osd_req_flags is overly general and doesn't suit its only user
(read_from_replica option) well:
- applying osd_req_flags in account_request() affects all OSD
requests, including linger (i.e. watch and notify). However,
linger requests should always go to the primary even though
some of them are reads (e.g. notify has side effects but it
is a read because it doesn't result in mutation on the OSDs).
- calls to class methods that are reads are allowed to go to
the replica, but most such calls issued for "rbd map" and/or
exclusive lock transitions are requested to be resent to the
primary via EAGAIN, doubling the latency.
Get rid of global osd_req_flags and set read_from_replica flag
only on specific OSD requests instead.
Fixes: 8ad44d5e0d ("libceph: read_from_replica option")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Allow hinting to bluestore if the data should/should not be compressed.
The default is to not hint (compression_hint=none).
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Allow indicating future I/O pattern via flags. This is supported since
Kraken (and bluestore persists flags together with expected_object_size
and expected_write_size).
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
rbd_notify_op_lock() isn't interested in a notify reply. Instead of
accepting that page vector just to free it, have watch-notify code take
care of it.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
rbd_dev->opts is used to distinguish between the image that is being
mapped and a parent. However, because we no longer establish watch for
read-only mappings, this test is imprecise and results in unnecessary
rbd_unregister_watch() calls.
Make it consistent with need_watch in rbd_dev_image_probe().
Fixes: b9ef2b8858 ("rbd: don't establish watch for read-only mappings")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
rbd_dev_unprobe() is supposed to undo most of rbd_dev_image_probe(),
including rbd_dev_header_info(), which means that rbd_dev_header_info()
isn't supposed to be called after rbd_dev_unprobe().
However, rbd_dev_image_release() calls rbd_dev_unprobe() before
rbd_unregister_watch(). This is racy because a header update notify
can sneak in:
"rbd unmap" thread ceph-watch-notify worker
rbd_dev_image_release()
rbd_dev_unprobe()
free and zero out header
rbd_watch_cb()
rbd_dev_refresh()
rbd_dev_header_info()
read in header
The same goes for "rbd map" because rbd_dev_image_probe() calls
rbd_dev_unprobe() on errors. In both cases this results in a memory
leak.
Fixes: fd22aef8b4 ("rbd: move rbd_unregister_watch() call into rbd_dev_image_release()")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
rbd_unregister_watch() flushes notifies and therefore cannot be called
under header_rwsem because a header update notify takes header_rwsem to
synchronize with "rbd map". If mapping an image fails after the watch
is established and a header update notify sneaks in, we deadlock when
erroring out from rbd_dev_image_probe().
Move watch registration and unregistration out of the critical section.
The only reason they were put there was to make header_rwsem management
slightly more obvious.
Fixes: 811c668877 ("rbd: fix rbd map vs notify races")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Allocate one queue per CPU and get a performance boost from
higher parallelism.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Avoid making allocations for !IMG_REQ_CHILD image requests. Only
IMG_REQ_CHILD image requests need to be freed now.
Move the initial request checks to rbd_queue_rq(). Unfortunately we
can't fill the image request and kick the state machine directly from
rbd_queue_rq() because ->queue_rq() isn't allowed to block.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Currently header_rwsem is acquired twice: once in rbd_dev_parent_get()
when the image request is being created and then in rbd_queue_workfn()
to capture mapping_size and snapc. Introduce rbd_img_capture_header()
and move image request allocation so that header_rwsem can be acquired
just once.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
The reference counter is never increased, so we can as well call
rbd_img_request_destroy() directly and drop the kref.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
IMG_REQ_LAYERED is set in rbd_img_request_create(), and tested and
cleared in rbd_img_request_destroy() when the image request is about to
be destroyed. The barriers are unnecessary.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Pull vfs file system parameter updates from Al Viro:
"Saner fs_parser.c guts and data structures. The system-wide registry
of syntax types (string/enum/int32/oct32/.../etc.) is gone and so is
the horror switch() in fs_parse() that would have to grow another case
every time something got added to that system-wide registry.
New syntax types can be added by filesystems easily now, and their
namespace is that of functions - not of system-wide enum members. IOW,
they can be shared or kept private and if some turn out to be widely
useful, we can make them common library helpers, etc., without having
to do anything whatsoever to fs_parse() itself.
