include, rados: Fixes for list-inconsistent-obj and librados

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit 4fc4326b36aa00092043514746aa39aedad06b0f)

rados: Balance format sections in same do_get_inconsistent_cmd()

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit de4681b80a5d24e72bbc7ac6dfee4775987fd834)

rados: Include epoch in the list-inconsistent-* command output

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit df4bc16c4d49489f7398eb1ecc7b9aef93258414)

rados: Improve error messages for list-inconsistent commands

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit cea7cf56ee3b9593c935d0c74ce6e4b28b14c2da)

10.2.1

10.2.2

Merge pull request #9692 from dillaman/wip-16292

jewel: librbd: remove should ignore mirror errors from older OSDs

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

librbd: remove should ignore mirror errors from older OSDs

Fixes: http://tracker.ceph.com/issues/16268
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 5a97a9b7ef238c4ceb16225a0b937f9a0cdbb2b6)

Merge pull request #9691 from dillaman/wip-16260-jewel

jewel: librbd: do not shut down exclusive lock while acquiring'

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9690 from dillaman/wip-16290

jewel: rbd-mirror: FAILED assert(!m_status_watcher)

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

librbd: track in-flight flush events for journaling

The tracking for flush events was accidentally dropped. This
can result in the journal being closed too early since it
might miss in-flight flush ops.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit a1ea1edce9332fa1730625629e3b4e16f02caaf4)

librbd: do not shut down exclusive lock while acquiring

Fixes: http://tracker.ceph.com/issues/16260
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit c5694fc6766fb8e213c4b65d2cd7b9d066b07ff7)

rbd-mirror: FAILED assert(!m_status_watcher)

Fixes: http://tracker.ceph.com/issues/16245
Signed-off-by: Mykola Golub <mgolub@mirantis.com>
(cherry picked from commit b8bfc0f2efb44d3eff7a5fd8bca24516c8815d62)

Merge pull request #9654 from dillaman/wip-16246

jewel: librbd: recursive lock possible when disabling journaling

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9673 from dillaman/wip-16265

jewel: rbd-mirror: do not re-use image id from mirror directory if creating image

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9674 from dillaman/wip-16267

jewel: qa/workunits: ensure replay has started before checking position

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

qa/workunits: ensure replay has started before checking position

Fixes: http://tracker.ceph.com/issues/16248
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 6a68b68e0d2741e6694fd28d31766808ba881815)

rbd-mirror: reset local image id when creating new image

Fixes: http://tracker.ceph.com/issues/16253
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 1948047acbc129de2982d53e130da1f577af224b)

rbd-mirror: handle an attempt to delete a non-mirrored image

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 45dda67b3ef3d5dc5abb327fc1f242140f004d90)

cls_rbd: disallow setting an empty mirror image entry to disabling

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 7b2f99a97ff64451f3ae72081186dff41c507271)

librbd: avoid holding owner_lock waiting for exclusive lock

Fixes: http://tracker.ceph.com/issues/16235
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit b96ab805eed49d4bdb494c8919d2e1ee7ad0a7e6)

Merge pull request #9656 from gregsfortytwo/wip-jewel-16024

Jewel client: fix bugs accidentally disabling readahead

Merge pull request #9655 from gregsfortytwo/wip-jewel-15508

Jewel mds: order directories by hash and fix simultaneous readdir races

client: fix simultaneous readdirs race

Current readdir code uses list to track the order of the dentries
in readdir replies.  When handling a readdir reply, it pushes the
resulting dentries to the back of directory's dentry_list. After
readdir finishes, the dentry_list reflects how MDS sorts dentries.

This method is racy when there are simultaneous readdirs. The fix
is use vector instead of list to trace how dentries are sorted in
its parent directory. As long as shared_gen doesn't change, each
dentry is at fixed position of the vector. So cocurrent readdirs
do not affect each other.

Fixes: http://tracker.ceph.com/issues/15508
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 9d297c5e98f814b282dadc379ab70dfa678db73e)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: move dir_{release,ordered}_count into class Inode

We close Inode::dir when it's empty. Once closing the dir, we lose
track of {release,ordered}_count. This causes direcotry to be wrongly
marked as complete. (dir is trimmed to empty in the middle of readdir)

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 235fcf68f5d96a4d9d6cc260d12da912fa7ea4a8)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

ceph_test_libcephfs: check order of entries in readdir result

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 9b17d145c15a469caa29853312b8dcb0b19f55f0)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

mds: don't reset readdir offset if client supports hash order dentry

Now the ordering of dentries is stable across directory fragmentation.
There is no need to reset readdir offset if directory get fragmented
in the middle of readdir.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 98a01af2cd8e01f14ab7be32c31a43571ef8bb87)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: using hash value to compose dentry offset

