 1];
struct lov_oinfo *lsm_oinfo[0];
};
-#define lsm_oi lsm_wire.lw_object_oi
-#define lsm_magic lsm_wire.lw_magic
-#define lsm_layout_gen lsm_wire.lw_layout_gen
-#define lsm_stripe_size lsm_wire.lw_stripe_size
-#define lsm_pattern lsm_wire.lw_pattern
-#define lsm_stripe_count lsm_wire.lw_stripe_count
-#define lsm_pool_name lsm_wire.lw_pool_name
-
static inline bool lsm_is_released(struct lov_stripe_md *lsm)
{
return !!(lsm->lsm_pattern & LOV_PATTERN_F_RELEASED);
diff --git a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
index ec40cf2..4d2b7d3 100644
--- a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
@@ -744,8 +744,8 @@ static inline struct lov_thread_info *lov_env_info(const struct lu_env *env)
static inline struct lov_layout_raid0 *lov_r0(struct lov_object *lov)
{
LASSERT(lov->lo_type == LLT_RAID0);
- LASSERT(lov->lo_lsm->lsm_wire.lw_magic == LOV_MAGIC ||
- lov->lo_lsm->lsm_wire.lw_magic == LOV_MAGIC_V3);
+ LASSERT(lov->lo_lsm->lsm_magic == LOV_MAGIC ||
+ lov->lo_lsm->lsm_magic == LOV_MAGIC_V3);
return &lov->u.raid0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 094/124] staging: lustre: lov: flatten struct lov_stripe_md
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Flatten out the lsm_wire struct from the middle of struct
lov_stripe_md and remove the member name macros.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-on: http://review.whamcloud.com/12581
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 30 +++++--------------
.../staging/lustre/lustre/lov/lov_cl_internal.h | 4 +-
2 files changed, 10 insertions(+), 24 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index d1c4ccc..8e426b8 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -88,31 +88,17 @@ struct lov_stripe_md {
/* maximum possible file size, might change as OSTs status changes,
* e.g. disconnected, deactivated
*/
- __u64 lsm_maxbytes;
- struct {
- /* Public members. */
- struct ost_id lw_object_oi; /* lov object id/seq */
-
- /* LOV-private members start here -- only for use in lov/. */
- __u32 lw_magic;
- __u32 lw_stripe_size; /* size of the stripe */
- __u32 lw_pattern; /* striping pattern (RAID0, RAID1) */
- __u16 lw_stripe_count; /* number of objects being striped over */
- __u16 lw_layout_gen; /* generation of the layout */
- char lw_pool_name[LOV_MAXPOOLNAME + 1]; /* pool name */
- } lsm_wire;
-
+ __u64 lsm_maxbytes;
+ struct ost_id lsm_oi;
+ __u32 lsm_magic;
+ __u32 lsm_stripe_size;
+ __u32 lsm_pattern; /* striping pattern (RAID0, RAID1) */
+ __u16 lsm_stripe_count;
+ __u16 lsm_layout_gen;
+ char lsm_pool_name[LOV_MAXPOOLNAME + 1];
struct lov_oinfo *lsm_oinfo[0];
};
-#define lsm_oi lsm_wire.lw_object_oi
-#define lsm_magic lsm_wire.lw_magic
-#define lsm_layout_gen lsm_wire.lw_layout_gen
-#define lsm_stripe_size lsm_wire.lw_stripe_size
-#define lsm_pattern lsm_wire.lw_pattern
-#define lsm_stripe_count lsm_wire.lw_stripe_count
-#define lsm_pool_name lsm_wire.lw_pool_name
-
static inline bool lsm_is_released(struct lov_stripe_md *lsm)
{
return !!(lsm->lsm_pattern & LOV_PATTERN_F_RELEASED);
diff --git a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
index ec40cf2..4d2b7d3 100644
--- a/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
+++ b/drivers/staging/lustre/lustre/lov/lov_cl_internal.h
@@ -744,8 +744,8 @@ static inline struct lov_thread_info *lov_env_info(const struct lu_env *env)
static inline struct lov_layout_raid0 *lov_r0(struct lov_object *lov)
{
LASSERT(lov->lo_type == LLT_RAID0);
- LASSERT(lov->lo_lsm->lsm_wire.lw_magic == LOV_MAGIC ||
- lov->lo_lsm->lsm_wire.lw_magic == LOV_MAGIC_V3);
+ LASSERT(lov->lo_lsm->lsm_magic == LOV_MAGIC ||
+ lov->lo_lsm->lsm_magic == LOV_MAGIC_V3);
return &lov->u.raid0;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 095/124] staging: lustre: ptlrpc: fix race between connect vs resend
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Boyko, James Simmons
From: Alexander Boyko <alexander_boyko@xyratex.com>
Buggy code at ptlrpc_connect_interpret()
finish:
rc = ptlrpc_import_recovery_state_machine(imp);
...
Set import connection flags
When import has FULL state ptlrpc_import_recovery_state_machine()
wakeup all waiters on import and all delayed request, which was
resented. And it could happened that request was send without
updated flags and AT is disabled. If such request is in progress
on the server, server drop the new instance, and could do early reply
for it. But this early reply confuse client, cause it wait real
reply(no AT for this request). Client try to touch buffer outside
reply and got EPROTO error.
The same bug existed for initital connect too. Import became FULL
before import connection flags was set.
Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5528
Xyratex-bug-id: MRP-2034
Reviewed-on: http://review.whamcloud.com/11723
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/import.c | 307 +++++++++++++------------
drivers/staging/lustre/lustre/ptlrpc/niobuf.c | 28 ++-
2 files changed, 179 insertions(+), 156 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 17f3fec..2e107e8 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -752,6 +752,153 @@ static int ptlrpc_busy_reconnect(int rc)
return (rc == -EBUSY) || (rc == -EAGAIN);
}
+static int ptlrpc_connect_set_flags(struct obd_import *imp,
+ struct obd_connect_data *ocd,
+ u64 old_connect_flags,
+ struct obd_export *exp, int init_connect)
+{
+ struct client_obd *cli = &imp->imp_obd->u.cli;
+ static bool warned;
+
+ if ((imp->imp_connect_flags_orig & OBD_CONNECT_IBITS) &&
+ !(ocd->ocd_connect_flags & OBD_CONNECT_IBITS)) {
+ LCONSOLE_WARN("%s: MDS %s does not support ibits lock, either very old or invalid: requested %#llx, replied %#llx\n",
+ imp->imp_obd->obd_name,
+ imp->imp_connection->c_remote_uuid.uuid,
+ imp->imp_connect_flags_orig,
+ ocd->ocd_connect_flags);
+ return -EPROTO;
+ }
+
+ spin_lock(&imp->imp_lock);
+ list_del(&imp->imp_conn_current->oic_item);
+ list_add(&imp->imp_conn_current->oic_item, &imp->imp_conn_list);
+ imp->imp_last_success_conn = imp->imp_conn_current->oic_last_attempt;
+
+ spin_unlock(&imp->imp_lock);
+
+ if (!warned && (ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
+ (ocd->ocd_version > LUSTRE_VERSION_CODE +
+ LUSTRE_VERSION_OFFSET_WARN ||
+ ocd->ocd_version < LUSTRE_VERSION_CODE -
+ LUSTRE_VERSION_OFFSET_WARN)) {
+ /*
+ * Sigh, some compilers do not like #ifdef in the middle
+ * of macro arguments
+ */
+ const char *older = "older than client. Consider upgrading server";
+ const char *newer = "newer than client. Consider recompiling application";
+
+ LCONSOLE_WARN("Server %s version (%d.%d.%d.%d) is much %s (%s)\n",
+ obd2cli_tgt(imp->imp_obd),
+ OBD_OCD_VERSION_MAJOR(ocd->ocd_version),
+ OBD_OCD_VERSION_MINOR(ocd->ocd_version),
+ OBD_OCD_VERSION_PATCH(ocd->ocd_version),
+ OBD_OCD_VERSION_FIX(ocd->ocd_version),
+ ocd->ocd_version > LUSTRE_VERSION_CODE ?
+ newer : older, LUSTRE_VERSION_STRING);
+ warned = true;
+ }
+
+#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
+ /*
+ * Check if server has LU-1252 fix applied to not always swab
+ * the IR MNE entries. Do this only once per connection. This
+ * fixup is version-limited, because we don't want to carry the
+ * OBD_CONNECT_MNE_SWAB flag around forever, just so long as we
+ * need interop with unpatched 2.2 servers. For newer servers,
+ * the client will do MNE swabbing only as needed. LU-1644
+ */
+ if (unlikely((ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
+ !(ocd->ocd_connect_flags & OBD_CONNECT_MNE_SWAB) &&
+ OBD_OCD_VERSION_MAJOR(ocd->ocd_version) == 2 &&
+ OBD_OCD_VERSION_MINOR(ocd->ocd_version) == 2 &&
+ OBD_OCD_VERSION_PATCH(ocd->ocd_version) < 55 &&
+ !strcmp(imp->imp_obd->obd_type->typ_name,
+ LUSTRE_MGC_NAME)))
+ imp->imp_need_mne_swab = 1;
+ else /* clear if server was upgraded since last connect */
+ imp->imp_need_mne_swab = 0;
+#endif
+
+ if (ocd->ocd_connect_flags & OBD_CONNECT_CKSUM) {
+ /*
+ * We sent to the server ocd_cksum_types with bits set
+ * for algorithms we understand. The server masked off
+ * the checksum types it doesn't support
+ */
+ if (!(ocd->ocd_cksum_types & cksum_types_supported_client())) {
+ LCONSOLE_WARN("The negotiation of the checksum algorithm to use with server %s failed (%x/%x), disabling checksums\n",
+ obd2cli_tgt(imp->imp_obd),
+ ocd->ocd_cksum_types,
+ cksum_types_supported_client());
+ cli->cl_checksum = 0;
+ cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
+ } else {
+ cli->cl_supp_cksum_types = ocd->ocd_cksum_types;
+ }
+ } else {
+ /*
+ * The server does not support OBD_CONNECT_CKSUM.
+ * Enforce ADLER for backward compatibility
+ */
+ cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
+ }
+ cli->cl_cksum_type = cksum_type_select(cli->cl_supp_cksum_types);
+
+ if (ocd->ocd_connect_flags & OBD_CONNECT_BRW_SIZE)
+ cli->cl_max_pages_per_rpc =
+ min(ocd->ocd_brw_size >> PAGE_SHIFT,
+ cli->cl_max_pages_per_rpc);
+ else if (imp->imp_connect_op == MDS_CONNECT ||
+ imp->imp_connect_op == MGS_CONNECT)
+ cli->cl_max_pages_per_rpc = 1;
+
+ LASSERT((cli->cl_max_pages_per_rpc <= PTLRPC_MAX_BRW_PAGES) &&
+ (cli->cl_max_pages_per_rpc > 0));
+
+ client_adjust_max_dirty(cli);
+
+ /*
+ * Reset ns_connect_flags only for initial connect. It might be
+ * changed in while using FS and if we reset it in reconnect
+ * this leads to losing user settings done before such as
+ * disable lru_resize, etc.
+ */
+ if (old_connect_flags != exp_connect_flags(exp) || init_connect) {
+ CDEBUG(D_HA, "%s: Resetting ns_connect_flags to server flags: %#llx\n",
+ imp->imp_obd->obd_name, ocd->ocd_connect_flags);
+ imp->imp_obd->obd_namespace->ns_connect_flags =
+ ocd->ocd_connect_flags;
+ imp->imp_obd->obd_namespace->ns_orig_connect_flags =
+ ocd->ocd_connect_flags;
+ }
+
+ if ((ocd->ocd_connect_flags & OBD_CONNECT_AT) &&
+ (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
+ /*
+ * We need a per-message support flag, because
+ * a. we don't know if the incoming connect reply
+ * supports AT or not (in reply_in_callback)
+ * until we unpack it.
+ * b. failovered server means export and flags are gone
+ * (in ptlrpc_send_reply).
+ * Can only be set when we know AT is supported at
+ * both ends
+ */
+ imp->imp_msghdr_flags |= MSGHDR_AT_SUPPORT;
+ else
+ imp->imp_msghdr_flags &= ~MSGHDR_AT_SUPPORT;
+
+ if ((ocd->ocd_connect_flags & OBD_CONNECT_FULL20) &&
+ (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
+ imp->imp_msghdr_flags |= MSGHDR_CKSUM_INCOMPAT18;
+ else
+ imp->imp_msghdr_flags &= ~MSGHDR_CKSUM_INCOMPAT18;
+
+ return 0;
+}
+
/**
* interpret_reply callback for connect RPCs.
* Looks into returned status of connect operation and decides
@@ -764,7 +911,6 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
{
struct ptlrpc_connect_async_args *aa = data;
struct obd_import *imp = request->rq_import;
- struct client_obd *cli = &imp->imp_obd->u.cli;
struct lustre_handle old_hdl;
__u64 old_connect_flags;
int msg_flags;
@@ -844,7 +990,6 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
old_connect_flags = exp_connect_flags(exp);
exp->exp_connect_data = *ocd;
imp->imp_obd->obd_self_export->exp_connect_data = *ocd;
- class_export_put(exp);
/*
* The net statistics after (re-)connect is not valid anymore,
@@ -854,6 +999,13 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
ptlrpc_at_adj_net_latency(request,
lustre_msg_get_service_time(request->rq_repmsg));
+ /* Import flags should be updated before waking import at FULL state */
+ rc = ptlrpc_connect_set_flags(imp, ocd, old_connect_flags, exp,
+ aa->pcaa_initial_connect);
+ class_export_put(exp);
+ if (rc)
+ goto out;
+
obd_import_event(imp->imp_obd, imp, IMP_EVENT_OCD);
if (aa->pcaa_initial_connect) {
@@ -997,150 +1149,13 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
finish:
rc = ptlrpc_import_recovery_state_machine(imp);
- if (rc != 0) {
- if (rc == -ENOTCONN) {
- CDEBUG(D_HA, "evicted/aborted by %s@%s during recovery; invalidating and reconnecting\n",
- obd2cli_tgt(imp->imp_obd),
- imp->imp_connection->c_remote_uuid.uuid);
- ptlrpc_connect_import(imp);
- imp->imp_connect_tried = 1;
- return 0;
- }
- } else {
- static bool warned;
-
- spin_lock(&imp->imp_lock);
- list_del(&imp->imp_conn_current->oic_item);
- list_add(&imp->imp_conn_current->oic_item, &imp->imp_conn_list);
- imp->imp_last_success_conn =
- imp->imp_conn_current->oic_last_attempt;
-
- spin_unlock(&imp->imp_lock);
-
- if ((imp->imp_connect_flags_orig & OBD_CONNECT_IBITS) &&
- !(ocd->ocd_connect_flags & OBD_CONNECT_IBITS)) {
- LCONSOLE_WARN("%s: MDS %s does not support ibits lock, either very old or invalid: requested %llx, replied %llx\n",
- imp->imp_obd->obd_name,
- imp->imp_connection->c_remote_uuid.uuid,
- imp->imp_connect_flags_orig,
- ocd->ocd_connect_flags);
- rc = -EPROTO;
- goto out;
- }
-
- if (!warned && (ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
- (ocd->ocd_version > LUSTRE_VERSION_CODE +
- LUSTRE_VERSION_OFFSET_WARN ||
- ocd->ocd_version < LUSTRE_VERSION_CODE -
- LUSTRE_VERSION_OFFSET_WARN)) {
- /* Sigh, some compilers do not like #ifdef in the middle
- * of macro arguments
- */
- const char *older = "older than client. Consider upgrading server";
- const char *newer = "newer than client. Consider recompiling application";
-
- LCONSOLE_WARN("Server %s version (%d.%d.%d.%d) is much %s (%s)\n",
- obd2cli_tgt(imp->imp_obd),
- OBD_OCD_VERSION_MAJOR(ocd->ocd_version),
- OBD_OCD_VERSION_MINOR(ocd->ocd_version),
- OBD_OCD_VERSION_PATCH(ocd->ocd_version),
- OBD_OCD_VERSION_FIX(ocd->ocd_version),
- ocd->ocd_version > LUSTRE_VERSION_CODE ?
- newer : older, LUSTRE_VERSION_STRING);
- warned = true;
- }
-
-#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
- /* Check if server has LU-1252 fix applied to not always swab
- * the IR MNE entries. Do this only once per connection. This
- * fixup is version-limited, because we don't want to carry the
- * OBD_CONNECT_MNE_SWAB flag around forever, just so long as we
- * need interop with unpatched 2.2 servers. For newer servers,
- * the client will do MNE swabbing only as needed. LU-1644
- */
- if (unlikely((ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
- !(ocd->ocd_connect_flags & OBD_CONNECT_MNE_SWAB) &&
- OBD_OCD_VERSION_MAJOR(ocd->ocd_version) == 2 &&
- OBD_OCD_VERSION_MINOR(ocd->ocd_version) == 2 &&
- OBD_OCD_VERSION_PATCH(ocd->ocd_version) < 55 &&
- strcmp(imp->imp_obd->obd_type->typ_name,
- LUSTRE_MGC_NAME) == 0))
- imp->imp_need_mne_swab = 1;
- else /* clear if server was upgraded since last connect */
- imp->imp_need_mne_swab = 0;
-#endif
-
- if (ocd->ocd_connect_flags & OBD_CONNECT_CKSUM) {
- /* We sent to the server ocd_cksum_types with bits set
- * for algorithms we understand. The server masked off
- * the checksum types it doesn't support
- */
- if ((ocd->ocd_cksum_types &
- cksum_types_supported_client()) == 0) {
- LCONSOLE_WARN("The negotiation of the checksum algorithm to use with server %s failed (%x/%x), disabling checksums\n",
- obd2cli_tgt(imp->imp_obd),
- ocd->ocd_cksum_types,
- cksum_types_supported_client());
- cli->cl_checksum = 0;
- cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
- } else {
- cli->cl_supp_cksum_types = ocd->ocd_cksum_types;
- }
- } else {
- /* The server does not support OBD_CONNECT_CKSUM.
- * Enforce ADLER for backward compatibility
- */
- cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
- }
- cli->cl_cksum_type = cksum_type_select(cli->cl_supp_cksum_types);
-
- if (ocd->ocd_connect_flags & OBD_CONNECT_BRW_SIZE)
- cli->cl_max_pages_per_rpc =
- min(ocd->ocd_brw_size >> PAGE_SHIFT,
- cli->cl_max_pages_per_rpc);
- else if (imp->imp_connect_op == MDS_CONNECT ||
- imp->imp_connect_op == MGS_CONNECT)
- cli->cl_max_pages_per_rpc = 1;
-
- /* Reset ns_connect_flags only for initial connect. It might be
- * changed in while using FS and if we reset it in reconnect
- * this leads to losing user settings done before such as
- * disable lru_resize, etc.
- */
- if (old_connect_flags != exp_connect_flags(exp) ||
- aa->pcaa_initial_connect) {
- CDEBUG(D_HA, "%s: Resetting ns_connect_flags to server flags: %#llx\n",
- imp->imp_obd->obd_name, ocd->ocd_connect_flags);
- imp->imp_obd->obd_namespace->ns_connect_flags =
- ocd->ocd_connect_flags;
- imp->imp_obd->obd_namespace->ns_orig_connect_flags =
- ocd->ocd_connect_flags;
- }
-
- if ((ocd->ocd_connect_flags & OBD_CONNECT_AT) &&
- (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
- /* We need a per-message support flag, because
- * a. we don't know if the incoming connect reply
- * supports AT or not (in reply_in_callback)
- * until we unpack it.
- * b. failovered server means export and flags are gone
- * (in ptlrpc_send_reply).
- * Can only be set when we know AT is supported at
- * both ends
- */
- imp->imp_msghdr_flags |= MSGHDR_AT_SUPPORT;
- else
- imp->imp_msghdr_flags &= ~MSGHDR_AT_SUPPORT;
-
- if ((ocd->ocd_connect_flags & OBD_CONNECT_FULL20) &&
- (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
- imp->imp_msghdr_flags |= MSGHDR_CKSUM_INCOMPAT18;
- else
- imp->imp_msghdr_flags &= ~MSGHDR_CKSUM_INCOMPAT18;
-
- LASSERT((cli->cl_max_pages_per_rpc <= PTLRPC_MAX_BRW_PAGES) &&
- (cli->cl_max_pages_per_rpc > 0));
- client_adjust_max_dirty(cli);
+ if (rc == -ENOTCONN) {
+ CDEBUG(D_HA, "evicted/aborted by %s@%s during recovery; invalidating and reconnecting\n",
+ obd2cli_tgt(imp->imp_obd),
+ imp->imp_connection->c_remote_uuid.uuid);
+ ptlrpc_connect_import(imp);
+ imp->imp_connect_tried = 1;
+ return 0;
}
out:
diff --git a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
index 44f4eae..5bab248 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
@@ -491,7 +491,8 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
struct ptlrpc_connection *connection;
lnet_handle_me_t reply_me_h;
lnet_md_t reply_md;
- struct obd_device *obd = request->rq_import->imp_obd;
+ struct obd_import *imp = request->rq_import;
+ struct obd_device *obd = imp->imp_obd;
if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_DROP_RPC))
return 0;
@@ -504,7 +505,7 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
*/
LASSERT(!request->rq_receiving_reply);
LASSERT(!((lustre_msg_get_flags(request->rq_reqmsg) & MSG_REPLAY) &&
- (request->rq_import->imp_state == LUSTRE_IMP_FULL)));
+ (imp->imp_state == LUSTRE_IMP_FULL)));
if (unlikely(obd && obd->obd_fail)) {
CDEBUG(D_HA, "muting rpc for failed imp obd %s\n",
@@ -517,15 +518,22 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
return -ENODEV;
}
- connection = request->rq_import->imp_connection;
+ connection = imp->imp_connection;
lustre_msg_set_handle(request->rq_reqmsg,
- &request->rq_import->imp_remote_handle);
+ &imp->imp_remote_handle);
lustre_msg_set_type(request->rq_reqmsg, PTL_RPC_MSG_REQUEST);
- lustre_msg_set_conn_cnt(request->rq_reqmsg,
- request->rq_import->imp_conn_cnt);
- lustre_msghdr_set_flags(request->rq_reqmsg,
- request->rq_import->imp_msghdr_flags);
+ lustre_msg_set_conn_cnt(request->rq_reqmsg, imp->imp_conn_cnt);
+ lustre_msghdr_set_flags(request->rq_reqmsg, imp->imp_msghdr_flags);
+
+ /**
+ * For enabled AT all request should have AT_SUPPORT in the
+ * FULL import state when OBD_CONNECT_AT is set
+ */
+ LASSERT(AT_OFF || imp->imp_state != LUSTRE_IMP_FULL ||
+ (imp->imp_msghdr_flags & MSGHDR_AT_SUPPORT) ||
+ !(imp->imp_connect_data.ocd_connect_flags &
+ OBD_CONNECT_AT));
if (request->rq_resend)
lustre_msg_add_flags(request->rq_reqmsg, MSG_RESENT);
@@ -629,7 +637,7 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
ptlrpc_request_addref(request);
if (obd && obd->obd_svc_stats)
lprocfs_counter_add(obd->obd_svc_stats, PTLRPC_REQACTIVE_CNTR,
- atomic_read(&request->rq_import->imp_inflight));
+ atomic_read(&imp->imp_inflight));
OBD_FAIL_TIMEOUT(OBD_FAIL_PTLRPC_DELAY_SEND, request->rq_timeout + 5);
@@ -641,7 +649,7 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
request->rq_deadline = request->rq_sent + request->rq_timeout +
ptlrpc_at_get_net_latency(request);
- ptlrpc_pinger_sending_on_import(request->rq_import);
+ ptlrpc_pinger_sending_on_import(imp);
DEBUG_REQ(D_INFO, request, "send flg=%x",
lustre_msg_get_flags(request->rq_reqmsg));
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 095/124] staging: lustre: ptlrpc: fix race between connect vs resend
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Alexander Boyko, James Simmons
From: Alexander Boyko <alexander_boyko@xyratex.com>
Buggy code at ptlrpc_connect_interpret()
finish:
rc = ptlrpc_import_recovery_state_machine(imp);
...
Set import connection flags
When import has FULL state ptlrpc_import_recovery_state_machine()
wakeup all waiters on import and all delayed request, which was
resented. And it could happened that request was send without
updated flags and AT is disabled. If such request is in progress
on the server, server drop the new instance, and could do early reply
for it. But this early reply confuse client, cause it wait real
reply(no AT for this request). Client try to touch buffer outside
reply and got EPROTO error.
The same bug existed for initital connect too. Import became FULL
before import connection flags was set.
Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5528
Xyratex-bug-id: MRP-2034
Reviewed-on: http://review.whamcloud.com/11723
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Alexander Boyko <alexander.boyko@seagate.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/import.c | 307 +++++++++++++------------
drivers/staging/lustre/lustre/ptlrpc/niobuf.c | 28 ++-
2 files changed, 179 insertions(+), 156 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 17f3fec..2e107e8 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -752,6 +752,153 @@ static int ptlrpc_busy_reconnect(int rc)
return (rc == -EBUSY) || (rc == -EAGAIN);
}
+static int ptlrpc_connect_set_flags(struct obd_import *imp,
+ struct obd_connect_data *ocd,
+ u64 old_connect_flags,
+ struct obd_export *exp, int init_connect)
+{
+ struct client_obd *cli = &imp->imp_obd->u.cli;
+ static bool warned;
+
+ if ((imp->imp_connect_flags_orig & OBD_CONNECT_IBITS) &&
+ !(ocd->ocd_connect_flags & OBD_CONNECT_IBITS)) {
+ LCONSOLE_WARN("%s: MDS %s does not support ibits lock, either very old or invalid: requested %#llx, replied %#llx\n",
+ imp->imp_obd->obd_name,
+ imp->imp_connection->c_remote_uuid.uuid,
+ imp->imp_connect_flags_orig,
+ ocd->ocd_connect_flags);
+ return -EPROTO;
+ }
+
+ spin_lock(&imp->imp_lock);
+ list_del(&imp->imp_conn_current->oic_item);
+ list_add(&imp->imp_conn_current->oic_item, &imp->imp_conn_list);
+ imp->imp_last_success_conn = imp->imp_conn_current->oic_last_attempt;
+
+ spin_unlock(&imp->imp_lock);
+
+ if (!warned && (ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
+ (ocd->ocd_version > LUSTRE_VERSION_CODE +
+ LUSTRE_VERSION_OFFSET_WARN ||
+ ocd->ocd_version < LUSTRE_VERSION_CODE -
+ LUSTRE_VERSION_OFFSET_WARN)) {
+ /*
+ * Sigh, some compilers do not like #ifdef in the middle
+ * of macro arguments
+ */
+ const char *older = "older than client. Consider upgrading server";
+ const char *newer = "newer than client. Consider recompiling application";
+
+ LCONSOLE_WARN("Server %s version (%d.%d.%d.%d) is much %s (%s)\n",
+ obd2cli_tgt(imp->imp_obd),
+ OBD_OCD_VERSION_MAJOR(ocd->ocd_version),
+ OBD_OCD_VERSION_MINOR(ocd->ocd_version),
+ OBD_OCD_VERSION_PATCH(ocd->ocd_version),
+ OBD_OCD_VERSION_FIX(ocd->ocd_version),
+ ocd->ocd_version > LUSTRE_VERSION_CODE ?
+ newer : older, LUSTRE_VERSION_STRING);
+ warned = true;
+ }
+
+#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
+ /*
+ * Check if server has LU-1252 fix applied to not always swab
+ * the IR MNE entries. Do this only once per connection. This
+ * fixup is version-limited, because we don't want to carry the
+ * OBD_CONNECT_MNE_SWAB flag around forever, just so long as we
+ * need interop with unpatched 2.2 servers. For newer servers,
+ * the client will do MNE swabbing only as needed. LU-1644
+ */
+ if (unlikely((ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
+ !(ocd->ocd_connect_flags & OBD_CONNECT_MNE_SWAB) &&
+ OBD_OCD_VERSION_MAJOR(ocd->ocd_version) == 2 &&
+ OBD_OCD_VERSION_MINOR(ocd->ocd_version) == 2 &&
+ OBD_OCD_VERSION_PATCH(ocd->ocd_version) < 55 &&
+ !strcmp(imp->imp_obd->obd_type->typ_name,
+ LUSTRE_MGC_NAME)))
+ imp->imp_need_mne_swab = 1;
+ else /* clear if server was upgraded since last connect */
+ imp->imp_need_mne_swab = 0;
+#endif
+
+ if (ocd->ocd_connect_flags & OBD_CONNECT_CKSUM) {
+ /*
+ * We sent to the server ocd_cksum_types with bits set
+ * for algorithms we understand. The server masked off
+ * the checksum types it doesn't support
+ */
+ if (!(ocd->ocd_cksum_types & cksum_types_supported_client())) {
+ LCONSOLE_WARN("The negotiation of the checksum algorithm to use with server %s failed (%x/%x), disabling checksums\n",
+ obd2cli_tgt(imp->imp_obd),
+ ocd->ocd_cksum_types,
+ cksum_types_supported_client());
+ cli->cl_checksum = 0;
+ cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
+ } else {
+ cli->cl_supp_cksum_types = ocd->ocd_cksum_types;
+ }
+ } else {
+ /*
+ * The server does not support OBD_CONNECT_CKSUM.
+ * Enforce ADLER for backward compatibility
+ */
+ cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
+ }
+ cli->cl_cksum_type = cksum_type_select(cli->cl_supp_cksum_types);
+
+ if (ocd->ocd_connect_flags & OBD_CONNECT_BRW_SIZE)
+ cli->cl_max_pages_per_rpc =
+ min(ocd->ocd_brw_size >> PAGE_SHIFT,
+ cli->cl_max_pages_per_rpc);
+ else if (imp->imp_connect_op == MDS_CONNECT ||
+ imp->imp_connect_op == MGS_CONNECT)
+ cli->cl_max_pages_per_rpc = 1;
+
+ LASSERT((cli->cl_max_pages_per_rpc <= PTLRPC_MAX_BRW_PAGES) &&
+ (cli->cl_max_pages_per_rpc > 0));
+
+ client_adjust_max_dirty(cli);
+
+ /*
+ * Reset ns_connect_flags only for initial connect. It might be
+ * changed in while using FS and if we reset it in reconnect
+ * this leads to losing user settings done before such as
+ * disable lru_resize, etc.
+ */
+ if (old_connect_flags != exp_connect_flags(exp) || init_connect) {
+ CDEBUG(D_HA, "%s: Resetting ns_connect_flags to server flags: %#llx\n",
+ imp->imp_obd->obd_name, ocd->ocd_connect_flags);
+ imp->imp_obd->obd_namespace->ns_connect_flags =
+ ocd->ocd_connect_flags;
+ imp->imp_obd->obd_namespace->ns_orig_connect_flags =
+ ocd->ocd_connect_flags;
+ }
+
+ if ((ocd->ocd_connect_flags & OBD_CONNECT_AT) &&
+ (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
+ /*
+ * We need a per-message support flag, because
+ * a. we don't know if the incoming connect reply
+ * supports AT or not (in reply_in_callback)
+ * until we unpack it.
+ * b. failovered server means export and flags are gone
+ * (in ptlrpc_send_reply).
+ * Can only be set when we know AT is supported at
+ * both ends
+ */
+ imp->imp_msghdr_flags |= MSGHDR_AT_SUPPORT;
+ else
+ imp->imp_msghdr_flags &= ~MSGHDR_AT_SUPPORT;
+
+ if ((ocd->ocd_connect_flags & OBD_CONNECT_FULL20) &&
+ (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
+ imp->imp_msghdr_flags |= MSGHDR_CKSUM_INCOMPAT18;
+ else
+ imp->imp_msghdr_flags &= ~MSGHDR_CKSUM_INCOMPAT18;
+
+ return 0;
+}
+
/**
* interpret_reply callback for connect RPCs.
* Looks into returned status of connect operation and decides
@@ -764,7 +911,6 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
{
struct ptlrpc_connect_async_args *aa = data;
struct obd_import *imp = request->rq_import;
- struct client_obd *cli = &imp->imp_obd->u.cli;
struct lustre_handle old_hdl;
__u64 old_connect_flags;
int msg_flags;
@@ -844,7 +990,6 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
old_connect_flags = exp_connect_flags(exp);
exp->exp_connect_data = *ocd;
imp->imp_obd->obd_self_export->exp_connect_data = *ocd;
- class_export_put(exp);
/*
* The net statistics after (re-)connect is not valid anymore,
@@ -854,6 +999,13 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
ptlrpc_at_adj_net_latency(request,
lustre_msg_get_service_time(request->rq_repmsg));
+ /* Import flags should be updated before waking import at FULL state */
+ rc = ptlrpc_connect_set_flags(imp, ocd, old_connect_flags, exp,
+ aa->pcaa_initial_connect);
+ class_export_put(exp);
+ if (rc)
+ goto out;
+
obd_import_event(imp->imp_obd, imp, IMP_EVENT_OCD);
if (aa->pcaa_initial_connect) {
@@ -997,150 +1149,13 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
finish:
rc = ptlrpc_import_recovery_state_machine(imp);
- if (rc != 0) {
- if (rc == -ENOTCONN) {
- CDEBUG(D_HA, "evicted/aborted by %s@%s during recovery; invalidating and reconnecting\n",
- obd2cli_tgt(imp->imp_obd),
- imp->imp_connection->c_remote_uuid.uuid);
- ptlrpc_connect_import(imp);
- imp->imp_connect_tried = 1;
- return 0;
- }
- } else {
- static bool warned;
-
- spin_lock(&imp->imp_lock);
- list_del(&imp->imp_conn_current->oic_item);
- list_add(&imp->imp_conn_current->oic_item, &imp->imp_conn_list);
- imp->imp_last_success_conn =
- imp->imp_conn_current->oic_last_attempt;
-
- spin_unlock(&imp->imp_lock);
-
- if ((imp->imp_connect_flags_orig & OBD_CONNECT_IBITS) &&
- !(ocd->ocd_connect_flags & OBD_CONNECT_IBITS)) {
- LCONSOLE_WARN("%s: MDS %s does not support ibits lock, either very old or invalid: requested %llx, replied %llx\n",
- imp->imp_obd->obd_name,
- imp->imp_connection->c_remote_uuid.uuid,
- imp->imp_connect_flags_orig,
- ocd->ocd_connect_flags);
- rc = -EPROTO;
- goto out;
- }
-
- if (!warned && (ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
- (ocd->ocd_version > LUSTRE_VERSION_CODE +
- LUSTRE_VERSION_OFFSET_WARN ||
- ocd->ocd_version < LUSTRE_VERSION_CODE -
- LUSTRE_VERSION_OFFSET_WARN)) {
- /* Sigh, some compilers do not like #ifdef in the middle
- * of macro arguments
- */
- const char *older = "older than client. Consider upgrading server";
- const char *newer = "newer than client. Consider recompiling application";
-
- LCONSOLE_WARN("Server %s version (%d.%d.%d.%d) is much %s (%s)\n",
- obd2cli_tgt(imp->imp_obd),
- OBD_OCD_VERSION_MAJOR(ocd->ocd_version),
- OBD_OCD_VERSION_MINOR(ocd->ocd_version),
- OBD_OCD_VERSION_PATCH(ocd->ocd_version),
- OBD_OCD_VERSION_FIX(ocd->ocd_version),
- ocd->ocd_version > LUSTRE_VERSION_CODE ?
- newer : older, LUSTRE_VERSION_STRING);
- warned = true;
- }
-
-#if OBD_OCD_VERSION(3, 0, 53, 0) > LUSTRE_VERSION_CODE
- /* Check if server has LU-1252 fix applied to not always swab
- * the IR MNE entries. Do this only once per connection. This
- * fixup is version-limited, because we don't want to carry the
- * OBD_CONNECT_MNE_SWAB flag around forever, just so long as we
- * need interop with unpatched 2.2 servers. For newer servers,
- * the client will do MNE swabbing only as needed. LU-1644
- */
- if (unlikely((ocd->ocd_connect_flags & OBD_CONNECT_VERSION) &&
- !(ocd->ocd_connect_flags & OBD_CONNECT_MNE_SWAB) &&
- OBD_OCD_VERSION_MAJOR(ocd->ocd_version) == 2 &&
- OBD_OCD_VERSION_MINOR(ocd->ocd_version) == 2 &&
- OBD_OCD_VERSION_PATCH(ocd->ocd_version) < 55 &&
- strcmp(imp->imp_obd->obd_type->typ_name,
- LUSTRE_MGC_NAME) == 0))
- imp->imp_need_mne_swab = 1;
- else /* clear if server was upgraded since last connect */
- imp->imp_need_mne_swab = 0;
-#endif
-
- if (ocd->ocd_connect_flags & OBD_CONNECT_CKSUM) {
- /* We sent to the server ocd_cksum_types with bits set
- * for algorithms we understand. The server masked off
- * the checksum types it doesn't support
- */
- if ((ocd->ocd_cksum_types &
- cksum_types_supported_client()) == 0) {
- LCONSOLE_WARN("The negotiation of the checksum algorithm to use with server %s failed (%x/%x), disabling checksums\n",
- obd2cli_tgt(imp->imp_obd),
- ocd->ocd_cksum_types,
- cksum_types_supported_client());
- cli->cl_checksum = 0;
- cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
- } else {
- cli->cl_supp_cksum_types = ocd->ocd_cksum_types;
- }
- } else {
- /* The server does not support OBD_CONNECT_CKSUM.
- * Enforce ADLER for backward compatibility
- */
- cli->cl_supp_cksum_types = OBD_CKSUM_ADLER;
- }
- cli->cl_cksum_type = cksum_type_select(cli->cl_supp_cksum_types);
-
- if (ocd->ocd_connect_flags & OBD_CONNECT_BRW_SIZE)
- cli->cl_max_pages_per_rpc =
- min(ocd->ocd_brw_size >> PAGE_SHIFT,
- cli->cl_max_pages_per_rpc);
- else if (imp->imp_connect_op == MDS_CONNECT ||
- imp->imp_connect_op == MGS_CONNECT)
- cli->cl_max_pages_per_rpc = 1;
-
- /* Reset ns_connect_flags only for initial connect. It might be
- * changed in while using FS and if we reset it in reconnect
- * this leads to losing user settings done before such as
- * disable lru_resize, etc.
- */
- if (old_connect_flags != exp_connect_flags(exp) ||
- aa->pcaa_initial_connect) {
- CDEBUG(D_HA, "%s: Resetting ns_connect_flags to server flags: %#llx\n",
- imp->imp_obd->obd_name, ocd->ocd_connect_flags);
- imp->imp_obd->obd_namespace->ns_connect_flags =
- ocd->ocd_connect_flags;
- imp->imp_obd->obd_namespace->ns_orig_connect_flags =
- ocd->ocd_connect_flags;
- }
-
- if ((ocd->ocd_connect_flags & OBD_CONNECT_AT) &&
- (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
- /* We need a per-message support flag, because
- * a. we don't know if the incoming connect reply
- * supports AT or not (in reply_in_callback)
- * until we unpack it.
- * b. failovered server means export and flags are gone
- * (in ptlrpc_send_reply).
- * Can only be set when we know AT is supported at
- * both ends
- */
- imp->imp_msghdr_flags |= MSGHDR_AT_SUPPORT;
- else
- imp->imp_msghdr_flags &= ~MSGHDR_AT_SUPPORT;
-
- if ((ocd->ocd_connect_flags & OBD_CONNECT_FULL20) &&
- (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
- imp->imp_msghdr_flags |= MSGHDR_CKSUM_INCOMPAT18;
- else
- imp->imp_msghdr_flags &= ~MSGHDR_CKSUM_INCOMPAT18;
-
- LASSERT((cli->cl_max_pages_per_rpc <= PTLRPC_MAX_BRW_PAGES) &&
- (cli->cl_max_pages_per_rpc > 0));
- client_adjust_max_dirty(cli);
+ if (rc == -ENOTCONN) {
+ CDEBUG(D_HA, "evicted/aborted by %s@%s during recovery; invalidating and reconnecting\n",
+ obd2cli_tgt(imp->imp_obd),
+ imp->imp_connection->c_remote_uuid.uuid);
+ ptlrpc_connect_import(imp);
+ imp->imp_connect_tried = 1;
+ return 0;
}
out:
diff --git a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
index 44f4eae..5bab248 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
@@ -491,7 +491,8 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
struct ptlrpc_connection *connection;
lnet_handle_me_t reply_me_h;
lnet_md_t reply_md;
- struct obd_device *obd = request->rq_import->imp_obd;
+ struct obd_import *imp = request->rq_import;
+ struct obd_device *obd = imp->imp_obd;
if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_DROP_RPC))
return 0;
@@ -504,7 +505,7 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
*/
LASSERT(!request->rq_receiving_reply);
LASSERT(!((lustre_msg_get_flags(request->rq_reqmsg) & MSG_REPLAY) &&
- (request->rq_import->imp_state == LUSTRE_IMP_FULL)));
+ (imp->imp_state == LUSTRE_IMP_FULL)));
if (unlikely(obd && obd->obd_fail)) {
CDEBUG(D_HA, "muting rpc for failed imp obd %s\n",
@@ -517,15 +518,22 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
return -ENODEV;
}
- connection = request->rq_import->imp_connection;
+ connection = imp->imp_connection;
lustre_msg_set_handle(request->rq_reqmsg,
- &request->rq_import->imp_remote_handle);
+ &imp->imp_remote_handle);
lustre_msg_set_type(request->rq_reqmsg, PTL_RPC_MSG_REQUEST);
- lustre_msg_set_conn_cnt(request->rq_reqmsg,
- request->rq_import->imp_conn_cnt);
- lustre_msghdr_set_flags(request->rq_reqmsg,
- request->rq_import->imp_msghdr_flags);
+ lustre_msg_set_conn_cnt(request->rq_reqmsg, imp->imp_conn_cnt);
+ lustre_msghdr_set_flags(request->rq_reqmsg, imp->imp_msghdr_flags);
+
+ /**
+ * For enabled AT all request should have AT_SUPPORT in the
+ * FULL import state when OBD_CONNECT_AT is set
+ */
+ LASSERT(AT_OFF || imp->imp_state != LUSTRE_IMP_FULL ||
+ (imp->imp_msghdr_flags & MSGHDR_AT_SUPPORT) ||
+ !(imp->imp_connect_data.ocd_connect_flags &
+ OBD_CONNECT_AT));
if (request->rq_resend)
lustre_msg_add_flags(request->rq_reqmsg, MSG_RESENT);
@@ -629,7 +637,7 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
ptlrpc_request_addref(request);
if (obd && obd->obd_svc_stats)
lprocfs_counter_add(obd->obd_svc_stats, PTLRPC_REQACTIVE_CNTR,
- atomic_read(&request->rq_import->imp_inflight));
+ atomic_read(&imp->imp_inflight));
OBD_FAIL_TIMEOUT(OBD_FAIL_PTLRPC_DELAY_SEND, request->rq_timeout + 5);
@@ -641,7 +649,7 @@ int ptl_send_rpc(struct ptlrpc_request *request, int noreply)
request->rq_deadline = request->rq_sent + request->rq_timeout +
ptlrpc_at_get_net_latency(request);
- ptlrpc_pinger_sending_on_import(request->rq_import);
+ ptlrpc_pinger_sending_on_import(imp);
DEBUG_REQ(D_INFO, request, "send flg=%x",
lustre_msg_get_flags(request->rq_reqmsg));
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 096/124] staging: lustre: osc: osc_object_ast_clear() LBUG
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
An OSC object could be destroyed with AGL locks waiting for granted,
so we'd get rid of the osc_object_ast_clear() assertion that its
dlm locks all getting granted.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6042
Reviewed-on: http://review.whamcloud.com/13163
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_object.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_object.c b/drivers/staging/lustre/lustre/osc/osc_object.c
index d211d19..5fc5479 100644
--- a/drivers/staging/lustre/lustre/osc/osc_object.c
+++ b/drivers/staging/lustre/lustre/osc/osc_object.c
@@ -195,7 +195,6 @@ static int osc_object_glimpse(const struct lu_env *env,
static int osc_object_ast_clear(struct ldlm_lock *lock, void *data)
{
- LASSERT(lock->l_granted_mode == lock->l_req_mode);
if (lock->l_ast_data == data)
lock->l_ast_data = NULL;
return LDLM_ITER_CONTINUE;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 096/124] staging: lustre: osc: osc_object_ast_clear() LBUG
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
An OSC object could be destroyed with AGL locks waiting for granted,
so we'd get rid of the osc_object_ast_clear() assertion that its
dlm locks all getting granted.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6042
Reviewed-on: http://review.whamcloud.com/13163
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_object.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_object.c b/drivers/staging/lustre/lustre/osc/osc_object.c
index d211d19..5fc5479 100644
--- a/drivers/staging/lustre/lustre/osc/osc_object.c
+++ b/drivers/staging/lustre/lustre/osc/osc_object.c
@@ -195,7 +195,6 @@ static int osc_object_glimpse(const struct lu_env *env,
static int osc_object_ast_clear(struct ldlm_lock *lock, void *data)
{
- LASSERT(lock->l_granted_mode == lock->l_req_mode);
if (lock->l_ast_data == data)
lock->l_ast_data = NULL;
return LDLM_ITER_CONTINUE;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 097/124] staging: lustre: osc: change cl_extent_tax and *grants to unsigned
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage and remove warnings.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12386
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/osc/osc_cache.c | 46 ++++++++++---------
.../staging/lustre/lustre/osc/osc_cl_internal.h | 2 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 4 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 4 +-
5 files changed, 30 insertions(+), 28 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 8e426b8..5f0f1e3 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -237,7 +237,7 @@ struct client_obd {
* the extent size. A chunk is max(PAGE_SIZE, OST block size)
*/
int cl_chunkbits;
- int cl_extent_tax; /* extent overhead, by bytes */
+ unsigned int cl_extent_tax; /* extent overhead, by bytes */
/* keep track of objects that have lois that contain pages which
* have been queued for async brw. this lock also protects the
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 31deab0..8152cc7 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -44,7 +44,7 @@ static int extent_debug; /* set it to be true for more debug */
static void osc_update_pending(struct osc_object *obj, int cmd, int delta);
static int osc_extent_wait(const struct lu_env *env, struct osc_extent *ext,
- int state);
+ enum osc_extent_state state);
static void osc_ap_completion(const struct lu_env *env, struct client_obd *cli,
struct osc_async_page *oap, int sent, int rc);
static int osc_make_ready(const struct lu_env *env, struct osc_async_page *oap,
@@ -177,7 +177,7 @@ static int osc_extent_sanity_check0(struct osc_extent *ext,
{
struct osc_object *obj = ext->oe_obj;
struct osc_async_page *oap;
- int page_count;
+ size_t page_count;
int rc = 0;
if (!osc_object_is_locked(obj)) {
@@ -632,7 +632,7 @@ static inline int overlapped(struct osc_extent *ex1, struct osc_extent *ex2)
*/
static struct osc_extent *osc_extent_find(const struct lu_env *env,
struct osc_object *obj, pgoff_t index,
- int *grants)
+ unsigned int *grants)
{
struct client_obd *cli = osc_cli(obj);
struct osc_lock *olck;
@@ -643,10 +643,10 @@ static struct osc_extent *osc_extent_find(const struct lu_env *env,
struct osc_extent *found = NULL;
pgoff_t chunk;
pgoff_t max_end;
- int max_pages; /* max_pages_per_rpc */
- int chunksize;
+ unsigned int max_pages; /* max_pages_per_rpc */
+ unsigned int chunksize;
int ppc_bits; /* pages per chunk bits */
- int chunk_mask;
+ pgoff_t chunk_mask;
int rc;
cur = osc_extent_alloc(obj);
@@ -700,8 +700,8 @@ restart:
if (!ext)
ext = first_extent(obj);
while (ext) {
- loff_t ext_chk_start = ext->oe_start >> ppc_bits;
- loff_t ext_chk_end = ext->oe_end >> ppc_bits;
+ pgoff_t ext_chk_start = ext->oe_start >> ppc_bits;
+ pgoff_t ext_chk_end = ext->oe_end >> ppc_bits;
LASSERT(sanity_check_nolock(ext) == 0);
if (chunk > ext_chk_end + 1)
@@ -913,7 +913,7 @@ int osc_extent_finish(const struct lu_env *env, struct osc_extent *ext,
return 0;
}
-static int extent_wait_cb(struct osc_extent *ext, int state)
+static int extent_wait_cb(struct osc_extent *ext, enum osc_extent_state state)
{
int ret;
@@ -928,7 +928,7 @@ static int extent_wait_cb(struct osc_extent *ext, int state)
* Wait for the extent's state to become @state.
