* [lustre-devel] [PATCH 01/20] lustre: nfs: don't store parent fid
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 02/20] lustre: sec: filename encryption - symlink support James Simmons
` (18 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lai Siyao, Lustre Development List
From: Lai Siyao <lai.siyao@whamcloud.com>
It's not necessary to store parent fid in lli_pfid, because MDT
can get it's parent fid from linkea, and now that DNE stripe
directory stores master inode fid in lli_pfid, stop storing parent
fid to avoid conflict.
WC-bug-id: https://jira.whamcloud.com/browse/LU-3544
Lustre-commit: 6512bfc74b152ef ("LU-3544 nfs: don't store parent fid")
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/10692
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Nathaniel Clark <nclark@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/llite/llite_nfs.c | 13 -------------
1 file changed, 13 deletions(-)
diff --git a/fs/lustre/llite/llite_nfs.c b/fs/lustre/llite/llite_nfs.c
index 6be2309..07fcad6 100644
--- a/fs/lustre/llite/llite_nfs.c
+++ b/fs/lustre/llite/llite_nfs.c
@@ -135,19 +135,6 @@ struct inode *search_inode_for_lustre(struct super_block *sb,
if (IS_ERR(result))
return result;
- /**
- * In case d_obtain_alias() found a disconnected dentry, always update
- * lli_pfid to allow later operation (normally open) have parent fid,
- * which may be used by MDS to create data.
- */
- if (parent) {
- struct ll_inode_info *lli = ll_i2info(inode);
-
- spin_lock(&lli->lli_lock);
- lli->lli_pfid = *parent;
- spin_unlock(&lli->lli_lock);
- }
-
/*
* Need to signal to the ll_intent_file_open that
* we came from NFS and so opencache needs to be
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 02/20] lustre: sec: filename encryption - symlink support
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 01/20] lustre: nfs: don't store parent fid James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 03/20] lustre: llite: support fallocate() on selected mirror James Simmons
` (17 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Sebastien Buisson <sbuisson@ddn.com>
On client side, call the appropriate fscrypt primitives from llite,
to proceed with symlink encryption before sending requests to servers
and symlink decryption upon request receipt.
The tricky part is that fscrypt needs an inode to encrypt the target
name. But by the time we prepare the symlink creation request to be
sent to the server with the target name (in ll_new_node), we do not
have an inode yet (it will be obtained only after we get the server
reply). So we create a fake inode and associate the right encryption
context to it, so that the symlink gets encrypted properly.
In order to report the correct size for an encrypted symlink (which is
ought to be the length of the symlink target), we need to read the
symlink target and decrypt or decode it in ->getattr(). This has a
performance hit, but given that the symlink target is cached in
->i_link (when the key is available), the symlink will not have to be
read and decrypted again later when it is actually followed,
readlink() is called, or lstat() is called again.
This part of the patch is adapted from kernel commit
d18760560593e5af921f51a8c9b64b6109d634c2
"fscrypt: add fscrypt_symlink_getattr() for computing st_size"
With encrypted file names, a symlink target is binary. So make sure
server side can handle that, by switching sp_symname to a
struct lu_name in struct md_op_spec.
WC-bug-id: https://jira.whamcloud.com/browse/LU-13717
Lustre-commit: e735298935b64541f ("LU-13717 sec: filename encryption - symlink support")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/43394
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/llite/namei.c | 97 ++++++++++++++++++++++++++++++++++++++---------
fs/lustre/llite/symlink.c | 85 ++++++++++++++++++++++++++++++++++++++---
2 files changed, 158 insertions(+), 24 deletions(-)
diff --git a/fs/lustre/llite/namei.c b/fs/lustre/llite/namei.c
index f0f10da..1812c09 100644
--- a/fs/lustre/llite/namei.c
+++ b/fs/lustre/llite/namei.c
@@ -1531,25 +1531,29 @@ static void ll_qos_mkdir_prep(struct md_op_data *op_data, struct inode *dir)
up_read(&rlli->lli_lsm_sem);
}
-static int ll_new_node(struct inode *dir, struct dentry *dentry,
- const char *tgt, umode_t mode, int rdev,
+static int ll_new_node(struct inode *dir, struct dentry *dchild,
+ const char *tgt, umode_t mode, u64 rdev,
u32 opc)
{
+ struct qstr *name = &dchild->d_name;
struct ptlrpc_request *request = NULL;
struct md_op_data *op_data = NULL;
struct inode *inode = NULL;
struct ll_sb_info *sbi = ll_i2sbi(dir);
- int tgt_len = 0;
+ struct fscrypt_str *disk_link = NULL;
bool encrypt = false;
int err;
- if (unlikely(tgt))
- tgt_len = strlen(tgt) + 1;
+ if (unlikely(tgt)) {
+ disk_link = (struct fscrypt_str *)rdev;
+ rdev = 0;
+ if (!disk_link)
+ return -EINVAL;
+ }
+
again:
- op_data = ll_prep_md_op_data(NULL, dir, NULL,
- dentry->d_name.name,
- dentry->d_name.len,
- 0, opc, NULL);
+ op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name,
+ name->len, 0, opc, NULL);
if (IS_ERR(op_data)) {
err = PTR_ERR(op_data);
goto err_exit;
@@ -1559,7 +1563,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
ll_qos_mkdir_prep(op_data, dir);
if (sbi->ll_flags & LL_SBI_FILE_SECCTX) {
- err = ll_dentry_init_security(dentry, mode, &dentry->d_name,
+ err = ll_dentry_init_security(dchild, mode, &dchild->d_name,
&op_data->op_file_secctx_name,
&op_data->op_file_secctx,
&op_data->op_file_secctx_size);
@@ -1585,9 +1589,40 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
err = fscrypt_inherit_context(dir, NULL, op_data, false);
if (err)
goto err_exit;
+
+ if (S_ISLNK(mode)) {
+ /* fscrypt needs inode to encrypt target name, so create
+ * a fake inode and associate encryption context got
+ * from fscrypt_inherit_context.
+ */
+ struct inode *fakeinode =
+ dchild->d_sb->s_op->alloc_inode(dchild->d_sb);
+
+ if (!fakeinode) {
+ err = -ENOMEM;
+ goto err_exit;
+ }
+ fakeinode->i_sb = dchild->d_sb;
+ fakeinode->i_mode |= S_IFLNK;
+ err = ll_set_encflags(fakeinode,
+ op_data->op_file_encctx,
+ op_data->op_file_encctx_size,
+ true);
+ if (!err)
+ err = __fscrypt_encrypt_symlink(fakeinode, tgt,
+ strlen(tgt),
+ disk_link);
+
+ ll_xattr_cache_destroy(fakeinode);
+ fscrypt_put_encryption_info(fakeinode);
+ dchild->d_sb->s_op->destroy_inode(fakeinode);
+ if (err)
+ goto err_exit;
+ }
}
- err = md_create(sbi->ll_md_exp, op_data, tgt, tgt_len, mode,
+ err = md_create(sbi->ll_md_exp, op_data, tgt ? disk_link->name : NULL,
+ tgt ? disk_link->len : 0, mode,
from_kuid(&init_user_ns, current_fsuid()),
from_kgid(&init_user_ns, current_fsgid()),
current_cap(), rdev, &request);
@@ -1687,16 +1722,32 @@ static int ll_new_node(struct inode *dir, struct dentry *dentry,
goto err_exit;
}
- d_instantiate(dentry, inode);
+ d_instantiate(dchild, inode);
if (encrypt) {
err = fscrypt_inherit_context(dir, inode, NULL, true);
if (err)
goto err_exit;
+
+ if (S_ISLNK(mode)) {
+ struct ll_inode_info *lli = ll_i2info(inode);
+
+ /* Cache the plaintext symlink target
+ * for later use by get_link()
+ */
+ lli->lli_symlink_name = kzalloc(strlen(tgt) + 1,
+ GFP_NOFS);
+ /* do not return an error if we cannot
+ * cache the symlink locally
+ */
+ if (lli->lli_symlink_name)
+ memcpy(lli->lli_symlink_name,
+ tgt, strlen(tgt) + 1);
+ }
}
if (!(sbi->ll_flags & LL_SBI_FILE_SECCTX))
- err = ll_inode_init_security(dentry, inode, dir);
+ err = ll_inode_init_security(dchild, inode, dir);
err_exit:
if (request)
ptlrpc_req_finished(request);
@@ -1894,17 +1945,27 @@ static int ll_rmdir(struct inode *dir, struct dentry *dchild)
return rc;
}
-static int ll_symlink(struct inode *dir, struct dentry *dentry,
- const char *oldname)
+static int ll_symlink(struct inode *dir, struct dentry *dchild,
+ const char *oldpath)
{
ktime_t kstart = ktime_get();
+ int len = strlen(oldpath);
+ struct fscrypt_str disk_link;
int err;
CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir=" DFID "(%p),target=%.*s\n",
- dentry, PFID(ll_inode2fid(dir)), dir, 3000, oldname);
+ dchild, PFID(ll_inode2fid(dir)), dir, 3000, oldpath);
+
+ err = fscrypt_prepare_symlink(dir, oldpath, len, dir->i_sb->s_blocksize,
+ &disk_link);
+ if (err)
+ return err;
+
+ err = ll_new_node(dir, dchild, oldpath, S_IFLNK | 0777,
+ (u64)&disk_link, LUSTRE_OPC_SYMLINK);
- err = ll_new_node(dir, dentry, oldname, S_IFLNK | 0777,
- 0, LUSTRE_OPC_SYMLINK);
+ if (disk_link.name != (unsigned char *)oldpath)
+ kfree(disk_link.name);
if (!err)
ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_SYMLINK,
diff --git a/fs/lustre/llite/symlink.c b/fs/lustre/llite/symlink.c
index cf5ad9e..8ea16bb 100644
--- a/fs/lustre/llite/symlink.c
+++ b/fs/lustre/llite/symlink.c
@@ -38,8 +38,13 @@
#include "llite_internal.h"
/* Must be called with lli_size_mutex locked */
+/* HAVE_IOP_GET_LINK is defined from kernel 4.5, whereas
+ * IS_ENCRYPTED is brought by kernel 4.14.
+ * So there is no need to handle encryption case otherwise.
+ */
static int ll_readlink_internal(struct inode *inode,
- struct ptlrpc_request **request, char **symname)
+ struct ptlrpc_request **request,
+ char **symname, struct delayed_call *done)
{
struct ll_inode_info *lli = ll_i2info(inode);
struct ll_sb_info *sbi = ll_i2sbi(inode);
@@ -97,7 +102,9 @@ static int ll_readlink_internal(struct inode *inode,
}
*symname = req_capsule_server_get(&(*request)->rq_pill, &RMF_MDT_MD);
- if (!*symname || strnlen(*symname, symlen) != symlen - 1) {
+ if (!*symname ||
+ (!IS_ENCRYPTED(inode) &&
+ strnlen(*symname, symlen) != symlen - 1)) {
/* not full/NULL terminated */
CERROR("%s: inode " DFID ": symlink not NULL terminated string of length %d\n",
ll_i2sbi(inode)->ll_fsname,
@@ -106,6 +113,21 @@ static int ll_readlink_internal(struct inode *inode,
goto failed;
}
+ if (IS_ENCRYPTED(inode)) {
+ const char *target = fscrypt_get_symlink(inode, *symname,
+ symlen, done);
+ if (IS_ERR(target))
+ return PTR_ERR(target);
+ symlen = strlen(target) + 1;
+ *symname = (char *)target;
+
+ /* Do not cache symlink targets encoded without the key,
+ * since those become outdated once the key is added.
+ */
+ if (!fscrypt_has_encryption_key(inode))
+ return 0;
+ }
+
lli->lli_symlink_name = kzalloc(symlen, GFP_NOFS);
/* do not return an error if we cannot cache the symlink locally */
if (lli->lli_symlink_name) {
@@ -131,12 +153,12 @@ static const char *ll_get_link(struct dentry *dentry,
int rc;
char *symname = NULL;
+ CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, inode="DFID"(%p)\n",
+ dentry, PFID(ll_inode2fid(inode)), inode);
if (!dentry)
return ERR_PTR(-ECHILD);
-
- CDEBUG(D_VFSTRACE, "VFS Op\n");
ll_inode_size_lock(inode);
- rc = ll_readlink_internal(inode, &request, &symname);
+ rc = ll_readlink_internal(inode, &request, &symname, done);
ll_inode_size_unlock(inode);
if (rc) {
ptlrpc_req_finished(request);
@@ -151,10 +173,61 @@ static const char *ll_get_link(struct dentry *dentry,
return symname;
}
+/**
+ * ll_getattr_link() - link-specific getattr to set the correct st_size
+ * for encrypted symlinks
+ *
+ * Override st_size of encrypted symlinks to be the length of the decrypted
+ * symlink target (or the no-key encoded symlink target, if the key is
+ * unavailable) rather than the length of the encrypted symlink target. This is
+ * necessary for st_size to match the symlink target that userspace actually
+ * sees. POSIX requires this, and some userspace programs depend on it.
+ *
+ * For non encrypted symlinks, this is a just calling ll_getattr().
+ * For encrypted symlinks, this additionally requires reading the symlink target
+ * from disk if needed, setting up the inode's encryption key if possible, and
+ * then decrypting or encoding the symlink target. This makes lstat() more
+ * heavyweight than is normally the case. However, decrypted symlink targets
+ * will be cached in ->i_link, so usually the symlink won't have to be read and
+ * decrypted again later if/when it is actually followed, readlink() is called,
+ * or lstat() is called again.
+ *
+ * Return: 0 on success, -errno on failure
+ */
+static int ll_getattr_link(const struct path *path, struct kstat *stat,
+ u32 request_mask, unsigned int flags)
+{
+ struct dentry *dentry = path->dentry;
+ struct inode *inode = d_inode(dentry);
+ DEFINE_DELAYED_CALL(done);
+ const char *link;
+ int rc;
+
+ rc = ll_getattr(path, stat, request_mask, flags);
+ if (rc || !IS_ENCRYPTED(inode))
+ return rc;
+
+ /*
+ * To get the symlink target that userspace will see (whether it's the
+ * decrypted target or the no-key encoded target), we can just get it
+ * in the same way the VFS does during path resolution and readlink().
+ */
+ link = READ_ONCE(inode->i_link);
+ if (!link) {
+ link = inode->i_op->get_link(dentry, inode, &done);
+ if (IS_ERR(link))
+ return PTR_ERR(link);
+ }
+ stat->size = strlen(link);
+ do_delayed_call(&done);
+ return 0;
+}
+
+
const struct inode_operations ll_fast_symlink_inode_operations = {
.setattr = ll_setattr,
.get_link = ll_get_link,
- .getattr = ll_getattr,
+ .getattr = ll_getattr_link,
.permission = ll_inode_permission,
.listxattr = ll_listxattr,
};
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 03/20] lustre: llite: support fallocate() on selected mirror
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 01/20] lustre: nfs: don't store parent fid James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 02/20] lustre: sec: filename encryption - symlink support James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 04/20] lustre: llite: move env contexts to ll_inode_info level James Simmons
` (16 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown
Cc: Mikhail Pershin, Lustre Development List
From: Mikhail Pershin <mpershin@whamcloud.com>
- add ability to do fallocate() on designated mirror in
FLR file
- add missing FALLOC_FL_KEEP_SIZE flag to fallocate() call
in llapi_hole_punch(). It was just not working without
that flag silently
Fixes: 7ce65bb0cd ("lustre: llite: mirror resync to keep sparseness")
WC-bug-id: https://jira.whamcloud.com/browse/LU-13397
Lustre-commit: 89736d502cc99f095 ("LU-13397 llite: support fallocate() on selected mirror")
Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/44721
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/llite/file.c | 9 ++++++---
fs/lustre/lov/lov_io.c | 3 ++-
2 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/fs/lustre/llite/file.c b/fs/lustre/llite/file.c
index f340d67..9dd5c8c 100644
--- a/fs/lustre/llite/file.c
+++ b/fs/lustre/llite/file.c
@@ -5220,7 +5220,8 @@ int ll_getattr(const struct path *path, struct kstat *stat,
false);
}
-int cl_falloc(struct inode *inode, int mode, loff_t offset, loff_t len)
+int cl_falloc(struct file *file, struct inode *inode, int mode, loff_t offset,
+ loff_t len)
{
struct lu_env *env;
struct cl_io *io;
@@ -5234,6 +5235,8 @@ int cl_falloc(struct inode *inode, int mode, loff_t offset, loff_t len)
io = vvp_env_thread_io(env);
io->ci_obj = ll_i2info(inode)->lli_clob;
+ ll_io_set_mirror(io, file);
+
io->ci_verify_layout = 1;
io->u.ci_setattr.sa_parent_fid = lu_object_fid(&io->ci_obj->co_lu);
io->u.ci_setattr.sa_falloc_mode = mode;
@@ -5272,7 +5275,7 @@ int cl_falloc(struct inode *inode, int mode, loff_t offset, loff_t len)
long ll_fallocate(struct file *filp, int mode, loff_t offset, loff_t len)
{
- struct inode *inode = filp->f_path.dentry->d_inode;
+ struct inode *inode = file_inode(filp);
int rc;
if (offset < 0 || len <= 0)
@@ -5298,7 +5301,7 @@ long ll_fallocate(struct file *filp, int mode, loff_t offset, loff_t len)
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_FALLOCATE, 1);
- rc = cl_falloc(inode, mode, offset, len);
+ rc = cl_falloc(filp, inode, mode, offset, len);
/*
* ENOTSUPP (524) is an NFSv3 specific error code erroneously
* used by Lustre in several places. Retuning it here would
diff --git a/fs/lustre/lov/lov_io.c b/fs/lustre/lov/lov_io.c
index eb71d7a..d5f895f 100644
--- a/fs/lustre/lov/lov_io.c
+++ b/fs/lustre/lov/lov_io.c
@@ -322,7 +322,8 @@ static int lov_io_mirror_init(struct lov_io *lio, struct lov_object *obj,
CDEBUG(D_LAYOUT, "designated I/O mirror state: %d\n",
lov_flr_state(obj));
- if ((cl_io_is_trunc(io) || io->ci_type == CIT_WRITE) &&
+ if ((cl_io_is_trunc(io) || io->ci_type == CIT_WRITE ||
+ cl_io_is_fallocate(io)) &&
(io->ci_layout_version != obj->lo_lsm->lsm_layout_gen)) {
/* For resync I/O, the ci_layout_version was the layout
* version when resync starts. If it doesn't match the
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 04/20] lustre: llite: move env contexts to ll_inode_info level
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (2 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 03/20] lustre: llite: support fallocate() on selected mirror James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 05/20] lustre: sec: do not expose security.c to listxattr/getxattr James Simmons
` (15 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Sebastien Buisson <sbuisson@ddn.com>
Contrary to file, inode is always available, so move the list of
env contexts from the file data to the ll_inode_info level.
This is needed because we will have to handle env properties in
ll_get_context() and ll_xattr_list()/ll_listxattr().
This also requires changing lli_lock from a spinlock to an rwlock.
