linux.git/fs/ceph/super.h, branch v5.2

ceph: avoid iput_final() while holding mutex or in dispatch thread

2019-06-05T18:34:39+00:00

iput_final() may wait for reahahead pages. The wait can cause deadlock.
For example:

  Workqueue: ceph-msgr ceph_con_workfn [libceph]
    Call Trace:
     schedule+0x36/0x80
     io_schedule+0x16/0x40
     __lock_page+0x101/0x140
     truncate_inode_pages_range+0x556/0x9f0
     truncate_inode_pages_final+0x4d/0x60
     evict+0x182/0x1a0
     iput+0x1d2/0x220
     iterate_session_caps+0x82/0x230 [ceph]
     dispatch+0x678/0xa80 [ceph]
     ceph_con_workfn+0x95b/0x1560 [libceph]
     process_one_work+0x14d/0x410
     worker_thread+0x4b/0x460
     kthread+0x105/0x140
     ret_from_fork+0x22/0x40

  Workqueue: ceph-msgr ceph_con_workfn [libceph]
    Call Trace:
     __schedule+0x3d6/0x8b0
     schedule+0x36/0x80
     schedule_preempt_disabled+0xe/0x10
     mutex_lock+0x2f/0x40
     ceph_check_caps+0x505/0xa80 [ceph]
     ceph_put_wrbuffer_cap_refs+0x1e5/0x2c0 [ceph]
     writepages_finish+0x2d3/0x410 [ceph]
     __complete_request+0x26/0x60 [libceph]
     handle_reply+0x6c8/0xa10 [libceph]
     dispatch+0x29a/0xbb0 [libceph]
     ceph_con_workfn+0x95b/0x1560 [libceph]
     process_one_work+0x14d/0x410
     worker_thread+0x4b/0x460
     kthread+0x105/0x140
     ret_from_fork+0x22/0x40

In above example, truncate_inode_pages_range() waits for readahead pages
while holding s_mutex. ceph_check_caps() waits for s_mutex and blocks
OSD dispatch thread. Later OSD replies (for readahead) can't be handled.

ceph_check_caps() also may lock snap_rwsem for read. So similar deadlock
can happen if iput_final() is called while holding snap_rwsem.

In general, it's not good to call iput_final() inside MDS/OSD dispatch
threads or while holding any mutex.

The fix is introducing ceph_async_iput(), which calls iput_final() in
workqueue.

Signed-off-by: "Yan, Zheng" 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

ceph: single workqueue for inode related works

2019-06-05T18:34:39+00:00

We have three workqueue for inode works. Later patch will introduce
one more work for inode. It's not good to introcuce more workqueue
and add more 'struct work_struct' to 'struct ceph_inode_info'.

Signed-off-by: "Yan, Zheng" 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

Merge tag 'ceph-for-5.2-rc1' of git://github.com/ceph/ceph-client

2019-05-16T23:24:01+00:00

Pull ceph updates from Ilya Dryomov:
 "On the filesystem side we have:

   - a fix to enforce quotas set above the mount point (Luis Henriques)

   - support for exporting snapshots through NFS (Zheng Yan)

   - proper statx implementation (Jeff Layton). statx flags are mapped
     to MDS caps, with AT_STATX_{DONT,FORCE}_SYNC taken into account.

   - some follow-up dentry name handling fixes, in particular
     elimination of our hand-rolled helper and the switch to __getname()
     as suggested by Al (Jeff Layton)

   - a set of MDS client cleanups in preparation for async MDS requests
     in the future (Jeff Layton)

   - a fix to sync the filesystem before remounting (Jeff Layton)

  On the rbd side, work is on-going on object-map and fast-diff image
  features"