*/
static int osc_extent_wait(const struct lu_env *env, struct osc_extent *ext,
- int state)
+ enum osc_extent_state state)
{
struct osc_object *obj = ext->oe_obj;
struct l_wait_info lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(600), NULL,
@@ -958,7 +958,7 @@ static int osc_extent_wait(const struct lu_env *env, struct osc_extent *ext,
rc = l_wait_event(ext->oe_waitq, extent_wait_cb(ext, state), &lwi);
if (rc == -ETIMEDOUT) {
OSC_EXTENT_DUMP(D_ERROR, ext,
- "%s: wait ext to %d timedout, recovery in progress?\n",
+ "%s: wait ext to %u timedout, recovery in progress?\n",
osc_export(obj)->exp_obd->obd_name, state);
lwi = LWI_INTR(NULL, NULL);
@@ -1176,7 +1176,8 @@ static int osc_extent_make_ready(const struct lu_env *env,
* called to expand the extent for the same IO. To expand the extent, the
* page index must be in the same or next chunk of ext->oe_end.
*/
-static int osc_extent_expand(struct osc_extent *ext, pgoff_t index, int *grants)
+static int osc_extent_expand(struct osc_extent *ext, pgoff_t index,
+ unsigned int *grants)
{
struct osc_object *obj = ext->oe_obj;
struct client_obd *cli = osc_cli(obj);
@@ -1185,7 +1186,7 @@ static int osc_extent_expand(struct osc_extent *ext, pgoff_t index, int *grants)
pgoff_t chunk = index >> ppc_bits;
pgoff_t end_chunk;
pgoff_t end_index;
- int chunksize = 1 << cli->cl_chunkbits;
+ unsigned int chunksize = 1 << cli->cl_chunkbits;
int rc = 0;
LASSERT(ext->oe_max_end >= index && ext->oe_start <= index);
@@ -1363,7 +1364,7 @@ static int osc_completion(const struct lu_env *env, struct osc_async_page *oap,
if (rc == 0 && srvlock) {
struct lu_device *ld = opg->ops_cl.cpl_obj->co_lu.lo_dev;
struct osc_stats *stats = &lu2osc_dev(ld)->od_stats;
- int bytes = oap->oap_count;
+ size_t bytes = oap->oap_count;
if (crt == CRT_READ)
stats->os_lockless_reads += bytes;
@@ -1494,7 +1495,7 @@ static void osc_unreserve_grant(struct client_obd *cli,
static void osc_free_grant(struct client_obd *cli, unsigned int nr_pages,
unsigned int lost_grant)
{
- int grant = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
+ unsigned long grant = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
spin_lock(&cli->cl_loi_list_lock);
atomic_long_sub(nr_pages, &obd_dirty_pages);
@@ -1884,7 +1885,7 @@ static void osc_ap_completion(const struct lu_env *env, struct client_obd *cli,
*/
static int try_to_add_extent_for_io(struct client_obd *cli,
struct osc_extent *ext, struct list_head *rpclist,
- int *pc, unsigned int *max_pages)
+ unsigned int *pc, unsigned int *max_pages)
{
struct osc_extent *tmp;
struct osc_async_page *oap = list_first_entry(&ext->oe_pages,
@@ -1937,12 +1938,13 @@ static int try_to_add_extent_for_io(struct client_obd *cli,
* 5. Traverse the extent tree from the 1st extent;
* 6. Above steps exit if there is no space in this RPC.
*/
-static int get_write_extents(struct osc_object *obj, struct list_head *rpclist)
+static unsigned int get_write_extents(struct osc_object *obj,
+ struct list_head *rpclist)
{
struct client_obd *cli = osc_cli(obj);
struct osc_extent *ext;
struct osc_extent *temp;
- int page_count = 0;
+ unsigned int page_count = 0;
unsigned int max_pages = cli->cl_max_pages_per_rpc;
LASSERT(osc_object_is_locked(obj));
@@ -2080,7 +2082,7 @@ osc_send_read_rpc(const struct lu_env *env, struct client_obd *cli,
struct osc_extent *ext;
struct osc_extent *next;
LIST_HEAD(rpclist);
- int page_count = 0;
+ unsigned int page_count = 0;
unsigned int max_pages = cli->cl_max_pages_per_rpc;
int rc = 0;
@@ -2295,7 +2297,7 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
struct client_obd *cli = oap->oap_cli;
struct osc_object *osc = oap->oap_obj;
pgoff_t index;
- int grants = 0;
+ unsigned int grants = 0, tmp;
int brw_flags = OBD_BRW_ASYNC;
int cmd = OBD_BRW_WRITE;
int need_release = 0;
@@ -2378,7 +2380,7 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
grants = 0;
need_release = 1;
} else if (ext->oe_end < index) {
- int tmp = grants;
+ tmp = grants;
/* try to expand this extent */
rc = osc_extent_expand(ext, index, &tmp);
if (rc < 0) {
@@ -2403,7 +2405,7 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
}
if (!ext) {
- int tmp = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
+ tmp = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
/* try to find new extent to cover this page */
LASSERT(!oio->oi_active);
diff --git a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
index 3068113..9c8de15 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
+++ b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
@@ -608,7 +608,7 @@ struct osc_extent {
/** link list of osc_object's oo_{hp|urgent|locking}_exts. */
struct list_head oe_link;
/** state of this extent */
- unsigned int oe_state;
+ enum osc_extent_state oe_state;
/** flags for this extent. */
unsigned int oe_intree:1,
/** 0 is write, 1 is read */
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 96beba4..8bbb2a6 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -109,11 +109,11 @@ static int osc_io_submit(const struct lu_env *env,
struct cl_page_list *qin = &queue->c2_qin;
struct cl_page_list *qout = &queue->c2_qout;
- int queued = 0;
+ unsigned int queued = 0;
int result = 0;
int cmd;
int brw_flags;
- int max_pages;
+ unsigned int max_pages;
LASSERT(qin->pl_nr > 0);
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 59ce7a8..3b59c2d 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1110,7 +1110,7 @@ static u32 osc_checksum_bulk(int nob, u32 pg_count,
}
while (nob > 0 && pg_count > 0) {
- int count = pga[i]->count > nob ? nob : pga[i]->count;
+ unsigned int count = pga[i]->count > nob ? nob : pga[i]->count;
/* corrupt the data before we compute the checksum, to
* simulate an OST->client data error
@@ -1120,7 +1120,7 @@ static u32 osc_checksum_bulk(int nob, u32 pg_count,
unsigned char *ptr = kmap(pga[i]->pg);
int off = pga[i]->off & ~PAGE_MASK;
- memcpy(ptr + off, "bad1", min(4, nob));
+ memcpy(ptr + off, "bad1", min_t(typeof(nob), 4, nob));
kunmap(pga[i]->pg);
}
cfs_crypto_hash_update_page(hdesc, pga[i]->pg,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 097/124] staging: lustre: osc: change cl_extent_tax and *grants to unsigned
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change the type accordant usage and remove warnings.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/12386
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/osc/osc_cache.c | 46 ++++++++++---------
.../staging/lustre/lustre/osc/osc_cl_internal.h | 2 +-
drivers/staging/lustre/lustre/osc/osc_io.c | 4 +-
drivers/staging/lustre/lustre/osc/osc_request.c | 4 +-
5 files changed, 30 insertions(+), 28 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 8e426b8..5f0f1e3 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -237,7 +237,7 @@ struct client_obd {
* the extent size. A chunk is max(PAGE_SIZE, OST block size)
*/
int cl_chunkbits;
- int cl_extent_tax; /* extent overhead, by bytes */
+ unsigned int cl_extent_tax; /* extent overhead, by bytes */
/* keep track of objects that have lois that contain pages which
* have been queued for async brw. this lock also protects the
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 31deab0..8152cc7 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -44,7 +44,7 @@ static int extent_debug; /* set it to be true for more debug */
static void osc_update_pending(struct osc_object *obj, int cmd, int delta);
static int osc_extent_wait(const struct lu_env *env, struct osc_extent *ext,
- int state);
+ enum osc_extent_state state);
static void osc_ap_completion(const struct lu_env *env, struct client_obd *cli,
struct osc_async_page *oap, int sent, int rc);
static int osc_make_ready(const struct lu_env *env, struct osc_async_page *oap,
@@ -177,7 +177,7 @@ static int osc_extent_sanity_check0(struct osc_extent *ext,
{
struct osc_object *obj = ext->oe_obj;
struct osc_async_page *oap;
- int page_count;
+ size_t page_count;
int rc = 0;
if (!osc_object_is_locked(obj)) {
@@ -632,7 +632,7 @@ static inline int overlapped(struct osc_extent *ex1, struct osc_extent *ex2)
*/
static struct osc_extent *osc_extent_find(const struct lu_env *env,
struct osc_object *obj, pgoff_t index,
- int *grants)
+ unsigned int *grants)
{
struct client_obd *cli = osc_cli(obj);
struct osc_lock *olck;
@@ -643,10 +643,10 @@ static struct osc_extent *osc_extent_find(const struct lu_env *env,
struct osc_extent *found = NULL;
pgoff_t chunk;
pgoff_t max_end;
- int max_pages; /* max_pages_per_rpc */
- int chunksize;
+ unsigned int max_pages; /* max_pages_per_rpc */
+ unsigned int chunksize;
int ppc_bits; /* pages per chunk bits */
- int chunk_mask;
+ pgoff_t chunk_mask;
int rc;
cur = osc_extent_alloc(obj);
@@ -700,8 +700,8 @@ restart:
if (!ext)
ext = first_extent(obj);
while (ext) {
- loff_t ext_chk_start = ext->oe_start >> ppc_bits;
- loff_t ext_chk_end = ext->oe_end >> ppc_bits;
+ pgoff_t ext_chk_start = ext->oe_start >> ppc_bits;
+ pgoff_t ext_chk_end = ext->oe_end >> ppc_bits;
LASSERT(sanity_check_nolock(ext) == 0);
if (chunk > ext_chk_end + 1)
@@ -913,7 +913,7 @@ int osc_extent_finish(const struct lu_env *env, struct osc_extent *ext,
return 0;
}
-static int extent_wait_cb(struct osc_extent *ext, int state)
+static int extent_wait_cb(struct osc_extent *ext, enum osc_extent_state state)
{
int ret;
@@ -928,7 +928,7 @@ static int extent_wait_cb(struct osc_extent *ext, int state)
* Wait for the extent's state to become @state.
*/
static int osc_extent_wait(const struct lu_env *env, struct osc_extent *ext,
- int state)
+ enum osc_extent_state state)
{
struct osc_object *obj = ext->oe_obj;
struct l_wait_info lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(600), NULL,
@@ -958,7 +958,7 @@ static int osc_extent_wait(const struct lu_env *env, struct osc_extent *ext,
rc = l_wait_event(ext->oe_waitq, extent_wait_cb(ext, state), &lwi);
if (rc == -ETIMEDOUT) {
OSC_EXTENT_DUMP(D_ERROR, ext,
- "%s: wait ext to %d timedout, recovery in progress?\n",
+ "%s: wait ext to %u timedout, recovery in progress?\n",
osc_export(obj)->exp_obd->obd_name, state);
lwi = LWI_INTR(NULL, NULL);
@@ -1176,7 +1176,8 @@ static int osc_extent_make_ready(const struct lu_env *env,
* called to expand the extent for the same IO. To expand the extent, the
* page index must be in the same or next chunk of ext->oe_end.
*/
-static int osc_extent_expand(struct osc_extent *ext, pgoff_t index, int *grants)
+static int osc_extent_expand(struct osc_extent *ext, pgoff_t index,
+ unsigned int *grants)
{
struct osc_object *obj = ext->oe_obj;
struct client_obd *cli = osc_cli(obj);
@@ -1185,7 +1186,7 @@ static int osc_extent_expand(struct osc_extent *ext, pgoff_t index, int *grants)
pgoff_t chunk = index >> ppc_bits;
pgoff_t end_chunk;
pgoff_t end_index;
- int chunksize = 1 << cli->cl_chunkbits;
+ unsigned int chunksize = 1 << cli->cl_chunkbits;
int rc = 0;
LASSERT(ext->oe_max_end >= index && ext->oe_start <= index);
@@ -1363,7 +1364,7 @@ static int osc_completion(const struct lu_env *env, struct osc_async_page *oap,
if (rc == 0 && srvlock) {
struct lu_device *ld = opg->ops_cl.cpl_obj->co_lu.lo_dev;
struct osc_stats *stats = &lu2osc_dev(ld)->od_stats;
- int bytes = oap->oap_count;
+ size_t bytes = oap->oap_count;
if (crt == CRT_READ)
stats->os_lockless_reads += bytes;
@@ -1494,7 +1495,7 @@ static void osc_unreserve_grant(struct client_obd *cli,
static void osc_free_grant(struct client_obd *cli, unsigned int nr_pages,
unsigned int lost_grant)
{
- int grant = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
+ unsigned long grant = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
spin_lock(&cli->cl_loi_list_lock);
atomic_long_sub(nr_pages, &obd_dirty_pages);
@@ -1884,7 +1885,7 @@ static void osc_ap_completion(const struct lu_env *env, struct client_obd *cli,
*/
static int try_to_add_extent_for_io(struct client_obd *cli,
struct osc_extent *ext, struct list_head *rpclist,
- int *pc, unsigned int *max_pages)
+ unsigned int *pc, unsigned int *max_pages)
{
struct osc_extent *tmp;
struct osc_async_page *oap = list_first_entry(&ext->oe_pages,
@@ -1937,12 +1938,13 @@ static int try_to_add_extent_for_io(struct client_obd *cli,
* 5. Traverse the extent tree from the 1st extent;
* 6. Above steps exit if there is no space in this RPC.
*/
-static int get_write_extents(struct osc_object *obj, struct list_head *rpclist)
+static unsigned int get_write_extents(struct osc_object *obj,
+ struct list_head *rpclist)
{
struct client_obd *cli = osc_cli(obj);
struct osc_extent *ext;
struct osc_extent *temp;
- int page_count = 0;
+ unsigned int page_count = 0;
unsigned int max_pages = cli->cl_max_pages_per_rpc;
LASSERT(osc_object_is_locked(obj));
@@ -2080,7 +2082,7 @@ osc_send_read_rpc(const struct lu_env *env, struct client_obd *cli,
struct osc_extent *ext;
struct osc_extent *next;
LIST_HEAD(rpclist);
- int page_count = 0;
+ unsigned int page_count = 0;
unsigned int max_pages = cli->cl_max_pages_per_rpc;
int rc = 0;
@@ -2295,7 +2297,7 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
struct client_obd *cli = oap->oap_cli;
struct osc_object *osc = oap->oap_obj;
pgoff_t index;
- int grants = 0;
+ unsigned int grants = 0, tmp;
int brw_flags = OBD_BRW_ASYNC;
int cmd = OBD_BRW_WRITE;
int need_release = 0;
@@ -2378,7 +2380,7 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
grants = 0;
need_release = 1;
} else if (ext->oe_end < index) {
- int tmp = grants;
+ tmp = grants;
/* try to expand this extent */
rc = osc_extent_expand(ext, index, &tmp);
if (rc < 0) {
@@ -2403,7 +2405,7 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
}
if (!ext) {
- int tmp = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
+ tmp = (1 << cli->cl_chunkbits) + cli->cl_extent_tax;
/* try to find new extent to cover this page */
LASSERT(!oio->oi_active);
diff --git a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
index 3068113..9c8de15 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
+++ b/drivers/staging/lustre/lustre/osc/osc_cl_internal.h
@@ -608,7 +608,7 @@ struct osc_extent {
/** link list of osc_object's oo_{hp|urgent|locking}_exts. */
struct list_head oe_link;
/** state of this extent */
- unsigned int oe_state;
+ enum osc_extent_state oe_state;
/** flags for this extent. */
unsigned int oe_intree:1,
/** 0 is write, 1 is read */
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 96beba4..8bbb2a6 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -109,11 +109,11 @@ static int osc_io_submit(const struct lu_env *env,
struct cl_page_list *qin = &queue->c2_qin;
struct cl_page_list *qout = &queue->c2_qout;
- int queued = 0;
+ unsigned int queued = 0;
int result = 0;
int cmd;
int brw_flags;
- int max_pages;
+ unsigned int max_pages;
LASSERT(qin->pl_nr > 0);
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 59ce7a8..3b59c2d 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1110,7 +1110,7 @@ static u32 osc_checksum_bulk(int nob, u32 pg_count,
}
while (nob > 0 && pg_count > 0) {
- int count = pga[i]->count > nob ? nob : pga[i]->count;
+ unsigned int count = pga[i]->count > nob ? nob : pga[i]->count;
/* corrupt the data before we compute the checksum, to
* simulate an OST->client data error
@@ -1120,7 +1120,7 @@ static u32 osc_checksum_bulk(int nob, u32 pg_count,
unsigned char *ptr = kmap(pga[i]->pg);
int off = pga[i]->off & ~PAGE_MASK;
- memcpy(ptr + off, "bad1", min(4, nob));
+ memcpy(ptr + off, "bad1", min_t(typeof(nob), 4, nob));
kunmap(pga[i]->pg);
}
cfs_crypto_hash_update_page(hdesc, pga[i]->pg,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 098/124] staging: lustre: lprocfs: cleanup stats locking code
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Andreas Dilger, James Simmons
From: Andreas Dilger <andreas.dilger@intel.com>
Add comment blocks on lprocfs_stats_lock() and lprocfs_stats_unlock().
Move common NOPERCPU code out of the switch() statements to reduce
code size and complexity, since it doesn't depend on the opc at all.
Replace switch() in lprocfs_stats_unlock() with a simple if/else,
since the lock opc was already checked in lprocfs_stats_lock().
Add an enum for the lprocfs_stats_lock() operations to make it clear
what the valid values are and allow compiler checking.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5946
Reviewed-on: http://review.whamcloud.com/12872
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/lprocfs_status.h | 132 +++++++++++---------
1 files changed, 74 insertions(+), 58 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lprocfs_status.h b/drivers/staging/lustre/lustre/include/lprocfs_status.h
index b1a8ca5..cc0713e 100644
--- a/drivers/staging/lustre/lustre/include/lprocfs_status.h
+++ b/drivers/staging/lustre/lustre/include/lprocfs_status.h
@@ -165,8 +165,10 @@ struct lprocfs_percpu {
struct lprocfs_counter lp_cntr[0];
};
-#define LPROCFS_GET_NUM_CPU 0x0001
-#define LPROCFS_GET_SMP_ID 0x0002
+enum lprocfs_stats_lock_ops {
+ LPROCFS_GET_NUM_CPU = 0x0001, /* number allocated per-CPU stats */
+ LPROCFS_GET_SMP_ID = 0x0002, /* current stat to be updated */
+};
enum lprocfs_stats_flags {
LPROCFS_STATS_FLAG_NONE = 0x0000, /* per cpu counter */
@@ -371,77 +373,91 @@ int lprocfs_write_frac_helper(const char __user *buffer,
int lprocfs_read_frac_helper(char *buffer, unsigned long count,
long val, int mult);
int lprocfs_stats_alloc_one(struct lprocfs_stats *stats, unsigned int cpuid);
-/*
- * \return value
- * < 0 : on error (only possible for opc as LPROCFS_GET_SMP_ID)
+
+/**
+ * Lock statistics structure for access, possibly only on this CPU.
+ *
+ * The statistics struct may be allocated with per-CPU structures for
+ * efficient concurrent update (usually only on server-wide stats), or
+ * as a single global struct (e.g. for per-client or per-job statistics),
+ * so the required locking depends on the type of structure allocated.
+ *
+ * For per-CPU statistics, pin the thread to the current cpuid so that
+ * will only access the statistics for that CPU. If the stats structure
+ * for the current CPU has not been allocated (or previously freed),
+ * allocate it now. The per-CPU statistics do not need locking since
+ * the thread is pinned to the CPU during update.
+ *
+ * For global statistics, lock the stats structure to prevent concurrent update.
+ *
+ * \param[in] stats statistics structure to lock
+ * \param[in] opc type of operation:
+ * LPROCFS_GET_SMP_ID: "lock" and return current CPU index
+ * for incrementing statistics for that CPU
+ * LPROCFS_GET_NUM_CPU: "lock" and return number of used
+ * CPU indices to iterate over all indices
+ * \param[out] flags CPU interrupt saved state for IRQ-safe locking
+ *
+ * \retval cpuid of current thread or number of allocated structs
+ * \retval negative on error (only for opc LPROCFS_GET_SMP_ID + per-CPU stats)
*/
-static inline int lprocfs_stats_lock(struct lprocfs_stats *stats, int opc,
+static inline int lprocfs_stats_lock(struct lprocfs_stats *stats,
+ enum lprocfs_stats_lock_ops opc,
unsigned long *flags)
{
- int rc = 0;
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
+ spin_lock_irqsave(&stats->ls_lock, *flags);
+ else
+ spin_lock(&stats->ls_lock);
+ return opc == LPROCFS_GET_NUM_CPU ? 1 : 0;
+ }
switch (opc) {
- default:
- LBUG();
+ case LPROCFS_GET_SMP_ID: {
+ unsigned int cpuid = get_cpu();
+
+ if (unlikely(!stats->ls_percpu[cpuid])) {
+ int rc = lprocfs_stats_alloc_one(stats, cpuid);
- case LPROCFS_GET_SMP_ID:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_lock_irqsave(&stats->ls_lock, *flags);
- else
- spin_lock(&stats->ls_lock);
- return 0;
- } else {
- unsigned int cpuid = get_cpu();
-
- if (unlikely(!stats->ls_percpu[cpuid])) {
- rc = lprocfs_stats_alloc_one(stats, cpuid);
- if (rc < 0) {
- put_cpu();
- return rc;
- }
+ if (rc < 0) {
+ put_cpu();
+ return rc;
}
- return cpuid;
}
-
+ return cpuid;
+ }
case LPROCFS_GET_NUM_CPU:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_lock_irqsave(&stats->ls_lock, *flags);
- else
- spin_lock(&stats->ls_lock);
- return 1;
- }
return stats->ls_biggest_alloc_num;
+ default:
+ LBUG();
}
}
-static inline void lprocfs_stats_unlock(struct lprocfs_stats *stats, int opc,
+/**
+ * Unlock statistics structure after access.
+ *
+ * Unlock the lock acquired via lprocfs_stats_lock() for global statistics,
+ * or unpin this thread from the current cpuid for per-CPU statistics.
+ *
+ * This function must be called using the same arguments as used when calling
+ * lprocfs_stats_lock() so that the correct operation can be performed.
+ *
+ * \param[in] stats statistics structure to unlock
+ * \param[in] opc type of operation (current cpuid or number of structs)
+ * \param[in] flags CPU interrupt saved state for IRQ-safe locking
+ */
+static inline void lprocfs_stats_unlock(struct lprocfs_stats *stats,
+ enum lprocfs_stats_lock_ops opc,
unsigned long *flags)
{
- switch (opc) {
- default:
- LBUG();
-
- case LPROCFS_GET_SMP_ID:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_unlock_irqrestore(&stats->ls_lock, *flags);
- else
- spin_unlock(&stats->ls_lock);
- } else {
- put_cpu();
- }
- return;
-
- case LPROCFS_GET_NUM_CPU:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_unlock_irqrestore(&stats->ls_lock, *flags);
- else
- spin_unlock(&stats->ls_lock);
- }
- return;
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
+ spin_unlock_irqrestore(&stats->ls_lock, *flags);
+ else
+ spin_unlock(&stats->ls_lock);
+ } else if (opc == LPROCFS_GET_SMP_ID) {
+ put_cpu();
}
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 098/124] staging: lustre: lprocfs: cleanup stats locking code
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
From: Andreas Dilger <andreas.dilger@intel.com>
Add comment blocks on lprocfs_stats_lock() and lprocfs_stats_unlock().
Move common NOPERCPU code out of the switch() statements to reduce
code size and complexity, since it doesn't depend on the opc at all.
Replace switch() in lprocfs_stats_unlock() with a simple if/else,
since the lock opc was already checked in lprocfs_stats_lock().
Add an enum for the lprocfs_stats_lock() operations to make it clear
what the valid values are and allow compiler checking.
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5946
Reviewed-on: http://review.whamcloud.com/12872
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/include/lprocfs_status.h | 132 +++++++++++---------
1 files changed, 74 insertions(+), 58 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lprocfs_status.h b/drivers/staging/lustre/lustre/include/lprocfs_status.h
index b1a8ca5..cc0713e 100644
--- a/drivers/staging/lustre/lustre/include/lprocfs_status.h
+++ b/drivers/staging/lustre/lustre/include/lprocfs_status.h
@@ -165,8 +165,10 @@ struct lprocfs_percpu {
struct lprocfs_counter lp_cntr[0];
};
-#define LPROCFS_GET_NUM_CPU 0x0001
-#define LPROCFS_GET_SMP_ID 0x0002
+enum lprocfs_stats_lock_ops {
+ LPROCFS_GET_NUM_CPU = 0x0001, /* number allocated per-CPU stats */
+ LPROCFS_GET_SMP_ID = 0x0002, /* current stat to be updated */
+};
enum lprocfs_stats_flags {
LPROCFS_STATS_FLAG_NONE = 0x0000, /* per cpu counter */
@@ -371,77 +373,91 @@ int lprocfs_write_frac_helper(const char __user *buffer,
int lprocfs_read_frac_helper(char *buffer, unsigned long count,
long val, int mult);
int lprocfs_stats_alloc_one(struct lprocfs_stats *stats, unsigned int cpuid);
-/*
- * \return value
- * < 0 : on error (only possible for opc as LPROCFS_GET_SMP_ID)
+
+/**
+ * Lock statistics structure for access, possibly only on this CPU.
+ *
+ * The statistics struct may be allocated with per-CPU structures for
+ * efficient concurrent update (usually only on server-wide stats), or
+ * as a single global struct (e.g. for per-client or per-job statistics),
+ * so the required locking depends on the type of structure allocated.
+ *
+ * For per-CPU statistics, pin the thread to the current cpuid so that
+ * will only access the statistics for that CPU. If the stats structure
+ * for the current CPU has not been allocated (or previously freed),
+ * allocate it now. The per-CPU statistics do not need locking since
+ * the thread is pinned to the CPU during update.
+ *
+ * For global statistics, lock the stats structure to prevent concurrent update.
+ *
+ * \param[in] stats statistics structure to lock
+ * \param[in] opc type of operation:
+ * LPROCFS_GET_SMP_ID: "lock" and return current CPU index
+ * for incrementing statistics for that CPU
+ * LPROCFS_GET_NUM_CPU: "lock" and return number of used
+ * CPU indices to iterate over all indices
+ * \param[out] flags CPU interrupt saved state for IRQ-safe locking
+ *
+ * \retval cpuid of current thread or number of allocated structs
+ * \retval negative on error (only for opc LPROCFS_GET_SMP_ID + per-CPU stats)
*/
-static inline int lprocfs_stats_lock(struct lprocfs_stats *stats, int opc,
+static inline int lprocfs_stats_lock(struct lprocfs_stats *stats,
+ enum lprocfs_stats_lock_ops opc,
unsigned long *flags)
{
- int rc = 0;
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
+ spin_lock_irqsave(&stats->ls_lock, *flags);
+ else
+ spin_lock(&stats->ls_lock);
+ return opc == LPROCFS_GET_NUM_CPU ? 1 : 0;
+ }
switch (opc) {
- default:
- LBUG();
+ case LPROCFS_GET_SMP_ID: {
+ unsigned int cpuid = get_cpu();
+
+ if (unlikely(!stats->ls_percpu[cpuid])) {
+ int rc = lprocfs_stats_alloc_one(stats, cpuid);
- case LPROCFS_GET_SMP_ID:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_lock_irqsave(&stats->ls_lock, *flags);
- else
- spin_lock(&stats->ls_lock);
- return 0;
- } else {
- unsigned int cpuid = get_cpu();
-
- if (unlikely(!stats->ls_percpu[cpuid])) {
- rc = lprocfs_stats_alloc_one(stats, cpuid);
- if (rc < 0) {
- put_cpu();
- return rc;
- }
+ if (rc < 0) {
+ put_cpu();
+ return rc;
}
- return cpuid;
}
-
+ return cpuid;
+ }
case LPROCFS_GET_NUM_CPU:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_lock_irqsave(&stats->ls_lock, *flags);
- else
- spin_lock(&stats->ls_lock);
- return 1;
- }
return stats->ls_biggest_alloc_num;
+ default:
+ LBUG();
}
}
-static inline void lprocfs_stats_unlock(struct lprocfs_stats *stats, int opc,
+/**
+ * Unlock statistics structure after access.
+ *
+ * Unlock the lock acquired via lprocfs_stats_lock() for global statistics,
+ * or unpin this thread from the current cpuid for per-CPU statistics.
+ *
+ * This function must be called using the same arguments as used when calling
+ * lprocfs_stats_lock() so that the correct operation can be performed.
+ *
+ * \param[in] stats statistics structure to unlock
+ * \param[in] opc type of operation (current cpuid or number of structs)
+ * \param[in] flags CPU interrupt saved state for IRQ-safe locking
+ */
+static inline void lprocfs_stats_unlock(struct lprocfs_stats *stats,
+ enum lprocfs_stats_lock_ops opc,
unsigned long *flags)
{
- switch (opc) {
- default:
- LBUG();
-
- case LPROCFS_GET_SMP_ID:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_unlock_irqrestore(&stats->ls_lock, *flags);
- else
- spin_unlock(&stats->ls_lock);
- } else {
- put_cpu();
- }
- return;
-
- case LPROCFS_GET_NUM_CPU:
- if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
- if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
- spin_unlock_irqrestore(&stats->ls_lock, *flags);
- else
- spin_unlock(&stats->ls_lock);
- }
- return;
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_NOPERCPU) {
+ if (stats->ls_flags & LPROCFS_STATS_FLAG_IRQ_SAFE)
+ spin_unlock_irqrestore(&stats->ls_lock, *flags);
+ else
+ spin_unlock(&stats->ls_lock);
+ } else if (opc == LPROCFS_GET_SMP_ID) {
+ put_cpu();
}
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 099/124] staging: lustre: llite: unlock inode size in ll_lov_setstripe_ea_info()
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In ll_lov_setstripe_ea_info() release the inode size lock on all
appropriate exit paths.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6059
Reviewed-on: http://review.whamcloud.com/13167
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/file.c | 13 ++++++-------
1 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 2cb0b65..ce96c9f 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1301,7 +1301,10 @@ int ll_lov_setstripe_ea_info(struct inode *inode, struct dentry *dentry,
int lum_size)
{
struct lov_stripe_md *lsm = NULL;
- struct lookup_intent oit = {.it_op = IT_OPEN, .it_flags = flags};
+ struct lookup_intent oit = {
+ .it_op = IT_OPEN,
+ .it_flags = flags | MDS_OPEN_BY_FID,
+ };
int rc = 0;
lsm = ccc_inode_lsm_get(inode);
@@ -1314,13 +1317,12 @@ int ll_lov_setstripe_ea_info(struct inode *inode, struct dentry *dentry,
}
ll_inode_size_lock(inode);
- oit.it_flags |= MDS_OPEN_BY_FID;
rc = ll_intent_file_open(dentry, lum, lum_size, &oit);
- if (rc)
+ if (rc < 0)
goto out_unlock;
rc = oit.it_status;
if (rc < 0)
- goto out_req_free;
+ goto out_unlock;
ll_release_openhandle(inode, &oit);
@@ -1330,9 +1332,6 @@ out_unlock:
ccc_inode_lsm_put(inode, lsm);
out:
return rc;
-out_req_free:
- ptlrpc_req_finished((struct ptlrpc_request *)oit.it_request);
- goto out;
}
int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 099/124] staging: lustre: llite: unlock inode size in ll_lov_setstripe_ea_info()
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
In ll_lov_setstripe_ea_info() release the inode size lock on all
appropriate exit paths.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6059
Reviewed-on: http://review.whamcloud.com/13167
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/file.c | 13 ++++++-------
1 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index 2cb0b65..ce96c9f 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -1301,7 +1301,10 @@ int ll_lov_setstripe_ea_info(struct inode *inode, struct dentry *dentry,
int lum_size)
{
struct lov_stripe_md *lsm = NULL;
- struct lookup_intent oit = {.it_op = IT_OPEN, .it_flags = flags};
+ struct lookup_intent oit = {
+ .it_op = IT_OPEN,
+ .it_flags = flags | MDS_OPEN_BY_FID,
+ };
int rc = 0;
lsm = ccc_inode_lsm_get(inode);
@@ -1314,13 +1317,12 @@ int ll_lov_setstripe_ea_info(struct inode *inode, struct dentry *dentry,
}
ll_inode_size_lock(inode);
- oit.it_flags |= MDS_OPEN_BY_FID;
rc = ll_intent_file_open(dentry, lum, lum_size, &oit);
- if (rc)
+ if (rc < 0)
goto out_unlock;
rc = oit.it_status;
if (rc < 0)
- goto out_req_free;
+ goto out_unlock;
ll_release_openhandle(inode, &oit);
@@ -1330,9 +1332,6 @@ out_unlock:
ccc_inode_lsm_put(inode, lsm);
out:
return rc;
-out_req_free:
- ptlrpc_req_finished((struct ptlrpc_request *)oit.it_request);
- goto out;
}
int ll_lov_getstripe_ea_info(struct inode *inode, const char *filename,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 100/124] staging: lustre: obd: change type of cl_conn_count to size_t
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change type of cl_conn_count to size_t.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/13125
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lib.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 5f0f1e3..986c6e9 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -202,7 +202,7 @@ struct client_obd {
struct rw_semaphore cl_sem;
struct obd_uuid cl_target_uuid;
struct obd_import *cl_import; /* ptlrpc connection state */
- int cl_conn_count;
+ size_t cl_conn_count;
/* max_mds_easize is purely a performance thing so we don't have to
* call obd_size_diskmd() all the time.
*/
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
index 9a5f56a..153e990 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
@@ -536,7 +536,7 @@ int client_disconnect_export(struct obd_export *exp)
imp = cli->cl_import;
down_write(&cli->cl_sem);
- CDEBUG(D_INFO, "disconnect %s - %d\n", obd->obd_name,
+ CDEBUG(D_INFO, "disconnect %s - %zu\n", obd->obd_name,
cli->cl_conn_count);
if (!cli->cl_conn_count) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 100/124] staging: lustre: obd: change type of cl_conn_count to size_t
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Dmitry Eremin, James Simmons
From: Dmitry Eremin <dmitry.eremin@intel.com>
Change type of cl_conn_count to size_t.
Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5577
Reviewed-on: http://review.whamcloud.com/13125
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 2 +-
drivers/staging/lustre/lustre/ldlm/ldlm_lib.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 5f0f1e3..986c6e9 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -202,7 +202,7 @@ struct client_obd {
struct rw_semaphore cl_sem;
struct obd_uuid cl_target_uuid;
struct obd_import *cl_import; /* ptlrpc connection state */
- int cl_conn_count;
+ size_t cl_conn_count;
/* max_mds_easize is purely a performance thing so we don't have to
* call obd_size_diskmd() all the time.
*/
diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
index 9a5f56a..153e990 100644
--- a/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
+++ b/drivers/staging/lustre/lustre/ldlm/ldlm_lib.c
@@ -536,7 +536,7 @@ int client_disconnect_export(struct obd_export *exp)
imp = cli->cl_import;
down_write(&cli->cl_sem);
- CDEBUG(D_INFO, "disconnect %s - %d\n", obd->obd_name,
+ CDEBUG(D_INFO, "disconnect %s - %zu\n", obd->obd_name,
cli->cl_conn_count);
if (!cli->cl_conn_count) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 101/124] staging: lustre: libcfs: check mask returned by cpumask_of_node
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Liang Zhen,
James Simmons
From: Liang Zhen <liang.zhen@intel.com>
cpumask_of_node can return NULL if NUMA node is unavailable,
in this case cfs_node_to_cpumask will try to copy from NULL
and cause kernel panic.
Signed-off-by: Liang Zhen <liang.zhen@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5751
Reviewed-on: http://review.whamcloud.com/13207
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lnet/libcfs/linux/linux-cpu.c | 17 ++++++++++++++---
1 files changed, 14 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c b/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c
index b52518c..e8b1a61 100644
--- a/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c
+++ b/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c
@@ -74,6 +74,17 @@ struct cfs_cpt_data {
static struct cfs_cpt_data cpt_data;
+static void
+cfs_node_to_cpumask(int node, cpumask_t *mask)
+{
+ const cpumask_t *tmp = cpumask_of_node(node);
+
+ if (tmp)
+ cpumask_copy(mask, tmp);
+ else
+ cpumask_clear(mask);
+}
+
void
cfs_cpt_table_free(struct cfs_cpt_table *cptab)
{
@@ -403,7 +414,7 @@ cfs_cpt_set_node(struct cfs_cpt_table *cptab, int cpt, int node)
mutex_lock(&cpt_data.cpt_mutex);
mask = cpt_data.cpt_cpumask;
- cpumask_copy(mask, cpumask_of_node(node));
+ cfs_node_to_cpumask(node, mask);
rc = cfs_cpt_set_cpumask(cptab, cpt, mask);
@@ -427,7 +438,7 @@ cfs_cpt_unset_node(struct cfs_cpt_table *cptab, int cpt, int node)
mutex_lock(&cpt_data.cpt_mutex);
mask = cpt_data.cpt_cpumask;
- cpumask_copy(mask, cpumask_of_node(node));
+ cfs_node_to_cpumask(node, mask);
cfs_cpt_unset_cpumask(cptab, cpt, mask);
@@ -749,7 +760,7 @@ cfs_cpt_table_create(int ncpt)
}
for_each_online_node(i) {
- cpumask_copy(mask, cpumask_of_node(i));
+ cfs_node_to_cpumask(i, mask);
while (!cpumask_empty(mask)) {
struct cfs_cpu_partition *part;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 101/124] staging: lustre: libcfs: check mask returned by cpumask_of_node
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Liang Zhen,
James Simmons
From: Liang Zhen <liang.zhen@intel.com>
cpumask_of_node can return NULL if NUMA node is unavailable,
in this case cfs_node_to_cpumask will try to copy from NULL
and cause kernel panic.
Signed-off-by: Liang Zhen <liang.zhen@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5751
Reviewed-on: http://review.whamcloud.com/13207
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lnet/libcfs/linux/linux-cpu.c | 17 ++++++++++++++---
1 files changed, 14 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c b/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c
index b52518c..e8b1a61 100644
--- a/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c
+++ b/drivers/staging/lustre/lnet/libcfs/linux/linux-cpu.c
@@ -74,6 +74,17 @@ struct cfs_cpt_data {
static struct cfs_cpt_data cpt_data;
+static void
+cfs_node_to_cpumask(int node, cpumask_t *mask)
+{
+ const cpumask_t *tmp = cpumask_of_node(node);
+
+ if (tmp)
+ cpumask_copy(mask, tmp);
+ else
+ cpumask_clear(mask);
+}
+
void
cfs_cpt_table_free(struct cfs_cpt_table *cptab)
{
@@ -403,7 +414,7 @@ cfs_cpt_set_node(struct cfs_cpt_table *cptab, int cpt, int node)
mutex_lock(&cpt_data.cpt_mutex);
mask = cpt_data.cpt_cpumask;
- cpumask_copy(mask, cpumask_of_node(node));
+ cfs_node_to_cpumask(node, mask);
rc = cfs_cpt_set_cpumask(cptab, cpt, mask);
@@ -427,7 +438,7 @@ cfs_cpt_unset_node(struct cfs_cpt_table *cptab, int cpt, int node)
mutex_lock(&cpt_data.cpt_mutex);
mask = cpt_data.cpt_cpumask;
- cpumask_copy(mask, cpumask_of_node(node));
+ cfs_node_to_cpumask(node, mask);
cfs_cpt_unset_cpumask(cptab, cpt, mask);
@@ -749,7 +760,7 @@ cfs_cpt_table_create(int ncpt)
}
for_each_online_node(i) {
- cpumask_copy(mask, cpumask_of_node(i));
+ cfs_node_to_cpumask(i, mask);
while (!cpumask_empty(mask)) {
struct cfs_cpu_partition *part;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 102/124] staging: lustre: remove lustre/include/linux/
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Merge the contents of lustre/include/linux/lvfs.h into
lustre/include/lvfs.h. Merge lustre/include/linux/lustre_user.h into
lustre/include/lustre/lustre_user.h. Move lustre_compat25.h and
lustre_patchless_compat.h from lustre/include/linux/ to
lustre/include/ and rename lustre_compat25.h to lustre_compat.h.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/13271
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 2 +-
.../lustre/lustre/include/linux/lustre_user.h | 66 --------------------
.../lustre/lustre/include/lustre/lustre_user.h | 27 ++++++++-
.../{linux/lustre_compat25.h => lustre_compat.h} | 6 +-
.../include/{linux => }/lustre_patchless_compat.h | 0
.../staging/lustre/lustre/include/obd_support.h | 2 +-
drivers/staging/lustre/lustre/llite/file.c | 2 +-
.../staging/lustre/lustre/llite/llite_internal.h | 3 +
drivers/staging/lustre/lustre/llite/llite_mmap.c | 1 -
drivers/staging/lustre/lustre/llite/rw.c | 1 -
drivers/staging/lustre/lustre/llite/rw26.c | 1 -
drivers/staging/lustre/lustre/obdclass/obd_mount.c | 2 +-
12 files changed, 36 insertions(+), 77 deletions(-)
delete mode 100644 drivers/staging/lustre/lustre/include/linux/lustre_user.h
rename drivers/staging/lustre/lustre/include/{linux/lustre_compat25.h => lustre_compat.h} (96%)
rename drivers/staging/lustre/lustre/include/{linux => }/lustre_patchless_compat.h (100%)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index c1bcff8..409a324 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -93,8 +93,8 @@
* super-class definitions.
*/
#include "lu_object.h"
+#include "lustre_compat.h"
#include <linux/atomic.h>
-#include "linux/lustre_compat25.h"
#include <linux/mutex.h>
#include <linux/radix-tree.h>
#include <linux/spinlock.h>
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_user.h b/drivers/staging/lustre/lustre/include/linux/lustre_user.h
deleted file mode 100644
index e967950..0000000
--- a/drivers/staging/lustre/lustre/include/linux/lustre_user.h
+++ /dev/null
@@ -1,66 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.gnu.org/licenses/gpl-2.0.html
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2012, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- *
- * lustre/include/linux/lustre_user.h
- *
- * Lustre public user-space interface definitions.
- */
-
-#ifndef _LINUX_LUSTRE_USER_H
-#define _LINUX_LUSTRE_USER_H
-
-# include <linux/quota.h>
-
-/*
- * asm-x86_64/processor.h on some SLES 9 distros seems to use
- * kernel-only typedefs. fortunately skipping it altogether is ok
- * (for now).
- */
-#define __ASM_X86_64_PROCESSOR_H
-
-#include <linux/string.h>
-
-/*
- * We need to always use 64bit version because the structure
- * is shared across entire cluster where 32bit and 64bit machines
- * are co-existing.