And we already get that kind of requests - the thing that finally
pushed me into doing that was "oh, and let's add one for timeouts -
things like 15s or 2h". If some filesystem really wants that, let them
do it. Without somebody having to play gatekeeper for the variants
blessed by direct support in fs_parse(), TYVM.
Quite a bit of boilerplate is gone. And IMO the data structures make a
lot more sense now. -200LoC, while we are at it"
* 'merge.nfs-fs_parse.1' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (25 commits)
tmpfs: switch to use of invalfc()
cgroup1: switch to use of errorfc() et.al.
procfs: switch to use of invalfc()
hugetlbfs: switch to use of invalfc()
cramfs: switch to use of errofc() et.al.
gfs2: switch to use of errorfc() et.al.
fuse: switch to use errorfc() et.al.
ceph: use errorfc() and friends instead of spelling the prefix out
prefix-handling analogues of errorf() and friends
turn fs_param_is_... into functions
fs_parse: handle optional arguments sanely
fs_parse: fold fs_parameter_desc/fs_parameter_spec
fs_parser: remove fs_parameter_description name field
add prefix to fs_context->log
ceph_parse_param(), ceph_parse_mon_ips(): switch to passing fc_log
new primitive: __fs_parse()
switch rbd and libceph to p_log-based primitives
struct p_log, variants of warnf() et.al. taking that one instead
teach logfc() to handle prefices, give it saner calling conventions
get rid of cg_invalf()
...
Unused now.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
fs_parse() analogue taking p_log instead of fs_context.
fs_parse() turned into a wrapper, callers in ceph_common and rbd
switched to __fs_parse().
As the result, fs_parse() never gets NULL fs_context and neither
do fs_context-based logging primitives
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
As it is, vfs_parse_fs_string() makes "foo" and "foo=" indistinguishable;
both get fs_value_is_string for ->type and NULL for ->string. To make
it even more unpleasant, that combination is impossible to produce with
fsconfig().
Much saner rules would be
"foo" => fs_value_is_flag, NULL
"foo=" => fs_value_is_string, ""
"foo=bar" => fs_value_is_string, "bar"
All cases are distinguishable, all results are expressable by fsconfig(),
->has_value checks are much simpler that way (to the point of the field
being useless) and quite a few regressions go away (gfs2 has no business
accepting -o nodebug=, for example).
Partially based upon patches from Miklos.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The rbd driver already provides additional information in sysfs
under /sys/bus/rbd, so we should set the 'device' link in the block
device to reference this information.
Signed-off-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
gcc -O3 warns about a dummy variable that is passed
down into rbd_img_fill_nodata without being initialized:
drivers/block/rbd.c: In function 'rbd_img_fill_nodata':
drivers/block/rbd.c:2573:13: error: 'dummy' is used uninitialized in this function [-Werror=uninitialized]
fctx->iter = *fctx->pos;
Since this is a dummy, I assume the warning is harmless, but
it's better to initialize it anyway and avoid the warning.
Fixes: mmtom ("init/Kconfig: enable -O3 for all arches")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Convert the ceph filesystem to the new internal mount API as the old
one will be obsoleted and removed. This allows greater flexibility in
communication of mount parameters between userspace, the VFS and the
filesystem.
See Documentation/filesystems/mount_api.txt for more information.
[ Numerous string handling, leak and regression fixes; rbd conversion
was particularly broken and had to be redone almost from scratch. ]
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
For a read-only mapping, ask for a set of features that make the image
only unwritable rather than both unreadable and unwritable by a client
that doesn't understand them. As of today, the difference between them
for krbd is journaling (JOURNALING) and live migration (MIGRATING).
get_features method supports read_only parameter since hammer, ceph.git
commit 6176ec5fde2a ("librbd: differentiate between R/O vs R/W RBD
features").
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Since infernalis, ceph.git commit 281f87f9ee52 ("cls_rbd: get_features
on snapshots returns HEAD image features"), querying and checking that
is pointless. Userspace support for manipulating image features after
image creation came also in infernalis, so a snapshot with a different
set of features wasn't ever possible.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
RBD_DEV_FLAG_EXISTS check in rbd_queue_workfn() is racy and leads to
inconsistent behaviour. If the object (or its snapshot) isn't there,
the OSD returns ENOENT. A read submitted before the snapshot removal
notification is processed would be zero-filled and ended with status
OK, while future reads would be failed with IOERR. It also doesn't
handle a case when an image that is mapped read-only is removed.