If MDS sorts dentries in dirfrag in hash order, we use hash value to
compose dentry offset. dentry offset is:

  (0xff << 52) | ((24 bits hash) << 28) |
  (the nth entry hash hash collision)

This offset is stable across directory fragmentation.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 680766ec131b95271e320f54dfe6d69ea8d4fbb3)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: record 'offset' for each entry of dir_result_t::buffer

This is preparation for using hash value as dentry 'offset'

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit bd6546e5e8bfb4ef8927807492d63d5285534618)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: fix cached readdir after seekdir

Client::seekdir doesn't reset dirp->at_cache_name for a forward seek
within same frag. So the dentry with name == at_cache_name may not be
the one prior to the readdir postion.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 0e32115bae2f1ac2c59b57852976b0de5587abac)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

mds: define end/complete in readdir reply as single u16 flags

so that we can introduce new flags for readdir reply.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 92cfbdf526d1030891da366299b635e589dd5f8e)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

mds: sort dentries in CDir in hash order

This gives us stable ordering of dentries. (Previously ordering of
dentries changes after directory gets fragmented)

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit f483224a0bc70b9eb5f62cd6cbb8b97df5b65953)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: save readdir result into dir_result_t directly

Current code saves the readdir result into MedaRequest, then updates
dir_result_t according to MetaRequest. I can't see any reason why
we need to do this.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit db5d60d1b8d01afd43da09a063f00aa6ab312903)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: don't allocate dir_result_t::buffer dynamically

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit c41ceb9f1425432295a52965de8d4a23b15fe5e5)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: simplify 'offset in frag'

don't distinguish leftmost frag from other frags. always use 2 as
first entry's offset.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 6572c2a24ea2abec4c7eecef92a292c2cc72c762)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: move readahead ref get/put to obj constr

The readahead context was incrementing the file handle reference count but only
decreasing in finish which is not called if readahead is a no-op.

This fixes a bug caught in testing where readahead was causing inode to become
disconnected:

    2016-06-10 19:46:48.953018 7f2a4351be80  1 client.4125 dump_inode: DISCONNECTED inode 10000000502 #10000000502 ref 110000000502.head(faked_ino=2307 ref=1 ll_ref=0 cap_refs={1024=0,2048=0,4096=0,8192=0} open={1=0,2=0} mode=100666 size=4194304/0 mtime=2016-06-10 19:29:45.107417 caps=-(0=pAsLsXsFscr) objectset[10000000502 ts 2/4012653 objects 0 dirty_or_tx 0] 0x7f2a24300d00)
    2016-06-10 19:46:48.953032 7f2a4351be80  2 client.4125 cache still has 0+155 items, waiting (for caps to release?)

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit b5a0773259ce9f2ef44e8e6925ab11dc1a5e55e6)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: use fh min readahead to enable readahead

The client was using the configured maximum readahead of 0 which was changed in
95ee69989129750fddce6a3b5644238c4b88ed74. This prevented readahead from ever
running (without setting a different default from 0).

Fixes: http://tracker.ceph.com/issues/16024
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 01c179bcce2dc8a269d78eb1be8198b12bd161f7)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: update comment

Default was changed to unlimited in 11e4680e7394e6e02612441c964e05207b97737b.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit bee02a073f4c551798bc490618a0e8d862893c25)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

client: use layout.get_period() helper

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit b331d12a43a4177369a003c3ab9d525da6a89950)

Signed-off-by: Greg Farnum <gfarnum@redhat.com

librbd: recursive lock possible when disabling journaling

If pool-level mirroring is enabled, deleting the journal
could cause a deadlock attempting to delete remote peer
sync-point snapshots.

Fixes: http://tracker.ceph.com/issues/16235
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit fb255e6c3cd44c8d24c53e3cd70395a11a712574)

Merge pull request #9553 from Abhishekvrshny/wip-16038-jewel

jewel: pybind: rbd API should default features parameter to None

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9552 from Abhishekvrshny/wip-15981-jewel

jewel: qa: dynamic_features.sh races with image deletion

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9629 from dillaman/wip-16231

jewel: rbd-mirror: fix deletion propagation edge cases

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9630 from dillaman/wip-16233

jewel: rbd-mirror: do not propagate deletions when pool unavailable

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9628 from dillaman/wip-16224

jewel: rbd-mirror: fix journal shut down ordering

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9627 from dillaman/wip-16217

jewel: rbd-mirror: refresh image after creating sync point

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9579 from Abhishekvrshny/wip-16006-jewel

jewel: BackoffThrottle spins unnecessarily with very small backoff while the throttle is full