- */
-#if __BITS_PER_LONG != 64 || defined(__ARCH_WANT_STAT64)
-typedef struct stat64 lstat_t;
-#define lstat_f lstat64
-#else
-typedef struct stat lstat_t;
-#define lstat_f lstat
-#endif
-
-#define HAVE_LOV_USER_MDS_DATA
-
-#endif /* _LUSTRE_USER_H */
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index ba679ad..6fc9855 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -42,8 +42,33 @@
* @{
*/
+#ifdef __KERNEL__
+# include <linux/quota.h>
+# include <linux/string.h> /* snprintf() */
+# include <linux/version.h>
+#else /* !__KERNEL__ */
+# define NEED_QUOTA_DEFS
+# include <stdio.h> /* snprintf() */
+# include <string.h>
+# include <sys/quota.h>
+# include <sys/stat.h>
+#endif /* __KERNEL__ */
#include "ll_fiemap.h"
-#include "../linux/lustre_user.h"
+
+/*
+ * We need to always use 64bit version because the structure
+ * is shared across entire cluster where 32bit and 64bit machines
+ * are co-existing.
+ */
+#if __BITS_PER_LONG != 64 || defined(__ARCH_WANT_STAT64)
+typedef struct stat64 lstat_t;
+#define lstat_f lstat64
+#else
+typedef struct stat lstat_t;
+#define lstat_f lstat
+#endif
+
+#define HAVE_LOV_USER_MDS_DATA
#define LUSTRE_EOF 0xffffffffffffffffULL
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h b/drivers/staging/lustre/lustre/include/lustre_compat.h
similarity index 96%
rename from drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
rename to drivers/staging/lustre/lustre/include/lustre_compat.h
index 1eb64ec..567c438 100644
--- a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
+++ b/drivers/staging/lustre/lustre/include/lustre_compat.h
@@ -30,8 +30,8 @@
* Lustre is a trademark of Sun Microsystems, Inc.
*/
-#ifndef _LINUX_COMPAT25_H
-#define _LINUX_COMPAT25_H
+#ifndef _LUSTRE_COMPAT_H
+#define _LUSTRE_COMPAT_H
#include <linux/fs_struct.h>
#include <linux/namei.h>
@@ -74,4 +74,4 @@
# define ext2_find_next_zero_bit find_next_zero_bit_le
#endif
-#endif /* _COMPAT25_H */
+#endif /* _LUSTRE_COMPAT_H */
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h b/drivers/staging/lustre/lustre/include/lustre_patchless_compat.h
similarity index 100%
rename from drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h
rename to drivers/staging/lustre/lustre/include/lustre_patchless_compat.h
diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h
index 3144cca..b346a7f 100644
--- a/drivers/staging/lustre/lustre/include/obd_support.h
+++ b/drivers/staging/lustre/lustre/include/obd_support.h
@@ -35,7 +35,7 @@
#include <linux/slab.h>
#include "../../include/linux/libcfs/libcfs.h"
-#include "linux/lustre_compat25.h"
+#include "lustre_compat.h"
#include "lprocfs_status.h"
/* global variables */
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index ce96c9f..e8963fe 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -42,11 +42,11 @@
#include <linux/file.h>
#include <linux/sched.h>
#include <linux/mount.h>
-#include "llite_internal.h"
#include "../include/lustre/ll_fiemap.h"
#include "../include/lustre/lustre_ioctl.h"
#include "../include/cl_object.h"
+#include "llite_internal.h"
static int
ll_put_grouplock(struct inode *inode, struct file *file, unsigned long arg);
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 05f2aa3..e5f1967 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -39,11 +39,14 @@
#include "../include/lustre_linkea.h"
/* for struct cl_lock_descr and struct cl_io */
+#include "../include/lustre_patchless_compat.h"
+#include "../include/lustre_compat.h"
#include "../include/cl_object.h"
#include "../include/lustre_lmv.h"
#include "../include/lustre_mdc.h"
#include "../include/lustre_intent.h"
#include <linux/compat.h>
+#include <linux/namei.h>
#include <linux/xattr.h>
#include <linux/posix_acl_xattr.h>
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/llite_mmap.c b/drivers/staging/lustre/lustre/llite/llite_mmap.c
index 0cd25bf..4366918 100644
--- a/drivers/staging/lustre/lustre/llite/llite_mmap.c
+++ b/drivers/staging/lustre/lustre/llite/llite_mmap.c
@@ -44,7 +44,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "llite_internal.h"
-#include "../include/linux/lustre_compat25.h"
static const struct vm_operations_struct ll_file_vm_ops;
diff --git a/drivers/staging/lustre/lustre/llite/rw.c b/drivers/staging/lustre/lustre/llite/rw.c
index 508cb46..50c0152 100644
--- a/drivers/staging/lustre/lustre/llite/rw.c
+++ b/drivers/staging/lustre/lustre/llite/rw.c
@@ -52,7 +52,6 @@
#include "../include/obd_cksum.h"
#include "llite_internal.h"
-#include "../include/linux/lustre_compat25.h"
static void ll_ra_stats_inc_sbi(struct ll_sb_info *sbi, enum ra_stat which);
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index df0318e..26f3a37 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -52,7 +52,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "llite_internal.h"
-#include "../include/linux/lustre_compat25.h"
/**
* Implements Linux VM address_space::invalidatepage() method. This method is
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_mount.c b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
index 0df1aa6..0d3a3b0 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_mount.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
@@ -41,7 +41,7 @@
#define PRINT_CMD CDEBUG
#include "../include/obd.h"
-#include "../include/linux/lustre_compat25.h"
+#include "../include/lustre_compat.h"
#include "../include/obd_class.h"
#include "../include/lustre/lustre_user.h"
#include "../include/lustre_log.h"
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 102/124] staging: lustre: remove lustre/include/linux/
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
John L. Hammond, James Simmons
From: John L. Hammond <john.hammond@intel.com>
Merge the contents of lustre/include/linux/lvfs.h into
lustre/include/lvfs.h. Merge lustre/include/linux/lustre_user.h into
lustre/include/lustre/lustre_user.h. Move lustre_compat25.h and
lustre_patchless_compat.h from lustre/include/linux/ to
lustre/include/ and rename lustre_compat25.h to lustre_compat.h.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2675
Reviewed-on: http://review.whamcloud.com/13271
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 2 +-
.../lustre/lustre/include/linux/lustre_user.h | 66 --------------------
.../lustre/lustre/include/lustre/lustre_user.h | 27 ++++++++-
.../{linux/lustre_compat25.h => lustre_compat.h} | 6 +-
.../include/{linux => }/lustre_patchless_compat.h | 0
.../staging/lustre/lustre/include/obd_support.h | 2 +-
drivers/staging/lustre/lustre/llite/file.c | 2 +-
.../staging/lustre/lustre/llite/llite_internal.h | 3 +
drivers/staging/lustre/lustre/llite/llite_mmap.c | 1 -
drivers/staging/lustre/lustre/llite/rw.c | 1 -
drivers/staging/lustre/lustre/llite/rw26.c | 1 -
drivers/staging/lustre/lustre/obdclass/obd_mount.c | 2 +-
12 files changed, 36 insertions(+), 77 deletions(-)
delete mode 100644 drivers/staging/lustre/lustre/include/linux/lustre_user.h
rename drivers/staging/lustre/lustre/include/{linux/lustre_compat25.h => lustre_compat.h} (96%)
rename drivers/staging/lustre/lustre/include/{linux => }/lustre_patchless_compat.h (100%)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index c1bcff8..409a324 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -93,8 +93,8 @@
* super-class definitions.
*/
#include "lu_object.h"
+#include "lustre_compat.h"
#include <linux/atomic.h>
-#include "linux/lustre_compat25.h"
#include <linux/mutex.h>
#include <linux/radix-tree.h>
#include <linux/spinlock.h>
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_user.h b/drivers/staging/lustre/lustre/include/linux/lustre_user.h
deleted file mode 100644
index e967950..0000000
--- a/drivers/staging/lustre/lustre/include/linux/lustre_user.h
+++ /dev/null
@@ -1,66 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.gnu.org/licenses/gpl-2.0.html
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2012, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- *
- * lustre/include/linux/lustre_user.h
- *
- * Lustre public user-space interface definitions.
- */
-
-#ifndef _LINUX_LUSTRE_USER_H
-#define _LINUX_LUSTRE_USER_H
-
-# include <linux/quota.h>
-
-/*
- * asm-x86_64/processor.h on some SLES 9 distros seems to use
- * kernel-only typedefs. fortunately skipping it altogether is ok
- * (for now).
- */
-#define __ASM_X86_64_PROCESSOR_H
-
-#include <linux/string.h>
-
-/*
- * We need to always use 64bit version because the structure
- * is shared across entire cluster where 32bit and 64bit machines
- * are co-existing.
- */
-#if __BITS_PER_LONG != 64 || defined(__ARCH_WANT_STAT64)
-typedef struct stat64 lstat_t;
-#define lstat_f lstat64
-#else
-typedef struct stat lstat_t;
-#define lstat_f lstat
-#endif
-
-#define HAVE_LOV_USER_MDS_DATA
-
-#endif /* _LUSTRE_USER_H */
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
index ba679ad..6fc9855 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_user.h
@@ -42,8 +42,33 @@
* @{
*/
+#ifdef __KERNEL__
+# include <linux/quota.h>
+# include <linux/string.h> /* snprintf() */
+# include <linux/version.h>
+#else /* !__KERNEL__ */
+# define NEED_QUOTA_DEFS
+# include <stdio.h> /* snprintf() */
+# include <string.h>
+# include <sys/quota.h>
+# include <sys/stat.h>
+#endif /* __KERNEL__ */
#include "ll_fiemap.h"
-#include "../linux/lustre_user.h"
+
+/*
+ * We need to always use 64bit version because the structure
+ * is shared across entire cluster where 32bit and 64bit machines
+ * are co-existing.
+ */
+#if __BITS_PER_LONG != 64 || defined(__ARCH_WANT_STAT64)
+typedef struct stat64 lstat_t;
+#define lstat_f lstat64
+#else
+typedef struct stat lstat_t;
+#define lstat_f lstat
+#endif
+
+#define HAVE_LOV_USER_MDS_DATA
#define LUSTRE_EOF 0xffffffffffffffffULL
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h b/drivers/staging/lustre/lustre/include/lustre_compat.h
similarity index 96%
rename from drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
rename to drivers/staging/lustre/lustre/include/lustre_compat.h
index 1eb64ec..567c438 100644
--- a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
+++ b/drivers/staging/lustre/lustre/include/lustre_compat.h
@@ -30,8 +30,8 @@
* Lustre is a trademark of Sun Microsystems, Inc.
*/
-#ifndef _LINUX_COMPAT25_H
-#define _LINUX_COMPAT25_H
+#ifndef _LUSTRE_COMPAT_H
+#define _LUSTRE_COMPAT_H
#include <linux/fs_struct.h>
#include <linux/namei.h>
@@ -74,4 +74,4 @@
# define ext2_find_next_zero_bit find_next_zero_bit_le
#endif
-#endif /* _COMPAT25_H */
+#endif /* _LUSTRE_COMPAT_H */
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h b/drivers/staging/lustre/lustre/include/lustre_patchless_compat.h
similarity index 100%
rename from drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h
rename to drivers/staging/lustre/lustre/include/lustre_patchless_compat.h
diff --git a/drivers/staging/lustre/lustre/include/obd_support.h b/drivers/staging/lustre/lustre/include/obd_support.h
index 3144cca..b346a7f 100644
--- a/drivers/staging/lustre/lustre/include/obd_support.h
+++ b/drivers/staging/lustre/lustre/include/obd_support.h
@@ -35,7 +35,7 @@
#include <linux/slab.h>
#include "../../include/linux/libcfs/libcfs.h"
-#include "linux/lustre_compat25.h"
+#include "lustre_compat.h"
#include "lprocfs_status.h"
/* global variables */
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index ce96c9f..e8963fe 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -42,11 +42,11 @@
#include <linux/file.h>
#include <linux/sched.h>
#include <linux/mount.h>
-#include "llite_internal.h"
#include "../include/lustre/ll_fiemap.h"
#include "../include/lustre/lustre_ioctl.h"
#include "../include/cl_object.h"
+#include "llite_internal.h"
static int
ll_put_grouplock(struct inode *inode, struct file *file, unsigned long arg);
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 05f2aa3..e5f1967 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -39,11 +39,14 @@
#include "../include/lustre_linkea.h"
/* for struct cl_lock_descr and struct cl_io */
+#include "../include/lustre_patchless_compat.h"
+#include "../include/lustre_compat.h"
#include "../include/cl_object.h"
#include "../include/lustre_lmv.h"
#include "../include/lustre_mdc.h"
#include "../include/lustre_intent.h"
#include <linux/compat.h>
+#include <linux/namei.h>
#include <linux/xattr.h>
#include <linux/posix_acl_xattr.h>
#include "vvp_internal.h"
diff --git a/drivers/staging/lustre/lustre/llite/llite_mmap.c b/drivers/staging/lustre/lustre/llite/llite_mmap.c
index 0cd25bf..4366918 100644
--- a/drivers/staging/lustre/lustre/llite/llite_mmap.c
+++ b/drivers/staging/lustre/lustre/llite/llite_mmap.c
@@ -44,7 +44,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "llite_internal.h"
-#include "../include/linux/lustre_compat25.h"
static const struct vm_operations_struct ll_file_vm_ops;
diff --git a/drivers/staging/lustre/lustre/llite/rw.c b/drivers/staging/lustre/lustre/llite/rw.c
index 508cb46..50c0152 100644
--- a/drivers/staging/lustre/lustre/llite/rw.c
+++ b/drivers/staging/lustre/lustre/llite/rw.c
@@ -52,7 +52,6 @@
#include "../include/obd_cksum.h"
#include "llite_internal.h"
-#include "../include/linux/lustre_compat25.h"
static void ll_ra_stats_inc_sbi(struct ll_sb_info *sbi, enum ra_stat which);
diff --git a/drivers/staging/lustre/lustre/llite/rw26.c b/drivers/staging/lustre/lustre/llite/rw26.c
index df0318e..26f3a37 100644
--- a/drivers/staging/lustre/lustre/llite/rw26.c
+++ b/drivers/staging/lustre/lustre/llite/rw26.c
@@ -52,7 +52,6 @@
#define DEBUG_SUBSYSTEM S_LLITE
#include "llite_internal.h"
-#include "../include/linux/lustre_compat25.h"
/**
* Implements Linux VM address_space::invalidatepage() method. This method is
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_mount.c b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
index 0df1aa6..0d3a3b0 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_mount.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_mount.c
@@ -41,7 +41,7 @@
#define PRINT_CMD CDEBUG
#include "../include/obd.h"
-#include "../include/linux/lustre_compat25.h"
+#include "../include/lustre_compat.h"
#include "../include/obd_class.h"
#include "../include/lustre/lustre_user.h"
#include "../include/lustre_log.h"
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 103/124] staging: lustre: llite: pack suppgid to MDS correctly
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Fan Yong,
James Simmons
From: Fan Yong <fan.yong@intel.com>
The ll_lookup_it() may trigger IT_OPEN RPC to open a file by name.
But at that time, the client does not know the target file's GID,
so it cannot pack the necessary supplementary group ID in the RPC.
Because of missing the supplementary group ID, the RPC maybe fail
for open permission check on the MDS. Under such case, MDS should
return the target file's GID, if the current thread on the client
in the right group (according to the file's GID), the client will
try the IT_OPEN RPC again with the right supplementary group ID.
This patch is also helpful if some other(s) changed the file's GID
after current RPC sent to the MDS with the suppgid as the original
GID by race.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5423
Reviewed-on: http://review.whamcloud.com/12476
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 1 +
drivers/staging/lustre/lustre/llite/namei.c | 36 ++++++++++++++++++-
2 files changed, 35 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index acb2806..72eaee9 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1987,6 +1987,7 @@ void lustre_swab_generic_32s(__u32 *val);
#define DISP_OPEN_LOCK 0x02000000
#define DISP_OPEN_LEASE 0x04000000
#define DISP_OPEN_STRIPE 0x08000000
+#define DISP_OPEN_DENY 0x10000000
/* INODE LOCK PARTS */
#define MDS_INODELOCK_LOOKUP 0x000001 /* For namespace, dentry etc, and also
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 1dd5df5..3960893 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -505,8 +505,8 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
struct lookup_intent lookup_it = { .it_op = IT_LOOKUP };
struct dentry *save = dentry, *retval;
struct ptlrpc_request *req = NULL;
+ struct md_op_data *op_data = NULL;
struct inode *inode;
- struct md_op_data *op_data;
__u32 opc;
int rc;
@@ -549,7 +549,36 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
rc = md_intent_lock(ll_i2mdexp(parent), op_data, it, &req,
&ll_md_blocking_ast, 0);
- ll_finish_md_op_data(op_data);
+ /*
+ * If the MDS allows the client to chgrp (CFS_SETGRP_PERM), but the
+ * client does not know which suppgid should be sent to the MDS, or
+ * some other(s) changed the target file's GID after this RPC sent
+ * to the MDS with the suppgid as the original GID, then we should
+ * try again with right suppgid.
+ */
+ if (rc == -EACCES && it->it_op & IT_OPEN &&
+ it_disposition(it, DISP_OPEN_DENY)) {
+ struct mdt_body *body;
+
+ LASSERT(req);
+
+ body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY);
+ if (op_data->op_suppgids[0] == body->mbo_gid ||
+ op_data->op_suppgids[1] == body->mbo_gid ||
+ !in_group_p(make_kgid(&init_user_ns, body->mbo_gid))) {
+ retval = ERR_PTR(-EACCES);
+ goto out;
+ }
+
+ fid_zero(&op_data->op_fid2);
+ op_data->op_suppgids[1] = body->mbo_gid;
+ ptlrpc_req_finished(req);
+ req = NULL;
+ ll_intent_release(it);
+ rc = md_intent_lock(ll_i2mdexp(parent), op_data, it, &req,
+ ll_md_blocking_ast, 0);
+ }
+
if (rc < 0) {
retval = ERR_PTR(rc);
goto out;
@@ -575,6 +604,9 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
else
retval = dentry;
out:
+ if (op_data && !IS_ERR(op_data))
+ ll_finish_md_op_data(op_data);
+
ptlrpc_req_finished(req);
return retval;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 103/124] staging: lustre: llite: pack suppgid to MDS correctly
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Fan Yong,
James Simmons
From: Fan Yong <fan.yong@intel.com>
The ll_lookup_it() may trigger IT_OPEN RPC to open a file by name.
But at that time, the client does not know the target file's GID,
so it cannot pack the necessary supplementary group ID in the RPC.
Because of missing the supplementary group ID, the RPC maybe fail
for open permission check on the MDS. Under such case, MDS should
return the target file's GID, if the current thread on the client
in the right group (according to the file's GID), the client will
try the IT_OPEN RPC again with the right supplementary group ID.
This patch is also helpful if some other(s) changed the file's GID
after current RPC sent to the MDS with the suppgid as the original
GID by race.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5423
Reviewed-on: http://review.whamcloud.com/12476
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 1 +
drivers/staging/lustre/lustre/llite/namei.c | 36 ++++++++++++++++++-
2 files changed, 35 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index acb2806..72eaee9 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1987,6 +1987,7 @@ void lustre_swab_generic_32s(__u32 *val);
#define DISP_OPEN_LOCK 0x02000000
#define DISP_OPEN_LEASE 0x04000000
#define DISP_OPEN_STRIPE 0x08000000
+#define DISP_OPEN_DENY 0x10000000
/* INODE LOCK PARTS */
#define MDS_INODELOCK_LOOKUP 0x000001 /* For namespace, dentry etc, and also
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 1dd5df5..3960893 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -505,8 +505,8 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
struct lookup_intent lookup_it = { .it_op = IT_LOOKUP };
struct dentry *save = dentry, *retval;
struct ptlrpc_request *req = NULL;
+ struct md_op_data *op_data = NULL;
struct inode *inode;
- struct md_op_data *op_data;
__u32 opc;
int rc;
@@ -549,7 +549,36 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
rc = md_intent_lock(ll_i2mdexp(parent), op_data, it, &req,
&ll_md_blocking_ast, 0);
- ll_finish_md_op_data(op_data);
+ /*
+ * If the MDS allows the client to chgrp (CFS_SETGRP_PERM), but the
+ * client does not know which suppgid should be sent to the MDS, or
+ * some other(s) changed the target file's GID after this RPC sent
+ * to the MDS with the suppgid as the original GID, then we should
+ * try again with right suppgid.
+ */
+ if (rc == -EACCES && it->it_op & IT_OPEN &&
+ it_disposition(it, DISP_OPEN_DENY)) {
+ struct mdt_body *body;
+
+ LASSERT(req);
+
+ body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY);
+ if (op_data->op_suppgids[0] == body->mbo_gid ||
+ op_data->op_suppgids[1] == body->mbo_gid ||
+ !in_group_p(make_kgid(&init_user_ns, body->mbo_gid))) {
+ retval = ERR_PTR(-EACCES);
+ goto out;
+ }
+
+ fid_zero(&op_data->op_fid2);
+ op_data->op_suppgids[1] = body->mbo_gid;
+ ptlrpc_req_finished(req);
+ req = NULL;
+ ll_intent_release(it);
+ rc = md_intent_lock(ll_i2mdexp(parent), op_data, it, &req,
+ ll_md_blocking_ast, 0);
+ }
+
if (rc < 0) {
retval = ERR_PTR(rc);
goto out;
@@ -575,6 +604,9 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
else
retval = dentry;
out:
+ if (op_data && !IS_ERR(op_data))
+ ll_finish_md_op_data(op_data);
+
ptlrpc_req_finished(req);
return retval;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 104/124] staging: lustre: clio: rename coo_attr_set to coo_attr_update
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
coo_attr_set() is used to update object's attribute but its name
makes confusion that people intuitively think that it is used to
pass object's attribute down to server sides.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1154
Reviewed-on: http://review.whamcloud.com/12888
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 8 ++++----
drivers/staging/lustre/lustre/llite/vvp_io.c | 2 +-
drivers/staging/lustre/lustre/llite/vvp_object.c | 6 +++---
drivers/staging/lustre/lustre/lov/lov_object.c | 6 +++---
drivers/staging/lustre/lustre/lov/lovsub_object.c | 6 +++---
drivers/staging/lustre/lustre/obdclass/cl_object.c | 17 +++++++++--------
drivers/staging/lustre/lustre/osc/osc_io.c | 9 +++++----
drivers/staging/lustre/lustre/osc/osc_lock.c | 4 ++--
drivers/staging/lustre/lustre/osc/osc_object.c | 6 +++---
drivers/staging/lustre/lustre/osc/osc_request.c | 2 +-
10 files changed, 34 insertions(+), 32 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 409a324..9403c0c 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -369,8 +369,8 @@ struct cl_object_operations {
* \return the same convention as for
* cl_object_operations::coo_attr_get() is used.
*/
- int (*coo_attr_set)(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid);
+ int (*coo_attr_update)(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid);
/**
* Update object configuration. Called top-to-bottom to modify object
* configuration.
@@ -2171,8 +2171,8 @@ void cl_object_attr_lock(struct cl_object *o);
void cl_object_attr_unlock(struct cl_object *o);
int cl_object_attr_get(const struct lu_env *env, struct cl_object *obj,
struct cl_attr *attr);
-int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid);
+int cl_object_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid);
int cl_object_glimpse(const struct lu_env *env, struct cl_object *obj,
struct ost_lvb *lvb);
int cl_conf_set(const struct lu_env *env, struct cl_object *obj,
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index 27fc0b0..2ab4503 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -627,7 +627,7 @@ static int vvp_io_setattr_time(const struct lu_env *env,
attr->cat_mtime = io->u.ci_setattr.sa_attr.lvb_mtime;
valid |= CAT_MTIME;
}
- result = cl_object_attr_set(env, obj, attr, valid);
+ result = cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
return result;
diff --git a/drivers/staging/lustre/lustre/llite/vvp_object.c b/drivers/staging/lustre/lustre/llite/vvp_object.c
index 6b6a63d..b57195d 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_object.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_object.c
@@ -101,8 +101,8 @@ static int vvp_attr_get(const struct lu_env *env, struct cl_object *obj,
return 0; /* layers below have to fill in the rest */
}
-static int vvp_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int vvp_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
struct inode *inode = vvp_object_inode(obj);
@@ -209,7 +209,7 @@ static const struct cl_object_operations vvp_ops = {
.coo_lock_init = vvp_lock_init,
.coo_io_init = vvp_io_init,
.coo_attr_get = vvp_attr_get,
- .coo_attr_set = vvp_attr_set,
+ .coo_attr_update = vvp_attr_update,
.coo_conf_set = vvp_conf_set,
.coo_prune = vvp_prune,
.coo_glimpse = vvp_object_glimpse
diff --git a/drivers/staging/lustre/lustre/lov/lov_object.c b/drivers/staging/lustre/lustre/lov/lov_object.c
index 01f95a9..52f7363 100644
--- a/drivers/staging/lustre/lustre/lov/lov_object.c
+++ b/drivers/staging/lustre/lustre/lov/lov_object.c
@@ -894,8 +894,8 @@ static int lov_attr_get(const struct lu_env *env, struct cl_object *obj,
return LOV_2DISPATCH_NOLOCK(cl2lov(obj), llo_getattr, env, obj, attr);
}
-static int lov_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int lov_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
/*
* No dispatch is required here, as no layout implements this.
@@ -932,7 +932,7 @@ static const struct cl_object_operations lov_ops = {
.coo_lock_init = lov_lock_init,
.coo_io_init = lov_io_init,
.coo_attr_get = lov_attr_get,
- .coo_attr_set = lov_attr_set,
+ .coo_attr_update = lov_attr_update,
.coo_conf_set = lov_conf_set,
.coo_getstripe = lov_object_getstripe
};
diff --git a/drivers/staging/lustre/lustre/lov/lovsub_object.c b/drivers/staging/lustre/lustre/lov/lovsub_object.c
index fb2f266..a2bac7a 100644
--- a/drivers/staging/lustre/lustre/lov/lovsub_object.c
+++ b/drivers/staging/lustre/lustre/lov/lovsub_object.c
@@ -98,8 +98,8 @@ static int lovsub_object_print(const struct lu_env *env, void *cookie,
return (*p)(env, cookie, "[%d]", los->lso_index);
}
-static int lovsub_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int lovsub_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
struct lov_object *lov = cl2lovsub(obj)->lso_super;
@@ -119,7 +119,7 @@ static int lovsub_object_glimpse(const struct lu_env *env,
static const struct cl_object_operations lovsub_ops = {
.coo_page_init = lovsub_page_init,
.coo_lock_init = lovsub_lock_init,
- .coo_attr_set = lovsub_attr_set,
+ .coo_attr_update = lovsub_attr_update,
.coo_glimpse = lovsub_object_glimpse
};
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_object.c b/drivers/staging/lustre/lustre/obdclass/cl_object.c
index ef5b1aa..e493c80 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_object.c
@@ -163,7 +163,7 @@ static spinlock_t *cl_object_attr_guard(struct cl_object *o)
*
* Prevents data-attributes from changing, until lock is released by
* cl_object_attr_unlock(). This has to be called before calls to
- * cl_object_attr_get(), cl_object_attr_set().
+ * cl_object_attr_get(), cl_object_attr_update().
*/
void cl_object_attr_lock(struct cl_object *o)
__acquires(cl_object_attr_guard(o))
@@ -217,11 +217,11 @@ EXPORT_SYMBOL(cl_object_attr_get);
* Updates data-attributes of an object \a obj.
*
* Only attributes, mentioned in a validness bit-mask \a v are
- * updated. Calls cl_object_operations::coo_attr_set() on every layer, bottom
- * to top.
+ * updated. Calls cl_object_operations::coo_attr_update() on every layer,
+ * bottom to top.
*/
-int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned v)
+int cl_object_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int v)
{
struct lu_object_header *top;
int result;
@@ -231,8 +231,9 @@ int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
top = obj->co_lu.lo_header;
result = 0;
list_for_each_entry_reverse(obj, &top->loh_layers, co_lu.lo_linkage) {
- if (obj->co_ops->coo_attr_set) {
- result = obj->co_ops->coo_attr_set(env, obj, attr, v);
+ if (obj->co_ops->coo_attr_update) {
+ result = obj->co_ops->coo_attr_update(env, obj, attr,
+ v);
if (result != 0) {
if (result > 0)
result = 0;
@@ -242,7 +243,7 @@ int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
}
return result;
}
-EXPORT_SYMBOL(cl_object_attr_set);
+EXPORT_SYMBOL(cl_object_attr_update);
/**
* Notifies layers (bottom-to-top) that glimpse AST was received.
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 8bbb2a6..42e9761 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -233,7 +233,7 @@ static void osc_page_touch_at(const struct lu_env *env,
attr->cat_size = kms;
valid |= CAT_SIZE;
}
- cl_object_attr_set(env, obj, attr, valid);
+ cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
@@ -476,7 +476,8 @@ static int osc_io_setattr_start(const struct lu_env *env,
attr->cat_ctime = lvb->lvb_ctime;
cl_valid |= CAT_CTIME;
}
- result = cl_object_attr_set(env, obj, attr, cl_valid);
+ result = cl_object_attr_update(env, obj, attr,
+ cl_valid);
}
cl_object_attr_unlock(obj);
}
@@ -564,7 +565,7 @@ static int osc_io_read_start(const struct lu_env *env,
if (!slice->cis_io->ci_noatime) {
cl_object_attr_lock(obj);
attr->cat_atime = ktime_get_real_seconds();
- rc = cl_object_attr_set(env, obj, attr, CAT_ATIME);
+ rc = cl_object_attr_update(env, obj, attr, CAT_ATIME);
cl_object_attr_unlock(obj);
}
return rc;
@@ -581,7 +582,7 @@ static int osc_io_write_start(const struct lu_env *env,
cl_object_attr_lock(obj);
attr->cat_ctime = ktime_get_real_seconds();
attr->cat_mtime = attr->cat_ctime;
- rc = cl_object_attr_set(env, obj, attr, CAT_MTIME | CAT_CTIME);
+ rc = cl_object_attr_update(env, obj, attr, CAT_MTIME | CAT_CTIME);
cl_object_attr_unlock(obj);
return rc;
diff --git a/drivers/staging/lustre/lustre/osc/osc_lock.c b/drivers/staging/lustre/lustre/osc/osc_lock.c
index 717d3ff..39a8a58 100644
--- a/drivers/staging/lustre/lustre/osc/osc_lock.c
+++ b/drivers/staging/lustre/lustre/osc/osc_lock.c
@@ -222,7 +222,7 @@ static void osc_lock_lvb_update(const struct lu_env *env,
ldlm_lock_allow_match_locked(dlmlock);
}
- cl_object_attr_set(env, obj, attr, valid);
+ cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
@@ -467,7 +467,7 @@ static int osc_dlm_blocking_ast0(const struct lu_env *env,
*/
attr->cat_kms = ldlm_extent_shift_kms(dlmlock, old_kms);
- cl_object_attr_set(env, obj, attr, CAT_KMS);
+ cl_object_attr_update(env, obj, attr, CAT_KMS);
cl_object_attr_unlock(obj);
unlock_res_and_lock(dlmlock);
diff --git a/drivers/staging/lustre/lustre/osc/osc_object.c b/drivers/staging/lustre/lustre/osc/osc_object.c
index 5fc5479..aae3a2d 100644
--- a/drivers/staging/lustre/lustre/osc/osc_object.c
+++ b/drivers/staging/lustre/lustre/osc/osc_object.c
@@ -159,8 +159,8 @@ static int osc_attr_get(const struct lu_env *env, struct cl_object *obj,
return 0;
}
-static int osc_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int osc_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
struct lov_oinfo *oinfo = cl2osc(obj)->oo_oinfo;
struct ost_lvb *lvb = &oinfo->loi_lvb;
@@ -261,7 +261,7 @@ static const struct cl_object_operations osc_ops = {
.coo_lock_init = osc_lock_init,
.coo_io_init = osc_io_init,
.coo_attr_get = osc_attr_get,
- .coo_attr_set = osc_attr_set,
+ .coo_attr_update = osc_attr_update,
.coo_glimpse = osc_object_glimpse,
.coo_prune = osc_object_prune
};
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 3b59c2d..68cf115 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1749,7 +1749,7 @@ static int brw_interpret(const struct lu_env *env,
}
if (valid != 0)
- cl_object_attr_set(env, obj, attr, valid);
+ cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
kmem_cache_free(obdo_cachep, aa->aa_oa);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 104/124] staging: lustre: clio: rename coo_attr_set to coo_attr_update
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
coo_attr_set() is used to update object's attribute but its name
makes confusion that people intuitively think that it is used to
pass object's attribute down to server sides.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1154
Reviewed-on: http://review.whamcloud.com/12888
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 8 ++++----
drivers/staging/lustre/lustre/llite/vvp_io.c | 2 +-
drivers/staging/lustre/lustre/llite/vvp_object.c | 6 +++---
drivers/staging/lustre/lustre/lov/lov_object.c | 6 +++---
drivers/staging/lustre/lustre/lov/lovsub_object.c | 6 +++---
drivers/staging/lustre/lustre/obdclass/cl_object.c | 17 +++++++++--------
drivers/staging/lustre/lustre/osc/osc_io.c | 9 +++++----
drivers/staging/lustre/lustre/osc/osc_lock.c | 4 ++--
drivers/staging/lustre/lustre/osc/osc_object.c | 6 +++---
drivers/staging/lustre/lustre/osc/osc_request.c | 2 +-
10 files changed, 34 insertions(+), 32 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 409a324..9403c0c 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -369,8 +369,8 @@ struct cl_object_operations {
* \return the same convention as for
* cl_object_operations::coo_attr_get() is used.
*/
- int (*coo_attr_set)(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid);
+ int (*coo_attr_update)(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid);
/**
* Update object configuration. Called top-to-bottom to modify object
* configuration.
@@ -2171,8 +2171,8 @@ void cl_object_attr_lock(struct cl_object *o);
void cl_object_attr_unlock(struct cl_object *o);
int cl_object_attr_get(const struct lu_env *env, struct cl_object *obj,
struct cl_attr *attr);
-int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid);
+int cl_object_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid);
int cl_object_glimpse(const struct lu_env *env, struct cl_object *obj,
struct ost_lvb *lvb);
int cl_conf_set(const struct lu_env *env, struct cl_object *obj,
diff --git a/drivers/staging/lustre/lustre/llite/vvp_io.c b/drivers/staging/lustre/lustre/llite/vvp_io.c
index 27fc0b0..2ab4503 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_io.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_io.c
@@ -627,7 +627,7 @@ static int vvp_io_setattr_time(const struct lu_env *env,
attr->cat_mtime = io->u.ci_setattr.sa_attr.lvb_mtime;
valid |= CAT_MTIME;
}
- result = cl_object_attr_set(env, obj, attr, valid);
+ result = cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
return result;
diff --git a/drivers/staging/lustre/lustre/llite/vvp_object.c b/drivers/staging/lustre/lustre/llite/vvp_object.c
index 6b6a63d..b57195d 100644
--- a/drivers/staging/lustre/lustre/llite/vvp_object.c
+++ b/drivers/staging/lustre/lustre/llite/vvp_object.c
@@ -101,8 +101,8 @@ static int vvp_attr_get(const struct lu_env *env, struct cl_object *obj,
return 0; /* layers below have to fill in the rest */
}
-static int vvp_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int vvp_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
struct inode *inode = vvp_object_inode(obj);
@@ -209,7 +209,7 @@ static const struct cl_object_operations vvp_ops = {
.coo_lock_init = vvp_lock_init,
.coo_io_init = vvp_io_init,
.coo_attr_get = vvp_attr_get,
- .coo_attr_set = vvp_attr_set,
+ .coo_attr_update = vvp_attr_update,
.coo_conf_set = vvp_conf_set,
.coo_prune = vvp_prune,
.coo_glimpse = vvp_object_glimpse
diff --git a/drivers/staging/lustre/lustre/lov/lov_object.c b/drivers/staging/lustre/lustre/lov/lov_object.c
index 01f95a9..52f7363 100644
--- a/drivers/staging/lustre/lustre/lov/lov_object.c
+++ b/drivers/staging/lustre/lustre/lov/lov_object.c
@@ -894,8 +894,8 @@ static int lov_attr_get(const struct lu_env *env, struct cl_object *obj,
return LOV_2DISPATCH_NOLOCK(cl2lov(obj), llo_getattr, env, obj, attr);
}
-static int lov_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int lov_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
/*
* No dispatch is required here, as no layout implements this.
@@ -932,7 +932,7 @@ static const struct cl_object_operations lov_ops = {
.coo_lock_init = lov_lock_init,
.coo_io_init = lov_io_init,
.coo_attr_get = lov_attr_get,
- .coo_attr_set = lov_attr_set,
+ .coo_attr_update = lov_attr_update,
.coo_conf_set = lov_conf_set,
.coo_getstripe = lov_object_getstripe
};
diff --git a/drivers/staging/lustre/lustre/lov/lovsub_object.c b/drivers/staging/lustre/lustre/lov/lovsub_object.c
index fb2f266..a2bac7a 100644
--- a/drivers/staging/lustre/lustre/lov/lovsub_object.c
+++ b/drivers/staging/lustre/lustre/lov/lovsub_object.c
@@ -98,8 +98,8 @@ static int lovsub_object_print(const struct lu_env *env, void *cookie,
return (*p)(env, cookie, "[%d]", los->lso_index);
}
-static int lovsub_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int lovsub_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
struct lov_object *lov = cl2lovsub(obj)->lso_super;
@@ -119,7 +119,7 @@ static int lovsub_object_glimpse(const struct lu_env *env,
static const struct cl_object_operations lovsub_ops = {
.coo_page_init = lovsub_page_init,
.coo_lock_init = lovsub_lock_init,
- .coo_attr_set = lovsub_attr_set,
+ .coo_attr_update = lovsub_attr_update,
.coo_glimpse = lovsub_object_glimpse
};
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_object.c b/drivers/staging/lustre/lustre/obdclass/cl_object.c
index ef5b1aa..e493c80 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_object.c
@@ -163,7 +163,7 @@ static spinlock_t *cl_object_attr_guard(struct cl_object *o)
*
* Prevents data-attributes from changing, until lock is released by
* cl_object_attr_unlock(). This has to be called before calls to
- * cl_object_attr_get(), cl_object_attr_set().
+ * cl_object_attr_get(), cl_object_attr_update().
*/
void cl_object_attr_lock(struct cl_object *o)
__acquires(cl_object_attr_guard(o))
@@ -217,11 +217,11 @@ EXPORT_SYMBOL(cl_object_attr_get);
* Updates data-attributes of an object \a obj.
*
* Only attributes, mentioned in a validness bit-mask \a v are
- * updated. Calls cl_object_operations::coo_attr_set() on every layer, bottom
- * to top.
+ * updated. Calls cl_object_operations::coo_attr_update() on every layer,
+ * bottom to top.
*/
-int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned v)
+int cl_object_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int v)
{
struct lu_object_header *top;
int result;
@@ -231,8 +231,9 @@ int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
top = obj->co_lu.lo_header;
result = 0;
list_for_each_entry_reverse(obj, &top->loh_layers, co_lu.lo_linkage) {
- if (obj->co_ops->coo_attr_set) {
- result = obj->co_ops->coo_attr_set(env, obj, attr, v);
+ if (obj->co_ops->coo_attr_update) {
+ result = obj->co_ops->coo_attr_update(env, obj, attr,
+ v);
if (result != 0) {
if (result > 0)
result = 0;
@@ -242,7 +243,7 @@ int cl_object_attr_set(const struct lu_env *env, struct cl_object *obj,
}
return result;
}
-EXPORT_SYMBOL(cl_object_attr_set);
+EXPORT_SYMBOL(cl_object_attr_update);
/**
* Notifies layers (bottom-to-top) that glimpse AST was received.
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 8bbb2a6..42e9761 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -233,7 +233,7 @@ static void osc_page_touch_at(const struct lu_env *env,
attr->cat_size = kms;
valid |= CAT_SIZE;
}
- cl_object_attr_set(env, obj, attr, valid);
+ cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
@@ -476,7 +476,8 @@ static int osc_io_setattr_start(const struct lu_env *env,
attr->cat_ctime = lvb->lvb_ctime;
cl_valid |= CAT_CTIME;
}
- result = cl_object_attr_set(env, obj, attr, cl_valid);
+ result = cl_object_attr_update(env, obj, attr,
+ cl_valid);
}
cl_object_attr_unlock(obj);
}
@@ -564,7 +565,7 @@ static int osc_io_read_start(const struct lu_env *env,
if (!slice->cis_io->ci_noatime) {
cl_object_attr_lock(obj);
attr->cat_atime = ktime_get_real_seconds();
- rc = cl_object_attr_set(env, obj, attr, CAT_ATIME);
+ rc = cl_object_attr_update(env, obj, attr, CAT_ATIME);
cl_object_attr_unlock(obj);
}
return rc;
@@ -581,7 +582,7 @@ static int osc_io_write_start(const struct lu_env *env,
cl_object_attr_lock(obj);
attr->cat_ctime = ktime_get_real_seconds();
attr->cat_mtime = attr->cat_ctime;
- rc = cl_object_attr_set(env, obj, attr, CAT_MTIME | CAT_CTIME);
+ rc = cl_object_attr_update(env, obj, attr, CAT_MTIME | CAT_CTIME);
cl_object_attr_unlock(obj);
return rc;
diff --git a/drivers/staging/lustre/lustre/osc/osc_lock.c b/drivers/staging/lustre/lustre/osc/osc_lock.c
index 717d3ff..39a8a58 100644
--- a/drivers/staging/lustre/lustre/osc/osc_lock.c
+++ b/drivers/staging/lustre/lustre/osc/osc_lock.c
@@ -222,7 +222,7 @@ static void osc_lock_lvb_update(const struct lu_env *env,
ldlm_lock_allow_match_locked(dlmlock);
}
- cl_object_attr_set(env, obj, attr, valid);
+ cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
@@ -467,7 +467,7 @@ static int osc_dlm_blocking_ast0(const struct lu_env *env,
*/
attr->cat_kms = ldlm_extent_shift_kms(dlmlock, old_kms);
- cl_object_attr_set(env, obj, attr, CAT_KMS);
+ cl_object_attr_update(env, obj, attr, CAT_KMS);
cl_object_attr_unlock(obj);
unlock_res_and_lock(dlmlock);
diff --git a/drivers/staging/lustre/lustre/osc/osc_object.c b/drivers/staging/lustre/lustre/osc/osc_object.c
index 5fc5479..aae3a2d 100644
--- a/drivers/staging/lustre/lustre/osc/osc_object.c
+++ b/drivers/staging/lustre/lustre/osc/osc_object.c
@@ -159,8 +159,8 @@ static int osc_attr_get(const struct lu_env *env, struct cl_object *obj,
return 0;
}
-static int osc_attr_set(const struct lu_env *env, struct cl_object *obj,
- const struct cl_attr *attr, unsigned valid)
+static int osc_attr_update(const struct lu_env *env, struct cl_object *obj,
+ const struct cl_attr *attr, unsigned int valid)
{
struct lov_oinfo *oinfo = cl2osc(obj)->oo_oinfo;
struct ost_lvb *lvb = &oinfo->loi_lvb;
@@ -261,7 +261,7 @@ static const struct cl_object_operations osc_ops = {
.coo_lock_init = osc_lock_init,
.coo_io_init = osc_io_init,
.coo_attr_get = osc_attr_get,
- .coo_attr_set = osc_attr_set,
+ .coo_attr_update = osc_attr_update,
.coo_glimpse = osc_object_glimpse,
.coo_prune = osc_object_prune
};
diff --git a/drivers/staging/lustre/lustre/osc/osc_request.c b/drivers/staging/lustre/lustre/osc/osc_request.c
index 3b59c2d..68cf115 100644
--- a/drivers/staging/lustre/lustre/osc/osc_request.c
+++ b/drivers/staging/lustre/lustre/osc/osc_request.c
@@ -1749,7 +1749,7 @@ static int brw_interpret(const struct lu_env *env,
}
if (valid != 0)
- cl_object_attr_set(env, obj, attr, valid);
+ cl_object_attr_update(env, obj, attr, valid);
cl_object_attr_unlock(obj);
}
kmem_cache_free(obdo_cachep, aa->aa_oa);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 105/124] staging: lustre: clio: pass fid for OST setattr
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
Store inode's fid in cl_setattr_ost() and OSC packs this info on the
wire (via lustre_set_wire_obdo) so that OST can use.
NOTE: currently lu_fid::f_ver and obdo::o_parent_ver are not used on
OFD device, and we use obdo::o_stripe_idx as
filter_fid::ff_parent::f_ver and save it to the device.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1154
Reviewed-on: http://review.whamcloud.com/12902
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 2 ++
drivers/staging/lustre/lustre/llite/lcommon_cl.c | 1 +
drivers/staging/lustre/lustre/lov/lov_io.c | 3 +++
drivers/staging/lustre/lustre/osc/osc_io.c | 6 ++++--
4 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 9403c0c..89292c9 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -1766,6 +1766,8 @@ struct cl_io {
struct cl_setattr_io {
struct ost_lvb sa_attr;
unsigned int sa_valid;
+ int sa_stripe_index;
+ struct lu_fid *sa_parent_fid;
} ci_setattr;
struct cl_fault_io {
/** page index within file. */
diff --git a/drivers/staging/lustre/lustre/llite/lcommon_cl.c b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
index 2ded6b5..084330d 100644
--- a/drivers/staging/lustre/lustre/llite/lcommon_cl.c
+++ b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
@@ -99,6 +99,7 @@ int cl_setattr_ost(struct inode *inode, const struct iattr *attr)
io->u.ci_setattr.sa_attr.lvb_ctime = LTIME_S(attr->ia_ctime);
io->u.ci_setattr.sa_attr.lvb_size = attr->ia_size;
io->u.ci_setattr.sa_valid = attr->ia_valid;
+ io->u.ci_setattr.sa_parent_fid = ll_inode2fid(inode);
again:
if (cl_io_init(env, io, CIT_SETATTR, io->ci_obj) == 0) {
diff --git a/drivers/staging/lustre/lustre/lov/lov_io.c b/drivers/staging/lustre/lustre/lov/lov_io.c
index 5d47a5a..d101579 100644
--- a/drivers/staging/lustre/lustre/lov/lov_io.c
+++ b/drivers/staging/lustre/lustre/lov/lov_io.c
@@ -87,6 +87,9 @@ static void lov_io_sub_inherit(struct cl_io *io, struct lov_io *lio,
case CIT_SETATTR: {
io->u.ci_setattr.sa_attr = parent->u.ci_setattr.sa_attr;
io->u.ci_setattr.sa_valid = parent->u.ci_setattr.sa_valid;
+ io->u.ci_setattr.sa_stripe_index = stripe;
+ io->u.ci_setattr.sa_parent_fid =
+ parent->u.ci_setattr.sa_parent_fid;
if (cl_io_is_trunc(io)) {
loff_t new_size = parent->u.ci_setattr.sa_attr.lvb_size;
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 42e9761..8a559cb 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -484,11 +484,13 @@ static int osc_io_setattr_start(const struct lu_env *env,
memset(oa, 0, sizeof(*oa));
if (result == 0) {
oa->o_oi = loi->loi_oi;
+ obdo_set_parent_fid(oa, io->u.ci_setattr.sa_parent_fid);
+ oa->o_stripe_idx = io->u.ci_setattr.sa_stripe_index;
oa->o_mtime = attr->cat_mtime;
oa->o_atime = attr->cat_atime;
oa->o_ctime = attr->cat_ctime;
- oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLATIME |
- OBD_MD_FLCTIME | OBD_MD_FLMTIME;
+ oa->o_valid |= OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLATIME |
+ OBD_MD_FLCTIME | OBD_MD_FLMTIME;
if (ia_valid & ATTR_SIZE) {
oa->o_size = size;
oa->o_blocks = OBD_OBJECT_EOF;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 105/124] staging: lustre: clio: pass fid for OST setattr
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Bobi Jam,
James Simmons
From: Bobi Jam <bobijam.xu@intel.com>
Store inode's fid in cl_setattr_ost() and OSC packs this info on the
wire (via lustre_set_wire_obdo) so that OST can use.