WC-bug-id: https://jira.whamcloud.com/browse/LU-14677
Lustre-commit: 4756af02e1297d145 ("LU-14677 llite: move env contexts to ll_inode_info level")
WC-bug-id: https://jira.whamcloud.com/browse/LU-15027
Lustre-commit: 3fb7b6271855c0b12 ("LU-15027 sec: initialize ll_inode_info for fake inode")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/44198
Reviewed-on: https://review.whamcloud.com/45023
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/llite/acl.c | 4 ++--
fs/lustre/llite/file.c | 9 +++------
fs/lustre/llite/llite_internal.h | 18 ++++++++----------
fs/lustre/llite/llite_lib.c | 4 +++-
fs/lustre/llite/llite_mmap.c | 17 +++++++----------
fs/lustre/llite/namei.c | 1 +
fs/lustre/llite/rw.c | 36 ++++++++++++++++++------------------
fs/lustre/llite/rw26.c | 5 +++--
fs/lustre/llite/xattr.c | 4 ++--
9 files changed, 47 insertions(+), 51 deletions(-)
diff --git a/fs/lustre/llite/acl.c b/fs/lustre/llite/acl.c
index f4cc149..bd045cc 100644
--- a/fs/lustre/llite/acl.c
+++ b/fs/lustre/llite/acl.c
@@ -41,10 +41,10 @@ struct posix_acl *ll_get_acl(struct inode *inode, int type)
struct ll_inode_info *lli = ll_i2info(inode);
struct posix_acl *acl = NULL;
- spin_lock(&lli->lli_lock);
+ read_lock(&lli->lli_lock);
/* VFS' acl_permission_check->check_acl will release the refcount */
acl = posix_acl_dup(lli->lli_posix_acl);
- spin_unlock(&lli->lli_lock);
+ read_unlock(&lli->lli_lock);
return acl;
}
diff --git a/fs/lustre/llite/file.c b/fs/lustre/llite/file.c
index 9dd5c8c..ad1c07e 100644
--- a/fs/lustre/llite/file.c
+++ b/fs/lustre/llite/file.c
@@ -751,10 +751,6 @@ static int ll_local_open(struct file *file, struct lookup_intent *it,
/* turn off the kernel's read-ahead */
file->f_ra.ra_pages = 0;
- /* ll_cl_context initialize */
- rwlock_init(&fd->fd_lock);
- INIT_LIST_HEAD(&fd->fd_lccs);
-
return 0;
}
@@ -1718,9 +1714,10 @@ static void ll_heat_add(struct inode *inode, enum cl_io_type iot,
range_locked = true;
}
- ll_cl_add(file, env, io, LCC_RW);
+ ll_cl_add(inode, env, io, LCC_RW);
rc = cl_io_loop(env, io);
- ll_cl_remove(file, env);
+ ll_cl_remove(inode, env);
+
if (range_locked && !is_parallel_dio) {
CDEBUG(D_VFSTRACE, "Range unlock [%llu, %llu]\n",
range.rl_start,
diff --git a/fs/lustre/llite/llite_internal.h b/fs/lustre/llite/llite_internal.h
index 25bd460..cfeec14 100644
--- a/fs/lustre/llite/llite_internal.h
+++ b/fs/lustre/llite/llite_internal.h
@@ -123,8 +123,7 @@ struct ll_trunc_sem {
struct ll_inode_info {
u32 lli_inode_magic;
-
- spinlock_t lli_lock;
+ rwlock_t lli_lock;
unsigned long lli_flags;
struct posix_acl *lli_posix_acl;
@@ -280,7 +279,8 @@ struct ll_inode_info {
struct rw_semaphore lli_xattrs_list_rwsem;
struct mutex lli_xattrs_enq_lock;
- struct list_head lli_xattrs;/* ll_xattr_entry->xe_list */
+ struct list_head lli_xattrs; /* ll_xattr_entry->xe_list */
+ struct list_head lli_lccs; /* list of ll_cl_context */
};
static inline void ll_trunc_sem_init(struct ll_trunc_sem *sem)
@@ -376,11 +376,11 @@ static inline void lli_clear_acl(struct ll_inode_info *lli)
static inline void lli_replace_acl(struct ll_inode_info *lli,
struct lustre_md *md)
{
- spin_lock(&lli->lli_lock);
+ write_lock(&lli->lli_lock);
if (lli->lli_posix_acl)
posix_acl_release(lli->lli_posix_acl);
lli->lli_posix_acl = md->posix_acl;
- spin_unlock(&lli->lli_lock);
+ write_unlock(&lli->lli_lock);
}
#else
static inline void lli_clear_acl(struct ll_inode_info *lli)
@@ -941,8 +941,6 @@ struct ll_file_data {
*/
bool fd_write_failed;
bool ll_lock_no_expand;
- rwlock_t fd_lock; /* protect lcc list */
- struct list_head fd_lccs; /* list of ll_cl_context */
/* Used by mirrored file to lead IOs to a specific mirror, usually
* for mirror resync. 0 means default.
*/
@@ -1107,10 +1105,10 @@ int ll_io_read_page(const struct lu_env *env, struct cl_io *io,
int vvp_io_write_commit(const struct lu_env *env, struct cl_io *io);
enum lcc_type;
-void ll_cl_add(struct file *file, const struct lu_env *env, struct cl_io *io,
+void ll_cl_add(struct inode *inode, const struct lu_env *env, struct cl_io *io,
enum lcc_type type);
-void ll_cl_remove(struct file *file, const struct lu_env *env);
-struct ll_cl_context *ll_cl_find(struct file *file);
+void ll_cl_remove(struct inode *inode, const struct lu_env *env);
+struct ll_cl_context *ll_cl_find(struct inode *inode);
static inline void mapping_clear_exiting(struct address_space *mapping)
{
diff --git a/fs/lustre/llite/llite_lib.c b/fs/lustre/llite/llite_lib.c
index 7a822b8..9ff881c 100644
--- a/fs/lustre/llite/llite_lib.c
+++ b/fs/lustre/llite/llite_lib.c
@@ -1080,7 +1080,7 @@ void ll_lli_init(struct ll_inode_info *lli)
{
lli->lli_inode_magic = LLI_INODE_MAGIC;
lli->lli_flags = 0;
- spin_lock_init(&lli->lli_lock);
+ rwlock_init(&lli->lli_lock);
lli->lli_posix_acl = NULL;
/* Do not set lli_fid, it has been initialized already. */
fid_zero(&lli->lli_pfid);
@@ -1132,6 +1132,8 @@ void ll_lli_init(struct ll_inode_info *lli)
}
mutex_init(&lli->lli_layout_mutex);
memset(lli->lli_jobid, 0, sizeof(lli->lli_jobid));
+ /* ll_cl_context initialize */
+ INIT_LIST_HEAD(&lli->lli_lccs);
}
int ll_fill_super(struct super_block *sb)
diff --git a/fs/lustre/llite/llite_mmap.c b/fs/lustre/llite/llite_mmap.c
index 8238a4e..8047786 100644
--- a/fs/lustre/llite/llite_mmap.c
+++ b/fs/lustre/llite/llite_mmap.c
@@ -270,6 +270,7 @@ static inline vm_fault_t to_fault_error(int result)
*/
static vm_fault_t __ll_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
{
+ struct inode *inode = file_inode(vma->vm_file);
struct lu_env *env;
struct cl_io *io;
struct vvp_io *vio = NULL;
@@ -282,15 +283,15 @@ static vm_fault_t __ll_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
if (IS_ERR(env))
return VM_FAULT_ERROR;
- if (ll_sbi_has_fast_read(ll_i2sbi(file_inode(vma->vm_file)))) {
+ if (ll_sbi_has_fast_read(ll_i2sbi(inode))) {
/* do fast fault */
bool has_retry = vmf->flags & FAULT_FLAG_RETRY_NOWAIT;
/* To avoid loops, instruct downstream to not drop mmap_sem */
vmf->flags |= FAULT_FLAG_RETRY_NOWAIT;
- ll_cl_add(vma->vm_file, env, NULL, LCC_MMAP);
+ ll_cl_add(inode, env, NULL, LCC_MMAP);
fault_ret = filemap_fault(vmf);
- ll_cl_remove(vma->vm_file, env);
+ ll_cl_remove(inode, env);
if (has_retry)
vmf->flags &= ~FAULT_FLAG_RETRY_NOWAIT;
@@ -318,8 +319,6 @@ static vm_fault_t __ll_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
result = io->ci_result;
if (result == 0) {
- struct file *vm_file = vma->vm_file;
-
vio = vvp_env_io(env);
vio->u.fault.ft_vma = vma;
vio->u.fault.ft_vmpage = NULL;
@@ -327,15 +326,13 @@ static vm_fault_t __ll_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
vio->u.fault.ft_flags = 0;
vio->u.fault.ft_flags_valid = false;
- get_file(vm_file);
-
/* May call ll_readpage() */
- ll_cl_add(vm_file, env, io, LCC_MMAP);
+ ll_cl_add(inode, env, io, LCC_MMAP);
result = cl_io_loop(env, io);
- ll_cl_remove(vm_file, env);
- fput(vm_file);
+ ll_cl_remove(inode, env);
+
/* ft_flags are only valid if we reached
* the call to filemap_fault
*/
diff --git a/fs/lustre/llite/namei.c b/fs/lustre/llite/namei.c
index 1812c09..781bb16 100644
--- a/fs/lustre/llite/namei.c
+++ b/fs/lustre/llite/namei.c
@@ -1604,6 +1604,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dchild,
}
fakeinode->i_sb = dchild->d_sb;
fakeinode->i_mode |= S_IFLNK;
+ ll_lli_init(ll_i2info(fakeinode));
err = ll_set_encflags(fakeinode,
op_data->op_file_encctx,
op_data->op_file_encctx_size,
diff --git a/fs/lustre/llite/rw.c b/fs/lustre/llite/rw.c
index 48984aa..c9f29ef 100644
--- a/fs/lustre/llite/rw.c
+++ b/fs/lustre/llite/rw.c
@@ -1567,28 +1567,28 @@ int ll_writepages(struct address_space *mapping, struct writeback_control *wbc)
return result;
}
-struct ll_cl_context *ll_cl_find(struct file *file)
+struct ll_cl_context *ll_cl_find(struct inode *inode)
{
- struct ll_file_data *fd = file->private_data;
+ struct ll_inode_info *lli = ll_i2info(inode);
struct ll_cl_context *lcc;
struct ll_cl_context *found = NULL;
- read_lock(&fd->fd_lock);
- list_for_each_entry(lcc, &fd->fd_lccs, lcc_list) {
+ read_lock(&lli->lli_lock);
+ list_for_each_entry(lcc, &lli->lli_lccs, lcc_list) {
if (lcc->lcc_cookie == current) {
found = lcc;
break;
}
}
- read_unlock(&fd->fd_lock);
+ read_unlock(&lli->lli_lock);
return found;
}
-void ll_cl_add(struct file *file, const struct lu_env *env, struct cl_io *io,
+void ll_cl_add(struct inode *inode, const struct lu_env *env, struct cl_io *io,
enum lcc_type type)
{
- struct ll_file_data *fd = file->private_data;
+ struct ll_inode_info *lli = ll_i2info(inode);
struct ll_cl_context *lcc = &ll_env_info(env)->lti_io_ctx;
memset(lcc, 0, sizeof(*lcc));
@@ -1598,19 +1598,19 @@ void ll_cl_add(struct file *file, const struct lu_env *env, struct cl_io *io,
lcc->lcc_io = io;
lcc->lcc_type = type;
- write_lock(&fd->fd_lock);
- list_add(&lcc->lcc_list, &fd->fd_lccs);
- write_unlock(&fd->fd_lock);
+ write_lock(&lli->lli_lock);
+ list_add(&lcc->lcc_list, &lli->lli_lccs);
+ write_unlock(&lli->lli_lock);
}
-void ll_cl_remove(struct file *file, const struct lu_env *env)
+void ll_cl_remove(struct inode *inode, const struct lu_env *env)
{
- struct ll_file_data *fd = file->private_data;
+ struct ll_inode_info *lli = ll_i2info(inode);
struct ll_cl_context *lcc = &ll_env_info(env)->lti_io_ctx;
- write_lock(&fd->fd_lock);
+ write_lock(&lli->lli_lock);
list_del_init(&lcc->lcc_list);
- write_unlock(&fd->fd_lock);
+ write_unlock(&lli->lli_lock);
}
int ll_io_read_page(const struct lu_env *env, struct cl_io *io,
@@ -1815,15 +1815,16 @@ static bool ll_use_fast_io(struct file *file,
int ll_readpage(struct file *file, struct page *vmpage)
{
- struct cl_object *clob = ll_i2info(file_inode(file))->lli_clob;
+ struct inode *inode = file_inode(file);
+ struct cl_object *clob = ll_i2info(inode)->lli_clob;
struct ll_cl_context *lcc;
const struct lu_env *env = NULL;
struct cl_io *io = NULL;
struct cl_page *page;
- struct ll_sb_info *sbi = ll_i2sbi(file_inode(file));
+ struct ll_sb_info *sbi = ll_i2sbi(inode);
int result;
- lcc = ll_cl_find(file);
+ lcc = ll_cl_find(inode);
if (lcc) {
env = lcc->lcc_env;
io = lcc->lcc_io;
@@ -1833,7 +1834,6 @@ int ll_readpage(struct file *file, struct page *vmpage)
struct ll_file_data *fd = file->private_data;
struct ll_readahead_state *ras = &fd->fd_ras;
struct lu_env *local_env = NULL;
- struct inode *inode = file_inode(file);
struct vvp_page *vpg;
result = -ENODATA;
diff --git a/fs/lustre/llite/rw26.c b/fs/lustre/llite/rw26.c
index e5d80cb..0a271b9 100644
--- a/fs/lustre/llite/rw26.c
+++ b/fs/lustre/llite/rw26.c
@@ -355,7 +355,7 @@ static ssize_t ll_direct_IO(struct kiocb *iocb, struct iov_iter *iter)
if (ll_iov_iter_alignment(iter) & ~PAGE_MASK)
return -EINVAL;
- lcc = ll_cl_find(file);
+ lcc = ll_cl_find(inode);
if (!lcc)
return -EIO;
@@ -518,6 +518,7 @@ static int ll_write_begin(struct file *file, struct address_space *mapping,
const struct lu_env *env = NULL;
struct cl_io *io = NULL;
struct cl_page *page = NULL;
+ struct inode *inode = file_inode(file);
struct cl_object *clob = ll_i2info(mapping->host)->lli_clob;
pgoff_t index = pos >> PAGE_SHIFT;
struct page *vmpage = NULL;
@@ -527,7 +528,7 @@ static int ll_write_begin(struct file *file, struct address_space *mapping,
CDEBUG(D_VFSTRACE, "Writing %lu of %d to %d bytes\n", index, from, len);
- lcc = ll_cl_find(file);
+ lcc = ll_cl_find(inode);
if (!lcc) {
vmpage = grab_cache_page_nowait(mapping, index);
result = ll_tiny_write_begin(vmpage, mapping);
diff --git a/fs/lustre/llite/xattr.c b/fs/lustre/llite/xattr.c
index cd973eb..001c828 100644
--- a/fs/lustre/llite/xattr.c
+++ b/fs/lustre/llite/xattr.c
@@ -453,9 +453,9 @@ static int ll_xattr_get_common(const struct xattr_handler *handler,
struct ll_inode_info *lli = ll_i2info(inode);
struct posix_acl *acl;
- spin_lock(&lli->lli_lock);
+ read_lock(&lli->lli_lock);
acl = posix_acl_dup(lli->lli_posix_acl);
- spin_unlock(&lli->lli_lock);
+ read_unlock(&lli->lli_lock);
if (!acl)
return -ENODATA;
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 05/20] lustre: sec: do not expose security.c to listxattr/getxattr
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (3 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 04/20] lustre: llite: move env contexts to ll_inode_info level James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 06/20] lustre: brw: log T10 GRD tags during checksum calcs James Simmons
` (14 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Sebastien Buisson <sbuisson@ddn.com>
security.c xattr, which contains encryption context, should not be
exposed by the xattr-related system calls such as listxattr() and
getxattr() because of its special semantics.
Update sanity-sec test_57 to test this.
WC-bug-id: https://jira.whamcloud.com/browse/LU-14677
Lustre-commit: efb66de719329ce4d ("LU-14677 sec: do not expose security.c to listxattr/getxattr")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/44101
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/llite/crypto.c | 16 ++++++++++++++++
fs/lustre/llite/llite_internal.h | 5 +++++
fs/lustre/llite/xattr.c | 32 +++++++++++++++++++++++++++++++-
3 files changed, 52 insertions(+), 1 deletion(-)
diff --git a/fs/lustre/llite/crypto.c b/fs/lustre/llite/crypto.c
index 5d99037..0fae9a5 100644
--- a/fs/lustre/llite/crypto.c
+++ b/fs/lustre/llite/crypto.c
@@ -32,10 +32,26 @@
static int ll_get_context(struct inode *inode, void *ctx, size_t len)
{
struct dentry *dentry = d_find_any_alias(inode);
+ struct lu_env *env;
+ u16 refcheck;
int rc;
+ env = cl_env_get(&refcheck);
+ if (IS_ERR(env))
+ return PTR_ERR(env);
+
+ /* Set lcc_getencctx=1 to allow this thread to read
+ * LL_XATTR_NAME_ENCRYPTION_CONTEXT xattr, as requested by llcrypt.
+ */
+ ll_cl_add(inode, env, NULL, LCC_RW);
+ ll_env_info(env)->lti_io_ctx.lcc_getencctx = 1;
+
rc = __vfs_getxattr(dentry, inode, LL_XATTR_NAME_ENCRYPTION_CONTEXT,
ctx, len);
+
+ ll_cl_remove(inode, env);
+ cl_env_put(env, &refcheck);
+
if (dentry)
dput(dentry);
diff --git a/fs/lustre/llite/llite_internal.h b/fs/lustre/llite/llite_internal.h
index cfeec14..e0fda00 100644
--- a/fs/lustre/llite/llite_internal.h
+++ b/fs/lustre/llite/llite_internal.h
@@ -1312,6 +1312,11 @@ struct ll_cl_context {
struct cl_io *lcc_io;
struct cl_page *lcc_page;
enum lcc_type lcc_type;
+ /**
+ * Get encryption context operation in progress,
+ * allow getxattr of LL_XATTR_NAME_ENCRYPTION_CONTEXT xattr
+ */
+ unsigned int lcc_getencctx:1;
};
struct ll_thread_info {
diff --git a/fs/lustre/llite/xattr.c b/fs/lustre/llite/xattr.c
index 001c828..59a1400 100644
--- a/fs/lustre/llite/xattr.c
+++ b/fs/lustre/llite/xattr.c
@@ -366,6 +366,21 @@ int ll_xattr_list(struct inode *inode, const char *name, int type, void *buffer,
void *xdata;
int rc;
+ /* Getting LL_XATTR_NAME_ENCRYPTION_CONTEXT xattr is only allowed
+ * when it comes from ll_get_context(), ie when llcrypt needs to
+ * know the encryption context.
+ * Otherwise, any direct reading of this xattr returns -EPERM.
+ */
+ if (type == XATTR_SECURITY_T &&
+ !strcmp(name, LL_XATTR_NAME_ENCRYPTION_CONTEXT)) {
+ struct ll_cl_context *lcc = ll_cl_find(inode);
+
+ if (!lcc || !lcc->lcc_getencctx) {
+ rc = -EPERM;
+ goto out_xattr;
+ }
+ }
+
if (sbi->ll_xattr_cache_enabled && type != XATTR_ACL_ACCESS_T &&
(type != XATTR_SECURITY_T || strcmp(name, "security.selinux"))) {
rc = ll_xattr_cache_get(inode, name, buffer, size, valid);
@@ -632,9 +647,24 @@ ssize_t ll_listxattr(struct dentry *dentry, char *buffer, size_t size)
rem = rc;
while (rem > 0) {
+ bool hide_xattr = false;
+
+ /* Listing xattrs should not expose
+ * LL_XATTR_NAME_ENCRYPTION_CONTEXT xattr, unless it comes
+ * from llcrypt.
+ */
+ if (get_xattr_type(xattr_name)->flags == XATTR_SECURITY_T &&
+ !strcmp(xattr_name, LL_XATTR_NAME_ENCRYPTION_CONTEXT)) {
+ struct ll_cl_context *lcc = ll_cl_find(inode);
+
+ if (!lcc || !lcc->lcc_getencctx)
+ hide_xattr = true;
+ }
+
len = strnlen(xattr_name, rem - 1) + 1;
rem -= len;
- if (!xattr_type_filter(sbi, get_xattr_type(xattr_name))) {
+ if (!xattr_type_filter(sbi, hide_xattr ? NULL :
+ get_xattr_type(xattr_name))) {
/* Skip OK xattr type, leave it in buffer. */
xattr_name += len;
continue;
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 06/20] lustre: brw: log T10 GRD tags during checksum calcs
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (4 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 05/20] lustre: sec: do not expose security.c to listxattr/getxattr James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 07/20] lustre: lov: prefer mirrors on non-rotational OSTs James Simmons
` (13 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Andreas Dilger <adilger@whamcloud.com>
Log the T10 guard tags during checksum calculation on the client and
target to help identify where checksum errors are being introduced.
The added debugging is only active on RPC resend, so will not add
overhead during the normal IO path.
