From: Ilya Dryomov <idryomov@gmail.com>
To: ceph-devel@vger.kernel.org
Subject: [PATCH 5/6] libceph: variable-sized ceph_object_id
Date: Thu, 19 May 2016 16:06:41 +0200 [thread overview]
Message-ID: <1463666802-25140-6-git-send-email-idryomov@gmail.com> (raw)
In-Reply-To: <1463666802-25140-1-git-send-email-idryomov@gmail.com>
Currently ceph_object_id can hold object names of up to 100
(CEPH_MAX_OID_NAME_LEN) characters. This is enough for all use cases,
expect one - long rbd image names:
- a format 1 header is named "<imgname>.rbd"
- an object that points to a format 2 header is named "rbd_id.<imgname>"
We operate on these potentially long-named objects during rbd map, and,
for format 1 images, during header refresh. (A format 2 header name is
a small system-generated string.)
Lift this 100 character limit by making ceph_object_id be able to point
to an externally-allocated string. Apart from being able to work with
almost arbitrarily-long named objects, this allows us to reduce the
size of ceph_object_id from >100 bytes to 64 bytes.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
---
drivers/block/rbd.c | 8 +++-
fs/ceph/addr.c | 6 +--
fs/ceph/file.c | 2 +-
fs/ceph/ioctl.c | 2 +-
include/linux/ceph/osdmap.h | 62 ++++++++++++++++++------------
net/ceph/debugfs.c | 2 +-
net/ceph/osd_client.c | 16 +++++---
net/ceph/osdmap.c | 93 ++++++++++++++++++++++++++++++++++++++++++++-
8 files changed, 150 insertions(+), 41 deletions(-)
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index a90c9291cbf9..a41a44d08397 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -1965,7 +1965,9 @@ static struct ceph_osd_request *rbd_osd_req_create(
osd_req->r_priv = obj_request;
osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout);
- ceph_oid_set_name(&osd_req->r_base_oid, obj_request->object_name);
+ if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s",
+ obj_request->object_name))
+ goto fail;
if (ceph_osdc_alloc_messages(osd_req, GFP_NOIO))
goto fail;
@@ -2017,7 +2019,9 @@ rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request)
osd_req->r_priv = obj_request;
osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout);
- ceph_oid_set_name(&osd_req->r_base_oid, obj_request->object_name);
+ if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s",
+ obj_request->object_name))
+ goto fail;
if (ceph_osdc_alloc_messages(osd_req, GFP_NOIO))
goto fail;
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
index 6fee7e0b8931..6f28dd9bacb2 100644
--- a/fs/ceph/addr.c
+++ b/fs/ceph/addr.c
@@ -1758,9 +1758,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool)
rd_req->r_flags = CEPH_OSD_FLAG_READ;
osd_req_op_init(rd_req, 0, CEPH_OSD_OP_STAT, 0);
rd_req->r_base_oloc.pool = pool;
- snprintf(rd_req->r_base_oid.name, sizeof(rd_req->r_base_oid.name),
- "%llx.00000000", ci->i_vino.ino);
- rd_req->r_base_oid.name_len = strlen(rd_req->r_base_oid.name);
+ ceph_oid_printf(&rd_req->r_base_oid, "%llx.00000000", ci->i_vino.ino);
err = ceph_osdc_alloc_messages(rd_req, GFP_NOFS);
if (err)
@@ -1777,7 +1775,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool)
CEPH_OSD_FLAG_ACK | CEPH_OSD_FLAG_ONDISK;
osd_req_op_init(wr_req, 0, CEPH_OSD_OP_CREATE, CEPH_OSD_OP_FLAG_EXCL);
wr_req->r_base_oloc.pool = pool;
- wr_req->r_base_oid = rd_req->r_base_oid;
+ ceph_oid_copy(&wr_req->r_base_oid, &rd_req->r_base_oid);
err = ceph_osdc_alloc_messages(wr_req, GFP_NOFS);
if (err)