NOTE: currently lu_fid::f_ver and obdo::o_parent_ver are not used on
OFD device, and we use obdo::o_stripe_idx as
filter_fid::ff_parent::f_ver and save it to the device.
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1154
Reviewed-on: http://review.whamcloud.com/12902
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/cl_object.h | 2 ++
drivers/staging/lustre/lustre/llite/lcommon_cl.c | 1 +
drivers/staging/lustre/lustre/lov/lov_io.c | 3 +++
drivers/staging/lustre/lustre/osc/osc_io.c | 6 ++++--
4 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/cl_object.h b/drivers/staging/lustre/lustre/include/cl_object.h
index 9403c0c..89292c9 100644
--- a/drivers/staging/lustre/lustre/include/cl_object.h
+++ b/drivers/staging/lustre/lustre/include/cl_object.h
@@ -1766,6 +1766,8 @@ struct cl_io {
struct cl_setattr_io {
struct ost_lvb sa_attr;
unsigned int sa_valid;
+ int sa_stripe_index;
+ struct lu_fid *sa_parent_fid;
} ci_setattr;
struct cl_fault_io {
/** page index within file. */
diff --git a/drivers/staging/lustre/lustre/llite/lcommon_cl.c b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
index 2ded6b5..084330d 100644
--- a/drivers/staging/lustre/lustre/llite/lcommon_cl.c
+++ b/drivers/staging/lustre/lustre/llite/lcommon_cl.c
@@ -99,6 +99,7 @@ int cl_setattr_ost(struct inode *inode, const struct iattr *attr)
io->u.ci_setattr.sa_attr.lvb_ctime = LTIME_S(attr->ia_ctime);
io->u.ci_setattr.sa_attr.lvb_size = attr->ia_size;
io->u.ci_setattr.sa_valid = attr->ia_valid;
+ io->u.ci_setattr.sa_parent_fid = ll_inode2fid(inode);
again:
if (cl_io_init(env, io, CIT_SETATTR, io->ci_obj) == 0) {
diff --git a/drivers/staging/lustre/lustre/lov/lov_io.c b/drivers/staging/lustre/lustre/lov/lov_io.c
index 5d47a5a..d101579 100644
--- a/drivers/staging/lustre/lustre/lov/lov_io.c
+++ b/drivers/staging/lustre/lustre/lov/lov_io.c
@@ -87,6 +87,9 @@ static void lov_io_sub_inherit(struct cl_io *io, struct lov_io *lio,
case CIT_SETATTR: {
io->u.ci_setattr.sa_attr = parent->u.ci_setattr.sa_attr;
io->u.ci_setattr.sa_valid = parent->u.ci_setattr.sa_valid;
+ io->u.ci_setattr.sa_stripe_index = stripe;
+ io->u.ci_setattr.sa_parent_fid =
+ parent->u.ci_setattr.sa_parent_fid;
if (cl_io_is_trunc(io)) {
loff_t new_size = parent->u.ci_setattr.sa_attr.lvb_size;
diff --git a/drivers/staging/lustre/lustre/osc/osc_io.c b/drivers/staging/lustre/lustre/osc/osc_io.c
index 42e9761..8a559cb 100644
--- a/drivers/staging/lustre/lustre/osc/osc_io.c
+++ b/drivers/staging/lustre/lustre/osc/osc_io.c
@@ -484,11 +484,13 @@ static int osc_io_setattr_start(const struct lu_env *env,
memset(oa, 0, sizeof(*oa));
if (result == 0) {
oa->o_oi = loi->loi_oi;
+ obdo_set_parent_fid(oa, io->u.ci_setattr.sa_parent_fid);
+ oa->o_stripe_idx = io->u.ci_setattr.sa_stripe_index;
oa->o_mtime = attr->cat_mtime;
oa->o_atime = attr->cat_atime;
oa->o_ctime = attr->cat_ctime;
- oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLATIME |
- OBD_MD_FLCTIME | OBD_MD_FLMTIME;
+ oa->o_valid |= OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLATIME |
+ OBD_MD_FLCTIME | OBD_MD_FLMTIME;
if (ia_valid & ATTR_SIZE) {
oa->o_size = size;
oa->o_blocks = OBD_OBJECT_EOF;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 106/124] staging: lustre: client: Fix mkdir -i 1 from DNE2 client to DNE1 server
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Artem Blagodarenko, James Simmons
From: Artem Blagodarenko <artem_blagodarenko@xyratex.com>
After DNE phase 2 has been added to client it sends
create request to slave MDT. DNT1-only server doesn't
expect request to slave MDT from client. It expects
only cross-mdt request from master MDT. Thus if DNE2
client tries to "mkdir -i 1" on DNE1 server, then
LBUG happened.
This patch adds OBD_CONNECT_DIR_STRIPE connection
flag check on client side. If striped directories are not
supported by server, then create requrest is sent to
master MDT.
Signed-off-by: Artem Blagodarenko <artem_blagodarenko@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6071
Xyratex-bug-id: MRP-2319
Reviewed-on: http://review.whamcloud.com/13189
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 20 ++++++++++++--------
1 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 185169c..5928605 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1690,15 +1690,19 @@ static int lmv_create(struct obd_export *exp, struct md_op_data *op_data,
if (rc)
return rc;
- /*
- * Send the create request to the MDT where the object
- * will be located
- */
- tgt = lmv_find_target(lmv, &op_data->op_fid2);
- if (IS_ERR(tgt))
- return PTR_ERR(tgt);
+ if (exp_connect_flags(exp) & OBD_CONNECT_DIR_STRIPE) {
+ /*
+ * Send the create request to the MDT where the object
+ * will be located
+ */
+ tgt = lmv_find_target(lmv, &op_data->op_fid2);
+ if (IS_ERR(tgt))
+ return PTR_ERR(tgt);
- op_data->op_mds = tgt->ltd_idx;
+ op_data->op_mds = tgt->ltd_idx;
+ } else {
+ CDEBUG(D_CONFIG, "Server doesn't support striped dirs\n");
+ }
CDEBUG(D_INODE, "CREATE obj "DFID" -> mds #%x\n",
PFID(&op_data->op_fid1), op_data->op_mds);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 106/124] staging: lustre: client: Fix mkdir -i 1 from DNE2 client to DNE1 server
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Artem Blagodarenko, James Simmons
From: Artem Blagodarenko <artem_blagodarenko@xyratex.com>
After DNE phase 2 has been added to client it sends
create request to slave MDT. DNT1-only server doesn't
expect request to slave MDT from client. It expects
only cross-mdt request from master MDT. Thus if DNE2
client tries to "mkdir -i 1" on DNE1 server, then
LBUG happened.
This patch adds OBD_CONNECT_DIR_STRIPE connection
flag check on client side. If striped directories are not
supported by server, then create requrest is sent to
master MDT.
Signed-off-by: Artem Blagodarenko <artem_blagodarenko@xyratex.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6071
Xyratex-bug-id: MRP-2319
Reviewed-on: http://review.whamcloud.com/13189
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: wang di <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 20 ++++++++++++--------
1 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 185169c..5928605 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1690,15 +1690,19 @@ static int lmv_create(struct obd_export *exp, struct md_op_data *op_data,
if (rc)
return rc;
- /*
- * Send the create request to the MDT where the object
- * will be located
- */
- tgt = lmv_find_target(lmv, &op_data->op_fid2);
- if (IS_ERR(tgt))
- return PTR_ERR(tgt);
+ if (exp_connect_flags(exp) & OBD_CONNECT_DIR_STRIPE) {
+ /*
+ * Send the create request to the MDT where the object
+ * will be located
+ */
+ tgt = lmv_find_target(lmv, &op_data->op_fid2);
+ if (IS_ERR(tgt))
+ return PTR_ERR(tgt);
- op_data->op_mds = tgt->ltd_idx;
+ op_data->op_mds = tgt->ltd_idx;
+ } else {
+ CDEBUG(D_CONFIG, "Server doesn't support striped dirs\n");
+ }
CDEBUG(D_INODE, "CREATE obj "DFID" -> mds #%x\n",
PFID(&op_data->op_fid1), op_data->op_mds);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 107/124] staging: lustre: lmv: Do not revalidate stripes with master lock
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Do not revalidate slave stripes while holding master lock.
Otherwise if the revalidating slaves are blocked, then the
master lock can not be released in time.
Remove some unnecesary merging in ll_revalidate_slave(), and
the attributes will be stored in each stripe, only
merging them if required.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6088
Reviewed-on: http://review.whamcloud.com/13432
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 5 +--
drivers/staging/lustre/lustre/include/obd_class.h | 15 +------
drivers/staging/lustre/lustre/llite/file.c | 19 +++------
.../staging/lustre/lustre/llite/llite_internal.h | 3 -
drivers/staging/lustre/lustre/llite/llite_lib.c | 7 +---
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 41 ++------------------
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 4 +-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 28 ++++++-------
8 files changed, 31 insertions(+), 91 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 986c6e9..c6937b2 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -1000,10 +1000,7 @@ struct md_ops {
int (*merge_attr)(struct obd_export *,
const struct lmv_stripe_md *lsm,
- struct cl_attr *attr);
-
- int (*update_lsm_md)(struct obd_export *, struct lmv_stripe_md *lsm,
- struct mdt_body *, ldlm_blocking_callback);
+ struct cl_attr *attr, ldlm_blocking_callback);
int (*set_open_replay_data)(struct obd_export *,
struct obd_client_handle *,
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 9836aed..16094db 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1497,23 +1497,14 @@ static inline int md_free_lustre_md(struct obd_export *exp,
return MDP(exp->exp_obd, free_lustre_md)(exp, md);
}
-static inline int md_update_lsm_md(struct obd_export *exp,
- struct lmv_stripe_md *lsm,
- struct mdt_body *body,
- ldlm_blocking_callback cb)
-{
- EXP_CHECK_MD_OP(exp, update_lsm_md);
- EXP_MD_COUNTER_INCREMENT(exp, update_lsm_md);
- return MDP(exp->exp_obd, update_lsm_md)(exp, lsm, body, cb);
-}
-
static inline int md_merge_attr(struct obd_export *exp,
const struct lmv_stripe_md *lsm,
- struct cl_attr *attr)
+ struct cl_attr *attr,
+ ldlm_blocking_callback cb)
{
EXP_CHECK_MD_OP(exp, merge_attr);
EXP_MD_COUNTER_INCREMENT(exp, merge_attr);
- return MDP(exp->exp_obd, merge_attr)(exp, lsm, attr);
+ return MDP(exp->exp_obd, merge_attr)(exp, lsm, attr, cb);
}
static inline int md_setxattr(struct obd_export *exp, const struct lu_fid *fid,
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index e8963fe..d8761b8 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -3045,12 +3045,13 @@ static int ll_merge_md_attr(struct inode *inode)
LASSERT(ll_i2info(inode)->lli_lsm_md);
rc = md_merge_attr(ll_i2mdexp(inode), ll_i2info(inode)->lli_lsm_md,
- &attr);
+ &attr, ll_md_blocking_ast);
if (rc)
return rc;
- ll_i2info(inode)->lli_stripe_dir_size = attr.cat_size;
- ll_i2info(inode)->lli_stripe_dir_nlink = attr.cat_nlink;
+ set_nlink(inode, attr.cat_nlink);
+ inode->i_blocks = attr.cat_blocks;
+ i_size_write(inode, attr.cat_size);
ll_i2info(inode)->lli_atime = attr.cat_atime;
ll_i2info(inode)->lli_mtime = attr.cat_mtime;
@@ -3123,16 +3124,10 @@ int ll_getattr(struct vfsmount *mnt, struct dentry *de, struct kstat *stat)
stat->mtime = inode->i_mtime;
stat->ctime = inode->i_ctime;
stat->blksize = 1 << inode->i_blkbits;
- stat->blocks = inode->i_blocks;
- if (S_ISDIR(inode->i_mode) &&
- ll_i2info(inode)->lli_lsm_md) {
- stat->nlink = lli->lli_stripe_dir_nlink;
- stat->size = lli->lli_stripe_dir_size;
- } else {
- stat->nlink = inode->i_nlink;
- stat->size = i_size_read(inode);
- }
+ stat->nlink = inode->i_nlink;
+ stat->size = i_size_read(inode);
+ stat->blocks = inode->i_blocks;
return 0;
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index e5f1967..51bf071 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -191,9 +191,6 @@ struct ll_inode_info {
unsigned int lli_sa_generation;
/* directory stripe information */
struct lmv_stripe_md *lli_lsm_md;
- /* striped directory size */
- loff_t lli_stripe_dir_size;
- u64 lli_stripe_dir_nlink;
};
/* for non-directory */
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index ff843e1..230868c 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -1117,12 +1117,7 @@ static int ll_init_lsm_md(struct inode *inode, struct lustre_md *md)
}
}
- /*
- * Here is where the lsm is being initialized(fill lmo_info) after
- * client retrieve MD stripe information from MDT.
- */
- return md_update_lsm_md(ll_i2mdexp(inode), lsm, md->body,
- ll_md_blocking_ast);
+ return 0;
}
static inline int lli_lsm_md_eq(const struct lmv_stripe_md *lsm_md1,
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 3197b6f..fe5c60a 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -148,8 +148,8 @@ out:
return rc;
}
-int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
- struct lmv_stripe_md *lsm,
+int lmv_revalidate_slaves(struct obd_export *exp,
+ const struct lmv_stripe_md *lsm,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags)
{
@@ -158,11 +158,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
struct ptlrpc_request *req = NULL;
struct mdt_body *body;
struct md_op_data *op_data;
- unsigned long size = 0;
- unsigned long nlink = 0;
- __s64 atime = 0;
- __s64 ctime = 0;
- __s64 mtime = 0;
int rc = 0, i;
/**
@@ -243,6 +238,7 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
}
i_size_write(inode, body->mbo_size);
+ inode->i_blocks = body->mbo_blocks;
set_nlink(inode, body->mbo_nlink);
LTIME_S(inode->i_atime) = body->mbo_atime;
LTIME_S(inode->i_ctime) = body->mbo_ctime;
@@ -251,41 +247,12 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
md_set_lock_data(tgt->ltd_exp, lockh, inode, NULL);
- if (i != 0)
- nlink += inode->i_nlink - 2;
- else
- nlink += inode->i_nlink;
-
- atime = LTIME_S(inode->i_atime) > atime ?
- LTIME_S(inode->i_atime) : atime;
- ctime = LTIME_S(inode->i_ctime) > ctime ?
- LTIME_S(inode->i_ctime) : ctime;
- mtime = LTIME_S(inode->i_mtime) > mtime ?
- LTIME_S(inode->i_mtime) : mtime;
-
if (it.it_lock_mode && lockh) {
ldlm_lock_decref(lockh, it.it_lock_mode);
it.it_lock_mode = 0;
}
-
- CDEBUG(D_INODE, "i %d "DFID" size %llu, nlink %u, atime %lu, mtime %lu, ctime %lu.\n",
- i, PFID(&fid), i_size_read(inode), inode->i_nlink,
- LTIME_S(inode->i_atime), LTIME_S(inode->i_mtime),
- LTIME_S(inode->i_ctime));
}
- /*
- * update attr of master request.
- */
- CDEBUG(D_INODE, "Return refreshed attrs: size = %lu nlink %lu atime %llu ctime %llu mtime %llu for " DFID"\n",
- size, nlink, atime, ctime, mtime,
- PFID(&lsm->lsm_md_oinfo[0].lmo_fid));
-
- if (mbody) {
- mbody->mbo_atime = atime;
- mbody->mbo_ctime = ctime;
- mbody->mbo_mtime = mtime;
- }
cleanup:
if (req)
ptlrpc_req_finished(req);
@@ -445,7 +412,7 @@ static int lmv_intent_lookup(struct obd_export *exp,
* during update_inode process (see ll_update_lsm_md)
*/
if (op_data->op_mea2) {
- rc = lmv_revalidate_slaves(exp, NULL, op_data->op_mea2,
+ rc = lmv_revalidate_slaves(exp, op_data->op_mea2,
cb_blocking,
extra_lock_flags);
if (rc != 0)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
index 8f703ea..52b0374 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h
+++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
@@ -57,8 +57,8 @@ int lmv_fid_alloc(const struct lu_env *env, struct obd_export *exp,
int lmv_unpack_md(struct obd_export *exp, struct lmv_stripe_md **lsmp,
const union lmv_mds_md *lmm, int stripe_count);
-int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
- struct lmv_stripe_md *lsm,
+int lmv_revalidate_slaves(struct obd_export *exp,
+ const struct lmv_stripe_md *lsm,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags);
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 5928605..941981d 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -3242,27 +3242,25 @@ static int lmv_quotacheck(struct obd_device *unused, struct obd_export *exp,
return rc;
}
-static int
-lmv_update_lsm_md(struct obd_export *exp, struct lmv_stripe_md *lsm,
- struct mdt_body *body, ldlm_blocking_callback cb_blocking)
+static int lmv_merge_attr(struct obd_export *exp,
+ const struct lmv_stripe_md *lsm,
+ struct cl_attr *attr,
+ ldlm_blocking_callback cb_blocking)
{
- return lmv_revalidate_slaves(exp, body, lsm, cb_blocking, 0);
-}
+ int rc, i;
-static int
-lmv_merge_attr(struct obd_export *exp, const struct lmv_stripe_md *lsm,
- struct cl_attr *attr)
-{
- int i;
+ rc = lmv_revalidate_slaves(exp, lsm, cb_blocking, 0);
+ if (rc < 0)
+ return rc;
for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
struct inode *inode = lsm->lsm_md_oinfo[i].lmo_root;
- CDEBUG(D_INFO, ""DFID" size %llu, nlink %u, atime %lu ctime %lu, mtime %lu.\n",
+ CDEBUG(D_INFO, ""DFID" size %llu, blocks %llu nlink %u, atime %lu ctime %lu, mtime %lu.\n",
PFID(&lsm->lsm_md_oinfo[i].lmo_fid),
- i_size_read(inode), inode->i_nlink,
- LTIME_S(inode->i_atime), LTIME_S(inode->i_ctime),
- LTIME_S(inode->i_mtime));
+ i_size_read(inode), (unsigned long long)inode->i_blocks,
+ inode->i_nlink, LTIME_S(inode->i_atime),
+ LTIME_S(inode->i_ctime), LTIME_S(inode->i_mtime));
/* for slave stripe, it needs to subtract nlink for . and .. */
if (i)
@@ -3271,6 +3269,7 @@ lmv_merge_attr(struct obd_export *exp, const struct lmv_stripe_md *lsm,
attr->cat_nlink = inode->i_nlink;
attr->cat_size += i_size_read(inode);
+ attr->cat_blocks += inode->i_blocks;
if (attr->cat_atime < LTIME_S(inode->i_atime))
attr->cat_atime = LTIME_S(inode->i_atime);
@@ -3328,7 +3327,6 @@ static struct md_ops lmv_md_ops = {
.lock_match = lmv_lock_match,
.get_lustre_md = lmv_get_lustre_md,
.free_lustre_md = lmv_free_lustre_md,
- .update_lsm_md = lmv_update_lsm_md,
.merge_attr = lmv_merge_attr,
.set_open_replay_data = lmv_set_open_replay_data,
.clear_open_replay_data = lmv_clear_open_replay_data,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 107/124] staging: lustre: lmv: Do not revalidate stripes with master lock
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Do not revalidate slave stripes while holding master lock.
Otherwise if the revalidating slaves are blocked, then the
master lock can not be released in time.
Remove some unnecesary merging in ll_revalidate_slave(), and
the attributes will be stored in each stripe, only
merging them if required.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6088
Reviewed-on: http://review.whamcloud.com/13432
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 5 +--
drivers/staging/lustre/lustre/include/obd_class.h | 15 +------
drivers/staging/lustre/lustre/llite/file.c | 19 +++------
.../staging/lustre/lustre/llite/llite_internal.h | 3 -
drivers/staging/lustre/lustre/llite/llite_lib.c | 7 +---
drivers/staging/lustre/lustre/lmv/lmv_intent.c | 41 ++------------------
drivers/staging/lustre/lustre/lmv/lmv_internal.h | 4 +-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 28 ++++++-------
8 files changed, 31 insertions(+), 91 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index 986c6e9..c6937b2 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -1000,10 +1000,7 @@ struct md_ops {
int (*merge_attr)(struct obd_export *,
const struct lmv_stripe_md *lsm,
- struct cl_attr *attr);
-
- int (*update_lsm_md)(struct obd_export *, struct lmv_stripe_md *lsm,
- struct mdt_body *, ldlm_blocking_callback);
+ struct cl_attr *attr, ldlm_blocking_callback);
int (*set_open_replay_data)(struct obd_export *,
struct obd_client_handle *,
diff --git a/drivers/staging/lustre/lustre/include/obd_class.h b/drivers/staging/lustre/lustre/include/obd_class.h
index 9836aed..16094db 100644
--- a/drivers/staging/lustre/lustre/include/obd_class.h
+++ b/drivers/staging/lustre/lustre/include/obd_class.h
@@ -1497,23 +1497,14 @@ static inline int md_free_lustre_md(struct obd_export *exp,
return MDP(exp->exp_obd, free_lustre_md)(exp, md);
}
-static inline int md_update_lsm_md(struct obd_export *exp,
- struct lmv_stripe_md *lsm,
- struct mdt_body *body,
- ldlm_blocking_callback cb)
-{
- EXP_CHECK_MD_OP(exp, update_lsm_md);
- EXP_MD_COUNTER_INCREMENT(exp, update_lsm_md);
- return MDP(exp->exp_obd, update_lsm_md)(exp, lsm, body, cb);
-}
-
static inline int md_merge_attr(struct obd_export *exp,
const struct lmv_stripe_md *lsm,
- struct cl_attr *attr)
+ struct cl_attr *attr,
+ ldlm_blocking_callback cb)
{
EXP_CHECK_MD_OP(exp, merge_attr);
EXP_MD_COUNTER_INCREMENT(exp, merge_attr);
- return MDP(exp->exp_obd, merge_attr)(exp, lsm, attr);
+ return MDP(exp->exp_obd, merge_attr)(exp, lsm, attr, cb);
}
static inline int md_setxattr(struct obd_export *exp, const struct lu_fid *fid,
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index e8963fe..d8761b8 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -3045,12 +3045,13 @@ static int ll_merge_md_attr(struct inode *inode)
LASSERT(ll_i2info(inode)->lli_lsm_md);
rc = md_merge_attr(ll_i2mdexp(inode), ll_i2info(inode)->lli_lsm_md,
- &attr);
+ &attr, ll_md_blocking_ast);
if (rc)
return rc;
- ll_i2info(inode)->lli_stripe_dir_size = attr.cat_size;
- ll_i2info(inode)->lli_stripe_dir_nlink = attr.cat_nlink;
+ set_nlink(inode, attr.cat_nlink);
+ inode->i_blocks = attr.cat_blocks;
+ i_size_write(inode, attr.cat_size);
ll_i2info(inode)->lli_atime = attr.cat_atime;
ll_i2info(inode)->lli_mtime = attr.cat_mtime;
@@ -3123,16 +3124,10 @@ int ll_getattr(struct vfsmount *mnt, struct dentry *de, struct kstat *stat)
stat->mtime = inode->i_mtime;
stat->ctime = inode->i_ctime;
stat->blksize = 1 << inode->i_blkbits;
- stat->blocks = inode->i_blocks;
- if (S_ISDIR(inode->i_mode) &&
- ll_i2info(inode)->lli_lsm_md) {
- stat->nlink = lli->lli_stripe_dir_nlink;
- stat->size = lli->lli_stripe_dir_size;
- } else {
- stat->nlink = inode->i_nlink;
- stat->size = i_size_read(inode);
- }
+ stat->nlink = inode->i_nlink;
+ stat->size = i_size_read(inode);
+ stat->blocks = inode->i_blocks;
return 0;
}
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index e5f1967..51bf071 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -191,9 +191,6 @@ struct ll_inode_info {
unsigned int lli_sa_generation;
/* directory stripe information */
struct lmv_stripe_md *lli_lsm_md;
- /* striped directory size */
- loff_t lli_stripe_dir_size;
- u64 lli_stripe_dir_nlink;
};
/* for non-directory */
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index ff843e1..230868c 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -1117,12 +1117,7 @@ static int ll_init_lsm_md(struct inode *inode, struct lustre_md *md)
}
}
- /*
- * Here is where the lsm is being initialized(fill lmo_info) after
- * client retrieve MD stripe information from MDT.
- */
- return md_update_lsm_md(ll_i2mdexp(inode), lsm, md->body,
- ll_md_blocking_ast);
+ return 0;
}
static inline int lli_lsm_md_eq(const struct lmv_stripe_md *lsm_md1,
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_intent.c b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
index 3197b6f..fe5c60a 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_intent.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_intent.c
@@ -148,8 +148,8 @@ out:
return rc;
}
-int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
- struct lmv_stripe_md *lsm,
+int lmv_revalidate_slaves(struct obd_export *exp,
+ const struct lmv_stripe_md *lsm,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags)
{
@@ -158,11 +158,6 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
struct ptlrpc_request *req = NULL;
struct mdt_body *body;
struct md_op_data *op_data;
- unsigned long size = 0;
- unsigned long nlink = 0;
- __s64 atime = 0;
- __s64 ctime = 0;
- __s64 mtime = 0;
int rc = 0, i;
/**
@@ -243,6 +238,7 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
}
i_size_write(inode, body->mbo_size);
+ inode->i_blocks = body->mbo_blocks;
set_nlink(inode, body->mbo_nlink);
LTIME_S(inode->i_atime) = body->mbo_atime;
LTIME_S(inode->i_ctime) = body->mbo_ctime;
@@ -251,41 +247,12 @@ int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
md_set_lock_data(tgt->ltd_exp, lockh, inode, NULL);
- if (i != 0)
- nlink += inode->i_nlink - 2;
- else
- nlink += inode->i_nlink;
-
- atime = LTIME_S(inode->i_atime) > atime ?
- LTIME_S(inode->i_atime) : atime;
- ctime = LTIME_S(inode->i_ctime) > ctime ?
- LTIME_S(inode->i_ctime) : ctime;
- mtime = LTIME_S(inode->i_mtime) > mtime ?
- LTIME_S(inode->i_mtime) : mtime;
-
if (it.it_lock_mode && lockh) {
ldlm_lock_decref(lockh, it.it_lock_mode);
it.it_lock_mode = 0;
}
-
- CDEBUG(D_INODE, "i %d "DFID" size %llu, nlink %u, atime %lu, mtime %lu, ctime %lu.\n",
- i, PFID(&fid), i_size_read(inode), inode->i_nlink,
- LTIME_S(inode->i_atime), LTIME_S(inode->i_mtime),
- LTIME_S(inode->i_ctime));
}
- /*
- * update attr of master request.
- */
- CDEBUG(D_INODE, "Return refreshed attrs: size = %lu nlink %lu atime %llu ctime %llu mtime %llu for " DFID"\n",
- size, nlink, atime, ctime, mtime,
- PFID(&lsm->lsm_md_oinfo[0].lmo_fid));
-
- if (mbody) {
- mbody->mbo_atime = atime;
- mbody->mbo_ctime = ctime;
- mbody->mbo_mtime = mtime;
- }
cleanup:
if (req)
ptlrpc_req_finished(req);
@@ -445,7 +412,7 @@ static int lmv_intent_lookup(struct obd_export *exp,
* during update_inode process (see ll_update_lsm_md)
*/
if (op_data->op_mea2) {
- rc = lmv_revalidate_slaves(exp, NULL, op_data->op_mea2,
+ rc = lmv_revalidate_slaves(exp, op_data->op_mea2,
cb_blocking,
extra_lock_flags);
if (rc != 0)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_internal.h b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
index 8f703ea..52b0374 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_internal.h
+++ b/drivers/staging/lustre/lustre/lmv/lmv_internal.h
@@ -57,8 +57,8 @@ int lmv_fid_alloc(const struct lu_env *env, struct obd_export *exp,
int lmv_unpack_md(struct obd_export *exp, struct lmv_stripe_md **lsmp,
const union lmv_mds_md *lmm, int stripe_count);
-int lmv_revalidate_slaves(struct obd_export *exp, struct mdt_body *mbody,
- struct lmv_stripe_md *lsm,
+int lmv_revalidate_slaves(struct obd_export *exp,
+ const struct lmv_stripe_md *lsm,
ldlm_blocking_callback cb_blocking,
int extra_lock_flags);
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 5928605..941981d 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -3242,27 +3242,25 @@ static int lmv_quotacheck(struct obd_device *unused, struct obd_export *exp,
return rc;
}
-static int
-lmv_update_lsm_md(struct obd_export *exp, struct lmv_stripe_md *lsm,
- struct mdt_body *body, ldlm_blocking_callback cb_blocking)
+static int lmv_merge_attr(struct obd_export *exp,
+ const struct lmv_stripe_md *lsm,
+ struct cl_attr *attr,
+ ldlm_blocking_callback cb_blocking)
{
- return lmv_revalidate_slaves(exp, body, lsm, cb_blocking, 0);
-}
+ int rc, i;
-static int
-lmv_merge_attr(struct obd_export *exp, const struct lmv_stripe_md *lsm,
- struct cl_attr *attr)
-{
- int i;
+ rc = lmv_revalidate_slaves(exp, lsm, cb_blocking, 0);
+ if (rc < 0)
+ return rc;
for (i = 0; i < lsm->lsm_md_stripe_count; i++) {
struct inode *inode = lsm->lsm_md_oinfo[i].lmo_root;
- CDEBUG(D_INFO, ""DFID" size %llu, nlink %u, atime %lu ctime %lu, mtime %lu.\n",
+ CDEBUG(D_INFO, ""DFID" size %llu, blocks %llu nlink %u, atime %lu ctime %lu, mtime %lu.\n",
PFID(&lsm->lsm_md_oinfo[i].lmo_fid),
- i_size_read(inode), inode->i_nlink,
- LTIME_S(inode->i_atime), LTIME_S(inode->i_ctime),
- LTIME_S(inode->i_mtime));
+ i_size_read(inode), (unsigned long long)inode->i_blocks,
+ inode->i_nlink, LTIME_S(inode->i_atime),
+ LTIME_S(inode->i_ctime), LTIME_S(inode->i_mtime));
/* for slave stripe, it needs to subtract nlink for . and .. */
if (i)
@@ -3271,6 +3269,7 @@ lmv_merge_attr(struct obd_export *exp, const struct lmv_stripe_md *lsm,
attr->cat_nlink = inode->i_nlink;
attr->cat_size += i_size_read(inode);
+ attr->cat_blocks += inode->i_blocks;
if (attr->cat_atime < LTIME_S(inode->i_atime))
attr->cat_atime = LTIME_S(inode->i_atime);
@@ -3328,7 +3327,6 @@ static struct md_ops lmv_md_ops = {
.lock_match = lmv_lock_match,
.get_lustre_md = lmv_get_lustre_md,
.free_lustre_md = lmv_free_lustre_md,
- .update_lsm_md = lmv_update_lsm_md,
.merge_attr = lmv_merge_attr,
.set_open_replay_data = lmv_set_open_replay_data,
.clear_open_replay_data = lmv_clear_open_replay_data,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 108/124] staging: lustre: grant: quiet message on grant waiting timeout
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Johann Lombardi, James Simmons
From: Johann Lombardi <johann.lombardi@intel.com>
Use at_max in osc_enter_cache() to bound how long we wait for grant
space before switching to synchronous I/Os. Do not print a message
on the console when the timeout is hit since such long wait can
be legitimate with flaky network (i.e. BRW is resent multiple times).
Signed-off-by: Johann Lombardi <johann.lombardi@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5521
Reviewed-on: http://review.whamcloud.com/12146
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_cache.c | 61 ++++++++++++++++---------
1 files changed, 40 insertions(+), 21 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 8152cc7..6b7677f 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1536,7 +1536,7 @@ static int osc_enter_cache_try(struct client_obd *cli,
{
int rc;
- OSC_DUMP_GRANT(D_CACHE, cli, "need:%d.\n", bytes);
+ OSC_DUMP_GRANT(D_CACHE, cli, "need:%d\n", bytes);
rc = osc_reserve_grant(cli, bytes);
if (rc < 0)
@@ -1581,11 +1581,13 @@ static int osc_enter_cache(const struct lu_env *env, struct client_obd *cli,
struct osc_object *osc = oap->oap_obj;
struct lov_oinfo *loi = osc->oo_oinfo;
struct osc_cache_waiter ocw;
- struct l_wait_info lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(600), NULL,
- LWI_ON_SIGNAL_NOOP, NULL);
+ struct l_wait_info lwi;
int rc = -EDQUOT;
- OSC_DUMP_GRANT(D_CACHE, cli, "need:%d.\n", bytes);
+ lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(AT_OFF ? obd_timeout : at_max),
+ NULL, LWI_ON_SIGNAL_NOOP, NULL);
+
+ OSC_DUMP_GRANT(D_CACHE, cli, "need:%d\n", bytes);
spin_lock(&cli->cl_loi_list_lock);
@@ -1595,12 +1597,14 @@ static int osc_enter_cache(const struct lu_env *env, struct client_obd *cli,
if (OBD_FAIL_CHECK(OBD_FAIL_OSC_NO_GRANT) ||
!cli->cl_dirty_max_pages || cli->cl_ar.ar_force_sync ||
loi->loi_ar.ar_force_sync) {
+ OSC_DUMP_GRANT(D_CACHE, cli, "forced sync i/o\n");
rc = -EDQUOT;
goto out;
}
/* Hopefully normal case - cache space and write credits available */
if (osc_enter_cache_try(cli, oap, bytes, 0)) {
+ OSC_DUMP_GRANT(D_CACHE, cli, "granted from cache\n");
rc = 0;
goto out;
}
@@ -1629,32 +1633,49 @@ static int osc_enter_cache(const struct lu_env *env, struct client_obd *cli,
spin_lock(&cli->cl_loi_list_lock);
- /* l_wait_event is interrupted by signal, or timed out */
if (rc < 0) {
- if (rc == -ETIMEDOUT) {
- OSC_DUMP_GRANT(D_ERROR, cli,
- "try to reserve %d.\n", bytes);
- osc_extent_tree_dump(D_ERROR, osc);
- rc = -EDQUOT;
- }
-
+ /* l_wait_event is interrupted by signal, or timed out */
list_del_init(&ocw.ocw_entry);
- goto out;
+ break;
}
-
LASSERT(list_empty(&ocw.ocw_entry));
rc = ocw.ocw_rc;
if (rc != -EDQUOT)
- goto out;
+ break;
if (osc_enter_cache_try(cli, oap, bytes, 0)) {
rc = 0;
- goto out;
+ break;
}
}
+
+ switch (rc) {
+ case 0:
+ OSC_DUMP_GRANT(D_CACHE, cli, "finally got grant space\n");
+ break;
+ case -ETIMEDOUT:
+ OSC_DUMP_GRANT(D_CACHE, cli,
+ "timeout, fall back to sync i/o\n");
+ osc_extent_tree_dump(D_CACHE, osc);
+ /* fall back to synchronous I/O */
+ rc = -EDQUOT;
+ break;
+ case -EINTR:
+ /* Ensures restartability - LU-3581 */
+ OSC_DUMP_GRANT(D_CACHE, cli, "interrupted\n");
+ rc = -ERESTARTSYS;
+ break;
+ case -EDQUOT:
+ OSC_DUMP_GRANT(D_CACHE, cli,
+ "no grant space, fall back to sync i/o\n");
+ break;
+ default:
+ CDEBUG(D_CACHE, "%s: event for cache space @ %p never arrived due to %d, fall back to sync i/o\n",
+ cli->cl_import->imp_obd->obd_name, &ocw, rc);
+ break;
+ }
out:
spin_unlock(&cli->cl_loi_list_lock);
- OSC_DUMP_GRANT(D_CACHE, cli, "returned %d.\n", rc);
return rc;
}
@@ -1679,10 +1700,8 @@ void osc_wake_cache_waiters(struct client_obd *cli)
goto wakeup;
}
- ocw->ocw_rc = 0;
- if (!osc_enter_cache_try(cli, ocw->ocw_oap, ocw->ocw_grant, 0))
- ocw->ocw_rc = -EDQUOT;
-
+ if (osc_enter_cache_try(cli, ocw->ocw_oap, ocw->ocw_grant, 0))
+ ocw->ocw_rc = 0;
wakeup:
CDEBUG(D_CACHE, "wake up %p for oap %p, avail grant %ld, %d\n",
ocw, ocw->ocw_oap, cli->cl_avail_grant, ocw->ocw_rc);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 108/124] staging: lustre: grant: quiet message on grant waiting timeout
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Johann Lombardi, James Simmons
From: Johann Lombardi <johann.lombardi@intel.com>
Use at_max in osc_enter_cache() to bound how long we wait for grant
space before switching to synchronous I/Os. Do not print a message
on the console when the timeout is hit since such long wait can
be legitimate with flaky network (i.e. BRW is resent multiple times).
Signed-off-by: Johann Lombardi <johann.lombardi@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5521
Reviewed-on: http://review.whamcloud.com/12146
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/osc/osc_cache.c | 61 ++++++++++++++++---------
1 files changed, 40 insertions(+), 21 deletions(-)
diff --git a/drivers/staging/lustre/lustre/osc/osc_cache.c b/drivers/staging/lustre/lustre/osc/osc_cache.c
index 8152cc7..6b7677f 100644
--- a/drivers/staging/lustre/lustre/osc/osc_cache.c
+++ b/drivers/staging/lustre/lustre/osc/osc_cache.c
@@ -1536,7 +1536,7 @@ static int osc_enter_cache_try(struct client_obd *cli,
{
int rc;
- OSC_DUMP_GRANT(D_CACHE, cli, "need:%d.\n", bytes);
+ OSC_DUMP_GRANT(D_CACHE, cli, "need:%d\n", bytes);
rc = osc_reserve_grant(cli, bytes);
if (rc < 0)
@@ -1581,11 +1581,13 @@ static int osc_enter_cache(const struct lu_env *env, struct client_obd *cli,
struct osc_object *osc = oap->oap_obj;
struct lov_oinfo *loi = osc->oo_oinfo;
struct osc_cache_waiter ocw;
- struct l_wait_info lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(600), NULL,
- LWI_ON_SIGNAL_NOOP, NULL);
+ struct l_wait_info lwi;
int rc = -EDQUOT;
- OSC_DUMP_GRANT(D_CACHE, cli, "need:%d.\n", bytes);
+ lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(AT_OFF ? obd_timeout : at_max),
+ NULL, LWI_ON_SIGNAL_NOOP, NULL);
+
+ OSC_DUMP_GRANT(D_CACHE, cli, "need:%d\n", bytes);
spin_lock(&cli->cl_loi_list_lock);
@@ -1595,12 +1597,14 @@ static int osc_enter_cache(const struct lu_env *env, struct client_obd *cli,
if (OBD_FAIL_CHECK(OBD_FAIL_OSC_NO_GRANT) ||
!cli->cl_dirty_max_pages || cli->cl_ar.ar_force_sync ||
loi->loi_ar.ar_force_sync) {
+ OSC_DUMP_GRANT(D_CACHE, cli, "forced sync i/o\n");
rc = -EDQUOT;
goto out;
}
/* Hopefully normal case - cache space and write credits available */
if (osc_enter_cache_try(cli, oap, bytes, 0)) {
+ OSC_DUMP_GRANT(D_CACHE, cli, "granted from cache\n");
rc = 0;
goto out;
}
@@ -1629,32 +1633,49 @@ static int osc_enter_cache(const struct lu_env *env, struct client_obd *cli,
spin_lock(&cli->cl_loi_list_lock);
- /* l_wait_event is interrupted by signal, or timed out */
if (rc < 0) {
- if (rc == -ETIMEDOUT) {
- OSC_DUMP_GRANT(D_ERROR, cli,
- "try to reserve %d.\n", bytes);
- osc_extent_tree_dump(D_ERROR, osc);
- rc = -EDQUOT;
- }
-
+ /* l_wait_event is interrupted by signal, or timed out */
list_del_init(&ocw.ocw_entry);
- goto out;
+ break;
}
-
LASSERT(list_empty(&ocw.ocw_entry));
rc = ocw.ocw_rc;
if (rc != -EDQUOT)
- goto out;
+ break;
if (osc_enter_cache_try(cli, oap, bytes, 0)) {
rc = 0;
- goto out;
+ break;
}
}
+
+ switch (rc) {
+ case 0:
+ OSC_DUMP_GRANT(D_CACHE, cli, "finally got grant space\n");
+ break;
+ case -ETIMEDOUT:
+ OSC_DUMP_GRANT(D_CACHE, cli,
+ "timeout, fall back to sync i/o\n");
+ osc_extent_tree_dump(D_CACHE, osc);
+ /* fall back to synchronous I/O */
+ rc = -EDQUOT;
+ break;
+ case -EINTR:
+ /* Ensures restartability - LU-3581 */
+ OSC_DUMP_GRANT(D_CACHE, cli, "interrupted\n");
+ rc = -ERESTARTSYS;
+ break;
+ case -EDQUOT:
+ OSC_DUMP_GRANT(D_CACHE, cli,
+ "no grant space, fall back to sync i/o\n");
+ break;
+ default:
+ CDEBUG(D_CACHE, "%s: event for cache space @ %p never arrived due to %d, fall back to sync i/o\n",
+ cli->cl_import->imp_obd->obd_name, &ocw, rc);
+ break;
+ }
out:
spin_unlock(&cli->cl_loi_list_lock);
- OSC_DUMP_GRANT(D_CACHE, cli, "returned %d.\n", rc);
return rc;
}
@@ -1679,10 +1700,8 @@ void osc_wake_cache_waiters(struct client_obd *cli)
goto wakeup;
}
- ocw->ocw_rc = 0;
- if (!osc_enter_cache_try(cli, ocw->ocw_oap, ocw->ocw_grant, 0))
- ocw->ocw_rc = -EDQUOT;
-
+ if (osc_enter_cache_try(cli, ocw->ocw_oap, ocw->ocw_grant, 0))
+ ocw->ocw_rc = 0;
wakeup:
CDEBUG(D_CACHE, "wake up %p for oap %p, avail grant %ld, %d\n",
ocw, ocw->ocw_oap, cli->cl_avail_grant, ocw->ocw_rc);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 109/124] staging: lustre: misc: remove unnecessary EXPORT_SYMBOL
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/13321
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/fid/fid_lib.c | 2 --
drivers/staging/lustre/lustre/fld/fld_request.c | 2 --
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 1 -
drivers/staging/lustre/lustre/lov/lov_dev.c | 1 -
4 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/fid/fid_lib.c b/drivers/staging/lustre/lustre/fid/fid_lib.c
index 99ae7eb..4e49cb3 100644
--- a/drivers/staging/lustre/lustre/fid/fid_lib.c
+++ b/drivers/staging/lustre/lustre/fid/fid_lib.c
@@ -63,14 +63,12 @@ const struct lu_seq_range LUSTRE_SEQ_SPACE_RANGE = {
FID_SEQ_NORMAL,
(__u64)~0ULL
};
-EXPORT_SYMBOL(LUSTRE_SEQ_SPACE_RANGE);
/* Zero range, used for init and other purposes. */
const struct lu_seq_range LUSTRE_SEQ_ZERO_RANGE = {
0,
0
};
-EXPORT_SYMBOL(LUSTRE_SEQ_ZERO_RANGE);
/* Lustre Big Fs Lock fid. */
const struct lu_fid LUSTRE_BFL_FID = { .f_seq = FID_SEQ_SPECIAL,
diff --git a/drivers/staging/lustre/lustre/fld/fld_request.c b/drivers/staging/lustre/lustre/fld/fld_request.c
index ed7962e..0de72b7 100644
--- a/drivers/staging/lustre/lustre/fld/fld_request.c
+++ b/drivers/staging/lustre/lustre/fld/fld_request.c
@@ -219,7 +219,6 @@ int fld_client_del_target(struct lu_client_fld *fld, __u64 idx)
spin_unlock(&fld->lcf_lock);
return -ENOENT;
}
-EXPORT_SYMBOL(fld_client_del_target);
static struct dentry *fld_debugfs_dir;
@@ -454,7 +453,6 @@ void fld_client_flush(struct lu_client_fld *fld)
{
fld_cache_flush(fld->lcf_cache);
}
-EXPORT_SYMBOL(fld_client_flush);
static int __init fld_init(void)
{
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 941981d..cccb645 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2804,7 +2804,6 @@ lmv_pack_md(union lmv_mds_md **lmmp, const struct lmv_stripe_md *lsm,
return lmm_size;
}
-EXPORT_SYMBOL(lmv_pack_md);
static int lmv_unpack_md_v1(struct obd_export *exp, struct lmv_stripe_md *lsm,
const struct lmv_mds_md_v1 *lmm1)
diff --git a/drivers/staging/lustre/lustre/lov/lov_dev.c b/drivers/staging/lustre/lustre/lov/lov_dev.c
index b1f260d..056ae2e 100644
--- a/drivers/staging/lustre/lustre/lov/lov_dev.c
+++ b/drivers/staging/lustre/lustre/lov/lov_dev.c
@@ -516,6 +516,5 @@ struct lu_device_type lov_device_type = {
.ldt_ops = &lov_device_type_ops,
.ldt_ctx_tags = LCT_CL_THREAD
};
-EXPORT_SYMBOL(lov_device_type);
/** @} lov */
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 109/124] staging: lustre: misc: remove unnecessary EXPORT_SYMBOL
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/13321
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/fid/fid_lib.c | 2 --
drivers/staging/lustre/lustre/fld/fld_request.c | 2 --
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 1 -
drivers/staging/lustre/lustre/lov/lov_dev.c | 1 -
4 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/drivers/staging/lustre/lustre/fid/fid_lib.c b/drivers/staging/lustre/lustre/fid/fid_lib.c
index 99ae7eb..4e49cb3 100644
--- a/drivers/staging/lustre/lustre/fid/fid_lib.c
+++ b/drivers/staging/lustre/lustre/fid/fid_lib.c
@@ -63,14 +63,12 @@ const struct lu_seq_range LUSTRE_SEQ_SPACE_RANGE = {
FID_SEQ_NORMAL,
(__u64)~0ULL
};
-EXPORT_SYMBOL(LUSTRE_SEQ_SPACE_RANGE);
/* Zero range, used for init and other purposes. */
const struct lu_seq_range LUSTRE_SEQ_ZERO_RANGE = {
0,
0
};
-EXPORT_SYMBOL(LUSTRE_SEQ_ZERO_RANGE);
/* Lustre Big Fs Lock fid. */
const struct lu_fid LUSTRE_BFL_FID = { .f_seq = FID_SEQ_SPECIAL,
diff --git a/drivers/staging/lustre/lustre/fld/fld_request.c b/drivers/staging/lustre/lustre/fld/fld_request.c
index ed7962e..0de72b7 100644
--- a/drivers/staging/lustre/lustre/fld/fld_request.c
+++ b/drivers/staging/lustre/lustre/fld/fld_request.c
@@ -219,7 +219,6 @@ int fld_client_del_target(struct lu_client_fld *fld, __u64 idx)
spin_unlock(&fld->lcf_lock);
return -ENOENT;
}
-EXPORT_SYMBOL(fld_client_del_target);
static struct dentry *fld_debugfs_dir;
@@ -454,7 +453,6 @@ void fld_client_flush(struct lu_client_fld *fld)
{
fld_cache_flush(fld->lcf_cache);
}
-EXPORT_SYMBOL(fld_client_flush);
static int __init fld_init(void)
{
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index 941981d..cccb645 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2804,7 +2804,6 @@ lmv_pack_md(union lmv_mds_md **lmmp, const struct lmv_stripe_md *lsm,
return lmm_size;
}
-EXPORT_SYMBOL(lmv_pack_md);
static int lmv_unpack_md_v1(struct obd_export *exp, struct lmv_stripe_md *lsm,
const struct lmv_mds_md_v1 *lmm1)
diff --git a/drivers/staging/lustre/lustre/lov/lov_dev.c b/drivers/staging/lustre/lustre/lov/lov_dev.c
index b1f260d..056ae2e 100644
--- a/drivers/staging/lustre/lustre/lov/lov_dev.c
+++ b/drivers/staging/lustre/lustre/lov/lov_dev.c
@@ -516,6 +516,5 @@ struct lu_device_type lov_device_type = {
.ldt_ops = &lov_device_type_ops,
.ldt_ctx_tags = LCT_CL_THREAD
};
-EXPORT_SYMBOL(lov_device_type);
/** @} lov */
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 110/124] staging: lustre: obdclass: remove unnecessary EXPORT_SYMBOL
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/13323
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/cl_io.c | 1 -
drivers/staging/lustre/lustre/obdclass/cl_page.c | 2 --
drivers/staging/lustre/lustre/obdclass/class_obd.c | 3 ---
drivers/staging/lustre/lustre/obdclass/genops.c | 8 --------
.../lustre/lustre/obdclass/linux/linux-module.c | 1 -
drivers/staging/lustre/lustre/obdclass/llog_obd.c | 1 -
drivers/staging/lustre/lustre/obdclass/llog_swab.c | 2 --
.../lustre/lustre/obdclass/lprocfs_status.c | 2 --
drivers/staging/lustre/lustre/obdclass/lu_object.c | 4 ----
.../staging/lustre/lustre/obdclass/lustre_peer.c | 1 -
10 files changed, 0 insertions(+), 25 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_io.c b/drivers/staging/lustre/lustre/obdclass/cl_io.c
index 7894cf9..bc4b7b6 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_io.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_io.c
@@ -73,7 +73,6 @@ int cl_io_is_going(const struct lu_env *env)
{
return cl_env_info(env)->clt_current_io != NULL;
}
-EXPORT_SYMBOL(cl_io_is_going);
/**
* cl_io invariant that holds at all times when exported cl_io_*() functions
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index 18784a8..63973ba 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -170,7 +170,6 @@ struct cl_page *cl_page_alloc(const struct lu_env *env,
}
return page;
}
-EXPORT_SYMBOL(cl_page_alloc);
/**
* Returns a cl_page with index \a idx at the object \a o, and associated with
@@ -1006,7 +1005,6 @@ int cl_page_cancel(const struct lu_env *env, struct cl_page *page)
(const struct lu_env *,
const struct cl_page_slice *));
}
-EXPORT_SYMBOL(cl_page_cancel);
/**
* Converts a byte offset within object \a obj into a page index.