On top of this, because watch is no longer established for read-only
mappings, we no longer get notifications, so rbd_exists_validate() is
effectively dead code. While failing requests rather than returning
zeros is a good thing, RBD_DEV_FLAG_EXISTS is not it.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
With exclusive lock out of the way, watch is the only thing left that
prevents a read-only mapping from being used with read-only OSD caps.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
A read-only mapping should be usable with read-only OSD caps, so
neither the header lock nor the object map lock can be acquired.
Unfortunately, this means that images mapped read-only lose the
advantage of the object map.
Snapshots, however, can take advantage of the object map without
any exclusionary locks, so if the object map is desired, snapshot
the image and map the snapshot instead of the image.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
If an image is mapped read-only, don't allow setting its partition(s)
to read-write via BLKROSET: with the previous patch all writes to such
images are failed anyway.
If an image is mapped read-write, its partition(s) can be set to
read-only (and back to read-write) as before. Note that at the rbd
level the image will remain writeable: anything sent down by the block
layer will be executed, including any write from internal kernel users.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Even though -o ro/-o read_only/--read-only options are very old, we
have never really treated them seriously (on par with snapshots). As
a first step, fail writes to images mapped read-only just like we do
for snapshots.
We need this check in rbd because the block layer basically ignores
read-only setting, see commit a32e236eb9 ("Partially revert "block:
fail op_is_write() requests to read-only partitions"").
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
rbd_dev->opts is not available for parent images, making checking
rbd_dev->opts->read_only in various places (rbd_dev_image_probe(),
need_exclusive_lock(), use_object_map() in the following patches)
harder than it needs to be.
Keeping rbd_dev_image_probe() in mind, move the initialization in
do_rbd_add() up. snap_id isn't filled in at that point, so replace
rbd_is_snap() with a snap_name comparison.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
There is a spelling mistake in a debug message. Fix it.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Some versions of gcc (so far 6.3 and 7.4) throw a warning:
drivers/block/rbd.c: In function 'rbd_object_map_callback':
drivers/block/rbd.c:2124:21: warning: 'current_state' may be used uninitialized in this function [-Wmaybe-uninitialized]
(current_state == OBJECT_EXISTS && state == OBJECT_EXISTS_CLEAN))
drivers/block/rbd.c:2092:23: note: 'current_state' was declared here
u8 state, new_state, current_state;
^~~~~~~~~~~~~
It's bogus because all current_state accesses are guarded by
has_current_state.
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
There is a warning message in my test with below steps:
# rbd bench --io-type write --io-size 4K --io-threads 1 --io-pattern rand test &
# sleep 5
# pkill -9 rbd
# rbd map test &
# sleep 5
# pkill rbd
The reason is that the rbd_add_acquire_lock() is interruptable,
that means, when we kill the waiting on ->acquire_wait, the lock_dwork
could be still running.
1. do_rbd_add() 2. lock_dwork
rbd_add_acquire_lock()
- queue_delayed_work()
lock_dwork queued
- wait_for_completion_killable_timeout() <-- kill happen
rbd_dev_image_unlock() <-- UNLOCKED now, nothing to do.
rbd_dev_device_release()
rbd_dev_image_release()
- ...
lock successed here
- cancel_delayed_work_sync(&rbd_dev->lock_dwork)
Then when we reach the rbd_dev_free(), WARN_ON is triggered because
lock_state is not RBD_LOCK_STATE_UNLOCKED.
To fix it, this commit make sure the lock_dwork was finished before
calling rbd_dev_image_unlock().
On the other hand, this would not happend in do_rbd_remove(), because
after rbd mapped, lock_dwork will only be queued for IO request, and
request will continue unless lock_dwork finished. when we call
rbd_dev_image_unlock() in do_rbd_remove(), all requests are done.
That means, lock_state should not be locked again after
rbd_dev_image_unlock().
[ Cancel lock_dwork in rbd_add_acquire_lock(), only if the wait is
interrupted. ]
Fixes: 637cd06053 ("rbd: new exclusive lock wait/wake code")
Signed-off-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Make it more informative: log op_type, offset and length for block
layer requests and initiating obj_req for child requests.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
rbd_dev_image_id() allocates space for length but passes a smaller
value to rbd_obj_method_sync(). rbd_dev_v2_object_prefix() doesn't
allocate space for length. Fix both to be consistent.