* tag 'ceph-for-5.2-rc1' of git://github.com/ceph/ceph-client: (29 commits)
  ceph: flush dirty inodes before proceeding with remount
  ceph: fix unaligned access in ceph_send_cap_releases
  libceph: make ceph_pr_addr take an struct ceph_entity_addr pointer
  libceph: fix unaligned accesses in ceph_entity_addr handling
  rbd: don't assert on writes to snapshots
  rbd: client_mutex is never nested
  ceph: print inode number in __caps_issued_mask debugging messages
  ceph: just call get_session in __ceph_lookup_mds_session
  ceph: simplify arguments and return semantics of try_get_cap_refs
  ceph: fix comment over ceph_drop_caps_for_unlink
  ceph: move wait for mds request into helper function
  ceph: have ceph_mdsc_do_request call ceph_mdsc_submit_request
  ceph: after an MDS request, do callback and completions
  ceph: use pathlen values returned by set_request_path_attr
  ceph: use __getname/__putname in ceph_mdsc_build_path
  ceph: use ceph_mdsc_build_path instead of clone_dentry_name
  ceph: fix potential use-after-free in ceph_mdsc_build_path
  ceph: dump granular cap info in "caps" debugfs file
  ceph: make iterate_session_caps a public symbol
  ceph: fix NULL pointer deref when debugging is enabled
  ...

ceph: quota: fix quota subdir mounts

2019-05-07T17:22:36+00:00

The CephFS kernel client does not enforce quotas set in a directory that
isn't visible from the mount point.  For example, given the path
'/dir1/dir2', if quotas are set in 'dir1' and the filesystem is mounted with

  mount -t ceph ::/dir1/ /mnt

then the client won't be able to access 'dir1' inode, even if 'dir2' belongs
to a quota realm that points to it.

This patch fixes this issue by simply doing an MDS LOOKUPINO operation for
unknown inodes.  Any inode reference obtained this way will be added to a
list in ceph_mds_client, and will only be released when the filesystem is
umounted.

Link: https://tracker.ceph.com/issues/38482
Reported-by: Hendrik Peyerl 
Signed-off-by: Luis Henriques 
Reviewed-by: "Yan, Zheng" 
Signed-off-by: Ilya Dryomov

ceph: factor out ceph_lookup_inode()

2019-05-07T17:22:36+00:00

This function will be used by __fh_to_dentry and by the quotas code, to
find quota realm inodes that are not visible in the mountpoint.

Signed-off-by: Luis Henriques 
Reviewed-by: "Yan, Zheng" 
Signed-off-by: Ilya Dryomov

ceph: use ->free_inode()

2019-05-02T02:43:26+00:00

a lot of non-delayed work in this case; all of that is left in
->destroy_inode()

Signed-off-by: Al Viro

ceph: add mount option to limit caps count

2019-03-05T17:55:17+00:00

If number of caps exceed the limit, ceph_trim_dentires() also trim
dentries with valid leases. Trimming dentry releases references to
associated inode, which may evict inode and release caps.

By default, there is no limit for caps count.

Signed-off-by: "Yan, Zheng" 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

ceph: periodically trim stale dentries

2019-03-05T17:55:17+00:00

Previous commit make VFS delete stale dentry when last reference is
dropped. Lease also can become invalid when corresponding dentry has
no reference. This patch make cephfs periodically scan lease list,
delete corresponding dentry if lease is invalid.

There are two types of lease, dentry lease and dir lease. dentry lease
has life time and applies to singe dentry. Dentry lease is added to tail
of a list when it's updated, leases at front of the list will expire
first. Dir lease is CEPH_CAP_FILE_SHARED on directory inode, it applies
to all dentries in the directory. Dentries have dir leases are added to
another list. Dentries in the list are periodically checked in a round
robin manner.

Signed-off-by: "Yan, Zheng" 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

ceph: delete stale dentry when last reference is dropped

2019-03-05T17:55:17+00:00

introduce ceph_d_delete(), which checks if dentry has valid lease.

Signed-off-by: "Yan, Zheng" 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov

ceph: send cap releases more aggressively

2019-03-05T17:55:16+00:00

When pending cap releases fill up one message, start a work to send
cap release message. (old way is sending cap releases every 5 seconds)

Signed-off-by: "Yan, Zheng" 
Reviewed-by: Jeff Layton 
Signed-off-by: Ilya Dryomov