diff --git a/drivers/staging/lustre/lustre/obdclass/class_obd.c b/drivers/staging/lustre/lustre/obdclass/class_obd.c
index 4c30d6b..76e1ee8 100644
--- a/drivers/staging/lustre/lustre/obdclass/class_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/class_obd.c
@@ -44,7 +44,6 @@
#include "llog_internal.h"
struct obd_device *obd_devs[MAX_OBD_DEVICES];
-EXPORT_SYMBOL(obd_devs);
struct list_head obd_types;
DEFINE_RWLOCK(obd_dev_lock);
@@ -79,8 +78,6 @@ atomic_long_t obd_dirty_transit_pages;
EXPORT_SYMBOL(obd_dirty_transit_pages);
char obd_jobid_var[JOBSTATS_JOBID_VAR_MAX_LEN + 1] = JOBSTATS_DISABLE;
-EXPORT_SYMBOL(obd_jobid_var);
-
char obd_jobid_node[LUSTRE_JOBID_SIZE + 1];
/* Get jobid of current process from stored variable or calculate
diff --git a/drivers/staging/lustre/lustre/obdclass/genops.c b/drivers/staging/lustre/lustre/obdclass/genops.c
index 0bc623e..cf8bb2a 100644
--- a/drivers/staging/lustre/lustre/obdclass/genops.c
+++ b/drivers/staging/lustre/lustre/obdclass/genops.c
@@ -133,7 +133,6 @@ void class_put_type(struct obd_type *type)
module_put(type->typ_dt_ops->owner);
spin_unlock(&type->obd_type_lock);
}
-EXPORT_SYMBOL(class_put_type);
#define CLASS_MAX_NAME 1024
@@ -391,7 +390,6 @@ int class_name2dev(const char *name)
return -1;
}
-EXPORT_SYMBOL(class_name2dev);
struct obd_device *class_name2obd(const char *name)
{
@@ -421,7 +419,6 @@ int class_uuid2dev(struct obd_uuid *uuid)
return -1;
}
-EXPORT_SYMBOL(class_uuid2dev);
/**
* Get obd device from ::obd_devs[]
@@ -450,7 +447,6 @@ struct obd_device *class_num2obd(int num)
return obd;
}
-EXPORT_SYMBOL(class_num2obd);
/* Search for a client OBD connected to tgt_uuid. If grp_uuid is
* specified, then only the client with that uuid is returned,
@@ -817,7 +813,6 @@ void class_unlink_export(struct obd_export *exp)
spin_unlock(&exp->exp_obd->obd_dev_lock);
class_export_put(exp);
}
-EXPORT_SYMBOL(class_unlink_export);
/* Import management functions */
static void class_import_destroy(struct obd_import *imp)
@@ -973,7 +968,6 @@ void __class_export_add_lock_ref(struct obd_export *exp, struct ldlm_lock *lock)
lock, exp, lock->l_exp_refs_nr);
spin_unlock(&exp->exp_locks_list_guard);
}
-EXPORT_SYMBOL(__class_export_add_lock_ref);
void __class_export_del_lock_ref(struct obd_export *exp, struct ldlm_lock *lock)
{
@@ -991,7 +985,6 @@ void __class_export_del_lock_ref(struct obd_export *exp, struct ldlm_lock *lock)
lock, exp, lock->l_exp_refs_nr);
spin_unlock(&exp->exp_locks_list_guard);
}
-EXPORT_SYMBOL(__class_export_del_lock_ref);
#endif
/* A connection defines an export context in which preallocation can
@@ -1100,7 +1093,6 @@ EXPORT_SYMBOL(class_fail_export);
#if LUSTRE_TRACKS_LOCK_EXP_REFS
void (*class_export_dump_hook)(struct obd_export *) = NULL;
-EXPORT_SYMBOL(class_export_dump_hook);
#endif
/* Total amount of zombies to be destroyed */
diff --git a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
index ae172c4..be09e04 100644
--- a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
+++ b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
@@ -158,7 +158,6 @@ int obd_ioctl_popdata(void __user *arg, void *data, int len)
err = copy_to_user(arg, data, len) ? -EFAULT : 0;
return err;
}
-EXPORT_SYMBOL(obd_ioctl_popdata);
/* opening /dev/obd */
static int obd_class_open(struct inode *inode, struct file *file)
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_obd.c b/drivers/staging/lustre/lustre/obdclass/llog_obd.c
index 6ace7e0..a4277d6 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_obd.c
@@ -210,7 +210,6 @@ LU_KEY_INIT_FINI(llog, struct llog_thread_info);
/* context key: llog_thread_key */
LU_CONTEXT_KEY_DEFINE(llog, LCT_MD_THREAD | LCT_MG_THREAD | LCT_LOCAL);
LU_KEY_INIT_GENERIC(llog);
-EXPORT_SYMBOL(llog_thread_key);
int llog_info_init(void)
{
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_swab.c b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
index 21ab90c..8c4c1b3 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_swab.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
@@ -347,7 +347,6 @@ void lustre_swab_lustre_cfg(struct lustre_cfg *lcfg)
print_lustre_cfg(lcfg);
}
-EXPORT_SYMBOL(lustre_swab_lustre_cfg);
/* used only for compatibility with old on-disk cfg_marker data */
struct cfg_marker32 {
@@ -407,4 +406,3 @@ void lustre_swab_cfg_marker(struct cfg_marker *marker, int swab, int size)
__swab64s(&marker->cm_canceltime);
}
}
-EXPORT_SYMBOL(lustre_swab_cfg_marker);
diff --git a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
index be6b6af..852a5ac 100644
--- a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
+++ b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
@@ -621,7 +621,6 @@ void lprocfs_stats_collect(struct lprocfs_stats *stats, int idx,
lprocfs_stats_unlock(stats, LPROCFS_GET_NUM_CPU, &flags);
}
-EXPORT_SYMBOL(lprocfs_stats_collect);
/**
* Append a space separated list of current set flags to str.
@@ -1049,7 +1048,6 @@ int lprocfs_stats_alloc_one(struct lprocfs_stats *stats, unsigned int cpuid)
}
return rc;
}
-EXPORT_SYMBOL(lprocfs_stats_alloc_one);
struct lprocfs_stats *lprocfs_alloc_stats(unsigned int num,
enum lprocfs_stats_flags flags)
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 3f1fb28..054e567 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -1300,7 +1300,6 @@ void lu_stack_fini(const struct lu_env *env, struct lu_device *top)
}
}
}
-EXPORT_SYMBOL(lu_stack_fini);
enum {
/**
@@ -1522,14 +1521,12 @@ void lu_context_key_quiesce(struct lu_context_key *key)
++key_set_version;
}
}
-EXPORT_SYMBOL(lu_context_key_quiesce);
void lu_context_key_revive(struct lu_context_key *key)
{
key->lct_tags &= ~LCT_QUIESCENT;
++key_set_version;
}
-EXPORT_SYMBOL(lu_context_key_revive);
static void keys_fini(struct lu_context *ctx)
{
@@ -1688,7 +1685,6 @@ int lu_context_refill(struct lu_context *ctx)
{
return likely(ctx->lc_version == key_set_version) ? 0 : keys_fill(ctx);
}
-EXPORT_SYMBOL(lu_context_refill);
/**
* lu_ctx_tags/lu_ses_tags will be updated if there are new types of
diff --git a/drivers/staging/lustre/lustre/obdclass/lustre_peer.c b/drivers/staging/lustre/lustre/obdclass/lustre_peer.c
index 5974a9b..ffa740a 100644
--- a/drivers/staging/lustre/lustre/obdclass/lustre_peer.c
+++ b/drivers/staging/lustre/lustre/obdclass/lustre_peer.c
@@ -139,7 +139,6 @@ int class_add_uuid(const char *uuid, __u64 nid)
}
return 0;
}
-EXPORT_SYMBOL(class_add_uuid);
/* Delete the nids for one uuid if specified, otherwise delete all */
int class_del_uuid(const char *uuid)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 110/124] staging: lustre: obdclass: remove unnecessary EXPORT_SYMBOL
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/13323
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/obdclass/cl_io.c | 1 -
drivers/staging/lustre/lustre/obdclass/cl_page.c | 2 --
drivers/staging/lustre/lustre/obdclass/class_obd.c | 3 ---
drivers/staging/lustre/lustre/obdclass/genops.c | 8 --------
.../lustre/lustre/obdclass/linux/linux-module.c | 1 -
drivers/staging/lustre/lustre/obdclass/llog_obd.c | 1 -
drivers/staging/lustre/lustre/obdclass/llog_swab.c | 2 --
.../lustre/lustre/obdclass/lprocfs_status.c | 2 --
drivers/staging/lustre/lustre/obdclass/lu_object.c | 4 ----
.../staging/lustre/lustre/obdclass/lustre_peer.c | 1 -
10 files changed, 0 insertions(+), 25 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_io.c b/drivers/staging/lustre/lustre/obdclass/cl_io.c
index 7894cf9..bc4b7b6 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_io.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_io.c
@@ -73,7 +73,6 @@ int cl_io_is_going(const struct lu_env *env)
{
return cl_env_info(env)->clt_current_io != NULL;
}
-EXPORT_SYMBOL(cl_io_is_going);
/**
* cl_io invariant that holds at all times when exported cl_io_*() functions
diff --git a/drivers/staging/lustre/lustre/obdclass/cl_page.c b/drivers/staging/lustre/lustre/obdclass/cl_page.c
index 18784a8..63973ba 100644
--- a/drivers/staging/lustre/lustre/obdclass/cl_page.c
+++ b/drivers/staging/lustre/lustre/obdclass/cl_page.c
@@ -170,7 +170,6 @@ struct cl_page *cl_page_alloc(const struct lu_env *env,
}
return page;
}
-EXPORT_SYMBOL(cl_page_alloc);
/**
* Returns a cl_page with index \a idx at the object \a o, and associated with
@@ -1006,7 +1005,6 @@ int cl_page_cancel(const struct lu_env *env, struct cl_page *page)
(const struct lu_env *,
const struct cl_page_slice *));
}
-EXPORT_SYMBOL(cl_page_cancel);
/**
* Converts a byte offset within object \a obj into a page index.
diff --git a/drivers/staging/lustre/lustre/obdclass/class_obd.c b/drivers/staging/lustre/lustre/obdclass/class_obd.c
index 4c30d6b..76e1ee8 100644
--- a/drivers/staging/lustre/lustre/obdclass/class_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/class_obd.c
@@ -44,7 +44,6 @@
#include "llog_internal.h"
struct obd_device *obd_devs[MAX_OBD_DEVICES];
-EXPORT_SYMBOL(obd_devs);
struct list_head obd_types;
DEFINE_RWLOCK(obd_dev_lock);
@@ -79,8 +78,6 @@ atomic_long_t obd_dirty_transit_pages;
EXPORT_SYMBOL(obd_dirty_transit_pages);
char obd_jobid_var[JOBSTATS_JOBID_VAR_MAX_LEN + 1] = JOBSTATS_DISABLE;
-EXPORT_SYMBOL(obd_jobid_var);
-
char obd_jobid_node[LUSTRE_JOBID_SIZE + 1];
/* Get jobid of current process from stored variable or calculate
diff --git a/drivers/staging/lustre/lustre/obdclass/genops.c b/drivers/staging/lustre/lustre/obdclass/genops.c
index 0bc623e..cf8bb2a 100644
--- a/drivers/staging/lustre/lustre/obdclass/genops.c
+++ b/drivers/staging/lustre/lustre/obdclass/genops.c
@@ -133,7 +133,6 @@ void class_put_type(struct obd_type *type)
module_put(type->typ_dt_ops->owner);
spin_unlock(&type->obd_type_lock);
}
-EXPORT_SYMBOL(class_put_type);
#define CLASS_MAX_NAME 1024
@@ -391,7 +390,6 @@ int class_name2dev(const char *name)
return -1;
}
-EXPORT_SYMBOL(class_name2dev);
struct obd_device *class_name2obd(const char *name)
{
@@ -421,7 +419,6 @@ int class_uuid2dev(struct obd_uuid *uuid)
return -1;
}
-EXPORT_SYMBOL(class_uuid2dev);
/**
* Get obd device from ::obd_devs[]
@@ -450,7 +447,6 @@ struct obd_device *class_num2obd(int num)
return obd;
}
-EXPORT_SYMBOL(class_num2obd);
/* Search for a client OBD connected to tgt_uuid. If grp_uuid is
* specified, then only the client with that uuid is returned,
@@ -817,7 +813,6 @@ void class_unlink_export(struct obd_export *exp)
spin_unlock(&exp->exp_obd->obd_dev_lock);
class_export_put(exp);
}
-EXPORT_SYMBOL(class_unlink_export);
/* Import management functions */
static void class_import_destroy(struct obd_import *imp)
@@ -973,7 +968,6 @@ void __class_export_add_lock_ref(struct obd_export *exp, struct ldlm_lock *lock)
lock, exp, lock->l_exp_refs_nr);
spin_unlock(&exp->exp_locks_list_guard);
}
-EXPORT_SYMBOL(__class_export_add_lock_ref);
void __class_export_del_lock_ref(struct obd_export *exp, struct ldlm_lock *lock)
{
@@ -991,7 +985,6 @@ void __class_export_del_lock_ref(struct obd_export *exp, struct ldlm_lock *lock)
lock, exp, lock->l_exp_refs_nr);
spin_unlock(&exp->exp_locks_list_guard);
}
-EXPORT_SYMBOL(__class_export_del_lock_ref);
#endif
/* A connection defines an export context in which preallocation can
@@ -1100,7 +1093,6 @@ EXPORT_SYMBOL(class_fail_export);
#if LUSTRE_TRACKS_LOCK_EXP_REFS
void (*class_export_dump_hook)(struct obd_export *) = NULL;
-EXPORT_SYMBOL(class_export_dump_hook);
#endif
/* Total amount of zombies to be destroyed */
diff --git a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
index ae172c4..be09e04 100644
--- a/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
+++ b/drivers/staging/lustre/lustre/obdclass/linux/linux-module.c
@@ -158,7 +158,6 @@ int obd_ioctl_popdata(void __user *arg, void *data, int len)
err = copy_to_user(arg, data, len) ? -EFAULT : 0;
return err;
}
-EXPORT_SYMBOL(obd_ioctl_popdata);
/* opening /dev/obd */
static int obd_class_open(struct inode *inode, struct file *file)
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_obd.c b/drivers/staging/lustre/lustre/obdclass/llog_obd.c
index 6ace7e0..a4277d6 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_obd.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_obd.c
@@ -210,7 +210,6 @@ LU_KEY_INIT_FINI(llog, struct llog_thread_info);
/* context key: llog_thread_key */
LU_CONTEXT_KEY_DEFINE(llog, LCT_MD_THREAD | LCT_MG_THREAD | LCT_LOCAL);
LU_KEY_INIT_GENERIC(llog);
-EXPORT_SYMBOL(llog_thread_key);
int llog_info_init(void)
{
diff --git a/drivers/staging/lustre/lustre/obdclass/llog_swab.c b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
index 21ab90c..8c4c1b3 100644
--- a/drivers/staging/lustre/lustre/obdclass/llog_swab.c
+++ b/drivers/staging/lustre/lustre/obdclass/llog_swab.c
@@ -347,7 +347,6 @@ void lustre_swab_lustre_cfg(struct lustre_cfg *lcfg)
print_lustre_cfg(lcfg);
}
-EXPORT_SYMBOL(lustre_swab_lustre_cfg);
/* used only for compatibility with old on-disk cfg_marker data */
struct cfg_marker32 {
@@ -407,4 +406,3 @@ void lustre_swab_cfg_marker(struct cfg_marker *marker, int swab, int size)
__swab64s(&marker->cm_canceltime);
}
}
-EXPORT_SYMBOL(lustre_swab_cfg_marker);
diff --git a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
index be6b6af..852a5ac 100644
--- a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
+++ b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
@@ -621,7 +621,6 @@ void lprocfs_stats_collect(struct lprocfs_stats *stats, int idx,
lprocfs_stats_unlock(stats, LPROCFS_GET_NUM_CPU, &flags);
}
-EXPORT_SYMBOL(lprocfs_stats_collect);
/**
* Append a space separated list of current set flags to str.
@@ -1049,7 +1048,6 @@ int lprocfs_stats_alloc_one(struct lprocfs_stats *stats, unsigned int cpuid)
}
return rc;
}
-EXPORT_SYMBOL(lprocfs_stats_alloc_one);
struct lprocfs_stats *lprocfs_alloc_stats(unsigned int num,
enum lprocfs_stats_flags flags)
diff --git a/drivers/staging/lustre/lustre/obdclass/lu_object.c b/drivers/staging/lustre/lustre/obdclass/lu_object.c
index 3f1fb28..054e567 100644
--- a/drivers/staging/lustre/lustre/obdclass/lu_object.c
+++ b/drivers/staging/lustre/lustre/obdclass/lu_object.c
@@ -1300,7 +1300,6 @@ void lu_stack_fini(const struct lu_env *env, struct lu_device *top)
}
}
}
-EXPORT_SYMBOL(lu_stack_fini);
enum {
/**
@@ -1522,14 +1521,12 @@ void lu_context_key_quiesce(struct lu_context_key *key)
++key_set_version;
}
}
-EXPORT_SYMBOL(lu_context_key_quiesce);
void lu_context_key_revive(struct lu_context_key *key)
{
key->lct_tags &= ~LCT_QUIESCENT;
++key_set_version;
}
-EXPORT_SYMBOL(lu_context_key_revive);
static void keys_fini(struct lu_context *ctx)
{
@@ -1688,7 +1685,6 @@ int lu_context_refill(struct lu_context *ctx)
{
return likely(ctx->lc_version == key_set_version) ? 0 : keys_fill(ctx);
}
-EXPORT_SYMBOL(lu_context_refill);
/**
* lu_ctx_tags/lu_ses_tags will be updated if there are new types of
diff --git a/drivers/staging/lustre/lustre/obdclass/lustre_peer.c b/drivers/staging/lustre/lustre/obdclass/lustre_peer.c
index 5974a9b..ffa740a 100644
--- a/drivers/staging/lustre/lustre/obdclass/lustre_peer.c
+++ b/drivers/staging/lustre/lustre/obdclass/lustre_peer.c
@@ -139,7 +139,6 @@ int class_add_uuid(const char *uuid, __u64 nid)
}
return 0;
}
-EXPORT_SYMBOL(class_add_uuid);
/* Delete the nids for one uuid if specified, otherwise delete all */
int class_del_uuid(const char *uuid)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 111/124] staging: lustre: llite: lock the inode to be migrated
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Because the inode and its connected dentries will be cleared
out of the cache after migration, the inode needs to be locked
during the migration.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4712
Reviewed-on: http://review.whamcloud.com/9689
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/file.c | 9 +++++++--
1 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index d8761b8..5d4d17f 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -2802,11 +2802,15 @@ int ll_migrate(struct inode *parent, struct file *file, int mdtidx,
qstr.name = name;
qstr.len = namelen;
dchild = d_lookup(file_dentry(file), &qstr);
- if (dchild && dchild->d_inode) {
+ if (dchild) {
op_data->op_fid3 = *ll_inode2fid(dchild->d_inode);
if (dchild->d_inode) {
child_inode = igrab(dchild->d_inode);
- ll_invalidate_aliases(child_inode);
+ if (child_inode) {
+ inode_lock(child_inode);
+ op_data->op_fid3 = *ll_inode2fid(child_inode);
+ ll_invalidate_aliases(child_inode);
+ }
}
dput(dchild);
} else {
@@ -2847,6 +2851,7 @@ int ll_migrate(struct inode *parent, struct file *file, int mdtidx,
out_free:
if (child_inode) {
clear_nlink(child_inode);
+ inode_unlock(child_inode);
iput(child_inode);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 111/124] staging: lustre: llite: lock the inode to be migrated
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Because the inode and its connected dentries will be cleared
out of the cache after migration, the inode needs to be locked
during the migration.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4712
Reviewed-on: http://review.whamcloud.com/9689
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/llite/file.c | 9 +++++++--
1 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/file.c b/drivers/staging/lustre/lustre/llite/file.c
index d8761b8..5d4d17f 100644
--- a/drivers/staging/lustre/lustre/llite/file.c
+++ b/drivers/staging/lustre/lustre/llite/file.c
@@ -2802,11 +2802,15 @@ int ll_migrate(struct inode *parent, struct file *file, int mdtidx,
qstr.name = name;
qstr.len = namelen;
dchild = d_lookup(file_dentry(file), &qstr);
- if (dchild && dchild->d_inode) {
+ if (dchild) {
op_data->op_fid3 = *ll_inode2fid(dchild->d_inode);
if (dchild->d_inode) {
child_inode = igrab(dchild->d_inode);
- ll_invalidate_aliases(child_inode);
+ if (child_inode) {
+ inode_lock(child_inode);
+ op_data->op_fid3 = *ll_inode2fid(child_inode);
+ ll_invalidate_aliases(child_inode);
+ }
}
dput(dchild);
} else {
@@ -2847,6 +2851,7 @@ int ll_migrate(struct inode *parent, struct file *file, int mdtidx,
out_free:
if (child_inode) {
clear_nlink(child_inode);
+ inode_unlock(child_inode);
iput(child_inode);
}
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 112/124] staging: lustre: ptlrpc: remove unnecessary EXPORT_SYMBOL
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/12510
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_net.h | 7 ---
drivers/staging/lustre/lustre/ptlrpc/client.c | 17 +-----
drivers/staging/lustre/lustre/ptlrpc/connection.c | 5 --
drivers/staging/lustre/lustre/ptlrpc/import.c | 1 -
drivers/staging/lustre/lustre/ptlrpc/niobuf.c | 5 --
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 57 --------------------
drivers/staging/lustre/lustre/ptlrpc/pinger.c | 1 -
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 5 ++
drivers/staging/lustre/lustre/ptlrpc/recover.c | 2 -
drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c | 4 --
drivers/staging/lustre/lustre/ptlrpc/sec_config.c | 1 -
drivers/staging/lustre/lustre/ptlrpc/sec_gc.c | 2 -
drivers/staging/lustre/lustre/ptlrpc/service.c | 1 -
13 files changed, 8 insertions(+), 100 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 21e2267..e9aba99 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -2408,7 +2408,6 @@ int ptlrpc_send_reply(struct ptlrpc_request *req, int flags);
int ptlrpc_reply(struct ptlrpc_request *req);
int ptlrpc_send_error(struct ptlrpc_request *req, int difficult);
int ptlrpc_error(struct ptlrpc_request *req);
-void ptlrpc_resend_req(struct ptlrpc_request *request);
int ptlrpc_at_get_net_latency(struct ptlrpc_request *req);
int ptl_send_rpc(struct ptlrpc_request *request, int noreply);
int ptlrpc_register_rqbd(struct ptlrpc_request_buffer_desc *rqbd);
@@ -2428,23 +2427,17 @@ struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid);
int ptlrpc_queue_wait(struct ptlrpc_request *req);
int ptlrpc_replay_req(struct ptlrpc_request *req);
-int ptlrpc_unregister_reply(struct ptlrpc_request *req, int async);
void ptlrpc_abort_inflight(struct obd_import *imp);
void ptlrpc_abort_set(struct ptlrpc_request_set *set);
struct ptlrpc_request_set *ptlrpc_prep_set(void);
struct ptlrpc_request_set *ptlrpc_prep_fcset(int max, set_producer_func func,
void *arg);
-int ptlrpc_set_next_timeout(struct ptlrpc_request_set *);
int ptlrpc_check_set(const struct lu_env *env, struct ptlrpc_request_set *set);
int ptlrpc_set_wait(struct ptlrpc_request_set *);
-int ptlrpc_expired_set(void *data);
-void ptlrpc_interrupted_set(void *data);
void ptlrpc_mark_interrupted(struct ptlrpc_request *req);
void ptlrpc_set_destroy(struct ptlrpc_request_set *);
void ptlrpc_set_add_req(struct ptlrpc_request_set *, struct ptlrpc_request *);
-void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc,
- struct ptlrpc_request *req);
void ptlrpc_free_rq_pool(struct ptlrpc_request_pool *pool);
int ptlrpc_add_rqs_to_pool(struct ptlrpc_request_pool *pool, int num_rq);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index f3914cc..18c7ff7 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -45,6 +45,7 @@
static int ptlrpc_send_new_req(struct ptlrpc_request *req);
static int ptlrpcd_check_work(struct ptlrpc_request *req);
+static int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async);
/**
* Initialize passed in client structure \a cl.
@@ -89,7 +90,6 @@ struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid)
return c;
}
-EXPORT_SYMBOL(ptlrpc_uuid_to_connection);
/**
* Allocate and initialize new bulk descriptor on the sender.
@@ -886,7 +886,6 @@ struct ptlrpc_request_set *ptlrpc_prep_fcset(int max, set_producer_func func,
return set;
}
-EXPORT_SYMBOL(ptlrpc_prep_fcset);
/**
* Wind down and free request set structure previously allocated with
@@ -1006,7 +1005,6 @@ void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc,
wake_up(&pc->pc_partners[i]->pc_set->set_waitq);
}
}
-EXPORT_SYMBOL(ptlrpc_set_add_new_req);
/**
* Based on the current state of the import, determine if the request
@@ -2005,7 +2003,6 @@ int ptlrpc_expired_set(void *data)
*/
return 1;
}
-EXPORT_SYMBOL(ptlrpc_expired_set);
/**
* Sets rq_intr flag in \a req under spinlock.
@@ -2022,7 +2019,7 @@ EXPORT_SYMBOL(ptlrpc_mark_interrupted);
* Interrupts (sets interrupted flag) all uncompleted requests in
* a set \a data. Callback for l_wait_event for interruptible waits.
*/
-void ptlrpc_interrupted_set(void *data)
+static void ptlrpc_interrupted_set(void *data)
{
struct ptlrpc_request_set *set = data;
struct list_head *tmp;
@@ -2040,7 +2037,6 @@ void ptlrpc_interrupted_set(void *data)
ptlrpc_mark_interrupted(req);
}
}
-EXPORT_SYMBOL(ptlrpc_interrupted_set);
/**
* Get the smallest timeout in the set; this does NOT set a timeout.
@@ -2084,7 +2080,6 @@ int ptlrpc_set_next_timeout(struct ptlrpc_request_set *set)
}
return timeout;
}
-EXPORT_SYMBOL(ptlrpc_set_next_timeout);
/**
* Send all unset request from the set and then wait until all
@@ -2335,7 +2330,7 @@ EXPORT_SYMBOL(ptlrpc_req_xid);
* The request owner (i.e. the thread doing the I/O) must call...
* Returns 0 on success or 1 if unregistering cannot be made.
*/
-int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
+static int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
{
int rc;
wait_queue_head_t *wq;
@@ -2400,7 +2395,6 @@ int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
}
return 0;
}
-EXPORT_SYMBOL(ptlrpc_unregister_reply);
static void ptlrpc_free_request(struct ptlrpc_request *req)
{
@@ -2555,7 +2549,6 @@ void ptlrpc_resend_req(struct ptlrpc_request *req)
ptlrpc_client_wake_req(req);
spin_unlock(&req->rq_lock);
}
-EXPORT_SYMBOL(ptlrpc_resend_req);
/**
* Grab additional reference on a request \a req
@@ -2624,7 +2617,6 @@ void ptlrpc_retain_replayable_request(struct ptlrpc_request *req,
list_add(&req->rq_replay_list, &imp->imp_replay_list);
}
-EXPORT_SYMBOL(ptlrpc_retain_replayable_request);
/**
* Send request and wait until it completes.
@@ -2797,7 +2789,6 @@ int ptlrpc_replay_req(struct ptlrpc_request *req)
ptlrpcd_add_req(req);
return 0;
}
-EXPORT_SYMBOL(ptlrpc_replay_req);
/**
* Aborts all in-flight request on import \a imp sending and delayed lists
@@ -2857,7 +2848,6 @@ void ptlrpc_abort_inflight(struct obd_import *imp)
spin_unlock(&imp->imp_lock);
}
-EXPORT_SYMBOL(ptlrpc_abort_inflight);
/**
* Abort all uncompleted requests in request set \a set
@@ -2943,7 +2933,6 @@ __u64 ptlrpc_next_xid(void)
return next;
}
-EXPORT_SYMBOL(ptlrpc_next_xid);
/**
* Get a glimpse at what next xid value might have been.
diff --git a/drivers/staging/lustre/lustre/ptlrpc/connection.c b/drivers/staging/lustre/lustre/ptlrpc/connection.c
index 177a379..7b020d6 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/connection.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/connection.c
@@ -82,7 +82,6 @@ out:
libcfs_nid2str(conn->c_peer.nid));
return conn;
}
-EXPORT_SYMBOL(ptlrpc_connection_get);
int ptlrpc_connection_put(struct ptlrpc_connection *conn)
{
@@ -118,7 +117,6 @@ int ptlrpc_connection_put(struct ptlrpc_connection *conn)
return rc;
}
-EXPORT_SYMBOL(ptlrpc_connection_put);
struct ptlrpc_connection *
ptlrpc_connection_addref(struct ptlrpc_connection *conn)
@@ -130,7 +128,6 @@ ptlrpc_connection_addref(struct ptlrpc_connection *conn)
return conn;
}
-EXPORT_SYMBOL(ptlrpc_connection_addref);
int ptlrpc_connection_init(void)
{
@@ -146,13 +143,11 @@ int ptlrpc_connection_init(void)
return 0;
}
-EXPORT_SYMBOL(ptlrpc_connection_init);
void ptlrpc_connection_fini(void)
{
cfs_hash_putref(conn_hash);
}
-EXPORT_SYMBOL(ptlrpc_connection_fini);
/*
* Hash operations for net_peer<->connection
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 20daa06..5333cda 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -425,7 +425,6 @@ void ptlrpc_fail_import(struct obd_import *imp, __u32 conn_cnt)
ptlrpc_pinger_force(imp);
}
}
-EXPORT_SYMBOL(ptlrpc_fail_import);
int ptlrpc_reconnect_import(struct obd_import *imp)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
index 5bab248..9c93739 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
@@ -295,7 +295,6 @@ int ptlrpc_unregister_bulk(struct ptlrpc_request *req, int async)
}
return 0;
}
-EXPORT_SYMBOL(ptlrpc_unregister_bulk);
static void ptlrpc_at_set_reply(struct ptlrpc_request *req, int flags)
{
@@ -434,7 +433,6 @@ out:
ptlrpc_connection_put(conn);
return rc;
}
-EXPORT_SYMBOL(ptlrpc_send_reply);
int ptlrpc_reply(struct ptlrpc_request *req)
{
@@ -442,7 +440,6 @@ int ptlrpc_reply(struct ptlrpc_request *req)
return 0;
return ptlrpc_send_reply(req, 0);
}
-EXPORT_SYMBOL(ptlrpc_reply);
/**
* For request \a req send an error reply back. Create empty
@@ -469,13 +466,11 @@ int ptlrpc_send_error(struct ptlrpc_request *req, int may_be_difficult)
rc = ptlrpc_send_reply(req, may_be_difficult);
return rc;
}
-EXPORT_SYMBOL(ptlrpc_send_error);
int ptlrpc_error(struct ptlrpc_request *req)
{
return ptlrpc_send_error(req, 0);
}
-EXPORT_SYMBOL(ptlrpc_error);
/**
* Send request \a request.
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index a1ea0c2..1349bf6 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -66,7 +66,6 @@ u32 lustre_msg_hdr_size(__u32 magic, u32 count)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_hdr_size);
void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
u32 index)
@@ -76,7 +75,6 @@ void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
else
lustre_set_rep_swabbed(req, index);
}
-EXPORT_SYMBOL(ptlrpc_buf_set_swabbed);
int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
u32 index)
@@ -88,7 +86,6 @@ int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
return (ptlrpc_rep_need_swab(req) &&
!lustre_rep_swabbed(req, index));
}
-EXPORT_SYMBOL(ptlrpc_buf_need_swab);
/* early reply size */
u32 lustre_msg_early_size(void)
@@ -151,7 +148,6 @@ u32 lustre_msg_size(__u32 magic, int count, __u32 *lens)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_size);
/* This is used to determine the size of a buffer that was already packed
* and will correctly handle the different message formats.