WC-bug-id: https://jira.whamcloud.com/browse/LU-14895
Lustre-commit: c628b1b441d0ee191 ("LU-14895 brw: log T10 GRD tags during checksum calcs")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/44655
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/osc/osc_request.c | 83 ++++++++++++++++++++++++++-------------------
1 file changed, 49 insertions(+), 34 deletions(-)
diff --git a/fs/lustre/osc/osc_request.c b/fs/lustre/osc/osc_request.c
index db73fce..def2ee7 100644
--- a/fs/lustre/osc/osc_request.c
+++ b/fs/lustre/osc/osc_request.c
@@ -1186,7 +1186,7 @@ static int osc_checksum_bulk_t10pi(const char *obd_name, int nob,
size_t pg_count, struct brw_page **pga,
int opc, obd_dif_csum_fn *fn,
int sector_size,
- u32 *check_sum)
+ u32 *check_sum, bool resend)
{
struct ahash_request *hdesc;
/* Used Adler as the default checksum type on top of DIF tags */
@@ -1219,6 +1219,10 @@ static int osc_checksum_bulk_t10pi(const char *obd_name, int nob,
buffer = kmap(__page);
guard_start = (u16 *)buffer;
guard_number = PAGE_SIZE / sizeof(*guard_start);
+ CDEBUG(D_PAGE | (resend ? D_HA : 0),
+ "GRD tags per page=%u, resend=%u, bytes=%u, pages=%zu\n",
+ guard_number, resend, nob, pg_count);
+
while (nob > 0 && pg_count > 0) {
unsigned int count = pga[i]->count > nob ? nob : pga[i]->count;
@@ -1245,6 +1249,12 @@ static int osc_checksum_bulk_t10pi(const char *obd_name, int nob,
guard_number - used_number,
&used, sector_size,
fn);
+ if (unlikely(resend))
+ CDEBUG(D_PAGE | D_HA,
+ "pga[%u]: used %u off %llu+%u gen checksum: %*phN\n",
+ i, used, pga[i]->off & ~PAGE_MASK, count,
+ (int)(used * sizeof(*guard_start)),
+ guard_start + used_number);
if (rc)
break;
@@ -1346,7 +1356,7 @@ static int osc_checksum_bulk_rw(const char *obd_name,
enum cksum_types cksum_type,
int nob, size_t pg_count,
struct brw_page **pga, int opc,
- u32 *check_sum)
+ u32 *check_sum, bool resend)
{
obd_dif_csum_fn *fn = NULL;
int sector_size = 0;
@@ -1356,7 +1366,8 @@ static int osc_checksum_bulk_rw(const char *obd_name,
if (fn)
rc = osc_checksum_bulk_t10pi(obd_name, nob, pg_count, pga,
- opc, fn, sector_size, check_sum);
+ opc, fn, sector_size, check_sum,
+ resend);
else
rc = osc_checksum_bulk(nob, pg_count, pga, opc, cksum_type,
check_sum);
@@ -1727,14 +1738,15 @@ static int osc_brw_prep_request(int cmd, struct client_obd *cli,
rc = osc_checksum_bulk_rw(obd_name, cksum_type,
requested_nob, page_count,
pga, OST_WRITE,
- &body->oa.o_cksum);
+ &body->oa.o_cksum, resend);
if (rc < 0) {
- CDEBUG(D_PAGE, "failed to checksum, rc = %d\n",
+ CDEBUG(D_PAGE, "failed to checksum: rc = %d\n",
rc);
goto out;
}
- CDEBUG(D_PAGE, "checksum at write origin: %x\n",
- body->oa.o_cksum);
+ CDEBUG(D_PAGE | (resend ? D_HA : 0),
+ "checksum at write origin: %x (%x)\n",
+ body->oa.o_cksum, cksum_type);
/* save this in 'oa', too, for later checking */
oa->o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
@@ -1814,6 +1826,7 @@ static void dump_all_bulk_pages(struct obdo *oa, u32 page_count,
pga[0]->off,
pga[page_count - 1]->off + pga[page_count - 1]->count - 1,
client_cksum, server_cksum);
+ CWARN("dumping checksum data to %s\n", dbgcksum_file_name);
filp = filp_open(dbgcksum_file_name,
O_CREAT | O_EXCL | O_WRONLY | O_LARGEFILE, 0600);
if (IS_ERR(filp)) {
@@ -1840,8 +1853,6 @@ static void dump_all_bulk_pages(struct obdo *oa, u32 page_count,
}
len -= rc;
buf += rc;
- CDEBUG(D_INFO, "%s: wrote %d bytes\n",
- dbgcksum_file_name, rc);
}
kunmap(pga[i]->pg);
}
@@ -1850,6 +1861,8 @@ static void dump_all_bulk_pages(struct obdo *oa, u32 page_count,
if (rc)
CERROR("%s: sync returns %d\n", dbgcksum_file_name, rc);
filp_close(filp, NULL);
+
+ libcfs_debug_dumplog();
}
static int check_write_checksum(struct obdo *oa,
@@ -1902,7 +1915,7 @@ static int check_write_checksum(struct obdo *oa,
rc = osc_checksum_bulk_t10pi(obd_name, aa->aa_requested_nob,
aa->aa_page_count, aa->aa_ppga,
OST_WRITE, fn, sector_size,
- &new_cksum);
+ &new_cksum, true);
else
rc = osc_checksum_bulk(aa->aa_requested_nob, aa->aa_page_count,
aa->aa_ppga, OST_WRITE, cksum_type,
@@ -2067,17 +2080,18 @@ static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
if (body->oa.o_valid & OBD_MD_FLCKSUM) {
static int cksum_counter;
u32 server_cksum = body->oa.o_cksum;
+ int nob = rc;
char *via = "";
char *router = "";
enum cksum_types cksum_type;
u32 o_flags = body->oa.o_valid & OBD_MD_FLFLAGS ?
- body->oa.o_flags : 0;
+ body->oa.o_flags : 0;
cksum_type = obd_cksum_type_unpack(o_flags);
- rc = osc_checksum_bulk_rw(obd_name, cksum_type, rc,
+ rc = osc_checksum_bulk_rw(obd_name, cksum_type, nob,
aa->aa_page_count, aa->aa_ppga,
- OST_READ, &client_cksum);
+ OST_READ, &client_cksum, false);
if (rc < 0)
goto out;
@@ -2090,7 +2104,11 @@ static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
if (server_cksum != client_cksum) {
u32 page_count = aa->aa_page_count;
struct ost_body *clbody;
+ u32 client_cksum2;
+ osc_checksum_bulk_rw(obd_name, cksum_type, nob,
+ page_count, aa->aa_ppga,
+ OST_READ, &client_cksum2, true);
clbody = req_capsule_client_get(&req->rq_pill,
&RMF_OST_BODY);
if (cli->cl_checksum_dump)
@@ -2098,26 +2116,23 @@ static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
aa->aa_ppga, server_cksum,
client_cksum);
- LCONSOLE_ERROR_MSG(
- 0x133,
- "%s: BAD READ CHECKSUM: from %s%s%s inode " DFID
- " object " DOSTID
- " extent [%llu-%llu], client %x, server %x, cksum_type %x\n",
- obd_name,
- libcfs_nid2str(peer->nid),
- via, router,
- clbody->oa.o_valid & OBD_MD_FLFID ?
- clbody->oa.o_parent_seq : (u64)0,
- clbody->oa.o_valid & OBD_MD_FLFID ?
- clbody->oa.o_parent_oid : 0,
- clbody->oa.o_valid & OBD_MD_FLFID ?
- clbody->oa.o_parent_ver : 0,
- POSTID(&body->oa.o_oi),
- aa->aa_ppga[0]->off,
- aa->aa_ppga[page_count - 1]->off +
- aa->aa_ppga[page_count - 1]->count - 1,
- client_cksum, server_cksum,
- cksum_type);
+ LCONSOLE_ERROR_MSG(0x133,
+ "%s: BAD READ CHECKSUM: from %s%s%s inode "DFID" object "DOSTID" extent [%llu-%llu], client %x/%x, server %x, cksum_type %x\n",
+ obd_name,
+ libcfs_nid2str(peer->nid),
+ via, router,
+ clbody->oa.o_valid & OBD_MD_FLFID ?
+ clbody->oa.o_parent_seq : (u64)0,
+ clbody->oa.o_valid & OBD_MD_FLFID ?
+ clbody->oa.o_parent_oid : 0,
+ clbody->oa.o_valid & OBD_MD_FLFID ?
+ clbody->oa.o_parent_ver : 0,
+ POSTID(&body->oa.o_oi),
+ aa->aa_ppga[0]->off,
+ aa->aa_ppga[page_count - 1]->off +
+ aa->aa_ppga[page_count - 1]->count - 1,
+ client_cksum, client_cksum2,
+ server_cksum, cksum_type);
cksum_counter = 0;
aa->aa_oa->o_cksum = client_cksum;
rc = -EAGAIN;
@@ -2356,7 +2371,7 @@ static int brw_interpret(const struct lu_env *env,
req->rq_import->imp_obd->obd_name,
POSTID(&aa->aa_oa->o_oi), rc);
} else if (rc == -EINPROGRESS ||
- client_should_resend(aa->aa_resends, aa->aa_cli)) {
+ client_should_resend(aa->aa_resends, aa->aa_cli)) {
rc = osc_brw_redo_request(req, aa, rc);
} else {
CERROR("%s: too many resent retries for object: %llu:%llu, rc = %d.\n",
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 07/20] lustre: lov: prefer mirrors on non-rotational OSTs
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (5 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 06/20] lustre: brw: log T10 GRD tags during checksum calcs James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 08/20] lustre: sec: access to enc file's xattrs James Simmons
` (12 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Alex Zhuravlev <bzzz@whamcloud.com>
consider non-rotational OSTs as preferred unless explicit prefer
flag is set on a mirror.
WC-bug-id: https://jira.whamcloud.com/browse/LU-14996
Lustre-commit: 8507472dd37ebc07 ("LU-14996 lov: prefer mirrors on non-rotational OSTs")
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/44883
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/lov/lov_cl_internal.h | 5 +++--
fs/lustre/lov/lov_object.c | 26 ++++++++++++++++++++++----
2 files changed, 25 insertions(+), 6 deletions(-)
diff --git a/fs/lustre/lov/lov_cl_internal.h b/fs/lustre/lov/lov_cl_internal.h
index 7fcc327..d48e2df3 100644
--- a/fs/lustre/lov/lov_cl_internal.h
+++ b/fs/lustre/lov/lov_cl_internal.h
@@ -225,6 +225,7 @@ struct lov_layout_dom {
struct lov_layout_entry {
u32 lle_type;
unsigned int lle_valid:1;
+ unsigned int lle_preference;
struct lu_extent *lle_extent;
struct lov_stripe_md_entry *lle_lsme;
struct lov_comp_layout_entry_ops *lle_comp_ops;
@@ -236,12 +237,12 @@ struct lov_layout_entry {
struct lov_mirror_entry {
unsigned short lre_mirror_id;
- unsigned short lre_preferred:1,
- lre_stale:1, /* set if any components is stale */
+ unsigned short lre_stale:1, /* set if any components is stale */
lre_valid:1, /* set if at least one of components
* in this mirror is valid
*/
lre_foreign:1; /* set if it is a foreign component */
+ int lre_preference; /* overall preference of this mirror */
unsigned short lre_start; /* index to lo_entries, start index of
* this mirror
diff --git a/fs/lustre/lov/lov_object.c b/fs/lustre/lov/lov_object.c
index 16fed09..ff0f7fa 100644
--- a/fs/lustre/lov/lov_object.c
+++ b/fs/lustre/lov/lov_object.c
@@ -234,6 +234,7 @@ static int lov_init_raid0(const struct lu_env *env, struct lov_device *dev,
struct lov_oinfo *oinfo = lse->lsme_oinfo[i];
int ost_idx = oinfo->loi_ost_idx;
struct cl_device *subdev;
+ struct obd_export *exp;
if (lov_oinfo_is_dummy(oinfo))
continue;
@@ -250,6 +251,13 @@ static int lov_init_raid0(const struct lu_env *env, struct lov_device *dev,
goto out;
}
+ exp = dev->ld_lov->lov_tgts[ost_idx]->ltd_exp;
+ if (likely(exp)) {
+ /* the more fast OSTs the better */
+ if (exp->exp_obd->obd_osfs.os_state & OS_STATFS_NONROT)
+ lle->lle_preference++;
+ }
+
subdev = lovsub2cl_dev(dev->ld_target[ost_idx]);
subconf->u.coc_oinfo = oinfo;
LASSERTF(subdev, "not init ost %d\n", ost_idx);
@@ -621,7 +629,7 @@ static int lov_init_composite(const struct lu_env *env, struct lov_device *dev,
unsigned int mirror_count;
int result = 0;
unsigned int seq;
- int i, j;
+ int i, j, preference;
bool dom_size = 0;
LASSERT(lsm->lsm_entry_count > 0);
@@ -661,6 +669,7 @@ static int lov_init_composite(const struct lu_env *env, struct lov_device *dev,
lle->lle_lsme = lsm->lsm_entries[i];
lle->lle_type = lov_entry_type(lle->lle_lsme);
+ lle->lle_preference = 0;
switch (lle->lle_type) {
case LOV_PATTERN_RAID0:
lle->lle_comp_ops = &raid0_ops;
@@ -722,8 +731,8 @@ static int lov_init_composite(const struct lu_env *env, struct lov_device *dev,
/* entries must be sorted by mirrors */
lre->lre_mirror_id = mirror_id;
lre->lre_start = lre->lre_end = i;
- lre->lre_preferred = !!(lle->lle_lsme->lsme_flags &
- LCME_FL_PREF_RD);
+ lre->lre_preference = lle->lle_lsme->lsme_flags &
+ LCME_FL_PREF_RD ? 1000 : 0;
lre->lre_valid = lle->lle_valid;
lre->lre_stale = !lle->lle_valid;
lre->lre_foreign = lsme_is_foreign(lle->lle_lsme);
@@ -771,6 +780,7 @@ static int lov_init_composite(const struct lu_env *env, struct lov_device *dev,
* so that different clients would use different mirrors for read.
*/
mirror_count = 0;
+ preference = -1;
seq = hash_long((unsigned long)lov, 8);
for (i = 0; i < comp->lo_mirror_count; i++) {
unsigned int idx = (i + seq) % comp->lo_mirror_count;
@@ -784,8 +794,16 @@ static int lov_init_composite(const struct lu_env *env, struct lov_device *dev,
mirror_count++; /* valid mirror */
- if (lre->lre_preferred || comp->lo_preferred_mirror < 0)
+ /* aggregated preference of all involved OSTs */
+ for (j = lre->lre_start; j <= lre->lre_end; j++) {
+ lre->lre_preference +=
+ comp->lo_entries[j].lle_preference;
+ }
+
+ if (lre->lre_preference > preference) {
+ preference = lre->lre_preference;
comp->lo_preferred_mirror = idx;
+ }
}
if (!mirror_count) {
CDEBUG(D_INODE, DFID
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 08/20] lustre: sec: access to enc file's xattrs
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (6 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 07/20] lustre: lov: prefer mirrors on non-rotational OSTs James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 09/20] lustre: update version to 2.14.55 James Simmons
` (11 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Sebastien Buisson <sbuisson@ddn.com>
Encryption context is stored in 'security.c' xattr. This is put in the
xattr cache via ll_xattr_cache_insert() to avoid sending a getxattr
request to the server. But this operation declares the xattr cache for
the inode as 'valid', with two consequences. It prevents any further
filling with other xattrs, and trying to read an xattr value will
directly return -ENODATA, without any attempt to fetch the xattr from
the server.
This is solved by adding a new ll_file_flags 'LLIF_XATTR_CACHE_FILLED'
that tells if the xattr cache for the inode has been filled. This bit
is set only by ll_xattr_cache_refill(), and 'valid' now just means the
xattr cache for the inode has been initialized.
Fixes: 71d77bbe7e ("lustre: sec: atomicity of encryption context getting/setting")
WC-bug-id: https://jira.whamcloud.com/browse/LU-14989
Lustre-commit: 1faf54e8bf19c28a4 ("LU-14989 sec: access to enc file's xattrs")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/44855
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/llite/llite_internal.h | 2 ++
fs/lustre/llite/statahead.c | 24 +++++++++++++++
fs/lustre/llite/xattr_cache.c | 65 +++++++++++++++++++++++++++++-----------
3 files changed, 73 insertions(+), 18 deletions(-)
diff --git a/fs/lustre/llite/llite_internal.h b/fs/lustre/llite/llite_internal.h
index e0fda00..afd5c7a 100644
--- a/fs/lustre/llite/llite_internal.h
+++ b/fs/lustre/llite/llite_internal.h
@@ -109,6 +109,8 @@ enum ll_file_flags {
LLIF_FOREIGN_REMOVABLE = 5,
/* setting encryption context in progress */
LLIF_SET_ENC_CTX = 6,
+ /* Xattr cache is filled */
+ LLIF_XATTR_CACHE_FILLED = 7,
};
/* See comment on trunc_sem_down_read_nowait */
diff --git a/fs/lustre/llite/statahead.c b/fs/lustre/llite/statahead.c
index cb435d5..15b95b7 100644
--- a/fs/lustre/llite/statahead.c
+++ b/fs/lustre/llite/statahead.c
@@ -666,6 +666,30 @@ static void sa_instantiate(struct ll_statahead_info *sai,
if (rc)
goto out;
+ /* If encryption context was returned by MDT, put it in
+ * inode now to save an extra getxattr.
+ */
+ if (body->mbo_valid & OBD_MD_ENCCTX) {
+ void *encctx = req_capsule_server_get(&req->rq_pill,
+ &RMF_FILE_ENCCTX);
+ u32 encctxlen = req_capsule_get_size(&req->rq_pill,
+ &RMF_FILE_ENCCTX,
+ RCL_SERVER);
+
+ if (encctxlen) {
+ CDEBUG(D_SEC,
+ "server returned encryption ctx for "DFID"\n",
+ PFID(ll_inode2fid(child)));
+ rc = ll_xattr_cache_insert(child,
+ LL_XATTR_NAME_ENCRYPTION_CONTEXT,
+ encctx, encctxlen);
+ if (rc)
+ CWARN("%s: cannot set enc ctx for "DFID": rc = %d\n",
+ ll_i2sbi(child)->ll_fsname,
+ PFID(ll_inode2fid(child)), rc);
+ }
+ }
+
CDEBUG(D_READA, "%s: setting %.*s" DFID " l_data to inode %p\n",
ll_i2sbi(dir)->ll_fsname, entry->se_qstr.len,
entry->se_qstr.name, PFID(ll_inode2fid(child)), child);
diff --git a/fs/lustre/llite/xattr_cache.c b/fs/lustre/llite/xattr_cache.c
index fee1cf5..0641f73 100644
--- a/fs/lustre/llite/xattr_cache.c
+++ b/fs/lustre/llite/xattr_cache.c
@@ -109,6 +109,12 @@ static int ll_xattr_cache_add(struct list_head *cache,
struct ll_xattr_entry *xattr;
if (ll_xattr_cache_find(cache, xattr_name, &xattr) == 0) {
+ if (!strcmp(xattr_name, LL_XATTR_NAME_ENCRYPTION_CONTEXT))
+ /* it means enc ctx was already in cache,
+ * ignore error as it cannot be modified
+ */
+ return 0;
+
CDEBUG(D_CACHE, "duplicate xattr: [%s]\n", xattr_name);
return -EPROTO;
}
@@ -211,7 +217,7 @@ static int ll_xattr_cache_list(struct list_head *cache,
}
/**
- * Check if the xattr cache is initialized (filled).
+ * Check if the xattr cache is initialized.
*
* Return: 0 @cache is not initialized
* 1 @cache is initialized
@@ -222,6 +228,17 @@ static int ll_xattr_cache_valid(struct ll_inode_info *lli)
}
/**
+ * Check if the xattr cache is filled.
+ *
+ * \retval 0 @cache is not filled
+ * \retval 1 @cache is filled
+ */
+static int ll_xattr_cache_filled(struct ll_inode_info *lli)
+{
+ return test_bit(LLIF_XATTR_CACHE_FILLED, &lli->lli_flags);
+}
+
+/**
* This finalizes the xattr cache.
*
* Free all xattr memory. @lli is the inode info pointer.
@@ -236,6 +253,7 @@ static int ll_xattr_cache_destroy_locked(struct ll_inode_info *lli)
while (ll_xattr_cache_del(&lli->lli_xattrs, NULL) == 0)
; /* empty loop */
+ clear_bit(LLIF_XATTR_CACHE_FILLED, &lli->lli_flags);
clear_bit(LLIF_XATTR_CACHE, &lli->lli_flags);
return 0;
@@ -259,7 +277,8 @@ int ll_xattr_cache_destroy(struct inode *inode)
* Find or request an LDLM lock with xattr data.
* Since LDLM does not provide API for atomic match_or_enqueue,
* the function handles it with a separate enq lock.
- * If successful, the function exits with the list lock held.
+ * If successful, the function exits with a write lock held
+ * on lli_xattrs_list_rwsem.
*
* Return: 0 no error occurred
* -ENOMEM not enough memory
@@ -280,7 +299,7 @@ static int ll_xattr_find_get_lock(struct inode *inode,
/* inode may have been shrunk and recreated, so data is gone, match lock
* only when data exists.
*/
- if (ll_xattr_cache_valid(lli)) {
+ if (ll_xattr_cache_filled(lli)) {
/* Try matching first. */
mode = ll_take_md_lock(inode, MDS_INODELOCK_XATTR, &lockh, 0,
LCK_PR);
@@ -324,7 +343,9 @@ static int ll_xattr_find_get_lock(struct inode *inode,
/**
* Refill the xattr cache.
*
- * Fetch and cache the whole of xattrs for @inode, acquiring a read lock.
+ * Fetch and cache the whole of xattrs for @inode, thanks to the write lock
+ * on lli_xattrs_list_rwsem obtained from ll_xattr_find_get_lock().
+ * If successful, this write lock is kept.