Signed-off-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
The parent image is read only up to the overlap point, the rest of
the buffer should be zeroed. This snuck in because as it turns out
the overlap test case has not been triggering this code path for
a while now.
Fixes: a9b67e6994 ("rbd: replace obj_req->tried_parent with obj_req->read_state")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
- support for rbd object-map and fast-diff features (myself). This
will speed up reads, discards and things like snap diffs on sparse
images.
- ceph.snap.btime vxattr to expose snapshot creation time (David
Disseldorp). This will be used to integrate with "Restore Previous
Versions" feature added in Windows 7 for folks who reexport ceph
through SMB.
- security xattrs for ceph (Zheng Yan). Only selinux is supported
for now due to the limitations of ->dentry_init_security().
- support for MSG_ADDR2, FS_BTIME and FS_CHANGE_ATTR features (Jeff
Layton). This is actually a single feature bit which was missing
because of the filesystem pieces. With this in, the kernel client
will finally be reported as "luminous" by "ceph features" -- it is
still being reported as "jewel" even though all required Luminous
features were implemented in 4.13.
- stop NULL-terminating ceph vxattrs (Jeff Layton). The convention
with xattrs is to not terminate and this was causing inconsistencies
with ceph-fuse.
- change filesystem time granularity from 1 us to 1 ns, again fixing
an inconsistency with ceph-fuse (Luis Henriques).
On top of this there are some additional dentry name handling and cap
flushing fixes from Zheng. Finally, Jeff is formally taking over for
Zheng as the filesystem maintainer.
-----BEGIN PGP SIGNATURE-----
iQFHBAABCAAxFiEEydHwtzie9C7TfviiSn/eOAIR84sFAl0u+X8THGlkcnlvbW92
QGdtYWlsLmNvbQAKCRBKf944AhHzi9byB/9fIxzoxtDMvixtJabuGSJRtlijDlWF
GlO6yIWCXl/8v8easR2PCF75U/xv0+QFQmze8PVi8u4Xz589P247NnEuyEZ9n84i
aCavARho6QLZPEL+B04NaqoHBl+ORKQTA6eKGhyKwRp/rn83z5Ubuw2tN7krHT3b
kCY61FuTQGxNY2o/WKv/iLwINYr7H23hCf0WwyyKH1bp7OegiQ14Ebn1NtfS3sMx
hS6h8Ya826vmUW0bCSS/9kzKYBCjksTig0HphUOHq6BoZJs++0b7GukIulRyuLfD
J9Gr9HGPoDCVzdmFfpn2FSlxdmqfO9amUSagd0ftLQfFlPlrpoULi0GW
=Bgxr
-----END PGP SIGNATURE-----
Merge tag 'ceph-for-5.3-rc1' of git://github.com/ceph/ceph-client
Pull ceph updates from Ilya Dryomov:
"Lots of exciting things this time!
- support for rbd object-map and fast-diff features (myself). This
will speed up reads, discards and things like snap diffs on sparse
images.
- ceph.snap.btime vxattr to expose snapshot creation time (David
Disseldorp). This will be used to integrate with "Restore Previous
Versions" feature added in Windows 7 for folks who reexport ceph
through SMB.
- security xattrs for ceph (Zheng Yan). Only selinux is supported for
now due to the limitations of ->dentry_init_security().
- support for MSG_ADDR2, FS_BTIME and FS_CHANGE_ATTR features (Jeff
Layton). This is actually a single feature bit which was missing
because of the filesystem pieces. With this in, the kernel client
will finally be reported as "luminous" by "ceph features" -- it is
still being reported as "jewel" even though all required Luminous
features were implemented in 4.13.
- stop NULL-terminating ceph vxattrs (Jeff Layton). The convention
with xattrs is to not terminate and this was causing
inconsistencies with ceph-fuse.
- change filesystem time granularity from 1 us to 1 ns, again fixing
an inconsistency with ceph-fuse (Luis Henriques).