@@ -166,7 +162,6 @@ u32 lustre_packed_msg_size(struct lustre_msg *msg)
return 0;
}
}
-EXPORT_SYMBOL(lustre_packed_msg_size);
void lustre_init_msg_v2(struct lustre_msg_v2 *msg, int count, __u32 *lens,
char **bufs)
@@ -227,7 +222,6 @@ int lustre_pack_request(struct ptlrpc_request *req, __u32 magic, int count,
/* only use new format, we don't need to be compatible with 1.4 */
return lustre_pack_request_v2(req, count, lens, bufs);
}
-EXPORT_SYMBOL(lustre_pack_request);
#if RS_DEBUG
LIST_HEAD(ptlrpc_rs_debug_lru);
@@ -369,7 +363,6 @@ int lustre_pack_reply_flags(struct ptlrpc_request *req, int count, __u32 *lens,
lustre_msg_size(req->rq_reqmsg->lm_magic, count, lens));
return rc;
}
-EXPORT_SYMBOL(lustre_pack_reply_flags);
int lustre_pack_reply(struct ptlrpc_request *req, int count, __u32 *lens,
char **bufs)
@@ -491,7 +484,6 @@ void lustre_free_reply_state(struct ptlrpc_reply_state *rs)
sptlrpc_svc_free_rs(rs);
}
-EXPORT_SYMBOL(lustre_free_reply_state);
static int lustre_unpack_msg_v2(struct lustre_msg_v2 *m, int len)
{
@@ -579,7 +571,6 @@ int ptlrpc_unpack_req_msg(struct ptlrpc_request *req, int len)
}
return rc;
}
-EXPORT_SYMBOL(ptlrpc_unpack_req_msg);
int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len)
{
@@ -592,7 +583,6 @@ int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len)
}
return rc;
}
-EXPORT_SYMBOL(ptlrpc_unpack_rep_msg);
static inline int lustre_unpack_ptlrpc_body_v2(struct ptlrpc_request *req,
const int inout, int offset)
@@ -685,7 +675,6 @@ u32 lustre_msg_bufcount(struct lustre_msg *m)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_bufcount);
char *lustre_msg_string(struct lustre_msg *m, u32 index, u32 max_len)
{
@@ -729,7 +718,6 @@ char *lustre_msg_string(struct lustre_msg *m, u32 index, u32 max_len)
return str;
}
-EXPORT_SYMBOL(lustre_msg_string);
/* Wrap up the normal fixed length cases */
static inline void *__lustre_swab_buf(struct lustre_msg *msg, u32 index,
@@ -832,7 +820,6 @@ void lustre_msg_set_flags(struct lustre_msg *msg, u32 flags)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_flags);
void lustre_msg_clear_flags(struct lustre_msg *msg, u32 flags)
{
@@ -866,7 +853,6 @@ __u32 lustre_msg_get_op_flags(struct lustre_msg *msg)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_get_op_flags);
void lustre_msg_add_op_flags(struct lustre_msg *msg, u32 flags)
{
@@ -901,7 +887,6 @@ struct lustre_handle *lustre_msg_get_handle(struct lustre_msg *msg)
return NULL;
}
}
-EXPORT_SYMBOL(lustre_msg_get_handle);
__u32 lustre_msg_get_type(struct lustre_msg *msg)
{
@@ -936,7 +921,6 @@ void lustre_msg_add_version(struct lustre_msg *msg, u32 version)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_add_version);
__u32 lustre_msg_get_opc(struct lustre_msg *msg)
{
@@ -1053,7 +1037,6 @@ __u64 lustre_msg_get_slv(struct lustre_msg *msg)
return -EINVAL;
}
}
-EXPORT_SYMBOL(lustre_msg_get_slv);
void lustre_msg_set_slv(struct lustre_msg *msg, __u64 slv)
{
@@ -1073,7 +1056,6 @@ void lustre_msg_set_slv(struct lustre_msg *msg, __u64 slv)
return;
}
}
-EXPORT_SYMBOL(lustre_msg_set_slv);
__u32 lustre_msg_get_limit(struct lustre_msg *msg)
{
@@ -1092,7 +1074,6 @@ __u32 lustre_msg_get_limit(struct lustre_msg *msg)
return -EINVAL;
}
}
-EXPORT_SYMBOL(lustre_msg_get_limit);
void lustre_msg_set_limit(struct lustre_msg *msg, __u64 limit)
{
@@ -1112,7 +1093,6 @@ void lustre_msg_set_limit(struct lustre_msg *msg, __u64 limit)
return;
}
}
-EXPORT_SYMBOL(lustre_msg_set_limit);
__u32 lustre_msg_get_conn_cnt(struct lustre_msg *msg)
{
@@ -1143,7 +1123,6 @@ __u32 lustre_msg_get_magic(struct lustre_msg *msg)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_get_magic);
__u32 lustre_msg_get_timeout(struct lustre_msg *msg)
{
@@ -1226,7 +1205,6 @@ void lustre_msg_set_handle(struct lustre_msg *msg, struct lustre_handle *handle)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_handle);
void lustre_msg_set_type(struct lustre_msg *msg, __u32 type)
{
@@ -1242,7 +1220,6 @@ void lustre_msg_set_type(struct lustre_msg *msg, __u32 type)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_type);
void lustre_msg_set_opc(struct lustre_msg *msg, __u32 opc)
{
@@ -1258,7 +1235,6 @@ void lustre_msg_set_opc(struct lustre_msg *msg, __u32 opc)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_opc);
void lustre_msg_set_versions(struct lustre_msg *msg, __u64 *versions)
{
@@ -1325,7 +1301,6 @@ void lustre_msg_set_conn_cnt(struct lustre_msg *msg, __u32 conn_cnt)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_conn_cnt);
void lustre_msg_set_timeout(struct lustre_msg *msg, __u32 timeout)
{
@@ -1490,7 +1465,6 @@ void lustre_swab_ptlrpc_body(struct ptlrpc_body *b)
*/
CLASSERT(offsetof(typeof(*b), pb_jobid) != 0);
}
-EXPORT_SYMBOL(lustre_swab_ptlrpc_body);
void lustre_swab_connect(struct obd_connect_data *ocd)
{
@@ -1590,7 +1564,6 @@ void lustre_swab_obd_statfs(struct obd_statfs *os)
CLASSERT(offsetof(typeof(*os), os_spare8) != 0);
CLASSERT(offsetof(typeof(*os), os_spare9) != 0);
}
-EXPORT_SYMBOL(lustre_swab_obd_statfs);
void lustre_swab_obd_ioobj(struct obd_ioobj *ioo)
{
@@ -1598,7 +1571,6 @@ void lustre_swab_obd_ioobj(struct obd_ioobj *ioo)
__swab32s(&ioo->ioo_max_brw);
__swab32s(&ioo->ioo_bufcnt);
}
-EXPORT_SYMBOL(lustre_swab_obd_ioobj);
void lustre_swab_niobuf_remote(struct niobuf_remote *nbr)
{
@@ -1606,25 +1578,21 @@ void lustre_swab_niobuf_remote(struct niobuf_remote *nbr)
__swab32s(&nbr->rnb_len);
__swab32s(&nbr->rnb_flags);
}
-EXPORT_SYMBOL(lustre_swab_niobuf_remote);
void lustre_swab_ost_body(struct ost_body *b)
{
lustre_swab_obdo(&b->oa);
}
-EXPORT_SYMBOL(lustre_swab_ost_body);
void lustre_swab_ost_last_id(u64 *id)
{
__swab64s(id);
}
-EXPORT_SYMBOL(lustre_swab_ost_last_id);
void lustre_swab_generic_32s(__u32 *val)
{
__swab32s(val);
}
-EXPORT_SYMBOL(lustre_swab_generic_32s);
void lustre_swab_gl_desc(union ldlm_gl_desc *desc)
{
@@ -1703,7 +1671,6 @@ void lustre_swab_mdt_body(struct mdt_body *b)
__swab32s(&b->mbo_gid_h);
CLASSERT(offsetof(typeof(*b), mbo_padding_5) != 0);
}
-EXPORT_SYMBOL(lustre_swab_mdt_body);
void lustre_swab_mdt_ioepoch(struct mdt_ioepoch *b)
{
@@ -1712,7 +1679,6 @@ void lustre_swab_mdt_ioepoch(struct mdt_ioepoch *b)
__swab32s(&b->flags);
CLASSERT(offsetof(typeof(*b), padding) != 0);
}
-EXPORT_SYMBOL(lustre_swab_mdt_ioepoch);
void lustre_swab_mgs_target_info(struct mgs_target_info *mti)
{
@@ -1728,7 +1694,6 @@ void lustre_swab_mgs_target_info(struct mgs_target_info *mti)
for (i = 0; i < MTI_NIDS_MAX; i++)
__swab64s(&mti->mti_nids[i]);
}
-EXPORT_SYMBOL(lustre_swab_mgs_target_info);
void lustre_swab_mgs_nidtbl_entry(struct mgs_nidtbl_entry *entry)
{
@@ -1759,14 +1724,12 @@ void lustre_swab_mgs_config_body(struct mgs_config_body *body)
__swab32s(&body->mcb_units);
__swab16s(&body->mcb_type);
}
-EXPORT_SYMBOL(lustre_swab_mgs_config_body);
void lustre_swab_mgs_config_res(struct mgs_config_res *body)
{
__swab64s(&body->mcr_offset);
__swab64s(&body->mcr_size);
}
-EXPORT_SYMBOL(lustre_swab_mgs_config_res);
static void lustre_swab_obd_dqinfo(struct obd_dqinfo *i)
{
@@ -1799,7 +1762,6 @@ void lustre_swab_obd_quotactl(struct obd_quotactl *q)
lustre_swab_obd_dqinfo(&q->qc_dqinfo);
lustre_swab_obd_dqblk(&q->qc_dqblk);
}
-EXPORT_SYMBOL(lustre_swab_obd_quotactl);
void lustre_swab_fid2path(struct getinfo_fid2path *gf)
{
@@ -1833,7 +1795,6 @@ void lustre_swab_fiemap(struct ll_user_fiemap *fiemap)
for (i = 0; i < fiemap->fm_mapped_extents; i++)
lustre_swab_fiemap_extent(&fiemap->fm_extents[i]);
}
-EXPORT_SYMBOL(lustre_swab_fiemap);
void lustre_swab_mdt_rec_reint (struct mdt_rec_reint *rr)
{
@@ -1862,7 +1823,6 @@ void lustre_swab_mdt_rec_reint (struct mdt_rec_reint *rr)
CLASSERT(offsetof(typeof(*rr), rr_padding_4) != 0);
};
-EXPORT_SYMBOL(lustre_swab_mdt_rec_reint);
void lustre_swab_lov_desc(struct lov_desc *ld)
{
@@ -1901,7 +1861,6 @@ void lustre_swab_lmv_mds_md(union lmv_mds_md *lmm)
break;
}
}
-EXPORT_SYMBOL(lustre_swab_lmv_mds_md);
void lustre_swab_lmv_user_md(struct lmv_user_md *lum)
{
@@ -1955,7 +1914,6 @@ void lustre_swab_lov_mds_md(struct lov_mds_md *lmm)
__swab16s(&lmm->lmm_stripe_count);
__swab16s(&lmm->lmm_layout_gen);
}
-EXPORT_SYMBOL(lustre_swab_lov_mds_md);
void lustre_swab_lov_user_md_objects(struct lov_user_ost_data *lod,
int stripe_count)
@@ -1995,7 +1953,6 @@ void lustre_swab_ldlm_intent(struct ldlm_intent *i)
{
__swab64s(&i->opc);
}
-EXPORT_SYMBOL(lustre_swab_ldlm_intent);
static void lustre_swab_ldlm_resource_desc(struct ldlm_resource_desc *r)
{
@@ -2019,7 +1976,6 @@ void lustre_swab_ldlm_request(struct ldlm_request *rq)
__swab32s(&rq->lock_count);
/* lock_handle[] opaque */
}
-EXPORT_SYMBOL(lustre_swab_ldlm_request);
void lustre_swab_ldlm_reply(struct ldlm_reply *r)
{
@@ -2030,7 +1986,6 @@ void lustre_swab_ldlm_reply(struct ldlm_reply *r)
__swab64s(&r->lock_policy_res1);
__swab64s(&r->lock_policy_res2);
}
-EXPORT_SYMBOL(lustre_swab_ldlm_reply);
/* Dump functions */
void dump_ioo(struct obd_ioobj *ioo)
@@ -2040,14 +1995,12 @@ void dump_ioo(struct obd_ioobj *ioo)
POSTID(&ioo->ioo_oid), ioo->ioo_max_brw,
ioo->ioo_bufcnt);
}
-EXPORT_SYMBOL(dump_ioo);
void dump_rniobuf(struct niobuf_remote *nb)
{
CDEBUG(D_RPCTRACE, "niobuf_remote: offset=%llu, len=%d, flags=%x\n",
nb->rnb_offset, nb->rnb_len, nb->rnb_flags);
}
-EXPORT_SYMBOL(dump_rniobuf);
static void dump_obdo(struct obdo *oa)
{
@@ -2115,13 +2068,11 @@ void dump_ost_body(struct ost_body *ob)
{
dump_obdo(&ob->oa);
}
-EXPORT_SYMBOL(dump_ost_body);
void dump_rcs(__u32 *rc)
{
CDEBUG(D_RPCTRACE, "rmf_rcs: %d\n", *rc);
}
-EXPORT_SYMBOL(dump_rcs);
static inline int req_ptlrpc_body_swabbed(struct ptlrpc_request *req)
{
@@ -2206,14 +2157,12 @@ void lustre_swab_lustre_capa(struct lustre_capa *c)
__swab32s(&c->lc_timeout);
__swab32s(&c->lc_expiry);
}
-EXPORT_SYMBOL(lustre_swab_lustre_capa);
void lustre_swab_hsm_user_state(struct hsm_user_state *state)
{
__swab32s(&state->hus_states);
__swab32s(&state->hus_archive_id);
}
-EXPORT_SYMBOL(lustre_swab_hsm_user_state);
void lustre_swab_hsm_state_set(struct hsm_state_set *hss)
{
@@ -2236,14 +2185,12 @@ void lustre_swab_hsm_current_action(struct hsm_current_action *action)
__swab32s(&action->hca_action);
lustre_swab_hsm_extent(&action->hca_location);
}
-EXPORT_SYMBOL(lustre_swab_hsm_current_action);
void lustre_swab_hsm_user_item(struct hsm_user_item *hui)
{
lustre_swab_lu_fid(&hui->hui_fid);
lustre_swab_hsm_extent(&hui->hui_extent);
}
-EXPORT_SYMBOL(lustre_swab_hsm_user_item);
void lustre_swab_layout_intent(struct layout_intent *li)
{
@@ -2252,7 +2199,6 @@ void lustre_swab_layout_intent(struct layout_intent *li)
__swab64s(&li->li_start);
__swab64s(&li->li_end);
}
-EXPORT_SYMBOL(lustre_swab_layout_intent);
void lustre_swab_hsm_progress_kernel(struct hsm_progress_kernel *hpk)
{
@@ -2263,7 +2209,6 @@ void lustre_swab_hsm_progress_kernel(struct hsm_progress_kernel *hpk)
__swab16s(&hpk->hpk_flags);
__swab16s(&hpk->hpk_errval);
}
-EXPORT_SYMBOL(lustre_swab_hsm_progress_kernel);
void lustre_swab_hsm_request(struct hsm_request *hr)
{
@@ -2273,7 +2218,6 @@ void lustre_swab_hsm_request(struct hsm_request *hr)
__swab32s(&hr->hr_itemcount);
__swab32s(&hr->hr_data_len);
}
-EXPORT_SYMBOL(lustre_swab_hsm_request);
void lustre_swab_swap_layouts(struct mdc_swap_layouts *msl)
{
@@ -2286,4 +2230,3 @@ void lustre_swab_close_data(struct close_data *cd)
lustre_swab_lu_fid(&cd->cd_fid);
__swab64s(&cd->cd_data_version);
}
-EXPORT_SYMBOL(lustre_swab_close_data);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pinger.c b/drivers/staging/lustre/lustre/ptlrpc/pinger.c
index c0529d8..5504fc2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pinger.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pinger.c
@@ -340,7 +340,6 @@ void ptlrpc_pinger_sending_on_import(struct obd_import *imp)
{
ptlrpc_update_next_ping(imp, 0);
}
-EXPORT_SYMBOL(ptlrpc_pinger_sending_on_import);
void ptlrpc_pinger_commit_expected(struct obd_import *imp)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
index b420aa8..f14d193 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
+++ b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
@@ -62,6 +62,11 @@ void ptlrpc_request_cache_fini(void);
struct ptlrpc_request *ptlrpc_request_cache_alloc(gfp_t flags);
void ptlrpc_request_cache_free(struct ptlrpc_request *req);
void ptlrpc_init_xid(void);
+void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc,
+ struct ptlrpc_request *req);
+int ptlrpc_expired_set(void *data);
+int ptlrpc_set_next_timeout(struct ptlrpc_request_set *);
+void ptlrpc_resend_req(struct ptlrpc_request *request);
/* events.c */
int ptlrpc_init_portals(void);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/recover.c b/drivers/staging/lustre/lustre/ptlrpc/recover.c
index 718b3a8..405faf0 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/recover.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/recover.c
@@ -201,7 +201,6 @@ int ptlrpc_resend(struct obd_import *imp)
return 0;
}
-EXPORT_SYMBOL(ptlrpc_resend);
/**
* Go through all requests in delayed list and wake their threads
@@ -221,7 +220,6 @@ void ptlrpc_wake_delayed(struct obd_import *imp)
}
spin_unlock(&imp->imp_lock);
}
-EXPORT_SYMBOL(ptlrpc_wake_delayed);
void ptlrpc_request_handle_notconn(struct ptlrpc_request *failed_req)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
index 6adf274..b2cc5ea 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
@@ -348,7 +348,6 @@ void sptlrpc_enc_pool_put_pages(struct ptlrpc_bulk_desc *desc)
kfree(desc->bd_enc_iov);
desc->bd_enc_iov = NULL;
}
-EXPORT_SYMBOL(sptlrpc_enc_pool_put_pages);
static inline void enc_pools_alloc(void)
{
@@ -457,13 +456,11 @@ const char *sptlrpc_get_hash_name(__u8 hash_alg)
{
return cfs_crypto_hash_name(cfs_hash_alg_id[hash_alg]);
}
-EXPORT_SYMBOL(sptlrpc_get_hash_name);
__u8 sptlrpc_get_hash_alg(const char *algname)
{
return cfs_crypto_hash_alg(algname);
}
-EXPORT_SYMBOL(sptlrpc_get_hash_alg);
int bulk_sec_desc_unpack(struct lustre_msg *msg, int offset, int swabbed)
{
@@ -544,4 +541,3 @@ int sptlrpc_get_bulk_checksum(struct ptlrpc_bulk_desc *desc, __u8 alg,
return err;
}
-EXPORT_SYMBOL(sptlrpc_get_bulk_checksum);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_config.c b/drivers/staging/lustre/lustre/ptlrpc/sec_config.c
index c140354..2181a85 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_config.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_config.c
@@ -58,7 +58,6 @@ enum lustre_sec_part sptlrpc_target_sec_part(struct obd_device *obd)
CERROR("unknown target %p(%s)\n", obd, type);
return LUSTRE_SP_ANY;
}
-EXPORT_SYMBOL(sptlrpc_target_sec_part);
/****************************************
* user supplied flavor string parsing *
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
index f79a889..8ffd000 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
@@ -71,7 +71,6 @@ void sptlrpc_gc_add_sec(struct ptlrpc_sec *sec)
CDEBUG(D_SEC, "added sec %p(%s)\n", sec, sec->ps_policy->sp_name);
}
-EXPORT_SYMBOL(sptlrpc_gc_add_sec);
void sptlrpc_gc_del_sec(struct ptlrpc_sec *sec)
{
@@ -95,7 +94,6 @@ void sptlrpc_gc_del_sec(struct ptlrpc_sec *sec)
CDEBUG(D_SEC, "del sec %p(%s)\n", sec, sec->ps_policy->sp_name);
}
-EXPORT_SYMBOL(sptlrpc_gc_del_sec);
static void sec_process_ctx_list(void)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/service.c b/drivers/staging/lustre/lustre/ptlrpc/service.c
index a2fce66..3edc9b1 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/service.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/service.c
@@ -2407,7 +2407,6 @@ int ptlrpc_start_threads(struct ptlrpc_service *svc)
ptlrpc_stop_all_threads(svc);
return rc;
}
-EXPORT_SYMBOL(ptlrpc_start_threads);
int ptlrpc_start_thread(struct ptlrpc_service_part *svcpt, int wait)
{
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 112/124] staging: lustre: ptlrpc: remove unnecessary EXPORT_SYMBOL
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, frank zago,
James Simmons
From: frank zago <fzago@cray.com>
A lot of symbols don't need to be exported at all because they are
only used in the module they belong to.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5829
Reviewed-on: http://review.whamcloud.com/12510
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_net.h | 7 ---
drivers/staging/lustre/lustre/ptlrpc/client.c | 17 +-----
drivers/staging/lustre/lustre/ptlrpc/connection.c | 5 --
drivers/staging/lustre/lustre/ptlrpc/import.c | 1 -
drivers/staging/lustre/lustre/ptlrpc/niobuf.c | 5 --
.../staging/lustre/lustre/ptlrpc/pack_generic.c | 57 --------------------
drivers/staging/lustre/lustre/ptlrpc/pinger.c | 1 -
.../staging/lustre/lustre/ptlrpc/ptlrpc_internal.h | 5 ++
drivers/staging/lustre/lustre/ptlrpc/recover.c | 2 -
drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c | 4 --
drivers/staging/lustre/lustre/ptlrpc/sec_config.c | 1 -
drivers/staging/lustre/lustre/ptlrpc/sec_gc.c | 2 -
drivers/staging/lustre/lustre/ptlrpc/service.c | 1 -
13 files changed, 8 insertions(+), 100 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 21e2267..e9aba99 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -2408,7 +2408,6 @@ int ptlrpc_send_reply(struct ptlrpc_request *req, int flags);
int ptlrpc_reply(struct ptlrpc_request *req);
int ptlrpc_send_error(struct ptlrpc_request *req, int difficult);
int ptlrpc_error(struct ptlrpc_request *req);
-void ptlrpc_resend_req(struct ptlrpc_request *request);
int ptlrpc_at_get_net_latency(struct ptlrpc_request *req);
int ptl_send_rpc(struct ptlrpc_request *request, int noreply);
int ptlrpc_register_rqbd(struct ptlrpc_request_buffer_desc *rqbd);
@@ -2428,23 +2427,17 @@ struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid);
int ptlrpc_queue_wait(struct ptlrpc_request *req);
int ptlrpc_replay_req(struct ptlrpc_request *req);
-int ptlrpc_unregister_reply(struct ptlrpc_request *req, int async);
void ptlrpc_abort_inflight(struct obd_import *imp);
void ptlrpc_abort_set(struct ptlrpc_request_set *set);
struct ptlrpc_request_set *ptlrpc_prep_set(void);
struct ptlrpc_request_set *ptlrpc_prep_fcset(int max, set_producer_func func,
void *arg);
-int ptlrpc_set_next_timeout(struct ptlrpc_request_set *);
int ptlrpc_check_set(const struct lu_env *env, struct ptlrpc_request_set *set);
int ptlrpc_set_wait(struct ptlrpc_request_set *);
-int ptlrpc_expired_set(void *data);
-void ptlrpc_interrupted_set(void *data);
void ptlrpc_mark_interrupted(struct ptlrpc_request *req);
void ptlrpc_set_destroy(struct ptlrpc_request_set *);
void ptlrpc_set_add_req(struct ptlrpc_request_set *, struct ptlrpc_request *);
-void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc,
- struct ptlrpc_request *req);
void ptlrpc_free_rq_pool(struct ptlrpc_request_pool *pool);
int ptlrpc_add_rqs_to_pool(struct ptlrpc_request_pool *pool, int num_rq);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index f3914cc..18c7ff7 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -45,6 +45,7 @@
static int ptlrpc_send_new_req(struct ptlrpc_request *req);
static int ptlrpcd_check_work(struct ptlrpc_request *req);
+static int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async);
/**
* Initialize passed in client structure \a cl.
@@ -89,7 +90,6 @@ struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid)
return c;
}
-EXPORT_SYMBOL(ptlrpc_uuid_to_connection);
/**
* Allocate and initialize new bulk descriptor on the sender.
@@ -886,7 +886,6 @@ struct ptlrpc_request_set *ptlrpc_prep_fcset(int max, set_producer_func func,
return set;
}
-EXPORT_SYMBOL(ptlrpc_prep_fcset);
/**
* Wind down and free request set structure previously allocated with
@@ -1006,7 +1005,6 @@ void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc,
wake_up(&pc->pc_partners[i]->pc_set->set_waitq);
}
}
-EXPORT_SYMBOL(ptlrpc_set_add_new_req);
/**
* Based on the current state of the import, determine if the request
@@ -2005,7 +2003,6 @@ int ptlrpc_expired_set(void *data)
*/
return 1;
}
-EXPORT_SYMBOL(ptlrpc_expired_set);
/**
* Sets rq_intr flag in \a req under spinlock.
@@ -2022,7 +2019,7 @@ EXPORT_SYMBOL(ptlrpc_mark_interrupted);
* Interrupts (sets interrupted flag) all uncompleted requests in
* a set \a data. Callback for l_wait_event for interruptible waits.
*/
-void ptlrpc_interrupted_set(void *data)
+static void ptlrpc_interrupted_set(void *data)
{
struct ptlrpc_request_set *set = data;
struct list_head *tmp;
@@ -2040,7 +2037,6 @@ void ptlrpc_interrupted_set(void *data)
ptlrpc_mark_interrupted(req);
}
}
-EXPORT_SYMBOL(ptlrpc_interrupted_set);
/**
* Get the smallest timeout in the set; this does NOT set a timeout.
@@ -2084,7 +2080,6 @@ int ptlrpc_set_next_timeout(struct ptlrpc_request_set *set)
}
return timeout;
}
-EXPORT_SYMBOL(ptlrpc_set_next_timeout);
/**
* Send all unset request from the set and then wait until all
@@ -2335,7 +2330,7 @@ EXPORT_SYMBOL(ptlrpc_req_xid);
* The request owner (i.e. the thread doing the I/O) must call...
* Returns 0 on success or 1 if unregistering cannot be made.
*/
-int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
+static int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
{
int rc;
wait_queue_head_t *wq;
@@ -2400,7 +2395,6 @@ int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
}
return 0;
}
-EXPORT_SYMBOL(ptlrpc_unregister_reply);
static void ptlrpc_free_request(struct ptlrpc_request *req)
{
@@ -2555,7 +2549,6 @@ void ptlrpc_resend_req(struct ptlrpc_request *req)
ptlrpc_client_wake_req(req);
spin_unlock(&req->rq_lock);
}
-EXPORT_SYMBOL(ptlrpc_resend_req);
/**
* Grab additional reference on a request \a req
@@ -2624,7 +2617,6 @@ void ptlrpc_retain_replayable_request(struct ptlrpc_request *req,
list_add(&req->rq_replay_list, &imp->imp_replay_list);
}
-EXPORT_SYMBOL(ptlrpc_retain_replayable_request);
/**
* Send request and wait until it completes.
@@ -2797,7 +2789,6 @@ int ptlrpc_replay_req(struct ptlrpc_request *req)
ptlrpcd_add_req(req);
return 0;
}
-EXPORT_SYMBOL(ptlrpc_replay_req);
/**
* Aborts all in-flight request on import \a imp sending and delayed lists
@@ -2857,7 +2848,6 @@ void ptlrpc_abort_inflight(struct obd_import *imp)
spin_unlock(&imp->imp_lock);
}
-EXPORT_SYMBOL(ptlrpc_abort_inflight);
/**
* Abort all uncompleted requests in request set \a set
@@ -2943,7 +2933,6 @@ __u64 ptlrpc_next_xid(void)
return next;
}
-EXPORT_SYMBOL(ptlrpc_next_xid);
/**
* Get a glimpse at what next xid value might have been.
diff --git a/drivers/staging/lustre/lustre/ptlrpc/connection.c b/drivers/staging/lustre/lustre/ptlrpc/connection.c
index 177a379..7b020d6 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/connection.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/connection.c
@@ -82,7 +82,6 @@ out:
libcfs_nid2str(conn->c_peer.nid));
return conn;
}
-EXPORT_SYMBOL(ptlrpc_connection_get);
int ptlrpc_connection_put(struct ptlrpc_connection *conn)
{
@@ -118,7 +117,6 @@ int ptlrpc_connection_put(struct ptlrpc_connection *conn)
return rc;
}
-EXPORT_SYMBOL(ptlrpc_connection_put);
struct ptlrpc_connection *
ptlrpc_connection_addref(struct ptlrpc_connection *conn)
@@ -130,7 +128,6 @@ ptlrpc_connection_addref(struct ptlrpc_connection *conn)
return conn;
}
-EXPORT_SYMBOL(ptlrpc_connection_addref);
int ptlrpc_connection_init(void)
{
@@ -146,13 +143,11 @@ int ptlrpc_connection_init(void)
return 0;
}
-EXPORT_SYMBOL(ptlrpc_connection_init);
void ptlrpc_connection_fini(void)
{
cfs_hash_putref(conn_hash);
}
-EXPORT_SYMBOL(ptlrpc_connection_fini);
/*
* Hash operations for net_peer<->connection
diff --git a/drivers/staging/lustre/lustre/ptlrpc/import.c b/drivers/staging/lustre/lustre/ptlrpc/import.c
index 20daa06..5333cda 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/import.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/import.c
@@ -425,7 +425,6 @@ void ptlrpc_fail_import(struct obd_import *imp, __u32 conn_cnt)
ptlrpc_pinger_force(imp);
}
}
-EXPORT_SYMBOL(ptlrpc_fail_import);
int ptlrpc_reconnect_import(struct obd_import *imp)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
index 5bab248..9c93739 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/niobuf.c
@@ -295,7 +295,6 @@ int ptlrpc_unregister_bulk(struct ptlrpc_request *req, int async)
}
return 0;
}
-EXPORT_SYMBOL(ptlrpc_unregister_bulk);
static void ptlrpc_at_set_reply(struct ptlrpc_request *req, int flags)
{
@@ -434,7 +433,6 @@ out:
ptlrpc_connection_put(conn);
return rc;
}
-EXPORT_SYMBOL(ptlrpc_send_reply);
int ptlrpc_reply(struct ptlrpc_request *req)
{
@@ -442,7 +440,6 @@ int ptlrpc_reply(struct ptlrpc_request *req)
return 0;
return ptlrpc_send_reply(req, 0);
}
-EXPORT_SYMBOL(ptlrpc_reply);
/**
* For request \a req send an error reply back. Create empty
@@ -469,13 +466,11 @@ int ptlrpc_send_error(struct ptlrpc_request *req, int may_be_difficult)
rc = ptlrpc_send_reply(req, may_be_difficult);
return rc;
}
-EXPORT_SYMBOL(ptlrpc_send_error);
int ptlrpc_error(struct ptlrpc_request *req)
{
return ptlrpc_send_error(req, 0);
}
-EXPORT_SYMBOL(ptlrpc_error);
/**
* Send request \a request.
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
index a1ea0c2..1349bf6 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pack_generic.c
@@ -66,7 +66,6 @@ u32 lustre_msg_hdr_size(__u32 magic, u32 count)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_hdr_size);
void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
u32 index)
@@ -76,7 +75,6 @@ void ptlrpc_buf_set_swabbed(struct ptlrpc_request *req, const int inout,
else
lustre_set_rep_swabbed(req, index);
}
-EXPORT_SYMBOL(ptlrpc_buf_set_swabbed);
int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
u32 index)
@@ -88,7 +86,6 @@ int ptlrpc_buf_need_swab(struct ptlrpc_request *req, const int inout,
return (ptlrpc_rep_need_swab(req) &&
!lustre_rep_swabbed(req, index));
}
-EXPORT_SYMBOL(ptlrpc_buf_need_swab);
/* early reply size */
u32 lustre_msg_early_size(void)
@@ -151,7 +148,6 @@ u32 lustre_msg_size(__u32 magic, int count, __u32 *lens)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_size);
/* This is used to determine the size of a buffer that was already packed
* and will correctly handle the different message formats.
@@ -166,7 +162,6 @@ u32 lustre_packed_msg_size(struct lustre_msg *msg)
return 0;
}
}
-EXPORT_SYMBOL(lustre_packed_msg_size);
void lustre_init_msg_v2(struct lustre_msg_v2 *msg, int count, __u32 *lens,
char **bufs)
@@ -227,7 +222,6 @@ int lustre_pack_request(struct ptlrpc_request *req, __u32 magic, int count,
/* only use new format, we don't need to be compatible with 1.4 */
return lustre_pack_request_v2(req, count, lens, bufs);
}
-EXPORT_SYMBOL(lustre_pack_request);
#if RS_DEBUG
LIST_HEAD(ptlrpc_rs_debug_lru);
@@ -369,7 +363,6 @@ int lustre_pack_reply_flags(struct ptlrpc_request *req, int count, __u32 *lens,
lustre_msg_size(req->rq_reqmsg->lm_magic, count, lens));
return rc;
}
-EXPORT_SYMBOL(lustre_pack_reply_flags);
int lustre_pack_reply(struct ptlrpc_request *req, int count, __u32 *lens,
char **bufs)
@@ -491,7 +484,6 @@ void lustre_free_reply_state(struct ptlrpc_reply_state *rs)
sptlrpc_svc_free_rs(rs);
}
-EXPORT_SYMBOL(lustre_free_reply_state);
static int lustre_unpack_msg_v2(struct lustre_msg_v2 *m, int len)
{
@@ -579,7 +571,6 @@ int ptlrpc_unpack_req_msg(struct ptlrpc_request *req, int len)
}
return rc;
}
-EXPORT_SYMBOL(ptlrpc_unpack_req_msg);
int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len)
{
@@ -592,7 +583,6 @@ int ptlrpc_unpack_rep_msg(struct ptlrpc_request *req, int len)
}
return rc;
}
-EXPORT_SYMBOL(ptlrpc_unpack_rep_msg);
static inline int lustre_unpack_ptlrpc_body_v2(struct ptlrpc_request *req,
const int inout, int offset)
@@ -685,7 +675,6 @@ u32 lustre_msg_bufcount(struct lustre_msg *m)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_bufcount);
char *lustre_msg_string(struct lustre_msg *m, u32 index, u32 max_len)
{
@@ -729,7 +718,6 @@ char *lustre_msg_string(struct lustre_msg *m, u32 index, u32 max_len)
return str;
}
-EXPORT_SYMBOL(lustre_msg_string);
/* Wrap up the normal fixed length cases */
static inline void *__lustre_swab_buf(struct lustre_msg *msg, u32 index,
@@ -832,7 +820,6 @@ void lustre_msg_set_flags(struct lustre_msg *msg, u32 flags)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_flags);
void lustre_msg_clear_flags(struct lustre_msg *msg, u32 flags)
{
@@ -866,7 +853,6 @@ __u32 lustre_msg_get_op_flags(struct lustre_msg *msg)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_get_op_flags);
void lustre_msg_add_op_flags(struct lustre_msg *msg, u32 flags)
{
@@ -901,7 +887,6 @@ struct lustre_handle *lustre_msg_get_handle(struct lustre_msg *msg)
return NULL;
}
}
-EXPORT_SYMBOL(lustre_msg_get_handle);
__u32 lustre_msg_get_type(struct lustre_msg *msg)
{
@@ -936,7 +921,6 @@ void lustre_msg_add_version(struct lustre_msg *msg, u32 version)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_add_version);
__u32 lustre_msg_get_opc(struct lustre_msg *msg)
{
@@ -1053,7 +1037,6 @@ __u64 lustre_msg_get_slv(struct lustre_msg *msg)
return -EINVAL;
}
}
-EXPORT_SYMBOL(lustre_msg_get_slv);
void lustre_msg_set_slv(struct lustre_msg *msg, __u64 slv)
{
@@ -1073,7 +1056,6 @@ void lustre_msg_set_slv(struct lustre_msg *msg, __u64 slv)
return;
}
}
-EXPORT_SYMBOL(lustre_msg_set_slv);
__u32 lustre_msg_get_limit(struct lustre_msg *msg)
{
@@ -1092,7 +1074,6 @@ __u32 lustre_msg_get_limit(struct lustre_msg *msg)
return -EINVAL;
}
}
-EXPORT_SYMBOL(lustre_msg_get_limit);
void lustre_msg_set_limit(struct lustre_msg *msg, __u64 limit)
{
@@ -1112,7 +1093,6 @@ void lustre_msg_set_limit(struct lustre_msg *msg, __u64 limit)
return;
}
}
-EXPORT_SYMBOL(lustre_msg_set_limit);
__u32 lustre_msg_get_conn_cnt(struct lustre_msg *msg)
{
@@ -1143,7 +1123,6 @@ __u32 lustre_msg_get_magic(struct lustre_msg *msg)
return 0;
}
}
-EXPORT_SYMBOL(lustre_msg_get_magic);
__u32 lustre_msg_get_timeout(struct lustre_msg *msg)
{
@@ -1226,7 +1205,6 @@ void lustre_msg_set_handle(struct lustre_msg *msg, struct lustre_handle *handle)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_handle);
void lustre_msg_set_type(struct lustre_msg *msg, __u32 type)
{
@@ -1242,7 +1220,6 @@ void lustre_msg_set_type(struct lustre_msg *msg, __u32 type)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_type);
void lustre_msg_set_opc(struct lustre_msg *msg, __u32 opc)
{
@@ -1258,7 +1235,6 @@ void lustre_msg_set_opc(struct lustre_msg *msg, __u32 opc)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_opc);
void lustre_msg_set_versions(struct lustre_msg *msg, __u64 *versions)
{
@@ -1325,7 +1301,6 @@ void lustre_msg_set_conn_cnt(struct lustre_msg *msg, __u32 conn_cnt)
LASSERTF(0, "incorrect message magic: %08x\n", msg->lm_magic);
}
}
-EXPORT_SYMBOL(lustre_msg_set_conn_cnt);
void lustre_msg_set_timeout(struct lustre_msg *msg, __u32 timeout)
{
@@ -1490,7 +1465,6 @@ void lustre_swab_ptlrpc_body(struct ptlrpc_body *b)
*/
CLASSERT(offsetof(typeof(*b), pb_jobid) != 0);
}
-EXPORT_SYMBOL(lustre_swab_ptlrpc_body);
void lustre_swab_connect(struct obd_connect_data *ocd)
{
@@ -1590,7 +1564,6 @@ void lustre_swab_obd_statfs(struct obd_statfs *os)
CLASSERT(offsetof(typeof(*os), os_spare8) != 0);
CLASSERT(offsetof(typeof(*os), os_spare9) != 0);
}
-EXPORT_SYMBOL(lustre_swab_obd_statfs);
void lustre_swab_obd_ioobj(struct obd_ioobj *ioo)
{
@@ -1598,7 +1571,6 @@ void lustre_swab_obd_ioobj(struct obd_ioobj *ioo)
__swab32s(&ioo->ioo_max_brw);
__swab32s(&ioo->ioo_bufcnt);
}
-EXPORT_SYMBOL(lustre_swab_obd_ioobj);
void lustre_swab_niobuf_remote(struct niobuf_remote *nbr)
{
@@ -1606,25 +1578,21 @@ void lustre_swab_niobuf_remote(struct niobuf_remote *nbr)
__swab32s(&nbr->rnb_len);
__swab32s(&nbr->rnb_flags);
}
-EXPORT_SYMBOL(lustre_swab_niobuf_remote);
void lustre_swab_ost_body(struct ost_body *b)
{
lustre_swab_obdo(&b->oa);
}
-EXPORT_SYMBOL(lustre_swab_ost_body);
void lustre_swab_ost_last_id(u64 *id)
{
__swab64s(id);
}
-EXPORT_SYMBOL(lustre_swab_ost_last_id);
void lustre_swab_generic_32s(__u32 *val)
{
__swab32s(val);
}
-EXPORT_SYMBOL(lustre_swab_generic_32s);
void lustre_swab_gl_desc(union ldlm_gl_desc *desc)
{
@@ -1703,7 +1671,6 @@ void lustre_swab_mdt_body(struct mdt_body *b)
__swab32s(&b->mbo_gid_h);
CLASSERT(offsetof(typeof(*b), mbo_padding_5) != 0);
}
-EXPORT_SYMBOL(lustre_swab_mdt_body);
void lustre_swab_mdt_ioepoch(struct mdt_ioepoch *b)
{
@@ -1712,7 +1679,6 @@ void lustre_swab_mdt_ioepoch(struct mdt_ioepoch *b)
__swab32s(&b->flags);
CLASSERT(offsetof(typeof(*b), padding) != 0);
}
-EXPORT_SYMBOL(lustre_swab_mdt_ioepoch);
void lustre_swab_mgs_target_info(struct mgs_target_info *mti)
{
@@ -1728,7 +1694,6 @@ void lustre_swab_mgs_target_info(struct mgs_target_info *mti)
for (i = 0; i < MTI_NIDS_MAX; i++)
__swab64s(&mti->mti_nids[i]);
}
-EXPORT_SYMBOL(lustre_swab_mgs_target_info);
void lustre_swab_mgs_nidtbl_entry(struct mgs_nidtbl_entry *entry)
{
@@ -1759,14 +1724,12 @@ void lustre_swab_mgs_config_body(struct mgs_config_body *body)
__swab32s(&body->mcb_units);
__swab16s(&body->mcb_type);
}
-EXPORT_SYMBOL(lustre_swab_mgs_config_body);
void lustre_swab_mgs_config_res(struct mgs_config_res *body)
{
__swab64s(&body->mcr_offset);
__swab64s(&body->mcr_size);
}
-EXPORT_SYMBOL(lustre_swab_mgs_config_res);
static void lustre_swab_obd_dqinfo(struct obd_dqinfo *i)
{
@@ -1799,7 +1762,6 @@ void lustre_swab_obd_quotactl(struct obd_quotactl *q)
lustre_swab_obd_dqinfo(&q->qc_dqinfo);
lustre_swab_obd_dqblk(&q->qc_dqblk);
}
-EXPORT_SYMBOL(lustre_swab_obd_quotactl);
void lustre_swab_fid2path(struct getinfo_fid2path *gf)
{
@@ -1833,7 +1795,6 @@ void lustre_swab_fiemap(struct ll_user_fiemap *fiemap)
for (i = 0; i < fiemap->fm_mapped_extents; i++)
lustre_swab_fiemap_extent(&fiemap->fm_extents[i]);
}
-EXPORT_SYMBOL(lustre_swab_fiemap);
void lustre_swab_mdt_rec_reint (struct mdt_rec_reint *rr)
{
@@ -1862,7 +1823,6 @@ void lustre_swab_mdt_rec_reint (struct mdt_rec_reint *rr)
CLASSERT(offsetof(typeof(*rr), rr_padding_4) != 0);
};
-EXPORT_SYMBOL(lustre_swab_mdt_rec_reint);
void lustre_swab_lov_desc(struct lov_desc *ld)
{
@@ -1901,7 +1861,6 @@ void lustre_swab_lmv_mds_md(union lmv_mds_md *lmm)
break;
}
}
-EXPORT_SYMBOL(lustre_swab_lmv_mds_md);
void lustre_swab_lmv_user_md(struct lmv_user_md *lum)
{
@@ -1955,7 +1914,6 @@ void lustre_swab_lov_mds_md(struct lov_mds_md *lmm)
__swab16s(&lmm->lmm_stripe_count);
__swab16s(&lmm->lmm_layout_gen);
}
-EXPORT_SYMBOL(lustre_swab_lov_mds_md);
void lustre_swab_lov_user_md_objects(struct lov_user_ost_data *lod,
int stripe_count)
@@ -1995,7 +1953,6 @@ void lustre_swab_ldlm_intent(struct ldlm_intent *i)
{
__swab64s(&i->opc);
}
-EXPORT_SYMBOL(lustre_swab_ldlm_intent);
static void lustre_swab_ldlm_resource_desc(struct ldlm_resource_desc *r)
{
@@ -2019,7 +1976,6 @@ void lustre_swab_ldlm_request(struct ldlm_request *rq)
__swab32s(&rq->lock_count);
/* lock_handle[] opaque */
}
-EXPORT_SYMBOL(lustre_swab_ldlm_request);
void lustre_swab_ldlm_reply(struct ldlm_reply *r)
{
@@ -2030,7 +1986,6 @@ void lustre_swab_ldlm_reply(struct ldlm_reply *r)
__swab64s(&r->lock_policy_res1);
__swab64s(&r->lock_policy_res2);
}
-EXPORT_SYMBOL(lustre_swab_ldlm_reply);
/* Dump functions */
void dump_ioo(struct obd_ioobj *ioo)
@@ -2040,14 +1995,12 @@ void dump_ioo(struct obd_ioobj *ioo)
POSTID(&ioo->ioo_oid), ioo->ioo_max_brw,
ioo->ioo_bufcnt);
}
-EXPORT_SYMBOL(dump_ioo);
void dump_rniobuf(struct niobuf_remote *nb)
{
CDEBUG(D_RPCTRACE, "niobuf_remote: offset=%llu, len=%d, flags=%x\n",
nb->rnb_offset, nb->rnb_len, nb->rnb_flags);
}
-EXPORT_SYMBOL(dump_rniobuf);
static void dump_obdo(struct obdo *oa)
{
@@ -2115,13 +2068,11 @@ void dump_ost_body(struct ost_body *ob)
{
dump_obdo(&ob->oa);
}
-EXPORT_SYMBOL(dump_ost_body);
void dump_rcs(__u32 *rc)
{
CDEBUG(D_RPCTRACE, "rmf_rcs: %d\n", *rc);
}
-EXPORT_SYMBOL(dump_rcs);
static inline int req_ptlrpc_body_swabbed(struct ptlrpc_request *req)
{
@@ -2206,14 +2157,12 @@ void lustre_swab_lustre_capa(struct lustre_capa *c)
__swab32s(&c->lc_timeout);
__swab32s(&c->lc_expiry);
}
-EXPORT_SYMBOL(lustre_swab_lustre_capa);
void lustre_swab_hsm_user_state(struct hsm_user_state *state)
{
__swab32s(&state->hus_states);
__swab32s(&state->hus_archive_id);
}
-EXPORT_SYMBOL(lustre_swab_hsm_user_state);
void lustre_swab_hsm_state_set(struct hsm_state_set *hss)
{
@@ -2236,14 +2185,12 @@ void lustre_swab_hsm_current_action(struct hsm_current_action *action)
__swab32s(&action->hca_action);
lustre_swab_hsm_extent(&action->hca_location);
}
-EXPORT_SYMBOL(lustre_swab_hsm_current_action);
void lustre_swab_hsm_user_item(struct hsm_user_item *hui)
{
lustre_swab_lu_fid(&hui->hui_fid);
lustre_swab_hsm_extent(&hui->hui_extent);
}
-EXPORT_SYMBOL(lustre_swab_hsm_user_item);
void lustre_swab_layout_intent(struct layout_intent *li)
{
@@ -2252,7 +2199,6 @@ void lustre_swab_layout_intent(struct layout_intent *li)
__swab64s(&li->li_start);
__swab64s(&li->li_end);
}
-EXPORT_SYMBOL(lustre_swab_layout_intent);
void lustre_swab_hsm_progress_kernel(struct hsm_progress_kernel *hpk)
{
@@ -2263,7 +2209,6 @@ void lustre_swab_hsm_progress_kernel(struct hsm_progress_kernel *hpk)
__swab16s(&hpk->hpk_flags);
__swab16s(&hpk->hpk_errval);
}
-EXPORT_SYMBOL(lustre_swab_hsm_progress_kernel);
void lustre_swab_hsm_request(struct hsm_request *hr)
{
@@ -2273,7 +2218,6 @@ void lustre_swab_hsm_request(struct hsm_request *hr)
__swab32s(&hr->hr_itemcount);
__swab32s(&hr->hr_data_len);
}
-EXPORT_SYMBOL(lustre_swab_hsm_request);
void lustre_swab_swap_layouts(struct mdc_swap_layouts *msl)
{
@@ -2286,4 +2230,3 @@ void lustre_swab_close_data(struct close_data *cd)
lustre_swab_lu_fid(&cd->cd_fid);
__swab64s(&cd->cd_data_version);
}
-EXPORT_SYMBOL(lustre_swab_close_data);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/pinger.c b/drivers/staging/lustre/lustre/ptlrpc/pinger.c
index c0529d8..5504fc2 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/pinger.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/pinger.c
@@ -340,7 +340,6 @@ void ptlrpc_pinger_sending_on_import(struct obd_import *imp)
{
ptlrpc_update_next_ping(imp, 0);
}
-EXPORT_SYMBOL(ptlrpc_pinger_sending_on_import);
void ptlrpc_pinger_commit_expected(struct obd_import *imp)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
index b420aa8..f14d193 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
+++ b/drivers/staging/lustre/lustre/ptlrpc/ptlrpc_internal.h
@@ -62,6 +62,11 @@ void ptlrpc_request_cache_fini(void);
struct ptlrpc_request *ptlrpc_request_cache_alloc(gfp_t flags);
void ptlrpc_request_cache_free(struct ptlrpc_request *req);
void ptlrpc_init_xid(void);
+void ptlrpc_set_add_new_req(struct ptlrpcd_ctl *pc,
+ struct ptlrpc_request *req);
+int ptlrpc_expired_set(void *data);
+int ptlrpc_set_next_timeout(struct ptlrpc_request_set *);
+void ptlrpc_resend_req(struct ptlrpc_request *request);
/* events.c */
int ptlrpc_init_portals(void);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/recover.c b/drivers/staging/lustre/lustre/ptlrpc/recover.c
index 718b3a8..405faf0 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/recover.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/recover.c
@@ -201,7 +201,6 @@ int ptlrpc_resend(struct obd_import *imp)
return 0;
}
-EXPORT_SYMBOL(ptlrpc_resend);
/**
* Go through all requests in delayed list and wake their threads
@@ -221,7 +220,6 @@ void ptlrpc_wake_delayed(struct obd_import *imp)
}
spin_unlock(&imp->imp_lock);
}
-EXPORT_SYMBOL(ptlrpc_wake_delayed);
void ptlrpc_request_handle_notconn(struct ptlrpc_request *failed_req)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
index 6adf274..b2cc5ea 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c
@@ -348,7 +348,6 @@ void sptlrpc_enc_pool_put_pages(struct ptlrpc_bulk_desc *desc)
kfree(desc->bd_enc_iov);
desc->bd_enc_iov = NULL;
}
-EXPORT_SYMBOL(sptlrpc_enc_pool_put_pages);
static inline void enc_pools_alloc(void)
{
@@ -457,13 +456,11 @@ const char *sptlrpc_get_hash_name(__u8 hash_alg)
{
return cfs_crypto_hash_name(cfs_hash_alg_id[hash_alg]);
}
-EXPORT_SYMBOL(sptlrpc_get_hash_name);
__u8 sptlrpc_get_hash_alg(const char *algname)
{
return cfs_crypto_hash_alg(algname);
}
-EXPORT_SYMBOL(sptlrpc_get_hash_alg);
int bulk_sec_desc_unpack(struct lustre_msg *msg, int offset, int swabbed)
{
@@ -544,4 +541,3 @@ int sptlrpc_get_bulk_checksum(struct ptlrpc_bulk_desc *desc, __u8 alg,
return err;
}
-EXPORT_SYMBOL(sptlrpc_get_bulk_checksum);
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_config.c b/drivers/staging/lustre/lustre/ptlrpc/sec_config.c
index c140354..2181a85 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_config.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_config.c
@@ -58,7 +58,6 @@ enum lustre_sec_part sptlrpc_target_sec_part(struct obd_device *obd)
CERROR("unknown target %p(%s)\n", obd, type);
return LUSTRE_SP_ANY;
}
-EXPORT_SYMBOL(sptlrpc_target_sec_part);
/****************************************
* user supplied flavor string parsing *
diff --git a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
index f79a889..8ffd000 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/sec_gc.c
@@ -71,7 +71,6 @@ void sptlrpc_gc_add_sec(struct ptlrpc_sec *sec)
CDEBUG(D_SEC, "added sec %p(%s)\n", sec, sec->ps_policy->sp_name);
}
-EXPORT_SYMBOL(sptlrpc_gc_add_sec);
void sptlrpc_gc_del_sec(struct ptlrpc_sec *sec)
{
@@ -95,7 +94,6 @@ void sptlrpc_gc_del_sec(struct ptlrpc_sec *sec)
CDEBUG(D_SEC, "del sec %p(%s)\n", sec, sec->ps_policy->sp_name);
}
-EXPORT_SYMBOL(sptlrpc_gc_del_sec);
static void sec_process_ctx_list(void)
{
diff --git a/drivers/staging/lustre/lustre/ptlrpc/service.c b/drivers/staging/lustre/lustre/ptlrpc/service.c
index a2fce66..3edc9b1 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/service.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/service.c
@@ -2407,7 +2407,6 @@ int ptlrpc_start_threads(struct ptlrpc_service *svc)
ptlrpc_stop_all_threads(svc);
return rc;
}
-EXPORT_SYMBOL(ptlrpc_start_threads);
int ptlrpc_start_thread(struct ptlrpc_service_part *svcpt, int wait)
{
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* Re: [PATCH 112/124] staging: lustre: ptlrpc: remove unnecessary EXPORT_SYMBOL
2016-09-18 20:38 ` [lustre-devel] " James Simmons
@ 2016-09-19 5:43 ` kbuild test robot
-1 siblings, 0 replies; 263+ messages in thread
From: kbuild test robot @ 2016-09-19 5:43 UTC (permalink / raw)
To: James Simmons
Cc: kbuild-all, Greg Kroah-Hartman, devel, Andreas Dilger,
Oleg Drokin, frank zago, Linux Kernel Mailing List,
Lustre Development List
[-- Attachment #1: Type: text/plain, Size: 1302 bytes --]
Hi frank,
[auto build test ERROR on staging/staging-testing]
[cannot apply to v4.8-rc7 next-20160916]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
[Suggest to use git(>=2.9.0) format-patch --base=<commit> (or --base=auto for convenience) to record what (public, well-known) commit your patch series was built on]
[Check https://git-scm.com/docs/git-format-patch for more information]
url: https://github.com/0day-ci/linux/commits/James-Simmons/missing-patches-from-Lustre-2-7-release/20160919-045619
config: s390-allyesconfig (attached as .config)
compiler: s390x-linux-gnu-gcc (Debian 6.1.1-9) 6.1.1 20160705
reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=s390
All errors (new ones prefixed by >>):
>> ERROR: "lustre_swab_lov_mds_md" [drivers/staging/lustre/lustre/lov/lov.ko] undefined!