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 5d46d106bbb7..9d470397e249 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -715,7 +715,7 @@ static void ceph_aio_retry_work(struct work_struct *work)
CEPH_OSD_FLAG_ONDISK |
CEPH_OSD_FLAG_WRITE;
req->r_base_oloc = orig_req->r_base_oloc;
- req->r_base_oid = orig_req->r_base_oid;
+ ceph_oid_copy(&req->r_base_oid, &orig_req->r_base_oid);
ret = ceph_osdc_alloc_messages(req, GFP_NOFS);
if (ret) {
diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index f851d8d70158..db296709784a 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -213,7 +213,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
ceph_ino(inode), dl.object_no);
oloc.pool = ceph_file_layout_pg_pool(ci->i_layout);
- ceph_oid_set_name(&oid, dl.object_name);
+ ceph_oid_printf(&oid, "%s", dl.object_name);
r = ceph_oloc_oid_to_pg(osdc->osdmap, &oloc, &oid, &pgid);
if (r < 0) {
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index e55c08bc3a96..777a29412706 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -64,11 +64,47 @@ struct ceph_object_locator {
*/
#define CEPH_MAX_OID_NAME_LEN 100
+/*
+ * 51-char inline_name is long enough for all cephfs and all but one
+ * rbd requests: <imgname> in "<imgname>.rbd"/"rbd_id.<imgname>" can be
+ * arbitrarily long (~PAGE_SIZE). It's done once during rbd map; all
+ * other rbd requests fit into inline_name.
+ *
+ * Makes ceph_object_id 64 bytes on 64-bit.
+ */
+#define CEPH_OID_INLINE_LEN 52
+
+/*
+ * Both inline and external buffers have space for a NUL-terminator,
+ * which is carried around. It's not required though - RADOS object
+ * names don't have to be NUL-terminated and may contain NULs.
+ */
struct ceph_object_id {
- char name[CEPH_MAX_OID_NAME_LEN];
+ char *name;
+ char inline_name[CEPH_OID_INLINE_LEN];
int name_len;
};
+static inline void ceph_oid_init(struct ceph_object_id *oid)
+{
+ oid->name = oid->inline_name;
+ oid->name_len = 0;
+}
+
+static inline bool ceph_oid_empty(const struct ceph_object_id *oid)
+{
+ return oid->name == oid->inline_name && !oid->name_len;
+}
+
+void ceph_oid_copy(struct ceph_object_id *dest,
+ const struct ceph_object_id *src);
+__printf(2, 3)
+void ceph_oid_printf(struct ceph_object_id *oid, const char *fmt, ...);
+__printf(3, 4)
+int ceph_oid_aprintf(struct ceph_object_id *oid, gfp_t gfp,
+ const char *fmt, ...);
+void ceph_oid_destroy(struct ceph_object_id *oid);
+
struct ceph_pg_mapping {
struct rb_node node;
struct ceph_pg pgid;
@@ -113,30 +149,6 @@ struct ceph_osdmap {
int crush_scratch_ary[CEPH_PG_MAX_SIZE * 3];
};
-static inline void ceph_oid_set_name(struct ceph_object_id *oid,
- const char *name)
-{
- int len;
-
- len = strlen(name);
- if (len > sizeof(oid->name)) {
- WARN(1, "ceph_oid_set_name '%s' len %d vs %zu, truncating\n",
- name, len, sizeof(oid->name));
- len = sizeof(oid->name);
- }
-
- memcpy(oid->name, name, len);
- oid->name_len = len;
-}
-
-static inline void ceph_oid_copy(struct ceph_object_id *dest,
- struct ceph_object_id *src)
-{
- BUG_ON(src->name_len > sizeof(dest->name));
- memcpy(dest->name, src->name, src->name_len);
- dest->name_len = src->name_len;
-}
-
static inline int ceph_osd_exists(struct ceph_osdmap *map, int osd)
{
return osd >= 0 && osd < map->max_osd &&
diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c
index b902fbc7863e..6f8413293d15 100644
--- a/net/ceph/debugfs.c
+++ b/net/ceph/debugfs.c
@@ -161,7 +161,7 @@ static int osdc_show(struct seq_file *s, void *pp)
