linux-stable.git/fs/ceph, branch v5.17.2

ceph: set pool_ns in new inode layout for async creates

2022-01-26T19:17:50+00:00

Dan reported that he was unable to write to files that had been
asynchronously created when the client's OSD caps are restricted to a
particular namespace.

The issue is that the layout for the new inode is only partially being
filled. Ensure that we populate the pool_ns_data and pool_ns_len in the
iinfo before calling ceph_fill_inode.

Cc: stable@vger.kernel.org
URL: https://tracker.ceph.com/issues/54013
Fixes: 9a8d03ca2e2c ("ceph: attempt to do async create when possible")
Reported-by: Dan van der Ster 
Signed-off-by: Jeff Layton 
Reviewed-by: Ilya Dryomov 
Signed-off-by: Ilya Dryomov

ceph: properly put ceph_string reference after async create attempt

2022-01-26T19:17:50+00:00

The reference acquired by try_prep_async_create is currently leaked.
Ensure we put it.

Cc: stable@vger.kernel.org
Fixes: 9a8d03ca2e2c ("ceph: attempt to do async create when possible")
Signed-off-by: Jeff Layton 
Reviewed-by: Ilya Dryomov 
Signed-off-by: Ilya Dryomov

ceph: put the requests/sessions when it fails to alloc memory

2022-01-26T19:17:50+00:00

When failing to allocate the sessions memory we should make sure
the req1 and req2 and the sessions get put. And also in case the
max_sessions decreased so when kreallocate the new memory some
sessions maybe missed being put.

And if the max_sessions is 0 krealloc will return ZERO_SIZE_PTR,
which will lead to a distinct access fault.

URL: https://tracker.ceph.com/issues/53819
Fixes: e1a4541ec0b9 ("ceph: flush the mdlog before waiting on unsafe reqs")
Signed-off-by: Xiubo Li 
Reviewed-by: Venky Shankar 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

Merge tag 'fscache-fixes-20220121' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs

2022-01-22T08:59:32+00:00

Pull more fscache updates from David Howells:
"A set of fixes and minor updates for the fscache rewrite:

- Fix mishandling of volume collisions (the wait condition is
inverted and so it was only waiting if the volume collision was
already resolved).

- Fix miscalculation of whether there's space available in
cachefiles.

- Make sure a default cache name is set on a cache if the user hasn't
set one by the time they bind the cache.

- Adjust the way the backing inode is presented in tracepoints, add a
tracepoint for mkdir and trace directory lookup.

- Add a tracepoint for failure to set the active file mark.

- Add an explanation of the checks made on the backing filesystem.

- Check that the backing filesystem supports tmpfile.

- Document how the page-release cancellation of the read-skip
optimisation works.

And I've included a change for netfslib:

- Make ops->init_rreq() optional"

* tag 'fscache-fixes-20220121' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
netfs: Make ops->init_rreq() optional
fscache: Add a comment explaining how page-release optimisation works
cachefiles: Check that the backing filesystem supports tmpfiles
cachefiles: Explain checks in a comment
cachefiles: Trace active-mark failure
cachefiles: Make some tracepoint adjustments
cachefiles: set default tag name if it's unspecified
cachefiles: Calculate the blockshift in terms of bytes, not pages
fscache: Fix the volume collision wait condition

netfs: Make ops->init_rreq() optional

2022-01-21T21:36:28+00:00

Make the ops->init_rreq() callback optional.  This isn't required for the
erofs changes I'm implementing to do on-demand read through fscache[1].
Further, ceph has an empty init_rreq method that can then be removed and
it's marked optional in the documentation.