>> ERROR: "lustre_swab_lov_mds_md" [drivers/staging/lustre/lustre/llite/lustre.ko] undefined!
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 42434 bytes --]
^ permalink raw reply [flat|nested] 263+ messages in thread
* [PATCH 113/124] staging: lustre: obd: use proper flags for call_usermodehelper
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
James Simmons, James Simmons
When a parameter is permanently changed on the MGS the
MGS send a changelog packet to the proper nodes that
are affected by the change. Once the nodes receive the
change they then call the userland utility lctl to
change its local value. When calling a userland
application from the kernel you specify a flag to
control the interaction with the application. Originally
by default the flag was set to 0 which is UMH_NO_WAIT
which meant lctl was being called asynchronously. In
older kernels this was fine since UHM_NO_WAIT and
UHM_WAIT_PROC had nearly the same logic. This changed
with newer kernels which broke updating our parameters.
Plus doing a UHM_NO_WAIT doesn't report back a error
if something goes wrong with lctl. The fix is to set
the flag to UHM_WAIT_PROC so kernel space waits until
lctl has finished and we get a proper error code if
something does go wrong with lctl.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6063
Reviewed-on: http://review.whamcloud.com/13677
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/obdclass/obd_config.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_config.c b/drivers/staging/lustre/lustre/obdclass/obd_config.c
index fddfc9c..bbed1b7 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_config.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_config.c
@@ -738,7 +738,7 @@ static int process_param2_config(struct lustre_cfg *lcfg)
}
start = ktime_get();
- rc = call_usermodehelper(argv[0], argv, NULL, 1);
+ rc = call_usermodehelper(argv[0], argv, NULL, UMH_WAIT_PROC);
end = ktime_get();
if (rc < 0) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 113/124] staging: lustre: obd: use proper flags for call_usermodehelper
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
James Simmons, James Simmons
When a parameter is permanently changed on the MGS the
MGS send a changelog packet to the proper nodes that
are affected by the change. Once the nodes receive the
change they then call the userland utility lctl to
change its local value. When calling a userland
application from the kernel you specify a flag to
control the interaction with the application. Originally
by default the flag was set to 0 which is UMH_NO_WAIT
which meant lctl was being called asynchronously. In
older kernels this was fine since UHM_NO_WAIT and
UHM_WAIT_PROC had nearly the same logic. This changed
with newer kernels which broke updating our parameters.
Plus doing a UHM_NO_WAIT doesn't report back a error
if something goes wrong with lctl. The fix is to set
the flag to UHM_WAIT_PROC so kernel space waits until
lctl has finished and we get a proper error code if
something does go wrong with lctl.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6063
Reviewed-on: http://review.whamcloud.com/13677
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/obdclass/obd_config.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/staging/lustre/lustre/obdclass/obd_config.c b/drivers/staging/lustre/lustre/obdclass/obd_config.c
index fddfc9c..bbed1b7 100644
--- a/drivers/staging/lustre/lustre/obdclass/obd_config.c
+++ b/drivers/staging/lustre/lustre/obdclass/obd_config.c
@@ -738,7 +738,7 @@ static int process_param2_config(struct lustre_cfg *lcfg)
}
start = ktime_get();
- rc = call_usermodehelper(argv[0], argv, NULL, 1);
+ rc = call_usermodehelper(argv[0], argv, NULL, UMH_WAIT_PROC);
end = ktime_get();
if (rc < 0) {
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 114/124] staging: lustre: ptlrpc: prevent request timeout grow due to recovery
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Mikhail Pershin, James Simmons
From: Mikhail Pershin <mike.pershin@intel.com>
Patch fixes the issue seen on the client with growing request
timeout which occurred after the server side patch landed for
LU-5079. While commit itself is correct, it reveals another
issue. If request is being processed for a long time on server
then client adaptive timeouts will adapt to that after receiving
reply and new requests will have bigger timeout. Another problem
is that server AT history is corrupted by recovery request
processing time which not pure service time but includes
also waiting time for clients to recover.
Patch prevents the AT stats update from early replies on client and
from recovering requests processing time on server.
The ptlrpc_at_recv_early_reply() still updates the current request
timeout as asked by server, but don't include this into AT stats.
The real reply will bring that data from server after all.
Signed-off-by: Mikhail Pershin <mike.pershin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6084
Reviewed-on: http://review.whamcloud.com/13520
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 30 +++++++++++++----------
drivers/staging/lustre/lustre/ptlrpc/service.c | 12 ++++-----
2 files changed, 22 insertions(+), 20 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index 18c7ff7..8c51d51 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -364,23 +364,27 @@ static int ptlrpc_at_recv_early_reply(struct ptlrpc_request *req)
}
rc = unpack_reply(early_req);
- if (rc == 0) {
- /* Expecting to increase the service time estimate here */
- ptlrpc_at_adj_service(req,
- lustre_msg_get_timeout(early_req->rq_repmsg));
- ptlrpc_at_adj_net_latency(req,
- lustre_msg_get_service_time(early_req->rq_repmsg));
- }
-
- sptlrpc_cli_finish_early_reply(early_req);
-
- if (rc != 0) {
+ if (rc) {
+ sptlrpc_cli_finish_early_reply(early_req);
spin_lock(&req->rq_lock);
return rc;
}
- /* Adjust the local timeout for this req */
- ptlrpc_at_set_req_timeout(req);
+ /*
+ * Use new timeout value just to adjust the local value for this
+ * request, don't include it into at_history. It is unclear yet why
+ * service time increased and should it be counted or skipped, e.g.
+ * that can be recovery case or some error or server, the real reply
+ * will add all new data if it is worth to add.
+ */
+ req->rq_timeout = lustre_msg_get_timeout(early_req->rq_repmsg);
+ lustre_msg_set_timeout(req->rq_reqmsg, req->rq_timeout);
+
+ /* Network latency can be adjusted, it is pure network delays */
+ ptlrpc_at_adj_net_latency(req,
+ lustre_msg_get_service_time(early_req->rq_repmsg));
+
+ sptlrpc_cli_finish_early_reply(early_req);
spin_lock(&req->rq_lock);
olddl = req->rq_deadline;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/service.c b/drivers/staging/lustre/lustre/ptlrpc/service.c
index 3edc9b1..72f3930 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/service.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/service.c
@@ -1015,6 +1015,7 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req)
struct ptlrpc_request *reqcopy;
struct lustre_msg *reqmsg;
long olddl = req->rq_deadline - ktime_get_real_seconds();
+ time64_t newdl;
int rc;
/* deadline is when the client expects us to reply, margin is the
@@ -1052,16 +1053,14 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req)
*/
at_measured(&svcpt->scp_at_estimate, at_extra +
ktime_get_real_seconds() - req->rq_arrival_time.tv_sec);
+ newdl = req->rq_arrival_time.tv_sec + at_get(&svcpt->scp_at_estimate);
/* Check to see if we've actually increased the deadline -
* we may be past adaptive_max
*/
- if (req->rq_deadline >= req->rq_arrival_time.tv_sec +
- at_get(&svcpt->scp_at_estimate)) {
+ if (req->rq_deadline >= newdl) {
DEBUG_REQ(D_WARNING, req, "Couldn't add any time (%ld/%lld), not sending early reply\n",
- olddl, req->rq_arrival_time.tv_sec +
- at_get(&svcpt->scp_at_estimate) -
- ktime_get_real_seconds());
+ olddl, newdl - ktime_get_real_seconds());
return -ETIMEDOUT;
}
@@ -1117,8 +1116,7 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req)
if (!rc) {
/* Adjust our own deadline to what we told the client */
- req->rq_deadline = req->rq_arrival_time.tv_sec +
- at_get(&svcpt->scp_at_estimate);
+ req->rq_deadline = newdl;
req->rq_early_count++; /* number sent, server side */
} else {
DEBUG_REQ(D_ERROR, req, "Early reply send failed %d", rc);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 114/124] staging: lustre: ptlrpc: prevent request timeout grow due to recovery
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Mikhail Pershin, James Simmons
From: Mikhail Pershin <mike.pershin@intel.com>
Patch fixes the issue seen on the client with growing request
timeout which occurred after the server side patch landed for
LU-5079. While commit itself is correct, it reveals another
issue. If request is being processed for a long time on server
then client adaptive timeouts will adapt to that after receiving
reply and new requests will have bigger timeout. Another problem
is that server AT history is corrupted by recovery request
processing time which not pure service time but includes
also waiting time for clients to recover.
Patch prevents the AT stats update from early replies on client and
from recovering requests processing time on server.
The ptlrpc_at_recv_early_reply() still updates the current request
timeout as asked by server, but don't include this into AT stats.
The real reply will bring that data from server after all.
Signed-off-by: Mikhail Pershin <mike.pershin@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6084
Reviewed-on: http://review.whamcloud.com/13520
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/ptlrpc/client.c | 30 +++++++++++++----------
drivers/staging/lustre/lustre/ptlrpc/service.c | 12 ++++-----
2 files changed, 22 insertions(+), 20 deletions(-)
diff --git a/drivers/staging/lustre/lustre/ptlrpc/client.c b/drivers/staging/lustre/lustre/ptlrpc/client.c
index 18c7ff7..8c51d51 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/client.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/client.c
@@ -364,23 +364,27 @@ static int ptlrpc_at_recv_early_reply(struct ptlrpc_request *req)
}
rc = unpack_reply(early_req);
- if (rc == 0) {
- /* Expecting to increase the service time estimate here */
- ptlrpc_at_adj_service(req,
- lustre_msg_get_timeout(early_req->rq_repmsg));
- ptlrpc_at_adj_net_latency(req,
- lustre_msg_get_service_time(early_req->rq_repmsg));
- }
-
- sptlrpc_cli_finish_early_reply(early_req);
-
- if (rc != 0) {
+ if (rc) {
+ sptlrpc_cli_finish_early_reply(early_req);
spin_lock(&req->rq_lock);
return rc;
}
- /* Adjust the local timeout for this req */
- ptlrpc_at_set_req_timeout(req);
+ /*
+ * Use new timeout value just to adjust the local value for this
+ * request, don't include it into at_history. It is unclear yet why
+ * service time increased and should it be counted or skipped, e.g.
+ * that can be recovery case or some error or server, the real reply
+ * will add all new data if it is worth to add.
+ */
+ req->rq_timeout = lustre_msg_get_timeout(early_req->rq_repmsg);
+ lustre_msg_set_timeout(req->rq_reqmsg, req->rq_timeout);
+
+ /* Network latency can be adjusted, it is pure network delays */
+ ptlrpc_at_adj_net_latency(req,
+ lustre_msg_get_service_time(early_req->rq_repmsg));
+
+ sptlrpc_cli_finish_early_reply(early_req);
spin_lock(&req->rq_lock);
olddl = req->rq_deadline;
diff --git a/drivers/staging/lustre/lustre/ptlrpc/service.c b/drivers/staging/lustre/lustre/ptlrpc/service.c
index 3edc9b1..72f3930 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/service.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/service.c
@@ -1015,6 +1015,7 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req)
struct ptlrpc_request *reqcopy;
struct lustre_msg *reqmsg;
long olddl = req->rq_deadline - ktime_get_real_seconds();
+ time64_t newdl;
int rc;
/* deadline is when the client expects us to reply, margin is the
@@ -1052,16 +1053,14 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req)
*/
at_measured(&svcpt->scp_at_estimate, at_extra +
ktime_get_real_seconds() - req->rq_arrival_time.tv_sec);
+ newdl = req->rq_arrival_time.tv_sec + at_get(&svcpt->scp_at_estimate);
/* Check to see if we've actually increased the deadline -
* we may be past adaptive_max
*/
- if (req->rq_deadline >= req->rq_arrival_time.tv_sec +
- at_get(&svcpt->scp_at_estimate)) {
+ if (req->rq_deadline >= newdl) {
DEBUG_REQ(D_WARNING, req, "Couldn't add any time (%ld/%lld), not sending early reply\n",
- olddl, req->rq_arrival_time.tv_sec +
- at_get(&svcpt->scp_at_estimate) -
- ktime_get_real_seconds());
+ olddl, newdl - ktime_get_real_seconds());
return -ETIMEDOUT;
}
@@ -1117,8 +1116,7 @@ static int ptlrpc_at_send_early_reply(struct ptlrpc_request *req)
if (!rc) {
/* Adjust our own deadline to what we told the client */
- req->rq_deadline = req->rq_arrival_time.tv_sec +
- at_get(&svcpt->scp_at_estimate);
+ req->rq_deadline = newdl;
req->rq_early_count++; /* number sent, server side */
} else {
DEBUG_REQ(D_ERROR, req, "Early reply send failed %d", rc);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 115/124] staging: lustre: mdt: add indexing option to default dir stripe
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Add indexing option to default dirstripe EA. If MDT find
out the client send the create req to the wrong MDT because
of default stripeEA, it will return -EREMOTE, then client
will retrieve default stripeEA through xattr cache, and
re-create the object.
Also merged patch for LU-6341 to resolve the following problem.
Use ll_dir_getstripe to get default stripeEA in ll_new_node(),
Because ll_getxattr_common requires admin rights for retrieving
default LMVEA (because of trusted- prefix), which might cause
mkdir (from normal user) failure.
If parent does not have default stripeEA, then child should always
be in the same MDT for mkdir. Otherwise MDT should return -EREMOTE,
then client will refresh the default stripe index, and recreate
the object.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5523
Reviewed-on: http://review.whamcloud.com/13360
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6341
Reviewed-on: http://review.whamcloud.com/13990
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 3 +
.../staging/lustre/lustre/llite/llite_internal.h | 7 +++
drivers/staging/lustre/lustre/llite/llite_lib.c | 7 +++-
drivers/staging/lustre/lustre/llite/namei.c | 45 ++++++++++++++++++-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 5 ++
5 files changed, 63 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index c6937b2..ef11534 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -773,6 +773,9 @@ struct md_op_data {
/* File object data version for HSM release, on client */
__u64 op_data_version;
struct lustre_handle op_lease_handle;
+
+ /* default stripe offset */
+ __u32 op_default_stripe_offset;
};
struct md_callback {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 51bf071..70ca3e1 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -191,6 +191,13 @@ struct ll_inode_info {
unsigned int lli_sa_generation;
/* directory stripe information */
struct lmv_stripe_md *lli_lsm_md;
+ /* default directory stripe offset. This is extracted
+ * from the "dmv" xattr in order to decide which MDT to
+ * create a subdirectory on. The MDS itself fetches
+ * "dmv" and gets the rest of the default layout itself
+ * (count, hash, etc).
+ */
+ __u32 lli_def_stripe_offset;
};
/* for non-directory */
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 230868c..465b315 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -802,6 +802,7 @@ void ll_lli_init(struct ll_inode_info *lli)
spin_lock_init(&lli->lli_sa_lock);
lli->lli_opendir_pid = 0;
lli->lli_sa_enabled = 0;
+ lli->lli_def_stripe_offset = -1;
} else {
mutex_init(&lli->lli_size_mutex);
lli->lli_symlink_name = NULL;
@@ -2342,8 +2343,12 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
ll_i2gids(op_data->op_suppgids, i1, i2);
op_data->op_fid1 = *ll_inode2fid(i1);
- if (S_ISDIR(i1->i_mode))
+ op_data->op_default_stripe_offset = -1;
+ if (S_ISDIR(i1->i_mode)) {
op_data->op_mea1 = ll_i2info(i1)->lli_lsm_md;
+ op_data->op_default_stripe_offset =
+ ll_i2info(i1)->lli_def_stripe_offset;
+ }
if (i2) {
op_data->op_fid2 = *ll_inode2fid(i2);
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 3960893..624966d 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -204,6 +204,8 @@ int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc,
}
if (bits & MDS_INODELOCK_XATTR) {
+ if (S_ISDIR(inode->i_mode))
+ ll_i2info(inode)->lli_def_stripe_offset = -1;
ll_xattr_cache_destroy(inode);
bits &= ~MDS_INODELOCK_XATTR;
}
@@ -833,7 +835,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
if (unlikely(tgt))
tgt_len = strlen(tgt) + 1;
-
+again:
op_data = ll_prep_md_op_data(NULL, dir, NULL,
dentry->d_name.name,
dentry->d_name.len,
@@ -848,9 +850,45 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
from_kgid(&init_user_ns, current_fsgid()),
cfs_curproc_cap_pack(), rdev, &request);
ll_finish_md_op_data(op_data);
- if (err)
+ if (err < 0 && err != -EREMOTE)
goto err_exit;
+ /*
+ * If the client doesn't know where to create a subdirectory (or
+ * in case of a race that sends the RPC to the wrong MDS), the
+ * MDS will return -EREMOTE and the client will fetch the layout
+ * of the directory, then create the directory on the right MDT.
+ */
+ if (unlikely(err == -EREMOTE)) {
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct lmv_user_md *lum;
+ int lumsize, err2;
+
+ ptlrpc_req_finished(request);
+ request = NULL;
+
+ err2 = ll_dir_getstripe(dir, (void **)&lum, &lumsize, &request,
+ OBD_MD_DEFAULT_MEA);
+ if (!err2) {
+ /* Update stripe_offset and retry */
+ lli->lli_def_stripe_offset = lum->lum_stripe_offset;
+ } else if (err2 == -ENODATA &&
+ lli->lli_def_stripe_offset != -1) {
+ /*
+ * If there are no default stripe EA on the MDT, but the
+ * client has default stripe, then it probably means
+ * default stripe EA has just been deleted.
+ */
+ lli->lli_def_stripe_offset = -1;
+ } else {
+ goto err_exit;
+ }
+
+ ptlrpc_req_finished(request);
+ request = NULL;
+ goto again;
+ }
+
ll_update_times(request, dir);
err = ll_prep_inode(&inode, request, dir->i_sb, NULL);
@@ -859,7 +897,8 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
d_instantiate(dentry, inode);
err_exit:
- ptlrpc_req_finished(request);
+ if (request)
+ ptlrpc_req_finished(request);
return err;
}
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index cccb645..d67d0e0 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1164,6 +1164,11 @@ static int lmv_placement_policy(struct obd_device *obd,
return 0;
}
+ if (op_data->op_default_stripe_offset != -1) {
+ *mds = op_data->op_default_stripe_offset;
+ return 0;
+ }
+
/**
* If stripe_offset is provided during setdirstripe
* (setdirstripe -i xx), xx MDS will be chosen.
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 115/124] staging: lustre: mdt: add indexing option to default dir stripe
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
Add indexing option to default dirstripe EA. If MDT find
out the client send the create req to the wrong MDT because
of default stripeEA, it will return -EREMOTE, then client
will retrieve default stripeEA through xattr cache, and
re-create the object.
Also merged patch for LU-6341 to resolve the following problem.
Use ll_dir_getstripe to get default stripeEA in ll_new_node(),
Because ll_getxattr_common requires admin rights for retrieving
default LMVEA (because of trusted- prefix), which might cause
mkdir (from normal user) failure.
If parent does not have default stripeEA, then child should always
be in the same MDT for mkdir. Otherwise MDT should return -EREMOTE,
then client will refresh the default stripe index, and recreate
the object.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5523
Reviewed-on: http://review.whamcloud.com/13360
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6341
Reviewed-on: http://review.whamcloud.com/13990
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/obd.h | 3 +
.../staging/lustre/lustre/llite/llite_internal.h | 7 +++
drivers/staging/lustre/lustre/llite/llite_lib.c | 7 +++-
drivers/staging/lustre/lustre/llite/namei.c | 45 ++++++++++++++++++-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 5 ++
5 files changed, 63 insertions(+), 4 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index c6937b2..ef11534 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -773,6 +773,9 @@ struct md_op_data {
/* File object data version for HSM release, on client */
__u64 op_data_version;
struct lustre_handle op_lease_handle;
+
+ /* default stripe offset */
+ __u32 op_default_stripe_offset;
};
struct md_callback {
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 51bf071..70ca3e1 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -191,6 +191,13 @@ struct ll_inode_info {
unsigned int lli_sa_generation;
/* directory stripe information */
struct lmv_stripe_md *lli_lsm_md;
+ /* default directory stripe offset. This is extracted
+ * from the "dmv" xattr in order to decide which MDT to
+ * create a subdirectory on. The MDS itself fetches
+ * "dmv" and gets the rest of the default layout itself
+ * (count, hash, etc).
+ */
+ __u32 lli_def_stripe_offset;
};
/* for non-directory */
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 230868c..465b315 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -802,6 +802,7 @@ void ll_lli_init(struct ll_inode_info *lli)
spin_lock_init(&lli->lli_sa_lock);
lli->lli_opendir_pid = 0;
lli->lli_sa_enabled = 0;
+ lli->lli_def_stripe_offset = -1;
} else {
mutex_init(&lli->lli_size_mutex);
lli->lli_symlink_name = NULL;
@@ -2342,8 +2343,12 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
ll_i2gids(op_data->op_suppgids, i1, i2);
op_data->op_fid1 = *ll_inode2fid(i1);
- if (S_ISDIR(i1->i_mode))
+ op_data->op_default_stripe_offset = -1;
+ if (S_ISDIR(i1->i_mode)) {
op_data->op_mea1 = ll_i2info(i1)->lli_lsm_md;
+ op_data->op_default_stripe_offset =
+ ll_i2info(i1)->lli_def_stripe_offset;
+ }
if (i2) {
op_data->op_fid2 = *ll_inode2fid(i2);
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 3960893..624966d 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -204,6 +204,8 @@ int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc,
}
if (bits & MDS_INODELOCK_XATTR) {
+ if (S_ISDIR(inode->i_mode))
+ ll_i2info(inode)->lli_def_stripe_offset = -1;
ll_xattr_cache_destroy(inode);
bits &= ~MDS_INODELOCK_XATTR;
}
@@ -833,7 +835,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
if (unlikely(tgt))
tgt_len = strlen(tgt) + 1;
-
+again:
op_data = ll_prep_md_op_data(NULL, dir, NULL,
dentry->d_name.name,
dentry->d_name.len,
@@ -848,9 +850,45 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
from_kgid(&init_user_ns, current_fsgid()),
cfs_curproc_cap_pack(), rdev, &request);
ll_finish_md_op_data(op_data);
- if (err)
+ if (err < 0 && err != -EREMOTE)
goto err_exit;
+ /*
+ * If the client doesn't know where to create a subdirectory (or
+ * in case of a race that sends the RPC to the wrong MDS), the
+ * MDS will return -EREMOTE and the client will fetch the layout
+ * of the directory, then create the directory on the right MDT.
+ */
+ if (unlikely(err == -EREMOTE)) {
+ struct ll_inode_info *lli = ll_i2info(dir);
+ struct lmv_user_md *lum;
+ int lumsize, err2;
+
+ ptlrpc_req_finished(request);
+ request = NULL;
+
+ err2 = ll_dir_getstripe(dir, (void **)&lum, &lumsize, &request,
+ OBD_MD_DEFAULT_MEA);
+ if (!err2) {
+ /* Update stripe_offset and retry */
+ lli->lli_def_stripe_offset = lum->lum_stripe_offset;
+ } else if (err2 == -ENODATA &&
+ lli->lli_def_stripe_offset != -1) {
+ /*
+ * If there are no default stripe EA on the MDT, but the
+ * client has default stripe, then it probably means
+ * default stripe EA has just been deleted.
+ */
+ lli->lli_def_stripe_offset = -1;
+ } else {
+ goto err_exit;
+ }
+
+ ptlrpc_req_finished(request);
+ request = NULL;
+ goto again;
+ }
+
ll_update_times(request, dir);
err = ll_prep_inode(&inode, request, dir->i_sb, NULL);
@@ -859,7 +897,8 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
d_instantiate(dentry, inode);
err_exit:
- ptlrpc_req_finished(request);
+ if (request)
+ ptlrpc_req_finished(request);
return err;
}
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index cccb645..d67d0e0 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1164,6 +1164,11 @@ static int lmv_placement_policy(struct obd_device *obd,
return 0;
}
+ if (op_data->op_default_stripe_offset != -1) {
+ *mds = op_data->op_default_stripe_offset;
+ return 0;
+ }
+
/**
* If stripe_offset is provided during setdirstripe
* (setdirstripe -i xx), xx MDS will be chosen.
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 116/124] staging: lustre: llite: make default_easize writeable in /sysfs
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Ned Bass,
James Simmons
From: Ned Bass <bass6@llnl.gov>
Allow default_easize to be tuned via /sysfs. A system administrator
might want this if a rare access to widely striped files drives up the
value on a filesystem where narrowly striped files are the more common
case. In practice, however, this is wanted primarily to facilitate
a test case for LU-5549.
- Plumb the necessary interfaces through the LMV and MDC layers
to expose write access to this value by higher layers.
- Add block comments to modified functions.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5549
Reviewed-on: http://review.whamcloud.com/13112
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/llite/llite_internal.h | 1 +
drivers/staging/lustre/lustre/llite/llite_lib.c | 33 +++++++++++++
drivers/staging/lustre/lustre/llite/lproc_llite.c | 51 +++++++++++++++++++-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 35 +++++++++++++-
drivers/staging/lustre/lustre/mdc/mdc_request.c | 6 ++
5 files changed, 124 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 70ca3e1..3e98bd6 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -839,6 +839,7 @@ int ll_prep_inode(struct inode **inode, struct ptlrpc_request *req,
int ll_obd_statfs(struct inode *inode, void __user *arg);
int ll_get_max_mdsize(struct ll_sb_info *sbi, int *max_mdsize);
int ll_get_default_mdsize(struct ll_sb_info *sbi, int *default_mdsize);
+int ll_set_default_mdsize(struct ll_sb_info *sbi, int default_mdsize);
int ll_process_config(struct lustre_cfg *lcfg);
enum {
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 465b315..dab08a0 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -582,6 +582,17 @@ int ll_get_max_mdsize(struct ll_sb_info *sbi, int *lmmsize)
return rc;
}
+/**
+ * Get the value of the default_easize parameter.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] sbi superblock info for this filesystem
+ * \param[out] lmmsize pointer to storage location for value
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize)
{
int size, rc;
@@ -595,6 +606,28 @@ int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize)
return rc;
}
+/**
+ * Set the default_easize parameter to the given value.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] sbi superblock info for this filesystem
+ * \param[in] lmmsize the size to set
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
+int ll_set_default_mdsize(struct ll_sb_info *sbi, int lmmsize)
+{
+ if (lmmsize < sizeof(struct lov_mds_md) || lmmsize > PAGE_SIZE)
+ return -EINVAL;
+
+ return obd_set_info_async(NULL, sbi->ll_md_exp,
+ sizeof(KEY_DEFAULT_EASIZE),
+ KEY_DEFAULT_EASIZE,
+ sizeof(int), &lmmsize, NULL);
+}
+
static void client_common_put_super(struct super_block *sb)
{
struct ll_sb_info *sbi = ll_s2sbi(sb);
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index 188fd37..6eae605 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -744,6 +744,18 @@ static ssize_t max_easize_show(struct kobject *kobj,
}
LUSTRE_RO_ATTR(max_easize);
+/**
+ * Get default_easize.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] kobj kernel object for sysfs tree
+ * \param[in] attr attribute of this kernel object
+ * \param[in] buf buffer to write data into
+ *
+ * \retval positive \a count on success
+ * \retval negative negated errno on failure
+ */
static ssize_t default_easize_show(struct kobject *kobj,
struct attribute *attr,
char *buf)
@@ -759,7 +771,44 @@ static ssize_t default_easize_show(struct kobject *kobj,
return sprintf(buf, "%u\n", ealen);
}
-LUSTRE_RO_ATTR(default_easize);
+
+/**
+ * Set default_easize.
+ *
+ * Range checking on the passed value is handled by
+ * ll_set_default_mdsize().
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] kobj kernel object for sysfs tree
+ * \param[in] attr attribute of this kernel object
+ * \param[in] buffer string passed from user space
+ * \param[in] count \a buffer length
+ *
+ * \retval positive \a count on success
+ * \retval negative negated errno on failure
+ */
+static ssize_t default_easize_store(struct kobject *kobj,
+ struct attribute *attr,
+ const char *buffer,
+ size_t count)
+{
+ struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
+ ll_kobj);
+ unsigned long val;
+ int rc;
+
+ rc = kstrtoul(buffer, 10, &val);
+ if (rc)
+ return rc;
+
+ rc = ll_set_default_mdsize(sbi, val);
+ if (rc)
+ return rc;
+
+ return count;
+}
+LUSTRE_RW_ATTR(default_easize);
static int ll_sbi_flags_seq_show(struct seq_file *m, void *v)
{
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index d67d0e0..e4cf6d3 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2625,6 +2625,22 @@ static int lmv_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
return 0;
}
+/**
+ * Get by key a value associated with a LMV device.
+ *
+ * Dispatch request to lower-layer devices as needed.
+ *
+ * \param[in] env execution environment for this thread
+ * \param[in] exp export for the LMV device
+ * \param[in] keylen length of key identifier
+ * \param[in] key identifier of key to get value for
+ * \param[in] vallen size of \a val
+ * \param[out] val pointer to storage location for value
+ * \param[in] lsm optional striping metadata of object
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
static int lmv_get_info(const struct lu_env *env, struct obd_export *exp,
__u32 keylen, void *key, __u32 *vallen, void *val,
struct lov_stripe_md *lsm)
@@ -2688,6 +2704,22 @@ static int lmv_get_info(const struct lu_env *env, struct obd_export *exp,
return -EINVAL;
}
+/**
+ * Asynchronously set by key a value associated with a LMV device.
+ *
+ * Dispatch request to lower-layer devices as needed.
+ *
+ * \param[in] env execution environment for this thread
+ * \param[in] exp export for the LMV device
+ * \param[in] keylen length of key identifier
+ * \param[in] key identifier of key to store value for
+ * \param[in] vallen size of value to store
+ * \param[in] val pointer to data to be stored
+ * \param[in] set optional list of related ptlrpc requests
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
static int lmv_set_info_async(const struct lu_env *env, struct obd_export *exp,
u32 keylen, void *key, u32 vallen,
void *val, struct ptlrpc_request_set *set)
@@ -2705,7 +2737,8 @@ static int lmv_set_info_async(const struct lu_env *env, struct obd_export *exp,
}
lmv = &obd->u.lmv;
- if (KEY_IS(KEY_READ_ONLY) || KEY_IS(KEY_FLUSH_CTX)) {
+ if (KEY_IS(KEY_READ_ONLY) || KEY_IS(KEY_FLUSH_CTX) ||
+ KEY_IS(KEY_DEFAULT_EASIZE)) {
int i, err = 0;
for (i = 0; i < lmv->desc.ld_tgt_count; i++) {
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 125d882..1c00fea 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -2516,6 +2516,12 @@ static int mdc_set_info_async(const struct lu_env *env,
rc = mdc_hsm_copytool_send(vallen, val);
return rc;
}
+ if (KEY_IS(KEY_DEFAULT_EASIZE)) {
+ u32 *default_easize = val;
+
+ exp->exp_obd->u.cli.cl_default_mds_easize = *default_easize;
+ return 0;
+ }
CERROR("Unknown key %s\n", (char *)key);
return -EINVAL;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 116/124] staging: lustre: llite: make default_easize writeable in /sysfs
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Ned Bass,
James Simmons
From: Ned Bass <bass6@llnl.gov>
Allow default_easize to be tuned via /sysfs. A system administrator
might want this if a rare access to widely striped files drives up the
value on a filesystem where narrowly striped files are the more common
case. In practice, however, this is wanted primarily to facilitate
a test case for LU-5549.
- Plumb the necessary interfaces through the LMV and MDC layers
to expose write access to this value by higher layers.
- Add block comments to modified functions.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5549
Reviewed-on: http://review.whamcloud.com/13112
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lustre/llite/llite_internal.h | 1 +
drivers/staging/lustre/lustre/llite/llite_lib.c | 33 +++++++++++++
drivers/staging/lustre/lustre/llite/lproc_llite.c | 51 +++++++++++++++++++-
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 35 +++++++++++++-
drivers/staging/lustre/lustre/mdc/mdc_request.c | 6 ++
5 files changed, 124 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 70ca3e1..3e98bd6 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -839,6 +839,7 @@ int ll_prep_inode(struct inode **inode, struct ptlrpc_request *req,
int ll_obd_statfs(struct inode *inode, void __user *arg);
int ll_get_max_mdsize(struct ll_sb_info *sbi, int *max_mdsize);
int ll_get_default_mdsize(struct ll_sb_info *sbi, int *default_mdsize);
+int ll_set_default_mdsize(struct ll_sb_info *sbi, int default_mdsize);
int ll_process_config(struct lustre_cfg *lcfg);
enum {
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 465b315..dab08a0 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -582,6 +582,17 @@ int ll_get_max_mdsize(struct ll_sb_info *sbi, int *lmmsize)
return rc;
}
+/**
+ * Get the value of the default_easize parameter.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] sbi superblock info for this filesystem
+ * \param[out] lmmsize pointer to storage location for value
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize)
{
int size, rc;
@@ -595,6 +606,28 @@ int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize)
return rc;
}
+/**
+ * Set the default_easize parameter to the given value.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] sbi superblock info for this filesystem
+ * \param[in] lmmsize the size to set
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
+int ll_set_default_mdsize(struct ll_sb_info *sbi, int lmmsize)
+{
+ if (lmmsize < sizeof(struct lov_mds_md) || lmmsize > PAGE_SIZE)
+ return -EINVAL;
+
+ return obd_set_info_async(NULL, sbi->ll_md_exp,
+ sizeof(KEY_DEFAULT_EASIZE),
+ KEY_DEFAULT_EASIZE,
+ sizeof(int), &lmmsize, NULL);
+}
+
static void client_common_put_super(struct super_block *sb)
{
struct ll_sb_info *sbi = ll_s2sbi(sb);
diff --git a/drivers/staging/lustre/lustre/llite/lproc_llite.c b/drivers/staging/lustre/lustre/llite/lproc_llite.c
index 188fd37..6eae605 100644
--- a/drivers/staging/lustre/lustre/llite/lproc_llite.c
+++ b/drivers/staging/lustre/lustre/llite/lproc_llite.c
@@ -744,6 +744,18 @@ static ssize_t max_easize_show(struct kobject *kobj,
}
LUSTRE_RO_ATTR(max_easize);
+/**
+ * Get default_easize.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] kobj kernel object for sysfs tree
+ * \param[in] attr attribute of this kernel object
+ * \param[in] buf buffer to write data into
+ *
+ * \retval positive \a count on success
+ * \retval negative negated errno on failure
+ */
static ssize_t default_easize_show(struct kobject *kobj,
struct attribute *attr,
char *buf)
@@ -759,7 +771,44 @@ static ssize_t default_easize_show(struct kobject *kobj,
return sprintf(buf, "%u\n", ealen);
}
-LUSTRE_RO_ATTR(default_easize);
+
+/**
+ * Set default_easize.
+ *
+ * Range checking on the passed value is handled by
+ * ll_set_default_mdsize().
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] kobj kernel object for sysfs tree
+ * \param[in] attr attribute of this kernel object
+ * \param[in] buffer string passed from user space
+ * \param[in] count \a buffer length
+ *
+ * \retval positive \a count on success
+ * \retval negative negated errno on failure
+ */
+static ssize_t default_easize_store(struct kobject *kobj,
+ struct attribute *attr,
+ const char *buffer,
+ size_t count)
+{
+ struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
+ ll_kobj);
+ unsigned long val;
+ int rc;
+
+ rc = kstrtoul(buffer, 10, &val);
+ if (rc)
+ return rc;
+
+ rc = ll_set_default_mdsize(sbi, val);
+ if (rc)
+ return rc;
+
+ return count;
+}
+LUSTRE_RW_ATTR(default_easize);
static int ll_sbi_flags_seq_show(struct seq_file *m, void *v)
{
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index d67d0e0..e4cf6d3 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -2625,6 +2625,22 @@ static int lmv_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
return 0;
}
+/**
+ * Get by key a value associated with a LMV device.
+ *
+ * Dispatch request to lower-layer devices as needed.
+ *
+ * \param[in] env execution environment for this thread
+ * \param[in] exp export for the LMV device
+ * \param[in] keylen length of key identifier
+ * \param[in] key identifier of key to get value for
+ * \param[in] vallen size of \a val
+ * \param[out] val pointer to storage location for value
+ * \param[in] lsm optional striping metadata of object
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
static int lmv_get_info(const struct lu_env *env, struct obd_export *exp,
__u32 keylen, void *key, __u32 *vallen, void *val,
struct lov_stripe_md *lsm)
@@ -2688,6 +2704,22 @@ static int lmv_get_info(const struct lu_env *env, struct obd_export *exp,
return -EINVAL;
}
+/**
+ * Asynchronously set by key a value associated with a LMV device.
+ *
+ * Dispatch request to lower-layer devices as needed.
+ *
+ * \param[in] env execution environment for this thread
+ * \param[in] exp export for the LMV device
+ * \param[in] keylen length of key identifier
+ * \param[in] key identifier of key to store value for
+ * \param[in] vallen size of value to store
+ * \param[in] val pointer to data to be stored
+ * \param[in] set optional list of related ptlrpc requests
+ *
+ * \retval 0 on success
+ * \retval negative negated errno on failure
+ */
static int lmv_set_info_async(const struct lu_env *env, struct obd_export *exp,
u32 keylen, void *key, u32 vallen,
void *val, struct ptlrpc_request_set *set)
@@ -2705,7 +2737,8 @@ static int lmv_set_info_async(const struct lu_env *env, struct obd_export *exp,
}
lmv = &obd->u.lmv;
- if (KEY_IS(KEY_READ_ONLY) || KEY_IS(KEY_FLUSH_CTX)) {
+ if (KEY_IS(KEY_READ_ONLY) || KEY_IS(KEY_FLUSH_CTX) ||
+ KEY_IS(KEY_DEFAULT_EASIZE)) {
int i, err = 0;
for (i = 0; i < lmv->desc.ld_tgt_count; i++) {
diff --git a/drivers/staging/lustre/lustre/mdc/mdc_request.c b/drivers/staging/lustre/lustre/mdc/mdc_request.c
index 125d882..1c00fea 100644
--- a/drivers/staging/lustre/lustre/mdc/mdc_request.c
+++ b/drivers/staging/lustre/lustre/mdc/mdc_request.c
@@ -2516,6 +2516,12 @@ static int mdc_set_info_async(const struct lu_env *env,
rc = mdc_hsm_copytool_send(vallen, val);
return rc;
}
+ if (KEY_IS(KEY_DEFAULT_EASIZE)) {
+ u32 *default_easize = val;
+
+ exp->exp_obd->u.cli.cl_default_mds_easize = *default_easize;
+ return 0;
+ }
CERROR("Unknown key %s\n", (char *)key);
return -EINVAL;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 117/124] staging: lustre: mdc: cl_default_mds_easize not refreshed
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Ned Bass,
Kyle Blatter, James Simmons
From: Ned Bass <bass6@llnl.gov>
The client_obd::cl_default_mds_easize field should track the largest
observed EA size advertised by the MDT, subject to a reasonable upper
bound. The MDC uses cl_default_mds_easize to calculate the initial
size of request buffers. The default value should be small enough to
avoid wasted memory and excessive use of vmalloc(), yet large enough
to accommodate the common use case.
In the current code, the default value is only updated if
client_obd::cl_max_mds_easize is strictly less than
mdt_body::mbo_max_mdsize. This condition is almost never met, because
client_obd::cl_max_mds_easize is computed at client mount-time based
on the number of OSTs in the filesystem, so the MDT won't ever observe
and advertise an EA size larger than that.
As a result, client_obd::cl_default_mds_easize indefinitely retains
its initial value, which is computed at client mount-time based on
the filesystem's default stripe width. Any getattr() requests for
widely striped files will consequently allocate a request buffer
that is too small, forcing reallocations on both the client and
server side. To avoid this, update client_obd::cl_default_mds_easize
independently of the value of client_obd::cl_max_mds_easize.
In addition, this patch includes these changes:
- Add comments to the client_obd structure to clarify what the
cl_{default,max}_mds_{cookie,ea}size values mean.
- Prevent mdc_get_info() from storing uninitialized data in
client_obd::cl_max_mds_cookiesize.
- Use 4096 as an upper bound for the default values. The former
bound of PAGE_CACHE_SIZE is too large on 64k-page platforms
(i.e. PPC), so it fails to prevent the vmalloc() spinlock
contention described in LU-3338. The new value was chosen to
be large enough to accommodate common use cases while staying
well below the 16k threshold at which allocations start using
vmalloc().
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Kyle Blatter <kyleblatter@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5549
Reviewed-on: http://review.whamcloud.com/11614
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_mdc.h | 37 ++++++++++++-----
drivers/staging/lustre/lustre/include/obd.h | 41 +++++++++++++++++++-
drivers/staging/lustre/lustre/llite/llite_lib.c | 3 +-
3 files changed, 67 insertions(+), 14 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_mdc.h b/drivers/staging/lustre/lustre/include/lustre_mdc.h
index 9549fb4..8fc2d3f 100644
--- a/drivers/staging/lustre/lustre/include/lustre_mdc.h
+++ b/drivers/staging/lustre/lustre/include/lustre_mdc.h
@@ -156,26 +156,41 @@ static inline void mdc_put_rpc_lock(struct mdc_rpc_lock *lck,
mutex_unlock(&lck->rpcl_mutex);
}
-/* Update the maximum observed easize and cookiesize. The default easize
- * and cookiesize is initialized to the minimum value but allowed to grow
- * up to a single page in size if required to handle the common case.
+/**
+ * Update the maximum possible easize and cookiesize.
+ *
+ * The values are learned from ptlrpc replies sent by the MDT. The
+ * default easize and cookiesize is initialized to the minimum value but
+ * allowed to grow up to a single page in size if required to handle the
+ * common case.
+ *
+ * \see client_obd::cl_default_mds_easize and
+ * client_obd::cl_default_mds_cookiesize
+ *
+ * \param[in] exp export for MDC device
+ * \param[in] body body of ptlrpc reply from MDT
+ *
*/
static inline void mdc_update_max_ea_from_body(struct obd_export *exp,
struct mdt_body *body)
{
if (body->mbo_valid & OBD_MD_FLMODEASIZE) {
struct client_obd *cli = &exp->exp_obd->u.cli;
+ u32 def_cookiesize, def_easize;
- if (cli->cl_max_mds_easize < body->mbo_max_mdsize) {
+ if (cli->cl_max_mds_easize < body->mbo_max_mdsize)
cli->cl_max_mds_easize = body->mbo_max_mdsize;
- cli->cl_default_mds_easize =
- min_t(__u32, body->mbo_max_mdsize, PAGE_SIZE);
- }
- if (cli->cl_max_mds_cookiesize < body->mbo_max_cookiesize) {
+
+ def_easize = min_t(__u32, body->mbo_max_mdsize,
+ OBD_MAX_DEFAULT_EA_SIZE);
+ cli->cl_default_mds_easize = def_easize;
+
+ if (cli->cl_max_mds_cookiesize < body->mbo_max_cookiesize)
cli->cl_max_mds_cookiesize = body->mbo_max_cookiesize;
- cli->cl_default_mds_cookiesize =
- min_t(__u32, body->mbo_max_cookiesize, PAGE_SIZE);
- }
+
+ def_cookiesize = min_t(__u32, body->mbo_max_cookiesize,
+ OBD_MAX_DEFAULT_COOKIE_SIZE);
+ cli->cl_default_mds_cookiesize = def_cookiesize;
}
}
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index ef11534..f6fc4dd 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -196,6 +196,16 @@ enum obd_cl_sem_lock_class {
OBD_CLI_SEM_MDCOSC,
};
+/*
+ * Limit reply buffer size for striping data to one x86_64 page. This
+ * value is chosen to fit the striping data for common use cases while
+ * staying well below the limit at which the buffer must be backed by
+ * vmalloc(). Excessive use of vmalloc() may cause spinlock contention
+ * on the MDS.
+ */
+#define OBD_MAX_DEFAULT_EA_SIZE 4096
+#define OBD_MAX_DEFAULT_COOKIE_SIZE 4096
+
struct mdc_rpc_lock;
struct obd_import;
struct client_obd {
@@ -203,12 +213,39 @@ struct client_obd {
struct obd_uuid cl_target_uuid;
struct obd_import *cl_import; /* ptlrpc connection state */
size_t cl_conn_count;
- /* max_mds_easize is purely a performance thing so we don't have to
- * call obd_size_diskmd() all the time.
+ /*
+ * Cache maximum and default values for easize and cookiesize. This is
+ * strictly a performance optimization to minimize calls to
+ * obd_size_diskmd(). The default values are used to calculate the
+ * initial size of a request buffer. The ptlrpc layer will resize the
+ * buffer as needed to accommodate a larger reply from the
+ * server. The default values should be small enough to avoid wasted
+ * memory and excessive use of vmalloc(), yet large enough to avoid
+ * reallocating the buffer in the common use case.
+ */
+ /*
+ * Default EA size for striping attributes. It is initialized at
+ * mount-time based on the default stripe width of the filesystem,
+ * then it tracks the largest observed EA size advertised by
+ * the MDT, up to a maximum value of OBD_MAX_DEFAULT_EA_SIZE.
*/
u32 cl_default_mds_easize;
+ /* Maximum possible EA size computed at mount-time based on
+ * the number of OSTs in the filesystem. May be increased at
+ * run-time if a larger observed size is advertised by the MDT.