*
* Return: 0 no error occurred
* -EPROTO network protocol error
@@ -346,7 +367,7 @@ static int ll_xattr_cache_refill(struct inode *inode)
goto err_req;
/* Do we have the data at this point? */
- if (ll_xattr_cache_valid(lli)) {
+ if (ll_xattr_cache_filled(lli)) {
ll_stats_ops_tally(sbi, LPROC_LL_GETXATTR_HITS, 1);
ll_intent_drop_lock(&oit);
rc = 0;
@@ -385,7 +406,8 @@ static int ll_xattr_cache_refill(struct inode *inode)
CDEBUG(D_CACHE, "caching: xdata=%p xtail=%p\n", xdata, xtail);
- ll_xattr_cache_init(lli);
+ if (!ll_xattr_cache_valid(lli))
+ ll_xattr_cache_init(lli);
for (i = 0; i < body->mbo_max_mdsize; i++) {
CDEBUG(D_CACHE, "caching [%s]=%.*s\n", xdata, *xsizes, xval);
@@ -422,6 +444,8 @@ static int ll_xattr_cache_refill(struct inode *inode)
if (xdata != xtail || xval != xvtail)
CERROR("a hole in xattr data\n");
+ else
+ set_bit(LLIF_XATTR_CACHE_FILLED, &lli->lli_flags);
ll_set_lock_data(sbi->ll_md_exp, inode, &oit, NULL);
ll_intent_drop_lock(&oit);
@@ -466,16 +490,29 @@ int ll_xattr_cache_get(struct inode *inode, const char *name, char *buffer,
LASSERT(!!(valid & OBD_MD_FLXATTR) ^ !!(valid & OBD_MD_FLXATTRLS));
down_read(&lli->lli_xattrs_list_rwsem);
- if (!ll_xattr_cache_valid(lli)) {
+ /* For performance reasons, we do not want to refill complete xattr
+ * cache if we are just interested in encryption context.
+ */
+ if ((valid & OBD_MD_FLXATTRLS ||
+ strcmp(name, LL_XATTR_NAME_ENCRYPTION_CONTEXT) != 0) &&
+ !ll_xattr_cache_valid(lli)) {
up_read(&lli->lli_xattrs_list_rwsem);
rc = ll_xattr_cache_refill(inode);
if (rc)
return rc;
+ /* Turn the write lock obtained in ll_xattr_cache_refill()
+ * into a read lock.
+ */
downgrade_write(&lli->lli_xattrs_list_rwsem);
} else {
ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_GETXATTR_HITS, 1);
}
+ if (!ll_xattr_cache_valid(lli)) {
+ rc = -ENODATA;
+ goto out;
+ }
+
if (valid & OBD_MD_FLXATTR) {
struct ll_xattr_entry *xattr;
@@ -521,18 +558,10 @@ int ll_xattr_cache_insert(struct inode *inode,
struct ll_inode_info *lli = ll_i2info(inode);
int rc;
- down_read(&lli->lli_xattrs_list_rwsem);
+ down_write(&lli->lli_xattrs_list_rwsem);
if (!ll_xattr_cache_valid(lli))
ll_xattr_cache_init(lli);
- rc = ll_xattr_cache_add(&lli->lli_xattrs, name, buffer,
- size);
- up_read(&lli->lli_xattrs_list_rwsem);
-
- if (rc == -EPROTO &&
- strcmp(name, LL_XATTR_NAME_ENCRYPTION_CONTEXT) == 0)
- /* it means enc ctx was already in cache,
- * ignore error as it cannot be modified
- */
- rc = 0;
+ rc = ll_xattr_cache_add(&lli->lli_xattrs, name, buffer, size);
+ up_write(&lli->lli_xattrs_list_rwsem);
return rc;
}
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 09/20] lustre: update version to 2.14.55
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (7 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 08/20] lustre: sec: access to enc file's xattrs James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 10/20] lustre: osc: Do not attempt sending empty pages James Simmons
` (10 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Oleg Drokin <green@whamcloud.com>
New tag 2.14.55
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
include/uapi/linux/lustre/lustre_ver.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/uapi/linux/lustre/lustre_ver.h b/include/uapi/linux/lustre/lustre_ver.h
index 90254ed..d4ca95e 100644
--- a/include/uapi/linux/lustre/lustre_ver.h
+++ b/include/uapi/linux/lustre/lustre_ver.h
@@ -3,9 +3,9 @@
#define LUSTRE_MAJOR 2
#define LUSTRE_MINOR 14
-#define LUSTRE_PATCH 54
+#define LUSTRE_PATCH 55
#define LUSTRE_FIX 0
-#define LUSTRE_VERSION_STRING "2.14.54"
+#define LUSTRE_VERSION_STRING "2.14.55"
#define OBD_OCD_VERSION(major, minor, patch, fix) \
(((major) << 24) + ((minor) << 16) + ((patch) << 8) + (fix))
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 10/20] lustre: osc: Do not attempt sending empty pages
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (8 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 09/20] lustre: update version to 2.14.55 James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 11/20] lustre: ptlrpc: handle reply and resend reorder James Simmons
` (9 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Oleg Drokin <green@whamcloud.com>
Do not crash if trying to send a lock-prolonging emtpy read
to an old server, if the server does not support short reads.
Otherwise the client crashes when access the NULL page.
Fixes: 1febc3615e2b ("lustre: osc: Notify server if cache discard takes a long time")
WC-bug-id: https://jira.whamcloud.com/browse/LU-14711
Lustre-commit: 1a409a3e6a7468597 ("LU-14711 osc: Do not attempt sending empty pages")
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/44654
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/osc/osc_request.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/fs/lustre/osc/osc_request.c b/fs/lustre/osc/osc_request.c
index def2ee7..e5b7453 100644
--- a/fs/lustre/osc/osc_request.c
+++ b/fs/lustre/osc/osc_request.c
@@ -1567,6 +1567,12 @@ static int osc_brw_prep_request(int cmd, struct client_obd *cli,
!imp_connect_shortio(cli->cl_import))
short_io_size = 0;
+ /* If this is an empty RPC to old server, just ignore it */
+ if (!short_io_size && !pga[0]->pg) {
+ ptlrpc_request_free(req);
+ return -ENODATA;
+ }
+
req_capsule_set_size(pill, &RMF_SHORT_IO, RCL_CLIENT,
opc == OST_READ ? 0 : short_io_size);
if (opc == OST_READ)
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 11/20] lustre: ptlrpc: handle reply and resend reorder
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (9 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 10/20] lustre: osc: Do not attempt sending empty pages James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 12/20] lustre: ptlrpc: use wait_woken() in ptlrpcd() James Simmons
` (8 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown
Cc: Alexey Lyashkov, Alexander Boyko, Lustre Development List
From: Alexander Boyko <alexander.boyko@hpe.com>
ptlrpc can't detect a bulk transfer timeout
if rpc and bulk are reordered on router.
We should fail a bulk for situations where bulk is not
completed (after bulk timeout LNET_EVENT_UNLINK is set).
HPE-bug-id: LUS-7445, LUS-7569
WC-bug-id: https://jira.whamcloud.com/browse/LU-12567
Lustre-commit: f7f31f8f969f410cc ("LU-12567 ptlrpc: handle reply and resend reorder")
Signed-off-by: Alexander Boyko <alexander.boyko@hpe.com>
Signed-off-by: Alexey Lyashkov <alexey.lyashkov@hpe.com>
Reviewed-on: https://review.whamcloud.com/35571
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/ptlrpc/client.c | 5 ++++-
fs/lustre/ptlrpc/events.c | 3 +--
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/fs/lustre/ptlrpc/client.c b/fs/lustre/ptlrpc/client.c
index 83d269c..e800000 100644
--- a/fs/lustre/ptlrpc/client.c
+++ b/fs/lustre/ptlrpc/client.c
@@ -2075,7 +2075,10 @@ int ptlrpc_check_set(const struct lu_env *env, struct ptlrpc_request_set *set)
* was good after getting the REPLY for her GET or
* the ACK for her PUT.
*/
- DEBUG_REQ(D_ERROR, req, "bulk transfer failed");
+ DEBUG_REQ(D_ERROR, req, "bulk transfer failed %d/%d/%d",
+ req->rq_status,
+ req->rq_bulk->bd_nob,
+ req->rq_bulk->bd_nob_transferred);
req->rq_status = -EIO;
}
diff --git a/fs/lustre/ptlrpc/events.c b/fs/lustre/ptlrpc/events.c
index c81181d..559d811 100644
--- a/fs/lustre/ptlrpc/events.c
+++ b/fs/lustre/ptlrpc/events.c
@@ -219,10 +219,9 @@ void client_bulk_callback(struct lnet_event *ev)
spin_lock(&req->rq_lock);
req->rq_net_err = 1;
spin_unlock(&req->rq_lock);
+ desc->bd_failure = 1;
}
- if (ev->status != 0)
- desc->bd_failure = 1;
/* NB don't unlock till after wakeup; desc can disappear under us
* otherwise
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 12/20] lustre: ptlrpc: use wait_woken() in ptlrpcd()
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (10 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 11/20] lustre: ptlrpc: handle reply and resend reorder James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 13/20] lustre: quota: fix quota with root squash enabled James Simmons
` (7 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Mr NeilBrown <neilb@suse.de>
Using wait_event() to wait for ptlrpcd_check() to succeed is
problematic. ptlrpcd_check() is complex and can wait for other
events. This nested waiting can behave differently to expectation and
generates a warning
do not call blocking ops when !TASK_RUNNING
This happens because the task state is set to TASK_IDLE before
ptlrpcd_check() is calls.
A better approach (introduce for precisely this use-case) is to use
wait_woken() and woken_wake_function().
When a wake_up is requested on the waitq, woken_wake_function() sets a
flag to record the wakeup. wait_woken() will wait until this flag is
set. This way, the task state doesn't need to be set until after
ptlrpcd_check() has completed.
WC-bug-id: https://jira.whamcloud.com/browse/LU-12362
Lustre-commit: 885b494632ca16d95 ("LU-12362 ptlrpc: use wait_woken() in ptlrpcd()")
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/45069
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: xinliang <xinliang.liu@linaro.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/ptlrpc/ptlrpcd.c | 23 ++++++++++++++++++-----
1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/fs/lustre/ptlrpc/ptlrpcd.c b/fs/lustre/ptlrpc/ptlrpcd.c
index ed3f0e1..9cd9d39 100644
--- a/fs/lustre/ptlrpc/ptlrpcd.c
+++ b/fs/lustre/ptlrpc/ptlrpcd.c
@@ -435,18 +435,31 @@ static int ptlrpcd(void *arg)
* new_req_list and ptlrpcd_check() moves them into the set.
*/
do {
+ DEFINE_WAIT_FUNC(wait, woken_wake_function);
time64_t timeout;
timeout = ptlrpc_set_next_timeout(set);
lu_context_enter(&env.le_ctx);
lu_context_enter(env.le_ses);
- /* If timeout==0, wait indefinitely */
- if (wait_event_idle_timeout(
- set->set_waitq,
- ptlrpcd_check(&env, pc),
- timeout ? (timeout * HZ) : MAX_SCHEDULE_TIMEOUT) == 0)
+
+ add_wait_queue(&set->set_waitq, &wait);
+ while (!ptlrpcd_check(&env, pc)) {
+ int ret;
+
+ if (timeout == 0)
+ ret = wait_woken(&wait, TASK_IDLE,
+ MAX_SCHEDULE_TIMEOUT);
+ else
+ ret = wait_woken(&wait, TASK_IDLE,
+ HZ * timeout);
+ if (ret != 0)
+ continue;
+ /* Timed out */
ptlrpc_expired_set(set);
+ break;
+ }
+ remove_wait_queue(&set->set_waitq, &wait);
lu_context_exit(&env.le_ctx);
lu_context_exit(env.le_ses);
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 13/20] lustre: quota: fix quota with root squash enabled
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (11 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 12/20] lustre: ptlrpc: use wait_woken() in ptlrpcd() James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 14/20] lustre: llite: harden ll_sbi ll_flags James Simmons
` (6 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown
Cc: Wang Shilong, Wang Shilong, Lustre Development List
From: Wang Shilong <wshilong@ddn.com>
This patch tries to fix several problems:
1. OSD will ignore quota if IO comes from client
cache or root, however since following change:
LU-12687 osc: consume grants for direct I/O
DIO now consumes grant too, following check for
sync IO is wrong now:
(lnb[i].lnb_flags & (OBD_BRW_FROM_GRANT | OBD_BRW_SYNC))
== OBD_BRW_FROM_GRANT)
This was originally added to support 1.8 client, it is
going to be 2.15 now, so let's remove this broken check.
2. Server side will clear OBD_BRW_NOQUOTA if root squash
is enabled, this will revert fixes from:
"LU-13228 clio: mmap write when overquota"
We need to separate @ci_noquota and @oi_cap_sys_resource cases,
introduce a new flag OBD_BRW_SYS_RESOURCE, and extend test_75
to cover this case.
3. LU-14739 missed case that DoM quota should be considered
as well.
4. If EDQUOT is returned for root, we check the new root squash
flag OBD_FL_ROOT_SQUASH from server side. If this flag is not set,
we bypass quota for root, otherwise all root writes become sync
writes.
5. Fix a leftover problem with LU-9671 for DOM
Fixes: cd633cfc960b63 ("lustre: quota: nodemap squashed root cannot bypass quota")
WC-bug-id: https://jira.whamcloud.com/browse/LU-14739
Lustre-commit: bbfdc7c1670c9274 ("LU-14739 quota: fix quota with root squash enabled")
Signed-off-by: Wang Shilong <wshilong@ddn.com>
Signed-off-by: Wang Shilong <wangshilong1991@gmail.com>
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/44347
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/include/obd.h | 2 ++
fs/lustre/osc/osc_cache.c | 23 ++++++++++++++++++++---
fs/lustre/osc/osc_page.c | 4 ++--
fs/lustre/osc/osc_quota.c | 1 +
fs/lustre/osc/osc_request.c | 4 +++-
fs/lustre/ptlrpc/wiretest.c | 2 ++
include/uapi/linux/lustre/lustre_idl.h | 4 +++-
7 files changed, 33 insertions(+), 7 deletions(-)
diff --git a/fs/lustre/include/obd.h b/fs/lustre/include/obd.h
index 7642973..b3ad511 100644
--- a/fs/lustre/include/obd.h
+++ b/fs/lustre/include/obd.h
@@ -233,6 +233,8 @@ struct client_obd {
struct list_head cl_grant_chain;
time64_t cl_grant_shrink_interval; /* seconds */
+ int cl_root_squash; /* if root squash enabled*/
+
/* A chunk is an optimal size used by osc_extent to determine
* the extent size. A chunk is max(PAGE_SIZE, OST block size)
*/
diff --git a/fs/lustre/osc/osc_cache.c b/fs/lustre/osc/osc_cache.c
index 69cf9ba..1211438 100644
--- a/fs/lustre/osc/osc_cache.c
+++ b/fs/lustre/osc/osc_cache.c
@@ -2374,11 +2374,16 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
/* Set the OBD_BRW_SRVLOCK before the page is queued. */
brw_flags |= ops->ops_srvlock ? OBD_BRW_SRVLOCK : 0;
- if (oio->oi_cap_sys_resource || io->ci_noquota) {
+ if (io->ci_noquota) {
brw_flags |= OBD_BRW_NOQUOTA;
cmd |= OBD_BRW_NOQUOTA;
}
+ if (oio->oi_cap_sys_resource) {
+ brw_flags |= OBD_BRW_SYS_RESOURCE;
+ cmd |= OBD_BRW_SYS_RESOURCE;
+ }
+
/* check if the file's owner/group is over quota */
if (!io->ci_noquota) {
struct cl_object *obj;
@@ -2395,8 +2400,20 @@ int osc_queue_async_io(const struct lu_env *env, struct cl_io *io,
qid[USRQUOTA] = attr->cat_uid;
qid[GRPQUOTA] = attr->cat_gid;
qid[PRJQUOTA] = attr->cat_projid;
- if (rc == 0)
- rc = osc_quota_chkdq(cli, qid);
+ /*
+ * if EDQUOT returned for root, we double check
+ * if root squash enabled or not updated from server side.
+ * without root squash, we should bypass quota for root.
+ */
+ if (rc == 0 && osc_quota_chkdq(cli, qid) == -EDQUOT) {
+ if (oio->oi_cap_sys_resource &&
+ !cli->cl_root_squash) {
+ io->ci_noquota = 1;
+ rc = 0;
+ } else {
+ rc = -EDQUOT;
+ }
+ }
if (rc)
return rc;
}
diff --git a/fs/lustre/osc/osc_page.c b/fs/lustre/osc/osc_page.c
index 8aa21ee..d471df2 100644
--- a/fs/lustre/osc/osc_page.c
+++ b/fs/lustre/osc/osc_page.c
@@ -314,8 +314,8 @@ void osc_page_submit(const struct lu_env *env, struct osc_page *opg,
oap->oap_brw_flags = OBD_BRW_SYNC | brw_flags;
if (oio->oi_cap_sys_resource) {
- oap->oap_brw_flags |= OBD_BRW_NOQUOTA;
- oap->oap_cmd |= OBD_BRW_NOQUOTA;
+ oap->oap_brw_flags |= OBD_BRW_SYS_RESOURCE;
+ oap->oap_cmd |= OBD_BRW_SYS_RESOURCE;
}
opg->ops_submit_time = submit_time;
diff --git a/fs/lustre/osc/osc_quota.c b/fs/lustre/osc/osc_quota.c
index 8ff803c..708ad3c 100644
--- a/fs/lustre/osc/osc_quota.c
+++ b/fs/lustre/osc/osc_quota.c
@@ -119,6 +119,7 @@ int osc_quota_setdq(struct client_obd *cli, u64 xid, const unsigned int qid[],
return 0;
mutex_lock(&cli->cl_quota_mutex);
+ cli->cl_root_squash = !!(flags & OBD_FL_ROOT_SQUASH);
/* still mark the quots is running out for the old request, because it
* could be processed after the new request at OST, the side effect is
* the following request will be processed synchronously, but it will
diff --git a/fs/lustre/osc/osc_request.c b/fs/lustre/osc/osc_request.c
index e5b7453..22b7e5e 100644
--- a/fs/lustre/osc/osc_request.c
+++ b/fs/lustre/osc/osc_request.c
@@ -1167,7 +1167,8 @@ static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
if (p1->flag != p2->flag) {
unsigned int mask = ~(OBD_BRW_FROM_GRANT | OBD_BRW_NOCACHE |
OBD_BRW_SYNC | OBD_BRW_ASYNC |
- OBD_BRW_NOQUOTA | OBD_BRW_SOFT_SYNC);
+ OBD_BRW_NOQUOTA | OBD_BRW_SOFT_SYNC |
+ OBD_BRW_SYS_RESOURCE);
/* warn if we try to combine flags that we don't know to be
* safe to combine
@@ -3548,6 +3549,7 @@ int osc_setup_common(struct obd_device *obd, struct lustre_cfg *lcfg)
goto out_ptlrpcd_work;
cli->cl_grant_shrink_interval = GRANT_SHRINK_INTERVAL;
+ cli->cl_root_squash = 0;
osc_update_next_shrink(cli);
return 0;
diff --git a/fs/lustre/ptlrpc/wiretest.c b/fs/lustre/ptlrpc/wiretest.c
index bf09341..a381af4 100644
--- a/fs/lustre/ptlrpc/wiretest.c
+++ b/fs/lustre/ptlrpc/wiretest.c
@@ -2058,6 +2058,8 @@ void lustre_assert_wire_constants(void)
OBD_BRW_OVER_PRJQUOTA);
LASSERTF(OBD_BRW_RDMA_ONLY == 0x20000, "found 0x%.8x\n",
OBD_BRW_RDMA_ONLY);
+ LASSERTF(OBD_BRW_SYS_RESOURCE == 0x40000, "found 0x%.8x\n",
+ OBD_BRW_SYS_RESOURCE);
/* Checks for struct ost_body */
LASSERTF((int)sizeof(struct ost_body) == 208, "found %lld\n",
diff --git a/include/uapi/linux/lustre/lustre_idl.h b/include/uapi/linux/lustre/lustre_idl.h
index 7d92264..ec25140 100644
--- a/include/uapi/linux/lustre/lustre_idl.h
+++ b/include/uapi/linux/lustre/lustre_idl.h
@@ -977,6 +977,7 @@ enum obdo_flags {
OBD_FL_NOSPC_BLK = 0x00100000, /* no more block space on OST */
OBD_FL_FLUSH = 0x00200000, /* flush pages on the OST */
OBD_FL_SHORT_IO = 0x00400000, /* short io request */
+ OBD_FL_ROOT_SQUASH = 0x00800000, /* root squash */
/* OBD_FL_LOCAL_MASK = 0xF0000000, was local-only flags until 2.10 */
/*
@@ -1249,7 +1250,7 @@ struct hsm_state_set {
#define OBD_BRW_FROM_GRANT 0x20 /* the osc manages this under llite */
#define OBD_BRW_GRANTED 0x40 /* the ost manages this */
#define OBD_BRW_NOCACHE 0x80 /* this page is a part of non-cached IO */
-#define OBD_BRW_NOQUOTA 0x100
+#define OBD_BRW_NOQUOTA 0x100 /* do not enforce quota */
#define OBD_BRW_SRVLOCK 0x200 /* Client holds no lock over this page */
#define OBD_BRW_ASYNC 0x400 /* Server may delay commit to disk */
#define OBD_BRW_MEMALLOC 0x800 /* Client runs in the "kswapd" context */
@@ -1262,6 +1263,7 @@ struct hsm_state_set {
*/
#define OBD_BRW_OVER_PRJQUOTA 0x8000 /* Running out of project quota */
#define OBD_BRW_RDMA_ONLY 0x20000 /* RPC contains RDMA-only pages*/
+#define OBD_BRW_SYS_RESOURCE 0x40000 /* page has CAP_SYS_RESOURCE */
#define OBD_MAX_GRANT 0x7fffffffUL /* Max grant allowed to one client: 2 GiB */
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 14/20] lustre: llite: harden ll_sbi ll_flags
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (12 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 13/20] lustre: quota: fix quota with root squash enabled James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 15/20] lustre: osc: use original cli for osc_lru_reclaim for debug msg James Simmons
` (5 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
For most file systems mount flags are straight forward but this
is not the case for Lustre. We have to consider if the server
backend supports a mount option. Additionally its possible to
disable or enable a feature using sysfs during run time. Some
features can't be managed with a mount option but still can
be managed with sysfs or based on what is enabled on the server
node. All these states are reported together in the debugfs
file sbi_flags. The mount specific options are reported in
the super block show_option ll_show_option().