Signed-off-by: Jeffle Xu 
Signed-off-by: David Howells 
Reviewed-by: Jeff Layton 
cc: linux-cachefs@redhat.com
Link: https://lore.kernel.org/r/20211227125444.21187-1-jefflexu@linux.alibaba.com/ [1]
Link: https://lore.kernel.org/r/20211228124419.103020-1-jefflexu@linux.alibaba.com
Link: https://lore.kernel.org/r/164251410387.3435901.2504600788262093313.stgit@warthog.procyon.org.uk/ # v1

Merge tag 'ceph-for-5.17-rc1' of git://github.com/ceph/ceph-client

2022-01-20T11:46:20+00:00

Pull ceph updates from Ilya Dryomov:
 "The highlight is the new mount "device" string syntax implemented by
  Venky Shankar. It solves some long-standing issues with using
  different auth entities and/or mounting different CephFS filesystems
  from the same cluster, remounting and also misleading /proc/mounts
  contents. The existing syntax of course remains to be maintained.

  On top of that, there is a couple of fixes for edge cases in quota and
  a new mount option for turning on unbuffered I/O mode globally instead
  of on a per-file basis with ioctl(CEPH_IOC_SYNCIO)"

* tag 'ceph-for-5.17-rc1' of git://github.com/ceph/ceph-client:
  ceph: move CEPH_SUPER_MAGIC definition to magic.h
  ceph: remove redundant Lsx caps check
  ceph: add new "nopagecache" option
  ceph: don't check for quotas on MDS stray dirs
  ceph: drop send metrics debug message
  rbd: make const pointer spaces a static const array
  ceph: Fix incorrect statfs report for small quota
  ceph: mount syntax module parameter
  doc: document new CephFS mount device syntax
  ceph: record updated mon_addr on remount
  ceph: new device mount syntax
  libceph: rename parse_fsid() to ceph_parse_fsid() and export
  libceph: generalize addr/ip parsing based on delimiter

ceph: move CEPH_SUPER_MAGIC definition to magic.h

2022-01-13T12:40:07+00:00

The uapi headers are missing the ceph definition. Move it there so
userland apps can ID cephfs.

Signed-off-by: Jeff Layton 
Reviewed-by: Ilya Dryomov 
Signed-off-by: Ilya Dryomov

ceph: remove redundant Lsx caps check

2022-01-13T12:40:07+00:00

The newcaps has already included the Ls, no need to check it again.

Signed-off-by: Xiubo Li 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

ceph: add new "nopagecache" option

2022-01-13T12:40:07+00:00

CephFS is a bit unlike most other filesystems in that it only
conditionally does buffered I/O based on the caps that it gets from the
MDS. In most cases, unless there is contended access for an inode the
MDS does give Fbc caps to the client, so the unbuffered codepaths are
only infrequently traveled and are difficult to test.

At one time, the "-o sync" mount option would give you this behavior,
but that was removed in commit 7ab9b3807097 ("ceph: Don't use
ceph-sync-mode for synchronous-fs.").

Add a new mount option to tell the client to ignore Fbc caps when doing
I/O, and to use the synchronous codepaths exclusively, even on
non-O_DIRECT file descriptors. We already have an ioctl that forces this
behavior on a per-file basis, so we can just always set the CEPH_F_SYNC
flag in the file description on such mounts.

Additionally, this patch also changes the client to not request Fbc when
doing direct I/O. We aren't using the cache with O_DIRECT so we don't
have any need for those caps.

Signed-off-by: Jeff Layton 
Acked-by: Greg Farnum 
Reviewed-by: Venky Shankar 
Signed-off-by: Ilya Dryomov

ceph: don't check for quotas on MDS stray dirs

2022-01-13T12:40:07+00:00

玮文 胡 reported seeing the WARN_RATELIMIT pop when writing to an
inode that had been transplanted into the stray dir. The client was
trying to look up the quotarealm info from the parent and that tripped
the warning.

Change the ceph_vino_is_reserved helper to not throw a warning for
MDS stray directories (0x100 - 0x1ff), only for reserved dirs that
are not in that range.

Also, fix ceph_has_realms_with_quotas to return false when encountering
a reserved inode.

URL: https://tracker.ceph.com/issues/53180
Reported-by: Hu Weiwen 
Signed-off-by: Jeff Layton 
Reviewed-by: Luis Henriques 
Reviewed-by: Xiubo Li 
Signed-off-by: Ilya Dryomov