+ */
u32 cl_max_mds_easize;
+ /* Default cookie size for llog cookies (see struct llog_cookie). It is
+ * initialized to zero at mount-time, then it tracks the largest
+ * observed cookie size advertised by the MDT, up to a maximum value of
+ * OBD_MAX_DEFAULT_COOKIE_SIZE. Note that llog_cookies are not
+ * used by clients communicating with MDS versions 2.4.0 and later.
+ */
u32 cl_default_mds_cookiesize;
+ /* Maximum possible cookie size computed at mount-time based on
+ * the number of OSTs in the filesystem. May be increased at
+ * run-time if a larger observed size is advertised by the MDT.
+ */
u32 cl_max_mds_cookiesize;
enum lustre_sec_part cl_sp_me;
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index dab08a0..5dcaca0 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -619,7 +619,8 @@ int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize)
*/
int ll_set_default_mdsize(struct ll_sb_info *sbi, int lmmsize)
{
- if (lmmsize < sizeof(struct lov_mds_md) || lmmsize > PAGE_SIZE)
+ if (lmmsize < sizeof(struct lov_mds_md) ||
+ lmmsize > OBD_MAX_DEFAULT_EA_SIZE)
return -EINVAL;
return obd_set_info_async(NULL, sbi->ll_md_exp,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 117/124] staging: lustre: mdc: cl_default_mds_easize not refreshed
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, Ned Bass,
Kyle Blatter, James Simmons
From: Ned Bass <bass6@llnl.gov>
The client_obd::cl_default_mds_easize field should track the largest
observed EA size advertised by the MDT, subject to a reasonable upper
bound. The MDC uses cl_default_mds_easize to calculate the initial
size of request buffers. The default value should be small enough to
avoid wasted memory and excessive use of vmalloc(), yet large enough
to accommodate the common use case.
In the current code, the default value is only updated if
client_obd::cl_max_mds_easize is strictly less than
mdt_body::mbo_max_mdsize. This condition is almost never met, because
client_obd::cl_max_mds_easize is computed at client mount-time based
on the number of OSTs in the filesystem, so the MDT won't ever observe
and advertise an EA size larger than that.
As a result, client_obd::cl_default_mds_easize indefinitely retains
its initial value, which is computed at client mount-time based on
the filesystem's default stripe width. Any getattr() requests for
widely striped files will consequently allocate a request buffer
that is too small, forcing reallocations on both the client and
server side. To avoid this, update client_obd::cl_default_mds_easize
independently of the value of client_obd::cl_max_mds_easize.
In addition, this patch includes these changes:
- Add comments to the client_obd structure to clarify what the
cl_{default,max}_mds_{cookie,ea}size values mean.
- Prevent mdc_get_info() from storing uninitialized data in
client_obd::cl_max_mds_cookiesize.
- Use 4096 as an upper bound for the default values. The former
bound of PAGE_CACHE_SIZE is too large on 64k-page platforms
(i.e. PPC), so it fails to prevent the vmalloc() spinlock
contention described in LU-3338. The new value was chosen to
be large enough to accommodate common use cases while staying
well below the 16k threshold at which allocations start using
vmalloc().
Signed-off-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Kyle Blatter <kyleblatter@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5549
Reviewed-on: http://review.whamcloud.com/11614
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_mdc.h | 37 ++++++++++++-----
drivers/staging/lustre/lustre/include/obd.h | 41 +++++++++++++++++++-
drivers/staging/lustre/lustre/llite/llite_lib.c | 3 +-
3 files changed, 67 insertions(+), 14 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_mdc.h b/drivers/staging/lustre/lustre/include/lustre_mdc.h
index 9549fb4..8fc2d3f 100644
--- a/drivers/staging/lustre/lustre/include/lustre_mdc.h
+++ b/drivers/staging/lustre/lustre/include/lustre_mdc.h
@@ -156,26 +156,41 @@ static inline void mdc_put_rpc_lock(struct mdc_rpc_lock *lck,
mutex_unlock(&lck->rpcl_mutex);
}
-/* Update the maximum observed easize and cookiesize. The default easize
- * and cookiesize is initialized to the minimum value but allowed to grow
- * up to a single page in size if required to handle the common case.
+/**
+ * Update the maximum possible easize and cookiesize.
+ *
+ * The values are learned from ptlrpc replies sent by the MDT. The
+ * default easize and cookiesize is initialized to the minimum value but
+ * allowed to grow up to a single page in size if required to handle the
+ * common case.
+ *
+ * \see client_obd::cl_default_mds_easize and
+ * client_obd::cl_default_mds_cookiesize
+ *
+ * \param[in] exp export for MDC device
+ * \param[in] body body of ptlrpc reply from MDT
+ *
*/
static inline void mdc_update_max_ea_from_body(struct obd_export *exp,
struct mdt_body *body)
{
if (body->mbo_valid & OBD_MD_FLMODEASIZE) {
struct client_obd *cli = &exp->exp_obd->u.cli;
+ u32 def_cookiesize, def_easize;
- if (cli->cl_max_mds_easize < body->mbo_max_mdsize) {
+ if (cli->cl_max_mds_easize < body->mbo_max_mdsize)
cli->cl_max_mds_easize = body->mbo_max_mdsize;
- cli->cl_default_mds_easize =
- min_t(__u32, body->mbo_max_mdsize, PAGE_SIZE);
- }
- if (cli->cl_max_mds_cookiesize < body->mbo_max_cookiesize) {
+
+ def_easize = min_t(__u32, body->mbo_max_mdsize,
+ OBD_MAX_DEFAULT_EA_SIZE);
+ cli->cl_default_mds_easize = def_easize;
+
+ if (cli->cl_max_mds_cookiesize < body->mbo_max_cookiesize)
cli->cl_max_mds_cookiesize = body->mbo_max_cookiesize;
- cli->cl_default_mds_cookiesize =
- min_t(__u32, body->mbo_max_cookiesize, PAGE_SIZE);
- }
+
+ def_cookiesize = min_t(__u32, body->mbo_max_cookiesize,
+ OBD_MAX_DEFAULT_COOKIE_SIZE);
+ cli->cl_default_mds_cookiesize = def_cookiesize;
}
}
diff --git a/drivers/staging/lustre/lustre/include/obd.h b/drivers/staging/lustre/lustre/include/obd.h
index ef11534..f6fc4dd 100644
--- a/drivers/staging/lustre/lustre/include/obd.h
+++ b/drivers/staging/lustre/lustre/include/obd.h
@@ -196,6 +196,16 @@ enum obd_cl_sem_lock_class {
OBD_CLI_SEM_MDCOSC,
};
+/*
+ * Limit reply buffer size for striping data to one x86_64 page. This
+ * value is chosen to fit the striping data for common use cases while
+ * staying well below the limit at which the buffer must be backed by
+ * vmalloc(). Excessive use of vmalloc() may cause spinlock contention
+ * on the MDS.
+ */
+#define OBD_MAX_DEFAULT_EA_SIZE 4096
+#define OBD_MAX_DEFAULT_COOKIE_SIZE 4096
+
struct mdc_rpc_lock;
struct obd_import;
struct client_obd {
@@ -203,12 +213,39 @@ struct client_obd {
struct obd_uuid cl_target_uuid;
struct obd_import *cl_import; /* ptlrpc connection state */
size_t cl_conn_count;
- /* max_mds_easize is purely a performance thing so we don't have to
- * call obd_size_diskmd() all the time.
+ /*
+ * Cache maximum and default values for easize and cookiesize. This is
+ * strictly a performance optimization to minimize calls to
+ * obd_size_diskmd(). The default values are used to calculate the
+ * initial size of a request buffer. The ptlrpc layer will resize the
+ * buffer as needed to accommodate a larger reply from the
+ * server. The default values should be small enough to avoid wasted
+ * memory and excessive use of vmalloc(), yet large enough to avoid
+ * reallocating the buffer in the common use case.
+ */
+ /*
+ * Default EA size for striping attributes. It is initialized at
+ * mount-time based on the default stripe width of the filesystem,
+ * then it tracks the largest observed EA size advertised by
+ * the MDT, up to a maximum value of OBD_MAX_DEFAULT_EA_SIZE.
*/
u32 cl_default_mds_easize;
+ /* Maximum possible EA size computed at mount-time based on
+ * the number of OSTs in the filesystem. May be increased at
+ * run-time if a larger observed size is advertised by the MDT.
+ */
u32 cl_max_mds_easize;
+ /* Default cookie size for llog cookies (see struct llog_cookie). It is
+ * initialized to zero at mount-time, then it tracks the largest
+ * observed cookie size advertised by the MDT, up to a maximum value of
+ * OBD_MAX_DEFAULT_COOKIE_SIZE. Note that llog_cookies are not
+ * used by clients communicating with MDS versions 2.4.0 and later.
+ */
u32 cl_default_mds_cookiesize;
+ /* Maximum possible cookie size computed at mount-time based on
+ * the number of OSTs in the filesystem. May be increased at
+ * run-time if a larger observed size is advertised by the MDT.
+ */
u32 cl_max_mds_cookiesize;
enum lustre_sec_part cl_sp_me;
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index dab08a0..5dcaca0 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -619,7 +619,8 @@ int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize)
*/
int ll_set_default_mdsize(struct ll_sb_info *sbi, int lmmsize)
{
- if (lmmsize < sizeof(struct lov_mds_md) || lmmsize > PAGE_SIZE)
+ if (lmmsize < sizeof(struct lov_mds_md) ||
+ lmmsize > OBD_MAX_DEFAULT_EA_SIZE)
return -EINVAL;
return obd_set_info_async(NULL, sbi->ll_md_exp,
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 118/124] staging: lustre: lmv: fix parent FID for migration
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
If the migrating directory is under striped directory, it needs
to set right stripe FID for its parent.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6263
Reviewed-on: http://review.whamcloud.com/13817
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 30 ++++++++++++++++++++++----
1 files changed, 25 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index e4cf6d3..00a32e8 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1581,6 +1581,8 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
struct lmv_tgt_desc *tgt;
if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_BAD_NAME_HASH)) {
+ if (cfs_fail_val >= lsm->lsm_md_stripe_count)
+ return ERR_PTR(-EBADF);
oinfo = &lsm->lsm_md_oinfo[cfs_fail_val];
} else {
oinfo = lsm_name_to_stripe_info(lsm, name, namelen);
@@ -1588,11 +1590,15 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
return ERR_CAST(oinfo);
}
- *fid = oinfo->lmo_fid;
- *mds = oinfo->lmo_mds;
- tgt = lmv_get_target(lmv, *mds, NULL);
+ if (fid)
+ *fid = oinfo->lmo_fid;
+ if (mds)
+ *mds = oinfo->lmo_mds;
+
+ tgt = lmv_get_target(lmv, oinfo->lmo_mds, NULL);
- CDEBUG(D_INFO, "locate on mds %u "DFID"\n", *mds, PFID(fid));
+ CDEBUG(D_INFO, "locate on mds %u " DFID "\n", oinfo->lmo_mds,
+ PFID(&oinfo->lmo_fid));
return tgt;
}
@@ -1955,10 +1961,24 @@ static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
if (op_data->op_cli_flags & CLI_MIGRATE) {
LASSERTF(fid_is_sane(&op_data->op_fid3), "invalid FID "DFID"\n",
PFID(&op_data->op_fid3));
+
+ if (op_data->op_mea1) {
+ struct lmv_stripe_md *lsm = op_data->op_mea1;
+ struct lmv_tgt_desc *tmp;
+
+ /* Fix the parent fid for striped dir */
+ tmp = lmv_locate_target_for_name(lmv, lsm, old,
+ oldlen,
+ &op_data->op_fid1,
+ NULL);
+ if (IS_ERR(tmp))
+ return PTR_ERR(tmp);
+ }
+
rc = lmv_fid_alloc(NULL, exp, &op_data->op_fid2, op_data);
if (rc)
return rc;
- src_tgt = lmv_locate_mds(lmv, op_data, &op_data->op_fid3);
+ src_tgt = lmv_find_target(lmv, &op_data->op_fid3);
} else {
if (op_data->op_mea1) {
struct lmv_stripe_md *lsm = op_data->op_mea1;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 118/124] staging: lustre: lmv: fix parent FID for migration
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, wang di,
James Simmons
From: wang di <di.wang@intel.com>
If the migrating directory is under striped directory, it needs
to set right stripe FID for its parent.
Signed-off-by: wang di <di.wang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6263
Reviewed-on: http://review.whamcloud.com/13817
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/lmv/lmv_obd.c | 30 ++++++++++++++++++++++----
1 files changed, 25 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lustre/lmv/lmv_obd.c b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
index e4cf6d3..00a32e8 100644
--- a/drivers/staging/lustre/lustre/lmv/lmv_obd.c
+++ b/drivers/staging/lustre/lustre/lmv/lmv_obd.c
@@ -1581,6 +1581,8 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
struct lmv_tgt_desc *tgt;
if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_BAD_NAME_HASH)) {
+ if (cfs_fail_val >= lsm->lsm_md_stripe_count)
+ return ERR_PTR(-EBADF);
oinfo = &lsm->lsm_md_oinfo[cfs_fail_val];
} else {
oinfo = lsm_name_to_stripe_info(lsm, name, namelen);
@@ -1588,11 +1590,15 @@ lmv_locate_target_for_name(struct lmv_obd *lmv, struct lmv_stripe_md *lsm,
return ERR_CAST(oinfo);
}
- *fid = oinfo->lmo_fid;
- *mds = oinfo->lmo_mds;
- tgt = lmv_get_target(lmv, *mds, NULL);
+ if (fid)
+ *fid = oinfo->lmo_fid;
+ if (mds)
+ *mds = oinfo->lmo_mds;
+
+ tgt = lmv_get_target(lmv, oinfo->lmo_mds, NULL);
- CDEBUG(D_INFO, "locate on mds %u "DFID"\n", *mds, PFID(fid));
+ CDEBUG(D_INFO, "locate on mds %u " DFID "\n", oinfo->lmo_mds,
+ PFID(&oinfo->lmo_fid));
return tgt;
}
@@ -1955,10 +1961,24 @@ static int lmv_rename(struct obd_export *exp, struct md_op_data *op_data,
if (op_data->op_cli_flags & CLI_MIGRATE) {
LASSERTF(fid_is_sane(&op_data->op_fid3), "invalid FID "DFID"\n",
PFID(&op_data->op_fid3));
+
+ if (op_data->op_mea1) {
+ struct lmv_stripe_md *lsm = op_data->op_mea1;
+ struct lmv_tgt_desc *tmp;
+
+ /* Fix the parent fid for striped dir */
+ tmp = lmv_locate_target_for_name(lmv, lsm, old,
+ oldlen,
+ &op_data->op_fid1,
+ NULL);
+ if (IS_ERR(tmp))
+ return PTR_ERR(tmp);
+ }
+
rc = lmv_fid_alloc(NULL, exp, &op_data->op_fid2, op_data);
if (rc)
return rc;
- src_tgt = lmv_locate_mds(lmv, op_data, &op_data->op_fid3);
+ src_tgt = lmv_find_target(lmv, &op_data->op_fid3);
} else {
if (op_data->op_mea1) {
struct lmv_stripe_md *lsm = op_data->op_mea1;
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 119/124] staging: lustre: lnet: potential deadlock in lnet
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Quentin Bouget, James Simmons
From: Quentin Bouget <quentin.bouget.ocre@cea.fr>
Fixes potential deadlock in LNetMDAttach
Signed-off-by: Quentin Bouget <quentin.bouget.ocre@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8249
Reviewed-on: http://review.whamcloud.com/20676
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/lib-md.c | 24 +++++++++++++-----------
1 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/lib-md.c b/drivers/staging/lustre/lnet/lnet/lib-md.c
index e0b2f16..eab53cd 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-md.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-md.c
@@ -292,11 +292,12 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
return -ENOMEM;
rc = lnet_md_build(md, &umd, unlink);
+ if (rc)
+ goto out_free;
+
cpt = lnet_cpt_of_cookie(meh.cookie);
lnet_res_lock(cpt);
- if (rc)
- goto failed;
me = lnet_handle2me(&meh);
if (!me)
@@ -307,7 +308,7 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
rc = lnet_md_link(md, umd.eq_handle, cpt);
if (rc)
- goto failed;
+ goto out_unlock;
/*
* attach this MD to portal of ME and check if it matches any
@@ -324,10 +325,10 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
return 0;
- failed:
- lnet_md_free(md);
-
+out_unlock:
lnet_res_unlock(cpt);
+out_free:
+ lnet_md_free(md);
return rc;
}
EXPORT_SYMBOL(LNetMDAttach);
@@ -370,24 +371,25 @@ LNetMDBind(lnet_md_t umd, lnet_unlink_t unlink, lnet_handle_md_t *handle)
return -ENOMEM;
rc = lnet_md_build(md, &umd, unlink);
+ if (rc)
+ goto out_free;
cpt = lnet_res_lock_current();
- if (rc)
- goto failed;
rc = lnet_md_link(md, umd.eq_handle, cpt);
if (rc)
- goto failed;
+ goto out_unlock;
lnet_md2handle(handle, md);
lnet_res_unlock(cpt);
return 0;
- failed:
+out_unlock:
+ lnet_res_unlock(cpt);
+out_free:
lnet_md_free(md);
- lnet_res_unlock(cpt);
return rc;
}
EXPORT_SYMBOL(LNetMDBind);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 119/124] staging: lustre: lnet: potential deadlock in lnet
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Quentin Bouget, James Simmons
From: Quentin Bouget <quentin.bouget.ocre@cea.fr>
Fixes potential deadlock in LNetMDAttach
Signed-off-by: Quentin Bouget <quentin.bouget.ocre@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8249
Reviewed-on: http://review.whamcloud.com/20676
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/lib-md.c | 24 +++++++++++++-----------
1 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/lib-md.c b/drivers/staging/lustre/lnet/lnet/lib-md.c
index e0b2f16..eab53cd 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-md.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-md.c
@@ -292,11 +292,12 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
return -ENOMEM;
rc = lnet_md_build(md, &umd, unlink);
+ if (rc)
+ goto out_free;
+
cpt = lnet_cpt_of_cookie(meh.cookie);
lnet_res_lock(cpt);
- if (rc)
- goto failed;
me = lnet_handle2me(&meh);
if (!me)
@@ -307,7 +308,7 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
rc = lnet_md_link(md, umd.eq_handle, cpt);
if (rc)
- goto failed;
+ goto out_unlock;
/*
* attach this MD to portal of ME and check if it matches any
@@ -324,10 +325,10 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
return 0;
- failed:
- lnet_md_free(md);
-
+out_unlock:
lnet_res_unlock(cpt);
+out_free:
+ lnet_md_free(md);
return rc;
}
EXPORT_SYMBOL(LNetMDAttach);
@@ -370,24 +371,25 @@ LNetMDBind(lnet_md_t umd, lnet_unlink_t unlink, lnet_handle_md_t *handle)
return -ENOMEM;
rc = lnet_md_build(md, &umd, unlink);
+ if (rc)
+ goto out_free;
cpt = lnet_res_lock_current();
- if (rc)
- goto failed;
rc = lnet_md_link(md, umd.eq_handle, cpt);
if (rc)
- goto failed;
+ goto out_unlock;
lnet_md2handle(handle, md);
lnet_res_unlock(cpt);
return 0;
- failed:
+out_unlock:
+ lnet_res_unlock(cpt);
+out_free:
lnet_md_free(md);
- lnet_res_unlock(cpt);
return rc;
}
EXPORT_SYMBOL(LNetMDBind);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 120/124] staging: lustre: lnet: check if ni is in current net namespace
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:38 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Sebastien Buisson, James Simmons
From: Sebastien Buisson <sbuisson@ddn.com>
Add new 'ni_net_ns' field to struct lnet_ni to hold a reference
to original net namespace in which ni is created.
In LNetDist(), check if ni was created in same net namespace as
current's one. If not, assign order above 0xffff0000, to make
this ni not a priority.
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7845
Reviewed-on: http://review.whamcloud.com/21884
Reviewed-by: Olaf Weber <olaf@sgi.com>
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/lnet/lib-types.h | 2 ++
drivers/staging/lustre/lnet/lnet/config.c | 14 ++++++++++++++
drivers/staging/lustre/lnet/lnet/lib-move.c | 11 +++++++++++
3 files changed, 27 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-types.h b/drivers/staging/lustre/include/linux/lnet/lib-types.h
index 640ff72..b84a5bb 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-types.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-types.h
@@ -275,6 +275,8 @@ typedef struct lnet_ni {
struct lnet_ioctl_config_lnd_tunables *ni_lnd_tunables;
/* equivalent interfaces to use */
char *ni_interfaces[LNET_MAX_INTERFACES];
+ /* original net namespace */
+ struct net *ni_net_ns;
} lnet_ni_t;
#define LNET_PROTO_PING_MATCHBITS 0x8000000000000000LL
diff --git a/drivers/staging/lustre/lnet/lnet/config.c b/drivers/staging/lustre/lnet/lnet/config.c
index a72afdf..9e2183f 100644
--- a/drivers/staging/lustre/lnet/lnet/config.c
+++ b/drivers/staging/lustre/lnet/lnet/config.c
@@ -31,6 +31,8 @@
*/
#define DEBUG_SUBSYSTEM S_LNET
+#include <linux/nsproxy.h>
+#include <net/net_namespace.h>
#include "../../include/linux/lnet/lib-lnet.h"
struct lnet_text_buf { /* tmp struct for parsing routes */
@@ -110,6 +112,11 @@ lnet_ni_free(struct lnet_ni *ni)
LIBCFS_FREE(ni->ni_interfaces[i],
strlen(ni->ni_interfaces[i]) + 1);
}
+
+ /* release reference to net namespace */
+ if (ni->ni_net_ns)
+ put_net(ni->ni_net_ns);
+
LIBCFS_FREE(ni, sizeof(*ni));
}
@@ -171,6 +178,13 @@ lnet_ni_alloc(__u32 net, struct cfs_expr_list *el, struct list_head *nilist)
/* LND will fill in the address part of the NID */
ni->ni_nid = LNET_MKNID(net, 0);
+
+ /* Store net namespace in which current ni is being created */
+ if (current->nsproxy->net_ns)
+ ni->ni_net_ns = get_net(current->nsproxy->net_ns);
+ else
+ ni->ni_net_ns = NULL;
+
ni->ni_last_alive = ktime_get_real_seconds();
list_add_tail(&ni->ni_list, nilist);
return ni;
diff --git a/drivers/staging/lustre/lnet/lnet/lib-move.c b/drivers/staging/lustre/lnet/lnet/lib-move.c
index f89c9fe..48e6f8f 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-move.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-move.c
@@ -37,6 +37,8 @@
#define DEBUG_SUBSYSTEM S_LNET
#include "../../include/linux/lnet/lib-lnet.h"
+#include <linux/nsproxy.h>
+#include <net/net_namespace.h>
static int local_nid_dist_zero = 1;
module_param(local_nid_dist_zero, int, 0444);
@@ -2320,6 +2322,15 @@ LNetDist(lnet_nid_t dstnid, lnet_nid_t *srcnidp, __u32 *orderp)
}
if (LNET_NIDNET(ni->ni_nid) == dstnet) {
+ /*
+ * Check if ni was originally created in
+ * current net namespace.
+ * If not, assign order above 0xffff0000,
+ * to make this ni not a priority.
+ */
+ if (!net_eq(ni->ni_net_ns, current->nsproxy->net_ns))
+ order += 0xffff0000;
+
if (srcnidp)
*srcnidp = ni->ni_nid;
if (orderp)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 120/124] staging: lustre: lnet: check if ni is in current net namespace
@ 2016-09-18 20:38 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:38 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Sebastien Buisson, James Simmons
From: Sebastien Buisson <sbuisson@ddn.com>
Add new 'ni_net_ns' field to struct lnet_ni to hold a reference
to original net namespace in which ni is created.
In LNetDist(), check if ni was created in same net namespace as
current's one. If not, assign order above 0xffff0000, to make
this ni not a priority.
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7845
Reviewed-on: http://review.whamcloud.com/21884
Reviewed-by: Olaf Weber <olaf@sgi.com>
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/lnet/lib-types.h | 2 ++
drivers/staging/lustre/lnet/lnet/config.c | 14 ++++++++++++++
drivers/staging/lustre/lnet/lnet/lib-move.c | 11 +++++++++++
3 files changed, 27 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-types.h b/drivers/staging/lustre/include/linux/lnet/lib-types.h
index 640ff72..b84a5bb 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-types.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-types.h
@@ -275,6 +275,8 @@ typedef struct lnet_ni {
struct lnet_ioctl_config_lnd_tunables *ni_lnd_tunables;
/* equivalent interfaces to use */
char *ni_interfaces[LNET_MAX_INTERFACES];
+ /* original net namespace */
+ struct net *ni_net_ns;
} lnet_ni_t;
#define LNET_PROTO_PING_MATCHBITS 0x8000000000000000LL
diff --git a/drivers/staging/lustre/lnet/lnet/config.c b/drivers/staging/lustre/lnet/lnet/config.c
index a72afdf..9e2183f 100644
--- a/drivers/staging/lustre/lnet/lnet/config.c
+++ b/drivers/staging/lustre/lnet/lnet/config.c
@@ -31,6 +31,8 @@
*/
#define DEBUG_SUBSYSTEM S_LNET
+#include <linux/nsproxy.h>
+#include <net/net_namespace.h>
#include "../../include/linux/lnet/lib-lnet.h"
struct lnet_text_buf { /* tmp struct for parsing routes */
@@ -110,6 +112,11 @@ lnet_ni_free(struct lnet_ni *ni)
LIBCFS_FREE(ni->ni_interfaces[i],
strlen(ni->ni_interfaces[i]) + 1);
}
+
+ /* release reference to net namespace */
+ if (ni->ni_net_ns)
+ put_net(ni->ni_net_ns);
+
LIBCFS_FREE(ni, sizeof(*ni));
}
@@ -171,6 +178,13 @@ lnet_ni_alloc(__u32 net, struct cfs_expr_list *el, struct list_head *nilist)
/* LND will fill in the address part of the NID */
ni->ni_nid = LNET_MKNID(net, 0);
+
+ /* Store net namespace in which current ni is being created */
+ if (current->nsproxy->net_ns)
+ ni->ni_net_ns = get_net(current->nsproxy->net_ns);
+ else
+ ni->ni_net_ns = NULL;
+
ni->ni_last_alive = ktime_get_real_seconds();
list_add_tail(&ni->ni_list, nilist);
return ni;
diff --git a/drivers/staging/lustre/lnet/lnet/lib-move.c b/drivers/staging/lustre/lnet/lnet/lib-move.c
index f89c9fe..48e6f8f 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-move.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-move.c
@@ -37,6 +37,8 @@
#define DEBUG_SUBSYSTEM S_LNET
#include "../../include/linux/lnet/lib-lnet.h"
+#include <linux/nsproxy.h>
+#include <net/net_namespace.h>
static int local_nid_dist_zero = 1;
module_param(local_nid_dist_zero, int, 0444);
@@ -2320,6 +2322,15 @@ LNetDist(lnet_nid_t dstnid, lnet_nid_t *srcnidp, __u32 *orderp)
}
if (LNET_NIDNET(ni->ni_nid) == dstnet) {
+ /*
+ * Check if ni was originally created in
+ * current net namespace.
+ * If not, assign order above 0xffff0000,
+ * to make this ni not a priority.
+ */
+ if (!net_eq(ni->ni_net_ns, current->nsproxy->net_ns))
+ order += 0xffff0000;
+
if (srcnidp)
*srcnidp = ni->ni_nid;
if (orderp)
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 121/124] staging: lustre: lnet: Ensure routing is turned on first time
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:39 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Doug Oucharek, James Simmons
From: Doug Oucharek <doug.s.oucharek@intel.com>
In lnet_rtrpools_enable(), a mistake was made and routing
was not being turned on when the rtrpools are being allocated
for the first time.
This patch fixes that routine so we remember to turn on
routing after allocating the rtrpools.
Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8501
Reviewed-on: http://review.whamcloud.com/21934
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/router.c | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/router.c b/drivers/staging/lustre/lnet/lnet/router.c
index 69819c9..063ad55 100644
--- a/drivers/staging/lustre/lnet/lnet/router.c
+++ b/drivers/staging/lustre/lnet/lnet/router.c
@@ -1693,7 +1693,7 @@ lnet_rtrpools_adjust(int tiny, int small, int large)
int
lnet_rtrpools_enable(void)
{
- int rc;
+ int rc = 0;
if (the_lnet.ln_routing)
return 0;
@@ -1706,9 +1706,9 @@ lnet_rtrpools_enable(void)
* if we are just configuring this for the first
* time.
*/
- return lnet_rtrpools_alloc(1);
-
- rc = lnet_rtrpools_adjust_helper(0, 0, 0);
+ rc = lnet_rtrpools_alloc(1);
+ else
+ rc = lnet_rtrpools_adjust_helper(0, 0, 0);
if (rc)
return rc;
@@ -1718,7 +1718,7 @@ lnet_rtrpools_enable(void)
the_lnet.ln_ping_info->pi_features &= ~LNET_PING_FEAT_RTE_DISABLED;
lnet_net_unlock(LNET_LOCK_EX);
- return 0;
+ return rc;
}
void
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 121/124] staging: lustre: lnet: Ensure routing is turned on first time
@ 2016-09-18 20:39 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Doug Oucharek, James Simmons
From: Doug Oucharek <doug.s.oucharek@intel.com>
In lnet_rtrpools_enable(), a mistake was made and routing
was not being turned on when the rtrpools are being allocated
for the first time.
This patch fixes that routine so we remember to turn on
routing after allocating the rtrpools.
Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8501
Reviewed-on: http://review.whamcloud.com/21934
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/router.c | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/router.c b/drivers/staging/lustre/lnet/lnet/router.c
index 69819c9..063ad55 100644
--- a/drivers/staging/lustre/lnet/lnet/router.c
+++ b/drivers/staging/lustre/lnet/lnet/router.c
@@ -1693,7 +1693,7 @@ lnet_rtrpools_adjust(int tiny, int small, int large)
int
lnet_rtrpools_enable(void)
{
- int rc;
+ int rc = 0;
if (the_lnet.ln_routing)
return 0;
@@ -1706,9 +1706,9 @@ lnet_rtrpools_enable(void)
* if we are just configuring this for the first
* time.
*/
- return lnet_rtrpools_alloc(1);
-
- rc = lnet_rtrpools_adjust_helper(0, 0, 0);
+ rc = lnet_rtrpools_alloc(1);
+ else
+ rc = lnet_rtrpools_adjust_helper(0, 0, 0);
if (rc)
return rc;
@@ -1718,7 +1718,7 @@ lnet_rtrpools_enable(void)
the_lnet.ln_ping_info->pi_features &= ~LNET_PING_FEAT_RTE_DISABLED;
lnet_net_unlock(LNET_LOCK_EX);
- return 0;
+ return rc;
}
void
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 122/124] staging: lustre: lnet: Enable setting per NI peer_credits
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:39 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Doug Oucharek, James Simmons, James Simmons
From: Doug Oucharek <doug.s.oucharek@intel.com>
The code to allow peer_credits to be set per NI was originally
"left inactive" because there were concerns about peer_credits
interfering with the ability for IB nodes to connect to each
other when peer_credits are not the same (peer_credits controls
the queue depth for IB). With LU-3322, the values do not have
to match so it is now safe to enable this code so peer_credits
can be set per NI.
This patch enables existing code for setting per NI peer_credits.
Second this patch fixes a long standing bug in that the conf data
was not being used to set variables in the lnet_ni structure until
after lnd_startup() was called which meant LND drivers were
ignoring struct lnet_ni tunable values being set. Now we change
struct lnet_ni data fields based on conf data before calling
lnd_startup().
Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8507
Reviewed-on: http://review.whamcloud.com/21948
Reviewed-by: Olaf Weber <olaf@sgi.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/api-ni.c | 46 ++++++++++++-----------------
1 files changed, 19 insertions(+), 27 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/api-ni.c b/drivers/staging/lustre/lnet/lnet/api-ni.c
index 346db89..4daf828 100644
--- a/drivers/staging/lustre/lnet/lnet/api-ni.c
+++ b/drivers/staging/lustre/lnet/lnet/api-ni.c
@@ -1286,6 +1286,25 @@ lnet_startup_lndni(struct lnet_ni *ni, struct lnet_ioctl_config_data *conf)
sizeof(*ni->ni_lnd_tunables));
}
+ /*
+ * If given some LND tunable parameters, parse those now to
+ * override the values in the NI structure.
+ */
+ if (conf) {
+ if (conf->cfg_config_u.cfg_net.net_peer_rtr_credits >= 0)
+ ni->ni_peerrtrcredits =
+ conf->cfg_config_u.cfg_net.net_peer_rtr_credits;
+ if (conf->cfg_config_u.cfg_net.net_peer_timeout >= 0)
+ ni->ni_peertimeout =
+ conf->cfg_config_u.cfg_net.net_peer_timeout;
+ if (conf->cfg_config_u.cfg_net.net_peer_tx_credits != -1)
+ ni->ni_peertxcredits =
+ conf->cfg_config_u.cfg_net.net_peer_tx_credits;
+ if (conf->cfg_config_u.cfg_net.net_max_tx_credits >= 0)
+ ni->ni_maxtxcredits =
+ conf->cfg_config_u.cfg_net.net_max_tx_credits;
+ }
+
rc = lnd->lnd_startup(ni);
mutex_unlock(&the_lnet.ln_lnd_mutex);
@@ -1299,33 +1318,6 @@ lnet_startup_lndni(struct lnet_ni *ni, struct lnet_ioctl_config_data *conf)
goto failed0;
}
- /*
- * If given some LND tunable parameters, parse those now to
- * override the values in the NI structure.
- */
- if (conf && conf->cfg_config_u.cfg_net.net_peer_rtr_credits >= 0) {
- ni->ni_peerrtrcredits =
- conf->cfg_config_u.cfg_net.net_peer_rtr_credits;
- }
- if (conf && conf->cfg_config_u.cfg_net.net_peer_timeout >= 0) {
- ni->ni_peertimeout =
- conf->cfg_config_u.cfg_net.net_peer_timeout;
- }
- /*
- * TODO
- * Note: For now, don't allow the user to change
- * peertxcredits as this number is used in the
- * IB LND to control queue depth.
- *
- * if (conf && conf->cfg_config_u.cfg_net.net_peer_tx_credits != -1)
- * ni->ni_peertxcredits =
- * conf->cfg_config_u.cfg_net.net_peer_tx_credits;
- */
- if (conf && conf->cfg_config_u.cfg_net.net_max_tx_credits >= 0) {
- ni->ni_maxtxcredits =
- conf->cfg_config_u.cfg_net.net_max_tx_credits;
- }
-
LASSERT(ni->ni_peertimeout <= 0 || lnd->lnd_query);
lnet_net_lock(LNET_LOCK_EX);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 122/124] staging: lustre: lnet: Enable setting per NI peer_credits
@ 2016-09-18 20:39 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Doug Oucharek, James Simmons, James Simmons
From: Doug Oucharek <doug.s.oucharek@intel.com>
The code to allow peer_credits to be set per NI was originally
"left inactive" because there were concerns about peer_credits
interfering with the ability for IB nodes to connect to each
other when peer_credits are not the same (peer_credits controls
the queue depth for IB). With LU-3322, the values do not have
to match so it is now safe to enable this code so peer_credits
can be set per NI.
This patch enables existing code for setting per NI peer_credits.
Second this patch fixes a long standing bug in that the conf data
was not being used to set variables in the lnet_ni structure until
after lnd_startup() was called which meant LND drivers were
ignoring struct lnet_ni tunable values being set. Now we change
struct lnet_ni data fields based on conf data before calling
lnd_startup().
Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8507
Reviewed-on: http://review.whamcloud.com/21948
Reviewed-by: Olaf Weber <olaf@sgi.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/api-ni.c | 46 ++++++++++++-----------------
1 files changed, 19 insertions(+), 27 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/api-ni.c b/drivers/staging/lustre/lnet/lnet/api-ni.c
index 346db89..4daf828 100644
--- a/drivers/staging/lustre/lnet/lnet/api-ni.c
+++ b/drivers/staging/lustre/lnet/lnet/api-ni.c
@@ -1286,6 +1286,25 @@ lnet_startup_lndni(struct lnet_ni *ni, struct lnet_ioctl_config_data *conf)
sizeof(*ni->ni_lnd_tunables));
}
+ /*
+ * If given some LND tunable parameters, parse those now to
+ * override the values in the NI structure.
+ */
+ if (conf) {
+ if (conf->cfg_config_u.cfg_net.net_peer_rtr_credits >= 0)
+ ni->ni_peerrtrcredits =
+ conf->cfg_config_u.cfg_net.net_peer_rtr_credits;
+ if (conf->cfg_config_u.cfg_net.net_peer_timeout >= 0)
+ ni->ni_peertimeout =
+ conf->cfg_config_u.cfg_net.net_peer_timeout;
+ if (conf->cfg_config_u.cfg_net.net_peer_tx_credits != -1)
+ ni->ni_peertxcredits =
+ conf->cfg_config_u.cfg_net.net_peer_tx_credits;
+ if (conf->cfg_config_u.cfg_net.net_max_tx_credits >= 0)
+ ni->ni_maxtxcredits =
+ conf->cfg_config_u.cfg_net.net_max_tx_credits;
+ }
+
rc = lnd->lnd_startup(ni);
mutex_unlock(&the_lnet.ln_lnd_mutex);
@@ -1299,33 +1318,6 @@ lnet_startup_lndni(struct lnet_ni *ni, struct lnet_ioctl_config_data *conf)
goto failed0;
}
- /*
- * If given some LND tunable parameters, parse those now to
- * override the values in the NI structure.
- */
- if (conf && conf->cfg_config_u.cfg_net.net_peer_rtr_credits >= 0) {
- ni->ni_peerrtrcredits =
- conf->cfg_config_u.cfg_net.net_peer_rtr_credits;
- }
- if (conf && conf->cfg_config_u.cfg_net.net_peer_timeout >= 0) {
- ni->ni_peertimeout =
- conf->cfg_config_u.cfg_net.net_peer_timeout;
- }
- /*
- * TODO
- * Note: For now, don't allow the user to change
- * peertxcredits as this number is used in the
- * IB LND to control queue depth.
- *
- * if (conf && conf->cfg_config_u.cfg_net.net_peer_tx_credits != -1)
- * ni->ni_peertxcredits =
- * conf->cfg_config_u.cfg_net.net_peer_tx_credits;
- */
- if (conf && conf->cfg_config_u.cfg_net.net_max_tx_credits >= 0) {
- ni->ni_maxtxcredits =
- conf->cfg_config_u.cfg_net.net_max_tx_credits;
- }
-
LASSERT(ni->ni_peertimeout <= 0 || lnd->lnd_query);
lnet_net_lock(LNET_LOCK_EX);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 123/124] staging: lustre: o2iblnd: Put back work queue check previously removed
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:39 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Doug Oucharek, James Simmons
From: Doug Oucharek <doug.s.oucharek@intel.com>
The previous patch, http://review.whamcloud.com/21304/, removed
a check needed until LU-5718 is properly addressed. With
the check, LU-5718 results in an error message and a lost
RDMA operation. Without it, we have memory corruption and
a crash (much harder to debug).
Putting the check back in case LU-5718 is not fixed soon.
Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7650
Reviewed-on: http://review.whamcloud.com/22281
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Olaf Weber <olaf@sgi.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c | 10 ++++++++++
1 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
index 3a86879..b27de88 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
@@ -1093,6 +1093,16 @@ kiblnd_init_rdma(struct kib_conn *conn, struct kib_tx *tx, int type,
break;
}
+ if (tx->tx_nwrq >= IBLND_MAX_RDMA_FRAGS) {
+ CERROR("RDMA has too many fragments for peer %s (%d), src idx/frags: %d/%d dst idx/frags: %d/%d\n",
+ libcfs_nid2str(conn->ibc_peer->ibp_nid),
+ IBLND_MAX_RDMA_FRAGS,
+ srcidx, srcrd->rd_nfrags,
+ dstidx, dstrd->rd_nfrags);
+ rc = -EMSGSIZE;
+ break;
+ }
+
wrknob = min(min(kiblnd_rd_frag_size(srcrd, srcidx),
kiblnd_rd_frag_size(dstrd, dstidx)),
(__u32)resid);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread* [lustre-devel] [PATCH 123/124] staging: lustre: o2iblnd: Put back work queue check previously removed
@ 2016-09-18 20:39 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List,
Doug Oucharek, James Simmons
From: Doug Oucharek <doug.s.oucharek@intel.com>
The previous patch, http://review.whamcloud.com/21304/, removed
a check needed until LU-5718 is properly addressed. With
the check, LU-5718 results in an error message and a lost
RDMA operation. Without it, we have memory corruption and
a crash (much harder to debug).
Putting the check back in case LU-5718 is not fixed soon.
Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7650
Reviewed-on: http://review.whamcloud.com/22281
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Olaf Weber <olaf@sgi.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c | 10 ++++++++++
1 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
index 3a86879..b27de88 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
@@ -1093,6 +1093,16 @@ kiblnd_init_rdma(struct kib_conn *conn, struct kib_tx *tx, int type,
break;
}
+ if (tx->tx_nwrq >= IBLND_MAX_RDMA_FRAGS) {
+ CERROR("RDMA has too many fragments for peer %s (%d), src idx/frags: %d/%d dst idx/frags: %d/%d\n",
+ libcfs_nid2str(conn->ibc_peer->ibp_nid),
+ IBLND_MAX_RDMA_FRAGS,
+ srcidx, srcrd->rd_nfrags,
+ dstidx, dstrd->rd_nfrags);
+ rc = -EMSGSIZE;
+ break;
+ }
+
wrknob = min(min(kiblnd_rd_frag_size(srcrd, srcidx),
kiblnd_rd_frag_size(dstrd, dstidx)),
(__u32)resid);
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [PATCH 124/124] staging: lustre: update version to 2.6.99
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-18 20:39 ` James Simmons
-1 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
With all of the the missing patches from the
lustre 2.7 version merged upstream its time to update
the upstream clients version.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_ver.h | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_ver.h b/drivers/staging/lustre/lustre/include/lustre_ver.h
index 414075f..19c9135 100644
--- a/drivers/staging/lustre/lustre/include/lustre_ver.h
+++ b/drivers/staging/lustre/lustre/include/lustre_ver.h
@@ -2,10 +2,10 @@
#define _LUSTRE_VER_H_
#define LUSTRE_MAJOR 2
-#define LUSTRE_MINOR 5
-#define LUSTRE_PATCH 60
+#define LUSTRE_MINOR 6
+#define LUSTRE_PATCH 99
#define LUSTRE_FIX 0
-#define LUSTRE_VERSION_STRING "2.5.99"
+#define LUSTRE_VERSION_STRING "2.6.99"
#define OBD_OCD_VERSION(major, minor, patch, fix) \
(((major) << 24) + ((minor) << 16) + ((patch) << 8) + (fix))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 124/124] staging: lustre: update version to 2.6.99
@ 2016-09-18 20:39 ` James Simmons
0 siblings, 0 replies; 263+ messages in thread
From: James Simmons @ 2016-09-18 20:39 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Andreas Dilger, Oleg Drokin
Cc: Linux Kernel Mailing List, Lustre Development List, James Simmons
With all of the the missing patches from the
lustre 2.7 version merged upstream its time to update
the upstream clients version.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lustre/include/lustre_ver.h | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre_ver.h b/drivers/staging/lustre/lustre/include/lustre_ver.h
index 414075f..19c9135 100644
--- a/drivers/staging/lustre/lustre/include/lustre_ver.h
+++ b/drivers/staging/lustre/lustre/include/lustre_ver.h
@@ -2,10 +2,10 @@
#define _LUSTRE_VER_H_
#define LUSTRE_MAJOR 2
-#define LUSTRE_MINOR 5
-#define LUSTRE_PATCH 60
+#define LUSTRE_MINOR 6
+#define LUSTRE_PATCH 99
#define LUSTRE_FIX 0
-#define LUSTRE_VERSION_STRING "2.5.99"
+#define LUSTRE_VERSION_STRING "2.6.99"
#define OBD_OCD_VERSION(major, minor, patch, fix) \
(((major) << 24) + ((minor) << 16) + ((patch) << 8) + (fix))
--
1.7.1
^ permalink raw reply related [flat|nested] 263+ messages in thread
* Re: [PATCH 000/124] missing patches from Lustre 2.7 release
2016-09-18 20:36 ` [lustre-devel] " James Simmons
@ 2016-09-19 8:10 ` Greg Kroah-Hartman
-1 siblings, 0 replies; 263+ messages in thread
From: Greg Kroah-Hartman @ 2016-09-19 8:10 UTC (permalink / raw)
To: James Simmons
Cc: devel, Andreas Dilger, Oleg Drokin, Linux Kernel Mailing List,
Lustre Development List
On Sun, Sep 18, 2016 at 04:36:59PM -0400, James Simmons wrote:
> This batch of patches includes various cleanups and fixes from
> the Lustre 2.7 release. Most of these patches are removal of
> cruft and unsigned to signed comparsion fixes. Also added in
> the latest LNet fixes as well.
All now queued up, thanks.
greg k-h
^ permalink raw reply [flat|nested] 263+ messages in thread
* [lustre-devel] [PATCH 000/124] missing patches from Lustre 2.7 release
@ 2016-09-19 8:10 ` Greg Kroah-Hartman
0 siblings, 0 replies; 263+ messages in thread
From: Greg Kroah-Hartman @ 2016-09-19 8:10 UTC (permalink / raw)
To: James Simmons
Cc: devel, Andreas Dilger, Oleg Drokin, Linux Kernel Mailing List,
Lustre Development List
On Sun, Sep 18, 2016 at 04:36:59PM -0400, James Simmons wrote:
> This batch of patches includes various cleanups and fixes from
> the Lustre 2.7 release. Most of these patches are removal of
> cruft and unsigned to signed comparsion fixes. Also added in
> the latest LNet fixes as well.
All now queued up, thanks.
greg k-h
^ permalink raw reply [flat|nested] 263+ messages in thread