With all this complexity it is easy for it to get out of sync
and report incorrect things. We consolidate this handling by
moving to using match_table_t that is used by various Linux
file system to parse options. LL_SBI_FLAGS is replaced by our
match_table_t, ll_sbi_flags_name, that can be used for mount
options as well as reporting the sbi_flags in debugfs. We take
advantage of the fact that mount option parse will stop at the
first NULL in ll_sbi_flags_name and after that NULL list the
other features flags that are managed with other methods besides
mount options.
The next change is the move of ll_flags to a bitmap which gives
us two advantages. The first is that we can support more than
32 flags in the future. Second is no need to use bit shifting
math since we can use th enum LL_SBI_* values directly with
clear_bit() / set_bit() / test_bit(). Allow these changes should
miminize future problems with keeping all these states in sync.
WC-bug-id: https://jira.whamcloud.com/browse/LU-12262
Lustre-commit: 47e6f6abdacd6a3c ("LU-12262 llite: harden ll_sbi ll_flags")
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44541
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
---
fs/lustre/llite/crypto.c | 15 +-
fs/lustre/llite/dir.c | 25 +-
fs/lustre/llite/file.c | 12 +-
fs/lustre/llite/llite_foreign_symlink.c | 12 +-
fs/lustre/llite/llite_internal.h | 125 +++-----
fs/lustre/llite/llite_lib.c | 490 ++++++++++++++++----------------
fs/lustre/llite/lproc_llite.c | 78 ++---
fs/lustre/llite/namei.c | 13 +-
fs/lustre/llite/statahead.c | 3 +-
fs/lustre/llite/xattr.c | 17 +-
10 files changed, 359 insertions(+), 431 deletions(-)
diff --git a/fs/lustre/llite/crypto.c b/fs/lustre/llite/crypto.c
index 0fae9a5..0388e360 100644
--- a/fs/lustre/llite/crypto.c
+++ b/fs/lustre/llite/crypto.c
@@ -133,7 +133,7 @@ static int ll_set_context(struct inode *inode, const void *ctx, size_t len,
bool ll_sbi_has_test_dummy_encryption(struct ll_sb_info *sbi)
{
- return unlikely(sbi->ll_flags & LL_SBI_TEST_DUMMY_ENCRYPTION);
+ return unlikely(test_bit(LL_SBI_TEST_DUMMY_ENCRYPTION, sbi->ll_flags));
}
static bool ll_dummy_context(struct inode *inode)
@@ -145,16 +145,17 @@ static bool ll_dummy_context(struct inode *inode)
bool ll_sbi_has_encrypt(struct ll_sb_info *sbi)
{
- return sbi->ll_flags & LL_SBI_ENCRYPT;
+ return test_bit(LL_SBI_ENCRYPT, sbi->ll_flags);
}
void ll_sbi_set_encrypt(struct ll_sb_info *sbi, bool set)
{
- if (set)
- sbi->ll_flags |= LL_SBI_ENCRYPT;
- else
- sbi->ll_flags &=
- ~(LL_SBI_ENCRYPT | LL_SBI_TEST_DUMMY_ENCRYPTION);
+ if (set) {
+ set_bit(LL_SBI_ENCRYPT, sbi->ll_flags);
+ } else {
+ clear_bit(LL_SBI_ENCRYPT, sbi->ll_flags);
+ clear_bit(LL_SBI_TEST_DUMMY_ENCRYPTION, sbi->ll_flags);
+ }
}
static bool ll_empty_dir(struct inode *inode)
diff --git a/fs/lustre/llite/dir.c b/fs/lustre/llite/dir.c
index b7dd2aa..ee49c90 100644
--- a/fs/lustre/llite/dir.c
+++ b/fs/lustre/llite/dir.c
@@ -182,7 +182,7 @@ int ll_dir_read(struct inode *inode, u64 *ppos, struct md_op_data *op_data,
struct ll_sb_info *sbi = ll_i2sbi(inode);
u64 pos = *ppos;
bool is_api32 = ll_need_32bit_api(sbi);
- bool is_hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH;
+ bool is_hash64 = test_bit(LL_SBI_64BIT_HASH, sbi->ll_flags);
struct fscrypt_str lltr = FSTR_INIT(NULL, 0);
struct page *page;
bool done = false;
@@ -300,7 +300,7 @@ static int ll_readdir(struct file *filp, struct dir_context *ctx)
struct ll_file_data *lfd = filp->private_data;
struct ll_sb_info *sbi = ll_i2sbi(inode);
u64 pos = lfd ? lfd->lfd_pos : 0;
- int hash64 = sbi->ll_flags & LL_SBI_64BIT_HASH;
+ bool hash64 = test_bit(LL_SBI_64BIT_HASH, sbi->ll_flags);
bool api32 = ll_need_32bit_api(sbi);
struct md_op_data *op_data;
struct lu_fid pfid = { 0 };
@@ -495,7 +495,7 @@ static int ll_dir_setdirstripe(struct dentry *dparent, struct lmv_user_md *lump,
encrypt = true;
}
- if (sbi->ll_flags & LL_SBI_FILE_SECCTX) {
+ if (test_bit(LL_SBI_FILE_SECCTX, sbi->ll_flags)) {
/*
* selinux_dentry_init_security() uses dentry->d_parent and name
* to determine the security context for the file. So our fake
@@ -534,7 +534,7 @@ static int ll_dir_setdirstripe(struct dentry *dparent, struct lmv_user_md *lump,
dentry.d_inode = inode;
- if (sbi->ll_flags & LL_SBI_FILE_SECCTX) {
+ if (test_bit(LL_SBI_FILE_SECCTX, sbi->ll_flags)) {
/* no need to protect selinux_inode_setsecurity() by
* inode_lock. Taking it would lead to a client deadlock
* LU-13617
@@ -1270,13 +1270,14 @@ int quotactl_ioctl(struct super_block *sb, struct if_quotactl *qctl)
int ll_rmfid(struct file *file, void __user *arg)
{
const struct fid_array __user *ufa = arg;
+ struct inode *inode = file_inode(file);
struct fid_array *lfa = NULL;
size_t size;
unsigned int nr;
int i, rc, *rcs = NULL;
if (!capable(CAP_DAC_READ_SEARCH) &&
- !(ll_i2sbi(file_inode(file))->ll_flags & LL_SBI_USER_FID2PATH))
+ !test_bit(LL_SBI_USER_FID2PATH, ll_i2sbi(inode)->ll_flags))
return -EPERM;
/* Only need to get the buflen */
if (get_user(nr, &ufa->fa_nr))
@@ -1719,6 +1720,7 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
u32 __user *lmmsizep = NULL;
struct lu_fid __user *fidp = NULL;
int lmmsize;
+ bool api32;
if (cmd == IOC_MDC_GETFILEINFO_V1 ||
cmd == IOC_MDC_GETFILEINFO_V2 ||
@@ -1791,6 +1793,7 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
}
rc = -EOVERFLOW;
}
+ api32 = test_bit(LL_SBI_32BIT_API, sbi->ll_flags);
if (cmd == IOC_MDC_GETFILEINFO_V1 ||
cmd == LL_IOC_MDC_GETINFO_V1) {
@@ -1808,9 +1811,7 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
st.st_atime = body->mbo_atime;
st.st_mtime = body->mbo_mtime;
st.st_ctime = body->mbo_ctime;
- st.st_ino = cl_fid_build_ino(&body->mbo_fid1,
- sbi->ll_flags &
- LL_SBI_32BIT_API);
+ st.st_ino = cl_fid_build_ino(&body->mbo_fid1, api32);
if (copy_to_user(statp, &st, sizeof(st))) {
rc = -EFAULT;
@@ -1827,8 +1828,7 @@ static long ll_dir_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
stx.stx_gid = body->mbo_gid;
stx.stx_mode = body->mbo_mode;
stx.stx_ino = cl_fid_build_ino(&body->mbo_fid1,
- sbi->ll_flags &
- LL_SBI_32BIT_API);
+ api32);
stx.stx_size = body->mbo_size;
stx.stx_blocks = body->mbo_blocks;
stx.stx_atime.tv_sec = body->mbo_atime;
@@ -2252,10 +2252,13 @@ static loff_t ll_dir_seek(struct file *file, loff_t offset, int origin)
((api32 && offset <= LL_DIR_END_OFF_32BIT) ||
(!api32 && offset <= LL_DIR_END_OFF))) {
if (offset != file->f_pos) {
+ bool hash64;
+
+ hash64 = test_bit(LL_SBI_64BIT_HASH, sbi->ll_flags);
if ((api32 && offset == LL_DIR_END_OFF_32BIT) ||
(!api32 && offset == LL_DIR_END_OFF))
fd->lfd_pos = MDS_DIR_END_OFF;
- else if (api32 && sbi->ll_flags & LL_SBI_64BIT_HASH)
+ else if (api32 && hash64)
fd->lfd_pos = offset << 32;
else
fd->lfd_pos = offset;
diff --git a/fs/lustre/llite/file.c b/fs/lustre/llite/file.c
index ad1c07e..1e4ff49 100644
--- a/fs/lustre/llite/file.c
+++ b/fs/lustre/llite/file.c
@@ -2625,7 +2625,7 @@ int ll_fid2path(struct inode *inode, void __user *arg)
int rc;
if (!capable(CAP_DAC_READ_SEARCH) &&
- !(ll_i2sbi(inode)->ll_flags & LL_SBI_USER_FID2PATH))
+ !test_bit(LL_SBI_USER_FID2PATH, ll_i2sbi(inode)->ll_flags))
return -EPERM;
/* Only need to get the buflen */
@@ -5393,9 +5393,8 @@ int ll_inode_permission(struct inode *inode, int mask)
squash = &sbi->ll_squash;
if (unlikely(squash->rsi_uid &&
uid_eq(current_fsuid(), GLOBAL_ROOT_UID) &&
- !(sbi->ll_flags & LL_SBI_NOROOTSQUASH))) {
+ !test_bit(LL_SBI_NOROOTSQUASH, sbi->ll_flags)))
squash_id = true;
- }
if (squash_id) {
CDEBUG(D_OTHER, "squash creds (%d:%d)=>(%d:%d)\n",
@@ -5494,9 +5493,9 @@ const struct file_operations *ll_select_file_operations(struct ll_sb_info *sbi)
{
const struct file_operations *fops = &ll_file_operations_noflock;
- if (sbi->ll_flags & LL_SBI_FLOCK)
+ if (test_bit(LL_SBI_FLOCK, sbi->ll_flags))
fops = &ll_file_operations_flock;
- else if (sbi->ll_flags & LL_SBI_LOCALFLOCK)
+ else if (test_bit(LL_SBI_LOCALFLOCK, sbi->ll_flags))
fops = &ll_file_operations;
return fops;
@@ -5787,7 +5786,8 @@ int ll_layout_refresh(struct inode *inode, u32 *gen)
int rc;
*gen = ll_layout_version_get(lli);
- if (!(sbi->ll_flags & LL_SBI_LAYOUT_LOCK) || *gen != CL_LAYOUT_GEN_NONE)
+ if (!test_bit(LL_SBI_LAYOUT_LOCK, sbi->ll_flags) ||
+ *gen != CL_LAYOUT_GEN_NONE)
return 0;
/* sanity checks */
diff --git a/fs/lustre/llite/llite_foreign_symlink.c b/fs/lustre/llite/llite_foreign_symlink.c
index 7ba33f4..bfade93 100644
--- a/fs/lustre/llite/llite_foreign_symlink.c
+++ b/fs/lustre/llite/llite_foreign_symlink.c
@@ -205,7 +205,7 @@ static int ll_foreign_symlink_parse(struct ll_sb_info *sbi,
* of foreign LOV is relative path of faked symlink destination,
* to be completed by prefix
*/
- if (!(sbi->ll_flags & LL_SBI_FOREIGN_SYMLINK_UPCALL))
+ if (!test_bit(LL_SBI_FOREIGN_SYMLINK_UPCALL, sbi->ll_flags))
rc = ll_foreign_symlink_default_parse(sbi, inode, lfm,
destname);
else /* upcall is available */
@@ -385,7 +385,7 @@ ssize_t foreign_symlink_enable_show(struct kobject *kobj,
ll_kset.kobj);
return snprintf(buf, PAGE_SIZE, "%d\n",
- !!(sbi->ll_flags & LL_SBI_FOREIGN_SYMLINK));
+ test_bit(LL_SBI_FOREIGN_SYMLINK, sbi->ll_flags));
}
/*
@@ -412,9 +412,9 @@ ssize_t foreign_symlink_enable_store(struct kobject *kobj,
return rc;
if (val)
- sbi->ll_flags |= LL_SBI_FOREIGN_SYMLINK;
+ set_bit(LL_SBI_FOREIGN_SYMLINK, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_FOREIGN_SYMLINK;
+ clear_bit(LL_SBI_FOREIGN_SYMLINK, sbi->ll_flags);
return count;
}
@@ -545,7 +545,7 @@ ssize_t foreign_symlink_upcall_store(struct kobject *kobj,
* order, we may end up using the format provided by a different
* upcall than the one set in ll_foreign_symlink_upcall
*/
- sbi->ll_flags &= ~LL_SBI_FOREIGN_SYMLINK_UPCALL;
+ clear_bit(LL_SBI_FOREIGN_SYMLINK_UPCALL, sbi->ll_flags);
up_write(&sbi->ll_foreign_symlink_sem);
if (strcmp(new, "none")) {
@@ -692,7 +692,7 @@ ssize_t foreign_symlink_upcall_info_store(struct kobject *kobj,
old_nb_items = sbi->ll_foreign_symlink_upcall_nb_items;
sbi->ll_foreign_symlink_upcall_items = new_items;
sbi->ll_foreign_symlink_upcall_nb_items = nb_items;
- sbi->ll_flags |= LL_SBI_FOREIGN_SYMLINK_UPCALL;
+ set_bit(LL_SBI_FOREIGN_SYMLINK_UPCALL, sbi->ll_flags);
up_write(&sbi->ll_foreign_symlink_sem);
/* free old_items */
diff --git a/fs/lustre/llite/llite_internal.h b/fs/lustre/llite/llite_internal.h
index afd5c7a..bd49228 100644
--- a/fs/lustre/llite/llite_internal.h
+++ b/fs/lustre/llite/llite_internal.h
@@ -45,6 +45,8 @@
#include <lustre_mdc.h>
#include <lustre_intent.h>
#include <linux/compat.h>
+#include <linux/aio.h>
+#include <linux/parser.h>
#include <lustre_crypto.h>
#include <range_lock.h>
#include <linux/namei.h>
@@ -602,82 +604,41 @@ enum stats_track_type {
STATS_TRACK_LAST,
};
-/* flags for sbi->ll_flags */
-#define LL_SBI_NOLCK 0x01 /* DLM locking disabled (directio-only) */
-#define LL_SBI_CHECKSUM 0x02 /* checksum each page as it's written */
-#define LL_SBI_FLOCK 0x04
-#define LL_SBI_USER_XATTR 0x08 /* support user xattr */
-#define LL_SBI_ACL 0x10 /* support ACL */
-/* LL_SBI_RMT_CLIENT 0x40 remote client */
-#define LL_SBI_MDS_CAPA 0x80 /* support mds capa, obsolete */
-#define LL_SBI_OSS_CAPA 0x100 /* support oss capa, obsolete */
-#define LL_SBI_LOCALFLOCK 0x200 /* Local flocks support by kernel */
-#define LL_SBI_LRU_RESIZE 0x400 /* lru resize support */
-#define LL_SBI_LAZYSTATFS 0x800 /* lazystatfs mount option */
-/* LL_SBI_SOM_PREVIEW 0x1000 SOM preview mount option, obsolete */
-#define LL_SBI_32BIT_API 0x2000 /* generate 32 bit inodes. */
-#define LL_SBI_64BIT_HASH 0x4000 /* support 64-bits dir hash/offset */
-#define LL_SBI_AGL_ENABLED 0x8000 /* enable agl */
-#define LL_SBI_VERBOSE 0x10000 /* verbose mount/umount */
-#define LL_SBI_LAYOUT_LOCK 0x20000 /* layout lock support */
-#define LL_SBI_USER_FID2PATH 0x40000 /* allow fid2path by unprivileged users */
-#define LL_SBI_XATTR_CACHE 0x80000 /* support for xattr cache */
-#define LL_SBI_NOROOTSQUASH 0x100000 /* do not apply root squash */
-#define LL_SBI_ALWAYS_PING 0x200000 /* always ping even if server
- * suppress_pings
- */
-#define LL_SBI_FAST_READ 0x400000 /* fast read support */
-#define LL_SBI_FILE_SECCTX 0x800000 /* set file security context at
- * create
- */
-#define LL_SBI_TINY_WRITE 0x2000000 /* tiny write support */
-#define LL_SBI_FILE_HEAT 0x4000000 /* file heat support */
-#define LL_SBI_TEST_DUMMY_ENCRYPTION 0x8000000 /* test dummy encryption */
-#define LL_SBI_ENCRYPT 0x10000000 /* client side encryption */
-#define LL_SBI_FOREIGN_SYMLINK 0x20000000 /* foreign fake-symlink support */
-/* foreign fake-symlink upcall registered */
-#define LL_SBI_FOREIGN_SYMLINK_UPCALL 0x40000000
-#define LL_SBI_PARALLEL_DIO 0x80000000 /* parallel (async) submission of
- * RPCs for DIO
- */
+enum ll_sbi_flags {
+ LL_SBI_NOLCK, /* DLM locking disabled directio-only */
+ LL_SBI_CHECKSUM, /* checksum each page as it's written */
+ LL_SBI_LOCALFLOCK, /* local flocks instead of fs-wide */
+ LL_SBI_FLOCK, /* flock enabled */
+ LL_SBI_USER_XATTR, /* support user xattr */
+ LL_SBI_LRU_RESIZE, /* lru resize support */
+ LL_SBI_LAZYSTATFS, /* lazystatfs mount option */
+ LL_SBI_32BIT_API, /* generate 32 bit inodes. */
+ LL_SBI_USER_FID2PATH, /* fid2path by unprivileged users */
+ LL_SBI_VERBOSE, /* verbose mount/umount */
+ LL_SBI_ALWAYS_PING, /* ping even if server suppress_pings */
+ LL_SBI_TEST_DUMMY_ENCRYPTION, /* test dummy encryption */
+ LL_SBI_ENCRYPT, /* client side encryption */
+ LL_SBI_FOREIGN_SYMLINK, /* foreign fake-symlink support */
+ LL_SBI_FOREIGN_SYMLINK_UPCALL, /* foreign fake-symlink upcall set */
+ LL_SBI_NUM_MOUNT_OPT,
+
+ LL_SBI_ACL, /* support ACL */
+ LL_SBI_AGL_ENABLED, /* enable agl */
+ LL_SBI_64BIT_HASH, /* support 64-bits dir hash/offset */
+ LL_SBI_LAYOUT_LOCK, /* layout lock support */
+ LL_SBI_XATTR_CACHE, /* support for xattr cache */
+ LL_SBI_NOROOTSQUASH, /* do not apply root squash */
+ LL_SBI_FAST_READ, /* fast read support */
+ LL_SBI_FILE_SECCTX, /* file security context at create */
+ LL_SBI_TINY_WRITE, /* tiny write support */
+ LL_SBI_FILE_HEAT, /* file heat support */
+ LL_SBI_PARALLEL_DIO, /* parallel (async) O_DIRECT RPCs */
+ LL_SBI_NUM_FLAGS
+};
-#define LL_SBI_FLAGS { \
- "nolck", \
- "checksum", \
- "flock", \
- "user_xattr", \
- "acl", \
- "???", \
- "???", \
- "mds_capa", \
- "oss_capa", \
- "flock", \
- "lru_resize", \
- "lazy_statfs", \
- "som", \
- "32bit_api", \
- "64bit_hash", \
- "agl", \
- "verbose", \
- "layout", \
- "user_fid2path",\
- "xattr_cache", \
- "norootsquash", \
- "always_ping", \
- "fast_read", \
- "file_secctx", \
- "pio", \
- "tiny_write", \
- "file_heat", \
- "test_dummy_encryption", \
- "noencrypt", \
- "foreign_symlink", \
- "foreign_symlink_upcall", \
- "parallel_dio", \
-}
+int ll_sbi_flags_seq_show(struct seq_file *m, void *v);
-/*
- * This is embedded into llite super-blocks to keep track of connect
+/* This is embedded into llite super-blocks to keep track of connect
* flags (capabilities) supported by all imports given mount is
* connected to.