On top of this there are some additional dentry name handling and cap
flushing fixes from Zheng. Finally, Jeff is formally taking over for
Zheng as the filesystem maintainer"
* tag 'ceph-for-5.3-rc1' of git://github.com/ceph/ceph-client: (71 commits)
ceph: fix end offset in truncate_inode_pages_range call
ceph: use generic_delete_inode() for ->drop_inode
ceph: use ceph_evict_inode to cleanup inode's resource
ceph: initialize superblock s_time_gran to 1
MAINTAINERS: take over for Zheng as CephFS kernel client maintainer
rbd: setallochint only if object doesn't exist
rbd: support for object-map and fast-diff
rbd: call rbd_dev_mapping_set() from rbd_dev_image_probe()
libceph: export osd_req_op_data() macro
libceph: change ceph_osdc_call() to take page vector for response
libceph: bump CEPH_MSG_MAX_DATA_LEN (again)
rbd: new exclusive lock wait/wake code
rbd: quiescing lock should wait for image requests
rbd: lock should be quiesced on reacquire
rbd: introduce copyup state machine
rbd: rename rbd_obj_setup_*() to rbd_obj_init_*()
rbd: move OSD request allocation into object request state machines
rbd: factor out __rbd_osd_setup_discard_ops()
rbd: factor out rbd_osd_setup_copyup()
rbd: introduce obj_req->osd_reqs list
...
setallochint is really only useful on object creation. Continue
hinting unconditionally if object map cannot be used.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Speed up reads, discards and zeroouts through RBD_OBJ_FLAG_MAY_EXIST
and RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT based on object map.
Invalid object maps are not trusted, but still updated. Note that we
never iterate, resize or invalidate object maps. If object-map feature
is enabled but object map fails to load, we just fail the requester
(either "rbd map" or I/O, by way of post-acquire action).
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Snapshot object map will be loaded in rbd_dev_image_probe(), so we need
to know snapshot's size (as opposed to HEAD's size) sooner.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
This will be used for loading object map. rbd_obj_read_sync() isn't
suitable because object map must be accessed through class methods.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
rbd_wait_state_locked() is built around rbd_dev->lock_waitq and blocks
rbd worker threads while waiting for the lock, potentially impacting
other rbd devices. There is no good way to pass an error code into
image request state machines when acquisition fails, hence the use of
RBD_DEV_FLAG_BLACKLISTED for everything and various other issues.
Introduce rbd_dev->acquiring_list and move acquisition into image
request state machine. Use rbd_img_schedule() for kicking and passing
error codes. No blocking occurs while waiting for the lock, but
rbd_dev->lock_rwsem is still held across lock, unlock and set_cookie
calls.
Always acquire the lock on "rbd map" to avoid associating the latency
of acquiring the lock with the first I/O request.
A slight regression is that lock_timeout is now respected only if lock
acquisition is triggered by "rbd map" and not by I/O. This is somewhat
compensated by the fact that we no longer block if the peer refuses to
release lock -- I/O is failed with EROFS right away.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Syncing OSD requests doesn't really work. A single image request may
be comprised of multiple object requests, each of which can go through
a series of OSD requests (original, copyups, etc). On top of that, the
OSD cliest may be shared with other rbd devices.
What we want is to ensure that all in-flight image requests complete.
Introduce rbd_dev->running_list and block in RBD_LOCK_STATE_RELEASING
until that happens. New OSD requests may be started during this time.
Note that __rbd_img_handle_request() acquires rbd_dev->lock_rwsem only
if need_exclusive_lock() returns true. This avoids a deadlock similar
to the one outlined in the previous commit between unlock and I/O that
doesn't require lock, such as a read with object-map feature disabled.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Quiesce exclusive lock at the top of rbd_reacquire_lock() instead
of only when ceph_cls_set_cookie() fails. This avoids a deadlock on
rbd_dev->lock_rwsem.
If rbd_dev->lock_rwsem is needed for I/O completion, set_cookie can
hang ceph-msgr worker thread if set_cookie reply ends up behind an I/O
reply, because, like lock and unlock requests, set_cookie is sent and
waited upon with rbd_dev->lock_rwsem held for write.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Both write and copyup paths will get more complex with object map.
Factor copyup code out into a separate state machine.
While at it, take advantage of obj_req->osd_reqs list and issue empty
and current snapc OSD requests together, one after another.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
These functions don't allocate and set up OSD requests anymore.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
Following submission, move initial OSD request allocation into object
request state machines. Everything that has to do with OSD requests is
now handled inside the state machine, all __rbd_img_fill_request() has
left is initialization.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>
With obj_req->xferred removed, obj_req->ex.oe_off and obj_req->ex.oe_len
can be updated if required for alignment. Previously the new offset and
length weren't stored anywhere beyond rbd_obj_setup_discard().
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Dongsheng Yang <dongsheng.yang@easystack.cn>