req->r_osd ? req->r_osd->o_osd : -1,
req->r_pgid.pool, req->r_pgid.seed);
- seq_printf(s, "%.*s", req->r_base_oid.name_len,
+ seq_printf(s, "%*pE", req->r_base_oid.name_len,
req->r_base_oid.name);
if (req->r_reassert_version.epoch)
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index 75e27bd3d372..95910aed8e2e 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -334,7 +334,10 @@ static void ceph_osdc_release_request(struct kref *kref)
for (which = 0; which < req->r_num_ops; which++)
osd_req_op_data_release(req, which);
+ ceph_oid_destroy(&req->r_base_oid);
+ ceph_oid_destroy(&req->r_target_oid);
ceph_put_snap_context(req->r_snapc);
+
if (req->r_mempool)
mempool_free(req, req->r_osdc->req_mempool);
else if (req->r_num_ops <= CEPH_OSD_SLAB_OPS)
@@ -401,7 +404,9 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
INIT_LIST_HEAD(&req->r_req_lru_item);
INIT_LIST_HEAD(&req->r_osd_item);
+ ceph_oid_init(&req->r_base_oid);
req->r_base_oloc.pool = -1;
+ ceph_oid_init(&req->r_target_oid);
req->r_target_oloc.pool = -1;
dout("%s req %p\n", __func__, req);
@@ -415,6 +420,8 @@ int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp)
struct ceph_msg *msg;
int msg_size;
+ WARN_ON(ceph_oid_empty(&req->r_base_oid));
+
/* create request message */
msg_size = 4 + 4 + 4; /* client_inc, osdmap_epoch, flags */
msg_size += 4 + 4 + 4 + 8; /* mtime, reassert_version */
@@ -859,10 +866,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc,
}
req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout);
-
- snprintf(req->r_base_oid.name, sizeof(req->r_base_oid.name),
- "%llx.%08llx", vino.ino, objnum);
- req->r_base_oid.name_len = strlen(req->r_base_oid.name);
+ ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum);
r = ceph_osdc_alloc_messages(req, GFP_NOFS);
if (r)
@@ -1410,7 +1414,7 @@ static int __calc_request_pg(struct ceph_osdmap *osdmap,
req->r_target_oloc = req->r_base_oloc; /* struct */
need_check_tiering = true;
}
- if (req->r_target_oid.name_len == 0) {
+ if (ceph_oid_empty(&req->r_target_oid)) {
ceph_oid_copy(&req->r_target_oid, &req->r_base_oid);
need_check_tiering = true;
}
@@ -2501,7 +2505,7 @@ void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
/* oid */
ceph_encode_32(&p, req->r_base_oid.name_len);
memcpy(p, req->r_base_oid.name, req->r_base_oid.name_len);
- dout("oid '%.*s' len %d\n", req->r_base_oid.name_len,
+ dout("oid %*pE len %d\n", req->r_base_oid.name_len,
req->r_base_oid.name, req->r_base_oid.name_len);
p += req->r_base_oid.name_len;
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 243574c8cf33..4668b871ca47 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -1381,8 +1381,99 @@ bad:
return ERR_PTR(err);
}
+void ceph_oid_copy(struct ceph_object_id *dest,
+ const struct ceph_object_id *src)
+{
+ WARN_ON(!ceph_oid_empty(dest));
+
+ if (src->name != src->inline_name) {
+ /* very rare, see ceph_object_id definition */
+ dest->name = kmalloc(src->name_len + 1,
+ GFP_NOIO | __GFP_NOFAIL);
+ }
+ memcpy(dest->name, src->name, src->name_len + 1);
+ dest->name_len = src->name_len;
+}
+EXPORT_SYMBOL(ceph_oid_copy);
+static __printf(2, 0)
+int oid_printf_vargs(struct ceph_object_id *oid, const char *fmt, va_list ap)
+{
+ int len;
+
+ WARN_ON(!ceph_oid_empty(oid));
+
+ len = vsnprintf(oid->inline_name, sizeof(oid->inline_name), fmt, ap);
+ if (len >= sizeof(oid->inline_name))
+ return len;
+
+ oid->name_len = len;
+ return 0;
+}
+
+/*
+ * If oid doesn't fit into inline buffer, BUG.