*/
@@ -708,7 +669,7 @@ struct ll_sb_info {
struct dentry *ll_debugfs_entry;
struct lu_fid ll_root_fid; /* root object fid */
- int ll_flags;
+ DECLARE_BITMAP(ll_flags, LL_SBI_NUM_FLAGS); /* enum ll_sbi_flags */
unsigned int ll_xattr_cache_enabled:1,
ll_xattr_cache_set:1, /* already set to 0/1 */
ll_client_common_fill_super_succeeded:1,
@@ -970,7 +931,7 @@ static inline bool ll_need_32bit_api(struct ll_sb_info *sbi)
#if BITS_PER_LONG == 32
return true;
#else
- if (unlikely(sbi->ll_flags & LL_SBI_32BIT_API))
+ if (unlikely(test_bit(LL_SBI_32BIT_API, sbi->ll_flags)))
return true;
#if defined(CONFIG_COMPAT)
@@ -991,27 +952,27 @@ static inline bool ll_need_32bit_api(struct ll_sb_info *sbi)
static inline bool ll_sbi_has_fast_read(struct ll_sb_info *sbi)
{
- return !!(sbi->ll_flags & LL_SBI_FAST_READ);
+ return test_bit(LL_SBI_FAST_READ, sbi->ll_flags);
}
static inline bool ll_sbi_has_tiny_write(struct ll_sb_info *sbi)
{
- return !!(sbi->ll_flags & LL_SBI_TINY_WRITE);
+ return test_bit(LL_SBI_TINY_WRITE, sbi->ll_flags);
}
static inline bool ll_sbi_has_file_heat(struct ll_sb_info *sbi)
{
- return !!(sbi->ll_flags & LL_SBI_FILE_HEAT);
+ return test_bit(LL_SBI_FILE_HEAT, sbi->ll_flags);
}
static inline bool ll_sbi_has_foreign_symlink(struct ll_sb_info *sbi)
{
- return !!(sbi->ll_flags & LL_SBI_FOREIGN_SYMLINK);
+ return test_bit(LL_SBI_FOREIGN_SYMLINK, sbi->ll_flags);
}
static inline bool ll_sbi_has_parallel_dio(struct ll_sb_info *sbi)
{
- return !!(sbi->ll_flags & LL_SBI_PARALLEL_DIO);
+ return test_bit(LL_SBI_PARALLEL_DIO, sbi->ll_flags);
}
void ll_ras_enter(struct file *f, loff_t pos, size_t count);
@@ -1615,7 +1576,7 @@ static inline int ll_file_nolock(const struct file *file)
struct inode *inode = file_inode(file);
return ((fd->fd_flags & LL_FILE_IGNORE_LOCK) ||
- (ll_i2sbi(inode)->ll_flags & LL_SBI_NOLCK));
+ test_bit(LL_SBI_NOLCK, ll_i2sbi(inode)->ll_flags));
}
static inline void ll_set_lock_data(struct obd_export *exp, struct inode *inode,
diff --git a/fs/lustre/llite/llite_lib.c b/fs/lustre/llite/llite_lib.c
index 9ff881c..abd470a 100644
--- a/fs/lustre/llite/llite_lib.c
+++ b/fs/lustre/llite/llite_lib.c
@@ -155,11 +155,11 @@ static struct ll_sb_info *ll_init_sbi(void)
sbi->ll_ra_info.ra_max_read_ahead_whole_pages = -1;
atomic_set(&sbi->ll_ra_info.ra_async_inflight, 0);
- sbi->ll_flags |= LL_SBI_VERBOSE;
- sbi->ll_flags |= LL_SBI_CHECKSUM;
- sbi->ll_flags |= LL_SBI_FLOCK;
- sbi->ll_flags |= LL_SBI_LRU_RESIZE;
- sbi->ll_flags |= LL_SBI_LAZYSTATFS;
+ set_bit(LL_SBI_VERBOSE, sbi->ll_flags);
+ set_bit(LL_SBI_CHECKSUM, sbi->ll_flags);
+ set_bit(LL_SBI_FLOCK, sbi->ll_flags);
+ set_bit(LL_SBI_LRU_RESIZE, sbi->ll_flags);
+ set_bit(LL_SBI_LAZYSTATFS, sbi->ll_flags);
for (i = 0; i <= LL_PROCESS_HIST_MAX; i++) {
struct per_process_info *pp_ext;
@@ -176,10 +176,10 @@ static struct ll_sb_info *ll_init_sbi(void)
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;
- sbi->ll_flags |= LL_SBI_FAST_READ;
- sbi->ll_flags |= LL_SBI_TINY_WRITE;
- sbi->ll_flags |= LL_SBI_PARALLEL_DIO;
+ set_bit(LL_SBI_AGL_ENABLED, sbi->ll_flags);
+ set_bit(LL_SBI_FAST_READ, sbi->ll_flags);
+ set_bit(LL_SBI_TINY_WRITE, sbi->ll_flags);
+ set_bit(LL_SBI_PARALLEL_DIO, sbi->ll_flags);
ll_sbi_set_encrypt(sbi, true);
/* root squash */
@@ -257,6 +257,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
struct lustre_md lmd;
u64 valid;
int size, err, checksum;
+ bool api32;
sbi->ll_md_obd = class_name2obd(md);
if (!sbi->ll_md_obd) {
@@ -320,7 +321,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
OBD_CONNECT2_REP_MBITS |
OBD_CONNECT2_ATOMIC_OPEN_LOCK;
- if (sbi->ll_flags & LL_SBI_LRU_RESIZE)
+ if (test_bit(LL_SBI_LRU_RESIZE, sbi->ll_flags))
data->ocd_connect_flags |= OBD_CONNECT_LRU_RESIZE;
data->ocd_connect_flags |= OBD_CONNECT_ACL_FLAGS;
@@ -337,7 +338,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
if (sb_rdonly(sb))
data->ocd_connect_flags |= OBD_CONNECT_RDONLY;
- if (sbi->ll_flags & LL_SBI_USER_XATTR)
+ if (test_bit(LL_SBI_USER_XATTR, sbi->ll_flags))
data->ocd_connect_flags |= OBD_CONNECT_XATTR;
/* Setting this indicates we correctly support S_NOSEC (See kernel
@@ -348,7 +349,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
sbi->ll_fop = ll_select_file_operations(sbi);
/* always ping even if server suppress_pings */
- if (sbi->ll_flags & LL_SBI_ALWAYS_PING)
+ if (test_bit(LL_SBI_ALWAYS_PING, sbi->ll_flags))
data->ocd_connect_flags &= ~OBD_CONNECT_PINGLESS;
obd_connect_set_secctx(data);
@@ -451,29 +452,29 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
sbi->ll_namelen = osfs->os_namelen;
sbi->ll_mnt.mnt = current->fs->root.mnt;
- if ((sbi->ll_flags & LL_SBI_USER_XATTR) &&
+ if (test_bit(LL_SBI_USER_XATTR, sbi->ll_flags) &&
!(data->ocd_connect_flags & OBD_CONNECT_XATTR)) {
LCONSOLE_INFO("Disabling user_xattr feature because it is not supported on the server\n");
- sbi->ll_flags &= ~LL_SBI_USER_XATTR;
+ clear_bit(LL_SBI_USER_XATTR, sbi->ll_flags);
}
if (data->ocd_connect_flags & OBD_CONNECT_ACL) {
sb->s_flags |= SB_POSIXACL;
- sbi->ll_flags |= LL_SBI_ACL;
+ set_bit(LL_SBI_ACL, sbi->ll_flags);
} else {
LCONSOLE_INFO("client wants to enable acl, but mdt not!\n");
sb->s_flags &= ~SB_POSIXACL;
- sbi->ll_flags &= ~LL_SBI_ACL;
+ clear_bit(LL_SBI_ACL, sbi->ll_flags);
}
if (data->ocd_connect_flags & OBD_CONNECT_64BITHASH)
- sbi->ll_flags |= LL_SBI_64BIT_HASH;
+ set_bit(LL_SBI_64BIT_HASH, sbi->ll_flags);
if (data->ocd_connect_flags & OBD_CONNECT_LAYOUTLOCK)
- sbi->ll_flags |= LL_SBI_LAYOUT_LOCK;
+ set_bit(LL_SBI_LAYOUT_LOCK, sbi->ll_flags);
if (obd_connect_has_secctx(data))
- sbi->ll_flags |= LL_SBI_FILE_SECCTX;
+ set_bit(LL_SBI_FILE_SECCTX, sbi->ll_flags);
if (ll_sbi_has_encrypt(sbi) && !obd_connect_has_enc(data)) {
if (ll_sbi_has_test_dummy_encryption(sbi))
@@ -492,7 +493,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
* during processing llog, it won't be enabled here.
*/
spin_lock(&sbi->ll_lock);
- sbi->ll_flags |= LL_SBI_XATTR_CACHE;
+ set_bit(LL_SBI_XATTR_CACHE, sbi->ll_flags);
spin_unlock(&sbi->ll_lock);
sbi->ll_xattr_cache_enabled = 1;
}
@@ -542,7 +543,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
data->ocd_connect_flags |= OBD_CONNECT_LRU_RESIZE;
/* always ping even if server suppress_pings */
- if (sbi->ll_flags & LL_SBI_ALWAYS_PING)
+ if (test_bit(LL_SBI_ALWAYS_PING, sbi->ll_flags))
data->ocd_connect_flags &= ~OBD_CONNECT_PINGLESS;
if (ll_sbi_has_encrypt(sbi))
@@ -630,7 +631,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
* XXX: move this to after cbd setup?
*/
valid = OBD_MD_FLGETATTR | OBD_MD_FLBLOCKS | OBD_MD_FLMODEASIZE;
- if (sbi->ll_flags & LL_SBI_ACL)
+ if (test_bit(LL_SBI_ACL, sbi->ll_flags))
valid |= OBD_MD_FLACL;
op_data = kzalloc(sizeof(*op_data), GFP_NOFS);
@@ -660,9 +661,8 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
}
LASSERT(fid_is_sane(&sbi->ll_root_fid));
- root = ll_iget(sb, cl_fid_build_ino(&sbi->ll_root_fid,
- sbi->ll_flags & LL_SBI_32BIT_API),
- &lmd);
+ api32 = test_bit(LL_SBI_32BIT_API, sbi->ll_flags);
+ root = ll_iget(sb, cl_fid_build_ino(&sbi->ll_root_fid, api32), &lmd);
md_free_lustre_md(sbi->ll_md_exp, &lmd);
ptlrpc_req_finished(request);
@@ -673,7 +673,7 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt)
goto out_root;
}
- checksum = sbi->ll_flags & LL_SBI_CHECKSUM;
+ checksum = test_bit(LL_SBI_CHECKSUM, sbi->ll_flags);
if (sbi->ll_checksum_set) {
err = obd_set_info_async(NULL, sbi->ll_dt_exp,
sizeof(KEY_CHECKSUM), KEY_CHECKSUM,
@@ -864,183 +864,193 @@ void ll_kill_super(struct super_block *sb)
}
}
-static inline int ll_set_opt(const char *opt, char *data, int fl)
+/* Since we use this table for ll_sbi_flags_seq_show make
+ * sure what you want displayed for a specific token that
+ * is listed more than once below be listed first. For
+ * example we want "checksum" displayed, not "nochecksum"
+ * for the sbi_flags.
+ */
+static const match_table_t ll_sbi_flags_name = {
+ {LL_SBI_NOLCK, "nolock"},
+ {LL_SBI_CHECKSUM, "checksum"},
+ {LL_SBI_CHECKSUM, "nochecksum"},
+ {LL_SBI_LOCALFLOCK, "localflock"},
+ {LL_SBI_FLOCK, "flock"},
+ {LL_SBI_FLOCK, "noflock"},
+ {LL_SBI_USER_XATTR, "user_xattr"},
+ {LL_SBI_USER_XATTR, "nouser_xattr"},
+ {LL_SBI_LRU_RESIZE, "lruresize"},
+ {LL_SBI_LRU_RESIZE, "nolruresize"},
+ {LL_SBI_LAZYSTATFS, "lazystatfs"},
+ {LL_SBI_LAZYSTATFS, "nolazystatfs"},
+ {LL_SBI_32BIT_API, "32bitapi"},
+ {LL_SBI_USER_FID2PATH, "user_fid2path"},
+ {LL_SBI_USER_FID2PATH, "nouser_fid2path"},
+ {LL_SBI_VERBOSE, "verbose"},
+ {LL_SBI_VERBOSE, "noverbose"},
+ {LL_SBI_ALWAYS_PING, "always_ping"},
+ {LL_SBI_TEST_DUMMY_ENCRYPTION, "test_dummy_encryption"},
+ {LL_SBI_ENCRYPT, "encrypt"},
+ {LL_SBI_ENCRYPT, "noencrypt"},
+ {LL_SBI_FOREIGN_SYMLINK, "foreign_symlink=%s"},
+ {LL_SBI_NUM_MOUNT_OPT, NULL},
+
+ {LL_SBI_ACL, "acl"},
+ {LL_SBI_AGL_ENABLED, "agl"},
+ {LL_SBI_64BIT_HASH, "64bit_hash"},
+ {LL_SBI_LAYOUT_LOCK, "layout"},
+ {LL_SBI_XATTR_CACHE, "xattr_cache"},
+ {LL_SBI_NOROOTSQUASH, "norootsquash"},
+ {LL_SBI_FAST_READ, "fast_read"},
+ {LL_SBI_FILE_SECCTX, "file_secctx"},
+ {LL_SBI_TINY_WRITE, "tiny_write"},
+ {LL_SBI_FILE_HEAT, "file_heat"},
+ {LL_SBI_PARALLEL_DIO, "parallel_dio"},
+};
+
+int ll_sbi_flags_seq_show(struct seq_file *m, void *v)
{
- if (strncmp(opt, data, strlen(opt)) != 0)
- return 0;
- else
- return fl;
+ struct super_block *sb = m->private;
+ int i;
+
+ for (i = 0; i < LL_SBI_NUM_FLAGS; i++) {
+ int j;
+
+ if (!test_bit(i, ll_s2sbi(sb)->ll_flags))
+ continue;
+
+ for (j = 0; j < ARRAY_SIZE(ll_sbi_flags_name); j++) {
+ if (ll_sbi_flags_name[j].token == i &&
+ ll_sbi_flags_name[j].pattern) {
+ seq_printf(m, "%s ",
+ ll_sbi_flags_name[j].pattern);
+ break;
+ }
+ }
+ }
+ seq_puts(m, "\b\n");
+ return 0;
}
/* non-client-specific mount options are parsed in lmd_parse */
-static int ll_options(char *options, struct ll_sb_info *sbi)
+static int ll_options(char *options, struct super_block *sb)
{
- int tmp;
- char *s1 = options, *s2;
- int *flags = &sbi->ll_flags;
+ struct ll_sb_info *sbi = ll_s2sbi(sb);
+ char *s2, *s1, *opts;
if (!options)
return 0;
+ /* Don't stomp on lmd_opts */
+ opts = kstrdup(options, GFP_KERNEL);
+ if (!opts)
+ return -ENOMEM;
+ s1 = opts;
+ s2 = opts;
+
CDEBUG(D_CONFIG, "Parsing opts %s\n", options);
- while (*s1) {
+ while ((s1 = strsep(&opts, ",")) != NULL) {
+ substring_t args[MAX_OPT_ARGS];
+ bool turn_off = false;
+ int token;
+
+ if (!*s1)
+ continue;
+
CDEBUG(D_SUPER, "next opt=%s\n", s1);
- tmp = ll_set_opt("nolock", s1, LL_SBI_NOLCK);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("flock", s1, LL_SBI_FLOCK);
- if (tmp) {
- *flags = (*flags & ~LL_SBI_LOCALFLOCK) | tmp;
- goto next;
- }
- tmp = ll_set_opt("localflock", s1, LL_SBI_LOCALFLOCK);
- if (tmp) {
- *flags = (*flags & ~LL_SBI_FLOCK) | tmp;
- goto next;
- }
- tmp = ll_set_opt("noflock", s1,
- LL_SBI_FLOCK | LL_SBI_LOCALFLOCK);
- if (tmp) {
- *flags &= ~tmp;
- goto next;
- }
- tmp = ll_set_opt("user_xattr", s1, LL_SBI_USER_XATTR);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("nouser_xattr", s1, LL_SBI_USER_XATTR);
- if (tmp) {
- *flags &= ~tmp;
- goto next;
- }
- tmp = ll_set_opt("context", s1, 1);
- if (tmp)
- goto next;
- tmp = ll_set_opt("fscontext", s1, 1);
- if (tmp)
- goto next;
- tmp = ll_set_opt("defcontext", s1, 1);
- if (tmp)
- goto next;
- tmp = ll_set_opt("rootcontext", s1, 1);
- if (tmp)
- goto next;
- tmp = ll_set_opt("user_fid2path", s1, LL_SBI_USER_FID2PATH);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("nouser_fid2path", s1, LL_SBI_USER_FID2PATH);
- if (tmp) {
- *flags &= ~tmp;
- goto next;
- }
+ if (strncmp(s1, "no", 2) == 0)
+ turn_off = true;
- tmp = ll_set_opt("checksum", s1, LL_SBI_CHECKSUM);
- if (tmp) {
- *flags |= tmp;
- sbi->ll_checksum_set = 1;
- goto next;
+ /*
+ * Initialize args struct so we know whether arg was
+ * found; some options take optional arguments.