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #9638 from ceph/wip-16211-jewel

Jewel: remove reliance on FLAG_OMAP for reads

Reviewed-by: Sage Weil <sage@redhat.com>

rbd-mirror: do not propagate deletions when pool unavailable

Fixes: http://tracker.ceph.com/issues/16229
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit d321634414fa4b90c66474557ec990dfc935ae29)

rbd-mirror: image deleter properly handles multiple snapshots

Fixes: http://tracker.ceph.com/issues/16226
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit d72a8cb67fbdc7361f8f9e265c659ddb96dc0f74)

rbd-mirror: propagate deletions even if image replayer is stopped

If an image deletion which causes the replayer to stop (due to an error)
before the deletion is detected, the deletion should still occur.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit a0c526a23783fe543737f6445a2d9438cc73b137)

qa/workunits/rbd: improve deletion propagation test case

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit de15b198cd7b5e94016605c8016a5cf38c07c6f6)

librbd: quickly check for duplicate snap name upon create

During journal replay, quickly check for a duplicate snap name
before allocating a snapshot sequence from the OSDs and
reverting due to the duplicate name.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 8374d45c7644666103f88c410a2b1b781f3a0cb8)

rbd-mirror: stop external replayer before closing remote journaler

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 88bd3002082184c045bb82547929b30da1526cfb)

librbd: stop journal recorder before starting external replay

Fixes: http://tracker.ceph.com/issues/16165
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 5aa6eb2782bdb5cde6a050a5f7a4e683800414c0)

rbd-mirror: refresh image after creating sync point snapshot

Fixes: http://tracker.ceph.com/issues/16196
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit c9a3e3be7ddc31992167d55de12bbe6849447b43)

rbd-mirror: consistently use dout/derr for debug logging

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 03f5e4a917b3c18f8ccdaa5ce676ed9731dd2a8f)

Merge pull request #9612 from dillaman/wip-16208

jewel: rbd-mirror: support bootstrap canceling

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9610 from dillaman/wip-16209

jewel: librbd: prevent error messages when journal externally disabled

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9611 from dillaman/wip-16210

jewel: librbd: journal IO error results in failed assertion in AioCompletion

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9580 from Abhishekvrshny/wip-15996-jewel

jewel: librbd: reuse ImageCtx::finisher and SafeTimer for lots of images case

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9609 from dillaman/wip-16216

jewel: librbd: refresh image if needed in mirror functions

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9608 from dillaman/wip-16189

jewel: cls_rbd: mirror image status summary should read full directory

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9556 from Abhishekvrshny/wip-16162-jewel

jewel: ceph_test_librbd_fsx crashes during journal replay shut down

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9554 from Abhishekvrshny/wip-16084-jewel

jewel: librbd: validate image metadata configuration overrides

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9555 from Abhishekvrshny/wip-16147-jewel

jewel: Object Map is showing as invalid, even when Object Map is disabled for that Image.

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9550 from Abhishekvrshny/wip-15969-jewel

jewel: Initial python APIs to support mirroring

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9548 from Abhishekvrshny/wip-15963-jewel

jewel: Disabling journaling feature results in "Transport endpoint is not connected" error

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #9424 from SUSE/wip-16107-jewel

jewel: ceph gid mismatch on upgrade from hammer with pre-existing ceph user (SUSE)

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #8802 from SUSE/wip-15606-jewel

jewel: ceph-{mds,mon,osd,radosgw} systemd unit files need wants=time-sync.target

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #9320 from SUSE/wip-16017-jewel

jewel: ceph: cli: exception when pool name has non-ascii characters

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #9270 from Abhishekvrshny/wip-15852-jewel

jewel: mon/Monitor: memory leak on Monitor::handle_ping()

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #9237 from SUSE/wip-15961-jewel

jewel: ceph-osd valgrind invalid reads/writes

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #9103 from Abhishekvrshny/wip-15797-jewel

jewel: deadlock in OSD::_committed_osd_maps

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #9101 from Abhishekvrshny/wip-15799-jewel

jewel: osd boot race with noup being set

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #9104 from Abhishekvrshny/wip-15795-jewel

jewel: LibRadosWatchNotifyPPTests/LibRadosWatchNotifyPP.WatchNotify2Timeout/1 segv

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #9427 from SUSE/wip-16104-jewel

jewel: ceph-disk: workaround gperftool hang

Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #9614 from ceph/wip-16113-jewel

osd: hobject_t::get_max() vs is_max() discrepancy

Reviewed-by: Sage Weil <sage@redhat.com>

rbd-mirror: don't return split-brain error if we still in MIRROR_PEER_STATE_SYNCING

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
(cherry picked from commit 3ac620c5d66d8516e68264fcdb5e97b0e803d0ec)

rbd-mirror: refactor split-brain detection code

to make it more readable and extendable.