+ */
+void ceph_oid_printf(struct ceph_object_id *oid, const char *fmt, ...)
+{
+ va_list ap;
+
+ va_start(ap, fmt);
+ BUG_ON(oid_printf_vargs(oid, fmt, ap));
+ va_end(ap);
+}
+EXPORT_SYMBOL(ceph_oid_printf);
+
+static __printf(3, 0)
+int oid_aprintf_vargs(struct ceph_object_id *oid, gfp_t gfp,
+ const char *fmt, va_list ap)
+{
+ va_list aq;
+ int len;
+
+ va_copy(aq, ap);
+ len = oid_printf_vargs(oid, fmt, aq);
+ va_end(aq);
+
+ if (len) {
+ char *external_name;
+
+ external_name = kmalloc(len + 1, gfp);
+ if (!external_name)
+ return -ENOMEM;
+
+ oid->name = external_name;
+ WARN_ON(vsnprintf(oid->name, len + 1, fmt, ap) != len);
+ oid->name_len = len;
+ }
+
+ return 0;
+}
+
+/*
+ * If oid doesn't fit into inline buffer, allocate.
+ */
+int ceph_oid_aprintf(struct ceph_object_id *oid, gfp_t gfp,
+ const char *fmt, ...)
+{
+ va_list ap;
+ int ret;
+
+ va_start(ap, fmt);
+ ret = oid_aprintf_vargs(oid, gfp, fmt, ap);
+ va_end(ap);
+
+ return ret;
+}
+EXPORT_SYMBOL(ceph_oid_aprintf);
+
+void ceph_oid_destroy(struct ceph_object_id *oid)
+{
+ if (oid->name != oid->inline_name)
+ kfree(oid->name);
+}
+EXPORT_SYMBOL(ceph_oid_destroy);
/*
* calculate file layout from given offset, length.
@@ -1474,7 +1565,7 @@ int ceph_oloc_oid_to_pg(struct ceph_osdmap *osdmap,
pg_out->seed = ceph_str_hash(pi->object_hash, oid->name,
oid->name_len);
- dout("%s '%.*s' pgid %llu.%x\n", __func__, oid->name_len, oid->name,
+ dout("%s %*pE pgid %llu.%x\n", __func__, oid->name_len, oid->name,
pg_out->pool, pg_out->seed);
return 0;
}
--
2.4.3
next prev parent reply other threads:[~2016-05-19 14:07 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-19 14:06 [PATCH 0/6] libceph: get rid of CEPH_MAX_OID_NAME_LEN limit Ilya Dryomov
2016-05-19 14:06 ` [PATCH 1/6] libceph: make ceph_osdc_put_request() accept NULL Ilya Dryomov
2016-05-19 14:06 ` [PATCH 2/6] libceph: grab snapc in ceph_osdc_alloc_request() Ilya Dryomov
2016-05-19 14:06 ` [PATCH 3/6] libceph: move message allocation out of ceph_osdc_alloc_request() Ilya Dryomov
2016-05-19 14:06 ` [PATCH 4/6] libceph: change how osd_op_reply message size is calculated Ilya Dryomov
2016-05-19 14:06 ` Ilya Dryomov [this message]
2016-05-19 14:06 ` [PATCH 6/6] rbd: use header_oid instead of header_name Ilya Dryomov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1463666802-25140-6-git-send-email-idryomov@gmail.com \
--to=idryomov@gmail.com \
--cc=ceph-devel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.