+ */
+ args[0].to = NULL;
+ args[0].from = NULL;
+ token = match_token(s1, ll_sbi_flags_name, args);
+ if (token == LL_SBI_NUM_MOUNT_OPT) {
+ if (match_wildcard("context", s1) ||
+ match_wildcard("fscontext", s1) ||
+ match_wildcard("defcontext", s1) ||
+ match_wildcard("rootcontext", s1))
+ continue;
+
+ LCONSOLE_ERROR_MSG(0x152,
+ "Unknown option '%s', won't mount.\n",
+ s1);
+ return -EINVAL;
}
- tmp = ll_set_opt("nochecksum", s1, LL_SBI_CHECKSUM);
- if (tmp) {
- *flags &= ~tmp;
+
+ switch (token) {
+ case LL_SBI_NOLCK:
+ case LL_SBI_32BIT_API:
+ case LL_SBI_64BIT_HASH:
+ case LL_SBI_ALWAYS_PING:
+ set_bit(token, sbi->ll_flags);
+ break;
+
+ case LL_SBI_FLOCK:
+ clear_bit(LL_SBI_LOCALFLOCK, sbi->ll_flags);
+ if (turn_off)
+ clear_bit(LL_SBI_FLOCK, sbi->ll_flags);
+ else
+ set_bit(token, sbi->ll_flags);
+ break;
+
+ case LL_SBI_LOCALFLOCK:
+ clear_bit(LL_SBI_FLOCK, sbi->ll_flags);
+ set_bit(token, sbi->ll_flags);
+ break;
+
+ case LL_SBI_CHECKSUM:
sbi->ll_checksum_set = 1;
- goto next;
- }
- tmp = ll_set_opt("lruresize", s1, LL_SBI_LRU_RESIZE);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("nolruresize", s1, LL_SBI_LRU_RESIZE);
- if (tmp) {
- *flags &= ~tmp;
- goto next;
- }
- tmp = ll_set_opt("lazystatfs", s1, LL_SBI_LAZYSTATFS);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("nolazystatfs", s1, LL_SBI_LAZYSTATFS);
- if (tmp) {
- *flags &= ~tmp;
- goto next;
- }
- tmp = ll_set_opt("32bitapi", s1, LL_SBI_32BIT_API);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("verbose", s1, LL_SBI_VERBOSE);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("noverbose", s1, LL_SBI_VERBOSE);
- if (tmp) {
- *flags &= ~tmp;
- goto next;
- }
- tmp = ll_set_opt("always_ping", s1, LL_SBI_ALWAYS_PING);
- if (tmp) {
- *flags |= tmp;
- goto next;
- }
- tmp = ll_set_opt("test_dummy_encryption", s1,
- LL_SBI_TEST_DUMMY_ENCRYPTION);
- if (tmp) {
+ /* fall through */
+ case LL_SBI_USER_XATTR:
+ case LL_SBI_USER_FID2PATH:
+ case LL_SBI_LRU_RESIZE:
+ case LL_SBI_LAZYSTATFS:
+ case LL_SBI_VERBOSE:
+ if (turn_off)
+ clear_bit(token, sbi->ll_flags);
+ else
+ set_bit(token, sbi->ll_flags);
+ break;
+ case LL_SBI_TEST_DUMMY_ENCRYPTION: {
#ifdef CONFIG_FS_ENCRYPTION
- *flags |= tmp;
+ set_bit(token, sbi->ll_flags);
#else
LCONSOLE_WARN("Test dummy encryption mount option ignored: encryption not supported\n");
#endif
- goto next;
+ break;
}
- tmp = ll_set_opt("noencrypt", s1, LL_SBI_ENCRYPT);
- if (tmp) {
+ case LL_SBI_ENCRYPT:
#ifdef CONFIG_FS_ENCRYPTION
- *flags &= ~tmp;
+ if (turn_off)
+ clear_bit(token, sbi->ll_flags);
+ else
+ set_bit(token, sbi->ll_flags);
#else
- LCONSOLE_WARN("noencrypt mount option ignored: encryption not supported\n");
+ LCONSOLE_WARN("noencrypt or encrypt mount option ignored: encryption not supported\n");
#endif
- goto next;
- }
- tmp = ll_set_opt("foreign_symlink", s1, LL_SBI_FOREIGN_SYMLINK);
- if (tmp) {
- int prefix_pos = sizeof("foreign_symlink=") - 1;
- int equal_pos = sizeof("foreign_symlink=") - 2;
-
+ break;
+ case LL_SBI_FOREIGN_SYMLINK:
/* non-default prefix provided ? */
- if (strlen(s1) >= sizeof("foreign_symlink=") &&
- *(s1 + equal_pos) == '=') {
- char *old = sbi->ll_foreign_symlink_prefix;
- size_t old_len =
- sbi->ll_foreign_symlink_prefix_size;
+ if (args->from) {
+ size_t old_len;
+ char *old;
/* path must be absolute */
- if (*(s1 + sizeof("foreign_symlink=") -
- 1) != '/') {
+ if (args->from[0] != '/') {
LCONSOLE_ERROR_MSG(0x152,
"foreign prefix '%s' must be an absolute path\n",
- s1 + prefix_pos);
+ args->from);
return -EINVAL;
}
- /* last option ? */
- s2 = strchrnul(s1 + prefix_pos, ',');
-
- if (sbi->ll_foreign_symlink_prefix) {
- sbi->ll_foreign_symlink_prefix = NULL;
- sbi->ll_foreign_symlink_prefix_size = 0;
- }
+ old_len = sbi->ll_foreign_symlink_prefix_size;
+ old = sbi->ll_foreign_symlink_prefix;
/* alloc for path length and '\0' */
- sbi->ll_foreign_symlink_prefix = kmalloc(s2 - (s1 + prefix_pos) + 1,
- GFP_KERNEL);
+ sbi->ll_foreign_symlink_prefix = match_strdup(args);
if (!sbi->ll_foreign_symlink_prefix) {
/* restore previous */
sbi->ll_foreign_symlink_prefix = old;
@@ -1048,31 +1058,22 @@ static int ll_options(char *options, struct ll_sb_info *sbi)
old_len;
return -ENOMEM;
}
- kfree(old);
- strncpy(sbi->ll_foreign_symlink_prefix,
- s1 + prefix_pos,
- s2 - (s1 + prefix_pos));
sbi->ll_foreign_symlink_prefix_size =
- s2 - (s1 + prefix_pos) + 1;
+ args->to - args->from + 1;
+ kfree(old);
+
+ /* enable foreign symlink support */
+ set_bit(token, sbi->ll_flags);
} else {
LCONSOLE_ERROR_MSG(0x152,
"invalid %s option\n", s1);
}
- /* enable foreign symlink support */
- *flags |= tmp;
- goto next;
- }
- LCONSOLE_ERROR_MSG(0x152, "Unknown option '%s', won't mount.\n",
- s1);
- return -EINVAL;
-
-next:
- /* Find next opt */
- s2 = strchr(s1, ',');
- if (!s2)
+ /* fall through */
+ default:
break;
- s1 = s2 + 1;
+ }
}
+ kfree(opts);
return 0;
}
@@ -1168,7 +1169,7 @@ int ll_fill_super(struct super_block *sb)
goto out_free;
}
- err = ll_options(lsi->lsi_lmd->lmd_opts, sbi);
+ err = ll_options(lsi->lsi_lmd->lmd_opts, sb);
if (err)
goto out_free;
@@ -1271,7 +1272,7 @@ int ll_fill_super(struct super_block *sb)
kfree(cfg);
if (err)
ll_put_super(sb);
- else if (sbi->ll_flags & LL_SBI_VERBOSE)
+ else if (test_bit(LL_SBI_VERBOSE, sbi->ll_flags))
LCONSOLE_WARN("Mounted %s\n", profilenm);
return err;
@@ -1339,7 +1340,7 @@ void ll_put_super(struct super_block *sb)
while ((obd = class_devices_in_group(&sbi->ll_sb_uuid, &next)))
class_manual_cleanup(obd);
- if (sbi->ll_flags & LL_SBI_VERBOSE)
+ if (test_bit(LL_SBI_VERBOSE, sbi->ll_flags))
LCONSOLE_WARN("Unmounted %s\n", profilenm ? profilenm : "");
if (profilenm)
@@ -1408,7 +1409,7 @@ static struct inode *ll_iget_anon_dir(struct super_block *sb,
ino_t ino;
LASSERT(md->lmv);
- ino = cl_fid_build_ino(fid, sbi->ll_flags & LL_SBI_32BIT_API);
+ ino = cl_fid_build_ino(fid, test_bit(LL_SBI_32BIT_API, sbi->ll_flags));
inode = iget_locked(sb, ino);
if (!inode) {
CERROR("%s: failed get simple inode " DFID ": rc = -ENOENT\n",
@@ -2207,7 +2208,7 @@ int ll_statfs_internal(struct ll_sb_info *sbi, struct obd_statfs *osfs,
max_age = ktime_get_seconds() - sbi->ll_statfs_max_age;
- if (sbi->ll_flags & LL_SBI_LAZYSTATFS)
+ if (test_bit(LL_SBI_LAZYSTATFS, sbi->ll_flags))
flags |= OBD_STATFS_NODELAY;
rc = obd_statfs(NULL, sbi->ll_md_exp, osfs, max_age, flags);
@@ -2383,6 +2384,7 @@ int ll_update_inode(struct inode *inode, struct lustre_md *md)
struct ll_inode_info *lli = ll_i2info(inode);
struct mdt_body *body = md->body;
struct ll_sb_info *sbi = ll_i2sbi(inode);
+ bool api32;
int rc = 0;
if (body->mbo_valid & OBD_MD_FLEASIZE) {
@@ -2400,8 +2402,8 @@ int ll_update_inode(struct inode *inode, struct lustre_md *md)
if (body->mbo_valid & OBD_MD_FLACL)
lli_replace_acl(lli, md);
- inode->i_ino = cl_fid_build_ino(&body->mbo_fid1,
- sbi->ll_flags & LL_SBI_32BIT_API);
+ api32 = test_bit(LL_SBI_32BIT_API, sbi->ll_flags);
+ inode->i_ino = cl_fid_build_ino(&body->mbo_fid1, api32);
inode->i_generation = cl_fid_build_gen(&body->mbo_fid1);
if (body->mbo_valid & OBD_MD_FLATIME) {
@@ -2782,7 +2784,7 @@ int ll_remount_fs(struct super_block *sb, int *flags, char *data)
else
sb->s_flags &= ~SB_RDONLY;
- if (sbi->ll_flags & LL_SBI_VERBOSE)
+ if (test_bit(LL_SBI_VERBOSE, sbi->ll_flags))
LCONSOLE_WARN("Remounted %s %s\n", profilenm,
read_only ? "read-only" : "read-write");
}
@@ -2853,24 +2855,23 @@ int ll_prep_inode(struct inode **inode, struct req_capsule *pill,
if (rc)
goto out;
} else {
+ bool api32 = test_bit(LL_SBI_32BIT_API, sbi->ll_flags);
+ struct lu_fid *fid1 = &md.body->mbo_fid1;
+
LASSERT(sb);
/*
* At this point server returns to client's same fid as client
* generated for creating. So using ->fid1 is okay here.
*/
- if (!fid_is_sane(&md.body->mbo_fid1)) {
+ if (!fid_is_sane(fid1)) {
CERROR("%s: Fid is insane " DFID "\n",
- sbi->ll_fsname,
- PFID(&md.body->mbo_fid1));
+ sbi->ll_fsname, PFID(fid1));
rc = -EINVAL;
goto out;
}
- *inode = ll_iget(sb,
- cl_fid_build_ino(&md.body->mbo_fid1,
- sbi->ll_flags & LL_SBI_32BIT_API),
- &md);
+ *inode = ll_iget(sb, cl_fid_build_ino(fid1, api32), &md);
if (IS_ERR(*inode)) {
lmd_clear_acl(&md);
rc = IS_ERR(*inode) ? PTR_ERR(*inode) : -ENOMEM;
@@ -3055,7 +3056,7 @@ struct md_op_data *ll_prep_md_op_data(struct md_op_data *op_data,
fid_zero(&op_data->op_fid2);
}
- if (ll_i2sbi(i1)->ll_flags & LL_SBI_64BIT_HASH)
+ if (test_bit(LL_SBI_64BIT_HASH, ll_i2sbi(i1)->ll_flags))
op_data->op_cli_flags |= CLI_HASH64;
if (ll_need_32bit_api(ll_i2sbi(i1)))
@@ -3132,47 +3133,40 @@ void ll_finish_md_op_data(struct md_op_data *op_data)
int ll_show_options(struct seq_file *seq, struct dentry *dentry)
{
struct ll_sb_info *sbi;
+ int i;
LASSERT(seq && dentry);
sbi = ll_s2sbi(dentry->d_sb);
- if (sbi->ll_flags & LL_SBI_NOLCK)
- seq_puts(seq, ",nolock");
-
- /* "flock" is the default since 2.13, but it wasn't for many years,
- * so it is still useful to print this to show it is enabled.
- * Start to print "noflock" so it is now clear when flock is disabled.
- */
- if (sbi->ll_flags & LL_SBI_FLOCK)
- seq_puts(seq, ",flock");
- else if (sbi->ll_flags & LL_SBI_LOCALFLOCK)
- seq_puts(seq, ",localflock");
- else
- seq_puts(seq, ",noflock");
-
- if (sbi->ll_flags & LL_SBI_USER_XATTR)
- seq_puts(seq, ",user_xattr");
-
- if (sbi->ll_flags & LL_SBI_LAZYSTATFS)
- seq_puts(seq, ",lazystatfs");
-
- if (sbi->ll_flags & LL_SBI_USER_FID2PATH)
- seq_puts(seq, ",user_fid2path");
-
- if (sbi->ll_flags & LL_SBI_ALWAYS_PING)
- seq_puts(seq, ",always_ping");
+ if (test_bit(LL_SBI_NOLCK, sbi->ll_flags))
+ seq_puts(seq, "nolock");
- if (ll_sbi_has_test_dummy_encryption(sbi))
- seq_puts(seq, ",test_dummy_encryption");
-
- if (ll_sbi_has_encrypt(sbi))
- seq_puts(seq, ",encrypt");
- else
- seq_puts(seq, ",noencrypt");
+ for (i = 1; ll_sbi_flags_name[i].token != LL_SBI_NUM_MOUNT_OPT; i++) {
+ /* match_table in some cases has patterns for both enabled and
+ * disabled cases. Ignore 'no'xxx versions if bit is set.
+ */
+ if (test_bit(ll_sbi_flags_name[i].token, sbi->ll_flags) &&
+ strncmp(ll_sbi_flags_name[i].pattern, "no", 2)) {
+ if (ll_sbi_flags_name[i].token ==
+ LL_SBI_FOREIGN_SYMLINK) {
+ seq_show_option(seq, "foreign_symlink",
+ sbi->ll_foreign_symlink_prefix);
+ } else {
+ seq_printf(seq, ",%s",
+ ll_sbi_flags_name[i].pattern);
+ }
- if (sbi->ll_flags & LL_SBI_FOREIGN_SYMLINK) {
- seq_puts(seq, ",foreign_symlink=");
- seq_puts(seq, sbi->ll_foreign_symlink_prefix);
+ /* You can have either localflock or flock but not
+ * both. If localflock is set don't print flock or
+ * noflock.
+ */
+ if (ll_sbi_flags_name[i].token == LL_SBI_LOCALFLOCK)
+ i += 2;
+ } else if (!test_bit(ll_sbi_flags_name[i].token, sbi->ll_flags) &&
+ !strncmp(ll_sbi_flags_name[i].pattern, "no", 2)) {
+ seq_printf(seq, ",%s",
+ ll_sbi_flags_name[i].pattern);
+ }
}
return 0;
@@ -3276,12 +3270,9 @@ void ll_compute_rootsquash_state(struct ll_sb_info *sbi)
/* Update norootsquash flag */
spin_lock(&squash->rsi_lock);
if (list_empty(&squash->rsi_nosquash_nids)) {
- spin_lock(&sbi->ll_lock);
- sbi->ll_flags &= ~LL_SBI_NOROOTSQUASH;
- spin_unlock(&sbi->ll_lock);
+ clear_bit(LL_SBI_NOROOTSQUASH, sbi->ll_flags);
} else {
- /*
- * Do not apply root squash as soon as one of our NIDs is
+ /* Do not apply root squash as soon as one of our NIDs is
* in the nosquash_nids list
*/
matched = false;
@@ -3297,10 +3288,9 @@ void ll_compute_rootsquash_state(struct ll_sb_info *sbi)
}
spin_lock(&sbi->ll_lock);
if (matched)
- sbi->ll_flags |= LL_SBI_NOROOTSQUASH;
+ set_bit(LL_SBI_NOROOTSQUASH, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_NOROOTSQUASH;
- spin_unlock(&sbi->ll_lock);
+ clear_bit(LL_SBI_NOROOTSQUASH, sbi->ll_flags);
}
spin_unlock(&squash->rsi_lock);
}
@@ -3375,7 +3365,7 @@ int ll_getparent(struct file *file, struct getparent __user *arg)
int rc;
if (!capable(CAP_DAC_READ_SEARCH) &&
- !(ll_i2sbi(inode)->ll_flags & LL_SBI_USER_FID2PATH))
+ !test_bit(LL_SBI_USER_FID2PATH, ll_i2sbi(inode)->ll_flags))
return -EPERM;
if (get_user(name_size, &arg->gp_name_size))
diff --git a/fs/lustre/llite/lproc_llite.c b/fs/lustre/llite/lproc_llite.c
index 3b4f60c..eac905d 100644
--- a/fs/lustre/llite/lproc_llite.c
+++ b/fs/lustre/llite/lproc_llite.c
@@ -610,7 +610,8 @@ static ssize_t checksums_show(struct kobject *kobj, struct attribute *attr,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kset.kobj);
- return sprintf(buf, "%u\n", (sbi->ll_flags & LL_SBI_CHECKSUM) ? 1 : 0);
+ return scnprintf(buf, PAGE_SIZE, "%u\n",
+ test_bit(LL_SBI_CHECKSUM, sbi->ll_flags));
}
static ssize_t checksums_store(struct kobject *kobj, struct attribute *attr,
@@ -630,12 +631,10 @@ static ssize_t checksums_store(struct kobject *kobj, struct attribute *attr,
if (rc)
return rc;
- spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_CHECKSUM;
+ set_bit(LL_SBI_CHECKSUM, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_CHECKSUM;
- spin_unlock(&sbi->ll_lock);
+ clear_bit(LL_SBI_CHECKSUM, sbi->ll_flags);
tmp = val;
rc = obd_set_info_async(NULL, sbi->ll_dt_exp, sizeof(KEY_CHECKSUM),
@@ -809,7 +808,8 @@ static ssize_t statahead_agl_show(struct kobject *kobj,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kset.kobj);
- return sprintf(buf, "%u\n", sbi->ll_flags & LL_SBI_AGL_ENABLED ? 1 : 0);
+ return scnprintf(buf, PAGE_SIZE, "%u\n",
+ test_bit(LL_SBI_AGL_ENABLED, sbi->ll_flags));
}
static ssize_t statahead_agl_store(struct kobject *kobj,
@@ -826,12 +826,10 @@ static ssize_t statahead_agl_store(struct kobject *kobj,
if (rc)
return rc;
- spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_AGL_ENABLED;
+ set_bit(LL_SBI_AGL_ENABLED, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_AGL_ENABLED;
- spin_unlock(&sbi->ll_lock);
+ clear_bit(LL_SBI_AGL_ENABLED, sbi->ll_flags);
return count;
}
@@ -861,7 +859,8 @@ static ssize_t lazystatfs_show(struct kobject *kobj,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kset.kobj);
- return sprintf(buf, "%u\n", sbi->ll_flags & LL_SBI_LAZYSTATFS ? 1 : 0);
+ return scnprintf(buf, PAGE_SIZE, "%u\n",
+ test_bit(LL_SBI_LAZYSTATFS, sbi->ll_flags));
}
static ssize_t lazystatfs_store(struct kobject *kobj,
@@ -878,12 +877,10 @@ static ssize_t lazystatfs_store(struct kobject *kobj,
if (rc)
return rc;
- spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_LAZYSTATFS;
+ set_bit(LL_SBI_LAZYSTATFS, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_LAZYSTATFS;
- spin_unlock(&sbi->ll_lock);
+ clear_bit(LL_SBI_LAZYSTATFS, sbi->ll_flags);
return count;
}
@@ -1006,29 +1003,6 @@ static ssize_t default_easize_store(struct kobject *kobj,
}
LUSTRE_RW_ATTR(default_easize);
-static int ll_sbi_flags_seq_show(struct seq_file *m, void *v)
-{
- const char *const str[] = LL_SBI_FLAGS;
- struct super_block *sb = m->private;
- int flags = ll_s2sbi(sb)->ll_flags;
- int i = 0;
-
- while (flags != 0) {
- if (ARRAY_SIZE(str) <= i) {
- CERROR("%s: Revise array LL_SBI_FLAGS to match sbi flags please.\n",
- ll_s2sbi(sb)->ll_fsname);
- return -EINVAL;
- }
-
- if (flags & 0x1)
- seq_printf(m, "%s ", str[i]);
- flags >>= 1;
- ++i;
- }
- seq_puts(m, "\b\n");
- return 0;
-}
-
LDEBUGFS_SEQ_FOPS_RO(ll_sbi_flags);
static ssize_t xattr_cache_show(struct kobject *kobj,
@@ -1055,7 +1029,7 @@ static ssize_t xattr_cache_store(struct kobject *kobj,
if (rc)
return rc;
- if (val && !(sbi->ll_flags & LL_SBI_XATTR_CACHE))
+ if (val && !test_bit(LL_SBI_XATTR_CACHE, sbi->ll_flags))
return -ENOTSUPP;
sbi->ll_xattr_cache_enabled = val;
@@ -1072,7 +1046,8 @@ static ssize_t tiny_write_show(struct kobject *kobj,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kset.kobj);
- return sprintf(buf, "%u\n", !!(sbi->ll_flags & LL_SBI_TINY_WRITE));
+ return scnprintf(buf, PAGE_SIZE, "%u\n",
+ test_bit(LL_SBI_TINY_WRITE, sbi->ll_flags));
}
static ssize_t tiny_write_store(struct kobject *kobj,
@@ -1091,9 +1066,9 @@ static ssize_t tiny_write_store(struct kobject *kobj,
spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_TINY_WRITE;
+ set_bit(LL_SBI_TINY_WRITE, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_TINY_WRITE;
+ clear_bit(LL_SBI_TINY_WRITE, sbi->ll_flags);
spin_unlock(&sbi->ll_lock);
return count;
@@ -1108,7 +1083,7 @@ static ssize_t parallel_dio_show(struct kobject *kobj,
ll_kset.kobj);
return snprintf(buf, PAGE_SIZE, "%u\n",
- !!(sbi->ll_flags & LL_SBI_PARALLEL_DIO));
+ test_bit(LL_SBI_PARALLEL_DIO, sbi->ll_flags));
}
static ssize_t parallel_dio_store(struct kobject *kobj,
@@ -1127,9 +1102,9 @@ static ssize_t parallel_dio_store(struct kobject *kobj,
spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_PARALLEL_DIO;
+ set_bit(LL_SBI_PARALLEL_DIO, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_PARALLEL_DIO;
+ clear_bit(LL_SBI_PARALLEL_DIO, sbi->ll_flags);
spin_unlock(&sbi->ll_lock);
return count;
@@ -1274,7 +1249,8 @@ static ssize_t fast_read_show(struct kobject *kobj,
struct ll_sb_info *sbi = container_of(kobj, struct ll_sb_info,
ll_kset.kobj);
- return sprintf(buf, "%u\n", !!(sbi->ll_flags & LL_SBI_FAST_READ));
+ return scnprintf(buf, PAGE_SIZE, "%u\n",
+ test_bit(LL_SBI_FAST_READ, sbi->ll_flags));
}
static ssize_t fast_read_store(struct kobject *kobj,
@@ -1293,9 +1269,9 @@ static ssize_t fast_read_store(struct kobject *kobj,
spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_FAST_READ;
+ set_bit(LL_SBI_FAST_READ, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_FAST_READ;
+ clear_bit(LL_SBI_FAST_READ, sbi->ll_flags);
spin_unlock(&sbi->ll_lock);
return count;
@@ -1310,7 +1286,7 @@ static ssize_t file_heat_show(struct kobject *kobj,
ll_kset.kobj);
return scnprintf(buf, PAGE_SIZE, "%u\n",
- !!(sbi->ll_flags & LL_SBI_FILE_HEAT));
+ test_bit(LL_SBI_FILE_HEAT, sbi->ll_flags));
}
static ssize_t file_heat_store(struct kobject *kobj,
@@ -1329,9 +1305,9 @@ static ssize_t file_heat_store(struct kobject *kobj,
spin_lock(&sbi->ll_lock);