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
(cherry picked from commit 039cd2d2e3bb6a9c5042c125d2c67ec9adc02c47)

rbd-mirror: support bootstrap canceling

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
(cherry picked from commit 21f895f48498e518c3675a9e851559f4d071f6dd)

rbd-mirror: return ECANCELED when ImageReplayer start is canceled

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
(cherry picked from commit f581dbcb2eca108ce9440629b8f00f3229c07b95)

librbd: flush journal commit positions before starting op

Ensure all IO has been properly flushed and committed to the
journal before starting an op that could affect the IO path.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 636129653c7cac9665d0c423f6aafad5fd0c480a)

librbd: leaked AioCompletion if IO issued after image shut down

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit d3d4e068396e01c58832e0ea1224ca34bafb808e)

librbd: resize and snap create can hang on journal replay failure

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 1032f19b3c9d9c7916669d04af909131574b3cab)

librbd: missing journal state for flushing/restarting replay

Fixes: http://tracker.ceph.com/issues/16077
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit a7f5bc181c9cd677d39f95ad78134e0f85c22acb)

librbd: track IO in the order issued

Fixes: http://tracker.ceph.com/issues/15034
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit eb8dda5296eb5c19aa27e96d6e92eaa0c352cb62)

librbd: AioCompletion should retrieve CephContext from ImageCtx

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 133b1ffb21c982003c7aa9dafd585062f9b2eb51)

librbd: only mark async completion as done after callback

This change in behavior was introduced in fde9f78.

Signed-off-by: Hector Martin <marcan@marcan.st>
(cherry picked from commit 77f7c1cf829e5417eb356ebf3bf1fbc8ff7a25a5)

librbd: Journal no longer requires AioCompletion for IO events

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 82587a31f1d8ba1765f164391796753cf97a1878)

librbd: bubble journal error up via existing AioObjectRequests

The journal cannot directly fail the AioCompletion because it
might have one-or-more outstanding requests associated with
it.

Fixes: http://tracker.ceph.com/issues/15791
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 79c934742fb15c803780519b36070408599a74c5)

librbd: potential duplicate snap removal can result in crash

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 276fed6b7078158c2cd04fba14a11531c27898e0)

ObjectCacher: do not merge uncommitted journal bhs

Avoid the possibility of an overwrite callback for regions
that were never associated with a journal write.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit c6060ade16205337de3941a10d05e234b738fa07)

ObjectCacher: don't loop the extents of OSDRead in map_read

Looping the extents of OSDRead is done in _readx. The looping in
map_read is not needed.

Signed-off-by: Zhiqiang Wang <zhiqiang@xsky.com>
(cherry picked from commit fae912c3856cab6d03183411eecab90c6e6732d2)

ObjectCacher: fix the bytes_written in writex

The BufferHead returned after map_write may include other dirty data to
the left and/or right. The actual length of data been written is in the
ObjectExtent.

Signed-off-by: Zhiqiang Wang <zhiqiang@xsky.com>
(cherry picked from commit ca6f1681e4baf61682e258c5761536d67bf6bdb8)

osdc: don't loop the extents of OSDWrite in map_write

We are looping all the extents of the OSDWrite in writex. The looping in
map_write is not needed. And if there are same object in these
extents,they will be mapped multiple times.

Signed-off-by: Zhiqiang Wang <zhiqiang@xsky.com>
(cherry picked from commit 322f1059670898762a890a01af6d3b0dd3ccf826)

librbd: notify image feature update before releasing lock

If a client is waiting for the lock, ensure it sees the latest
version of the image features.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit f601ea835bb78554ac767ccb11789df960765bf9)

librbd: refresh image after acquiring exclusive lock

It's possible that the object map or journaling features have
been disabled while the lock was not owned.

Fixes: http://tracker.ceph.com/issues/16114
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 03c54f52d15d6283c630bac6f75427e6829f7d0a)

librbd: refresh image if needed in mirror functions

Fixes: http://tracker.ceph.com/issues/16096
Signed-off-by: Jon Bernard <jbernard@tuxion.com>
(cherry picked from commit a355c8f6d9e792517a8d0e096a99a64474369cfc)

cls_rbd: mirror image status summary should read full directory

Previously only retrieved the status for the first 64 images in
the rbd_mirroring directory.

Fixes: http://tracker.ceph.com/issues/16178
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 61f86e7ec14233ea0dee5b3370595ce050ee28c3)