if (val)
- sbi->ll_flags |= LL_SBI_FILE_HEAT;
+ set_bit(LL_SBI_FILE_HEAT, sbi->ll_flags);
else
- sbi->ll_flags &= ~LL_SBI_FILE_HEAT;
+ clear_bit(LL_SBI_FILE_HEAT, sbi->ll_flags);
spin_unlock(&sbi->ll_lock);
return count;
diff --git a/fs/lustre/llite/namei.c b/fs/lustre/llite/namei.c
index 781bb16..f942179 100644
--- a/fs/lustre/llite/namei.c
+++ b/fs/lustre/llite/namei.c
@@ -880,7 +880,7 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry,
it->it_create_mode &= ~current_umask();
if (it->it_op & IT_CREAT &&
- ll_i2sbi(parent)->ll_flags & LL_SBI_FILE_SECCTX) {
+ test_bit(LL_SBI_FILE_SECCTX, ll_i2sbi(parent)->ll_flags)) {
rc = ll_dentry_init_security(dentry, it->it_create_mode,
&dentry->d_name,
&op_data->op_file_secctx_name,
@@ -1424,7 +1424,8 @@ static int ll_create_it(struct inode *dir, struct dentry *dentry,
if (IS_ERR(inode))
return PTR_ERR(inode);
- if ((ll_i2sbi(inode)->ll_flags & LL_SBI_FILE_SECCTX) && secctx) {
+ if (test_bit(LL_SBI_FILE_SECCTX, ll_i2sbi(inode)->ll_flags) &&
+ secctx) {
/* must be done before d_instantiate, because it calls
* security_d_instantiate, which means a getxattr if security
* context is not set yet
@@ -1446,7 +1447,7 @@ static int ll_create_it(struct inode *dir, struct dentry *dentry,
return rc;
}
- if (!(ll_i2sbi(inode)->ll_flags & LL_SBI_FILE_SECCTX)) {
+ if (!test_bit(LL_SBI_FILE_SECCTX, ll_i2sbi(inode)->ll_flags)) {
rc = ll_inode_init_security(dentry, inode, dir);
if (rc)
return rc;
@@ -1562,7 +1563,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dchild,
if (S_ISDIR(mode))
ll_qos_mkdir_prep(op_data, dir);
- if (sbi->ll_flags & LL_SBI_FILE_SECCTX) {
+ if (test_bit(LL_SBI_FILE_SECCTX, sbi->ll_flags)) {
err = ll_dentry_init_security(dchild, mode, &dchild->d_name,
&op_data->op_file_secctx_name,
&op_data->op_file_secctx,
@@ -1707,7 +1708,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dchild,
if (err)
goto err_exit;
- if (sbi->ll_flags & LL_SBI_FILE_SECCTX) {
+ if (test_bit(LL_SBI_FILE_SECCTX, sbi->ll_flags)) {
/* must be done before d_instantiate, because it calls
* security_d_instantiate, which means a getxattr if security
* context is not set yet
@@ -1747,7 +1748,7 @@ static int ll_new_node(struct inode *dir, struct dentry *dchild,
}
}
- if (!(sbi->ll_flags & LL_SBI_FILE_SECCTX))
+ if (!test_bit(LL_SBI_FILE_SECCTX, sbi->ll_flags))
err = ll_inode_init_security(dchild, inode, dir);
err_exit:
if (request)
diff --git a/fs/lustre/llite/statahead.c b/fs/lustre/llite/statahead.c
index 15b95b7..4806e99 100644
--- a/fs/lustre/llite/statahead.c
+++ b/fs/lustre/llite/statahead.c
@@ -1667,7 +1667,8 @@ static int start_statahead_thread(struct inode *dir, struct dentry *dentry,
goto out;
}
- if (ll_i2sbi(parent->d_inode)->ll_flags & LL_SBI_AGL_ENABLED && agl)
+ if (test_bit(LL_SBI_AGL_ENABLED, ll_i2sbi(parent->d_inode)->ll_flags) &&
+ agl)
ll_start_agl(parent, sai);
atomic_inc(&ll_i2sbi(parent->d_inode)->ll_sa_total);
diff --git a/fs/lustre/llite/xattr.c b/fs/lustre/llite/xattr.c
index 59a1400..b67b822 100644
--- a/fs/lustre/llite/xattr.c
+++ b/fs/lustre/llite/xattr.c
@@ -67,11 +67,11 @@ static int xattr_type_filter(struct ll_sb_info *sbi,
if ((handler->flags == XATTR_ACL_ACCESS_T ||
handler->flags == XATTR_ACL_DEFAULT_T) &&
- !(sbi->ll_flags & LL_SBI_ACL))
+ !test_bit(LL_SBI_ACL, sbi->ll_flags))
return -EOPNOTSUPP;
if (handler->flags == XATTR_USER_T &&
- !(sbi->ll_flags & LL_SBI_USER_XATTR))
+ !test_bit(LL_SBI_USER_XATTR, sbi->ll_flags))
return -EOPNOTSUPP;
if (handler->flags == XATTR_TRUSTED_T &&
@@ -153,9 +153,7 @@ static int ll_xattr_set_common(const struct xattr_handler *handler,
if (rc) {
if (rc == -EOPNOTSUPP && handler->flags == XATTR_USER_T) {
LCONSOLE_INFO("Disabling user_xattr feature because it is not supported on the server\n");
- spin_lock(&sbi->ll_lock);
- sbi->ll_flags &= ~LL_SBI_USER_XATTR;
- spin_unlock(&sbi->ll_lock);
+ clear_bit(LL_SBI_USER_XATTR, sbi->ll_flags);
}
return rc;
}
@@ -431,12 +429,9 @@ int ll_xattr_list(struct inode *inode, const char *name, int type, void *buffer,
out_xattr:
if (rc == -EOPNOTSUPP && type == XATTR_USER_T) {
- LCONSOLE_INFO(
- "%s: disabling user_xattr feature because it is not supported on the server: rc = %d\n",
- sbi->ll_fsname, rc);
- spin_lock(&sbi->ll_lock);
- sbi->ll_flags &= ~LL_SBI_USER_XATTR;
- spin_unlock(&sbi->ll_lock);
+ LCONSOLE_INFO("%s: disabling user_xattr feature because it is not supported on the server: rc = %d\n",
+ sbi->ll_fsname, rc);
+ clear_bit(LL_SBI_USER_XATTR, sbi->ll_flags);
}
out:
ptlrpc_req_finished(req);
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 15/20] lustre: osc: use original cli for osc_lru_reclaim for debug msg
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (13 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 14/20] lustre: llite: harden ll_sbi ll_flags James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 16/20] lustre: obdclass: lu_ref_add() called in atomic context James Simmons
` (4 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
Before the list cleanup introduced in osc_lru_reclaim() the
variable cli was both passed in and used to scan the
cl_client_cache. After the scan was done then we use cli in
a debug message. It appears to be the original intent was to
use the original cli passed in for the debug message, not the
last scanned item. After the list cleanup patch landed now
cli can be NULL which can crash the node. The fix is to use
a separate struct client_obd variable for the scan and use
the original cli passed in for the debug message.
Fixes: ce96138f3692 ("lustre: use list_first_entry() in lustre subdirectory.")
WC-bug-id: https://jira.whamcloud.com/browse/LU-15013
Lustre-commit: 3c6a1e94c652685fac7c ("LU-15013 osc: use original cli for osc_lru_reclaim for debug msg")
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44966
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
---
fs/lustre/osc/osc_page.c | 19 ++++++++++---------
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/fs/lustre/osc/osc_page.c b/fs/lustre/osc/osc_page.c
index d471df2..cba5d02 100644
--- a/fs/lustre/osc/osc_page.c
+++ b/fs/lustre/osc/osc_page.c
@@ -696,6 +696,7 @@ static long osc_lru_reclaim(struct client_obd *cli, unsigned long npages)
{
struct lu_env *env;
struct cl_client_cache *cache = cli->cl_cache;
+ struct client_obd *scan;
int max_scans;
u16 refcheck;
long rc = 0;
@@ -735,20 +736,20 @@ static long osc_lru_reclaim(struct client_obd *cli, unsigned long npages)
max_scans = refcount_read(&cache->ccc_users) - 2;
while (--max_scans > 0 &&
- (cli = list_first_entry_or_null(&cache->ccc_lru,
- struct client_obd,
- cl_lru_osc)) != NULL) {
+ (scan = list_first_entry_or_null(&cache->ccc_lru,
+ struct client_obd,
+ cl_lru_osc)) != NULL) {
CDEBUG(D_CACHE, "%s: cli %p LRU pages: %ld, busy: %ld.\n",
- cli_name(cli), cli,
- atomic_long_read(&cli->cl_lru_in_list),
- atomic_long_read(&cli->cl_lru_busy));
+ cli_name(scan), scan,
+ atomic_long_read(&scan->cl_lru_in_list),
+ atomic_long_read(&scan->cl_lru_busy));
- list_move_tail(&cli->cl_lru_osc, &cache->ccc_lru);
- if (osc_cache_too_much(cli) > 0) {
+ list_move_tail(&scan->cl_lru_osc, &cache->ccc_lru);
+ if (osc_cache_too_much(scan) > 0) {
spin_unlock(&cache->ccc_lru_lock);
- rc = osc_lru_shrink(env, cli, npages, true);
+ rc = osc_lru_shrink(env, scan, npages, true);
spin_lock(&cache->ccc_lru_lock);
if (rc >= npages)
break;
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 16/20] lustre: obdclass: lu_ref_add() called in atomic context
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (14 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 15/20] lustre: osc: use original cli for osc_lru_reclaim for debug msg James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 17/20] lnet: Ensure round robin selection of local NIs James Simmons
` (3 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
For the native Linux client testing I turn on lu_ref
debugging. When turned on the following errors occur:
[ 2885.946815] Call Trace:
[ 2885.951240] dump_stack+0x68/0x9b
[ 2885.956523] ___might_sleep+0x205/0x260
[ 2885.962245] lu_ref_add+0x25/0x40 [obdclass]
[ 2885.968442] vvp_pgcache_current+0x101/0x1a0 [lustre]
[ 2885.975370] seq_read+0x1ab/0x3c0
and
[ 7042.102529] dump_stack+0x68/0x9b
[ 7042.107328] ___might_sleep+0x205/0x260
[ 7042.112647] lu_ref_add+0x25/0x40 [obdclass]
[ 7042.118385] mdc_lock_upcall+0x154/0x4d0 [mdc]
[ 7042.124275] mdc_enqueue_send+0x508/0x580 [mdc]
[ 7042.130225] ? mdc_lock_lvb_update+0x280/0x280 [mdc]
This is easily fixed with introducing a lu_object_ref_add_atomic()
function.
WC-bug-id: https://jira.whamcloud.com/browse/LU-15014
Lustre-commit: 5a37bc9577d4c871 ("LU-15014 obdclass: lu_ref_add() called in atomic context")
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44969
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
---
fs/lustre/include/lu_object.h | 7 +++++++
fs/lustre/llite/vvp_dev.c | 2 +-
fs/lustre/mdc/mdc_dev.c | 2 +-
3 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/fs/lustre/include/lu_object.h b/fs/lustre/include/lu_object.h
index 84e0489..146398a 100644
--- a/fs/lustre/include/lu_object.h
+++ b/fs/lustre/include/lu_object.h
@@ -829,6 +829,13 @@ static inline u32 lu_object_attr(const struct lu_object *o)
return o->lo_header->loh_attr;
}
+static inline void lu_object_ref_add_atomic(struct lu_object *o,
+ const char *scope,
+ const void *source)
+{
+ lu_ref_add_atomic(&o->lo_header->loh_reference, scope, source);
+}
+
static inline void lu_object_ref_add(struct lu_object *o,
const char *scope,
const void *source)
diff --git a/fs/lustre/llite/vvp_dev.c b/fs/lustre/llite/vvp_dev.c
index fda48bb..0c417d8 100644
--- a/fs/lustre/llite/vvp_dev.c
+++ b/fs/lustre/llite/vvp_dev.c
@@ -399,7 +399,7 @@ static struct page *vvp_pgcache_current(struct vvp_seq_private *priv)
continue;
priv->vsp_clob = lu2cl(lu_obj);
- lu_object_ref_add(lu_obj, "dump", current);
+ lu_object_ref_add_atomic(lu_obj, "dump", current);
priv->vsp_page_index = 0;
}
diff --git a/fs/lustre/mdc/mdc_dev.c b/fs/lustre/mdc/mdc_dev.c
index 4777b47..b2f60ea 100644
--- a/fs/lustre/mdc/mdc_dev.c
+++ b/fs/lustre/mdc/mdc_dev.c
@@ -461,7 +461,7 @@ static void mdc_lock_granted(const struct lu_env *env, struct osc_lock *oscl,
/* lock reference taken by ldlm_handle2lock_long() is
* owned by osc_lock and released in osc_lock_detach()
*/
- lu_ref_add(&dlmlock->l_reference, "osc_lock", oscl);
+ lu_ref_add_atomic(&dlmlock->l_reference, "osc_lock", oscl);
oscl->ols_has_ref = 1;
LASSERT(!oscl->ols_dlmlock);
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 17/20] lnet: Ensure round robin selection of local NIs
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (15 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 16/20] lustre: obdclass: lu_ref_add() called in atomic context James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 18/20] lnet: Ensure round robin selection of peer NIs James Simmons
` (2 subsequent siblings)
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown
Cc: Chris Horn, Lustre Development List
From: Chris Horn <chris.horn@hpe.com>
Use the net sequence number to set the NI sequence number to ensure
round robin selection of NIs on each net.
HPE-bug-id: LUS-10349
WC-bug-id: https://jira.whamcloud.com/browse/LU-13575
Lustre-commit: a18c4a16246e6185 ("LU-13575 lnet: Ensure round robin selection of local NIs")
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Reviewed-on: https://review.whamcloud.com/45003
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
net/lnet/lnet/lib-move.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/lnet/lnet/lib-move.c b/net/lnet/lnet/lib-move.c
index b9b322a..0f26001 100644
--- a/net/lnet/lnet/lib-move.c
+++ b/net/lnet/lnet/lib-move.c
@@ -1652,8 +1652,8 @@ void lnet_usr_translate_stats(struct lnet_ioctl_element_msg_stats *msg_stats,
*/
best_lpni->lpni_seq++;
best_lpni->lpni_peer_net->lpn_seq++;
- best_ni->ni_seq++;
best_ni->ni_net->net_seq++;
+ best_ni->ni_seq = best_ni->ni_net->net_seq;
CDEBUG(D_NET,
"%s NI seq info: [%d:%d:%d:%u] %s LPNI seq info [%d:%d:%d:%u]\n",
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 18/20] lnet: Ensure round robin selection of peer NIs
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (16 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 17/20] lnet: Ensure round robin selection of local NIs James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 19/20] lustre: mdc: update max_easize on reconnect James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 20/20] lnet: include linux/ethtool.h James Simmons
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown
Cc: Chris Horn, Lustre Development List
From: Chris Horn <chris.horn@hpe.com>
Use the peer net sequence number to set the peer NI sequence number to
ensure round robin selection of peer NIs on each peer net.
HPE-bug-id: LUS-10349
WC-bug-id: https://jira.whamcloud.com/browse/LU-13575
Lustre-commit: c51763948abfdbdc8 ("LU-13575 lnet: Ensure round robin selection of peer NIs")
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Reviewed-on: https://review.whamcloud.com/45004
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
net/lnet/lnet/lib-move.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/lnet/lnet/lib-move.c b/net/lnet/lnet/lib-move.c
index 0f26001..2b38480 100644
--- a/net/lnet/lnet/lib-move.c
+++ b/net/lnet/lnet/lib-move.c
@@ -1650,8 +1650,8 @@ void lnet_usr_translate_stats(struct lnet_ioctl_element_msg_stats *msg_stats,
* local ni and local net so that we pick the next ones
* in Round Robin.
*/
- best_lpni->lpni_seq++;
best_lpni->lpni_peer_net->lpn_seq++;
+ best_lpni->lpni_seq = best_lpni->lpni_peer_net->lpn_seq;
best_ni->ni_net->net_seq++;
best_ni->ni_seq = best_ni->ni_net->net_seq;
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 19/20] lustre: mdc: update max_easize on reconnect
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (17 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 18/20] lnet: Ensure round robin selection of peer NIs James Simmons
@ 2021-10-11 17:40 ` James Simmons
2021-10-11 17:40 ` [lustre-devel] [PATCH 20/20] lnet: include linux/ethtool.h James Simmons
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown
Cc: Sergey Cheremencev, Lustre Development List
From: Sergey Cheremencev <sergey.cheremencev@hpe.com>
If MDS was restarted to enable ea_inode, clients should get new
max_easize value. However, cl_max_mds_easize is not updated. This may
cause lfs getstripe to fail if file has huge stripe number
(2000 for example):
*** Error in `lfs': free(): invalid pointer: 0x0000000000de09d0 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x81299)[0x7f0623c03299]
/lib64/libc.so.6(closedir+0xd)[0x7f0623c42ddd]
/lib/liblustreapi.so.1(+0xa557)[0x7f06248b5557]
/lib/liblustreapi.so.1(+0xad74)[0x7f06248b5d74]
lfs[0x4105b3]
/lib/liblustreapi.so.1(Parser_execarg+0x51)[0x7f06248c88e1]
lfs[0x40448e]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f0623ba4555]
lfs[0x4044fc]
HPE-bug-id: LUS-9478
WC-bug-id: https://jira.whamcloud.com/browse/LU-15040
Lustre-commit: 5f15be0edea5c2d31 ("LU-15040 mdc: update max_easize on reconnect")
Reviewed-on: https://es-gerrit.dev.cray.com/158100
Reviewed-by: Alexey Lyashkov <c17817@cray.com>
Reviewed-by: Nikitas Angelinas <nangelinas@cray.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45073
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
fs/lustre/mdc/mdc_request.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/fs/lustre/mdc/mdc_request.c b/fs/lustre/mdc/mdc_request.c
index 8b94f6c..626f493 100644
--- a/fs/lustre/mdc/mdc_request.c
+++ b/fs/lustre/mdc/mdc_request.c
@@ -2720,6 +2720,7 @@ static int mdc_import_event(struct obd_device *obd, struct obd_import *imp,
if (OCD_HAS_FLAG(ocd, GRANT))
osc_init_grant(cli, ocd);
+ md_init_ea_size(obd->obd_self_export, ocd->ocd_max_easize, 0);
rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD);
break;
}
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread* [lustre-devel] [PATCH 20/20] lnet: include linux/ethtool.h
2021-10-11 17:40 [lustre-devel] [PATCH 00/20] lustre: sync to OpenSFS Oct 11, 2021 James Simmons
` (18 preceding siblings ...)
2021-10-11 17:40 ` [lustre-devel] [PATCH 19/20] lustre: mdc: update max_easize on reconnect James Simmons
@ 2021-10-11 17:40 ` James Simmons
19 siblings, 0 replies; 21+ messages in thread
From: James Simmons @ 2021-10-11 17:40 UTC (permalink / raw)
To: Andreas Dilger, Oleg Drokin, NeilBrown; +Cc: Lustre Development List
From: Jian Yu <yujian@whamcloud.com>
Kernel 5.11+ removes including linux/ethtool.h from
linux/netdevice.h, which caused the following build error:
dereferencing pointer to incomplete type 'const struct ethtool_ops'
This patch fixes the above issue by adding the include into
the file that uses the structure.
WC-bug-id: https://jira.whamcloud.com/browse/LU-15052
Lustre-commit: b2503cf65a0cb3583 ("LU-15052 lnet: include linux/ethtool.h")
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/45109
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
net/lnet/klnds/o2iblnd/o2iblnd.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/net/lnet/klnds/o2iblnd/o2iblnd.c b/net/lnet/klnds/o2iblnd/o2iblnd.c
index fd807c2..36d26b2 100644
--- a/net/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/net/lnet/klnds/o2iblnd/o2iblnd.c
@@ -36,9 +36,11 @@
#include <asm/div64.h>
#include <asm/page.h>
-#include "o2iblnd.h"
+#include <linux/ethtool.h>
#include <linux/inetdevice.h>
+#include "o2iblnd.h"
+
static struct lnet_lnd the_o2iblnd;
struct kib_data kiblnd_data;
--
1.8.3.1
_______________________________________________
lustre-devel mailing list
lustre-devel@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
^ permalink raw reply related [flat|nested] 21+ messages in thread