tree c8e5139e55f01a2932f48d9fd8f832a4b02112d3
parent 7db83c070bd29e73c8bb42d4b48c976be76f1dbe
parent 27fb5a72f50aa770dd38b0478c07acacef97e3e7
author Linus Torvalds <torvalds@linux-foundation.org> 1585857727 -0700
committer Linus Torvalds <torvalds@linux-foundation.org> 1585857727 -0700
mergetag object 27fb5a72f50aa770dd38b0478c07acacef97e3e7
 type commit
 tag xfs-5.7-merge-8
 tagger Darrick J. Wong <darrick.wong@oracle.com> 1585237761 -0700
 
 New code for 5.7:
  - Fix a hard to trigger race between iclog error checking and log shutdown.
  - Strengthen the AGF verifier.
  - Ratelimit some of the more spammy error messages.
  - Remove the icdinode uid/gid members and just use the ones in the vfs inode.
  - Hold ILOCK across insert/collapse range.
  - Clean up the extended attribute interfaces.
  - Clean up the attr flags mess.
  - Restore PF_MEMALLOC after exiting xfsaild thread to avoid triggering
    warnings in the process accounting code.
  - Remove the flexibly-sized array from struct xfs_agfl to eliminate
    compiler warnings about unaligned pointers and packed structures.
  - Various macro and typedef removals.
  - Stale metadata buffers if we decide they're corrupt outside of a
    verifier.
  - Check directory data/block/free block owners.
  - Fix a UAF when aborting inactivation of a corrupt xattr fork.
  - Teach online scrub to report failed directory and attr name lookups
    as a metadata corruption instead of a runtime error.
  - Avoid potential buffer overflows in sysfs files by using scnprintf.
  - Fix a regression in getdents lookups due to a mistake in pointer
    arithmetic.
  - Refactor btree cursor private data structures to use anonymous
    unions.
  - Cleanups in the log unmounting code.
  - Fix a potential mishandling of ENOMEM errors on multi-block directory
    buffer lookups.
  - Fix an incorrect test in the block allocation code.
  - Cleanups and name prefix shortening in the scrub code.
  - Introduce btree bulk loading code for online repair and scrub.
  - Fix a quotaoff log item leak (and hang) when the fs goes down midway
    through a quotaoff operation.
  - Remove di_version from the incore inode.
  - Refactor some of the log shutdown checking code.
  - Record the forcing of the log unmount records in the log force
    counters.
  - Fix a longstanding bug where quotacheck would purge the
    administrator's default quota grace interval and warning limits.
  - Reduce memory usage when scrubbing directory and xattr trees.
  - Don't let fsfreeze race with GETFSMAP or online scrub.
  - Handle bio_add_page failures more gracefully in xlog_write_iclog.
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCgAdFiEEUzaAxoMeQq6m2jMV+H93GTRKtOsFAl58z3AACgkQ+H93GTRK
 tOukrRAAhJmowV5+Req5YMYawRjafkIbCDH3WlFy9AdpFFA6pXSfX6YCtIKwKfq8
 +yRj/BFRGoMc6SouXo+J0i3YMS6yQZTjcmVWrQPVnj/+DGVjh+Y70gKExtz2CyjO
 ItGGxpRwOhpw49zVYmcH6Mrw8sBztHR0VsM0cq6YfJrkNcm0BsnAC+W6zQNaDG24
 UO1ivehBOooVh0C8pv0smVcPtBL2N+RRyS3XRT5hGFozUJgLLGDqnHAl1d+KOrWp
 hPQhUlDw9luiHPBxWkxUuFDr79gjUi7kyHILNt7TIkByyRcTUO9jhS2VpZd4oXlj
 /J3i1AS+9lhP1yGVxw2RHQhKMvdYBQiLADSCpzkA1dMma99cFGyzMMA6rG0WRMJ4
 erXxhAEoM4um3gxDka6+HJxySLOT8E22FesJbn6YIv4QSAkXDBPWz/9hPbjJuJQm
 6Y/YkFOZLp3c+xJM0tpCWxWaWW7A+t2OMRIFISSsXesrySpalpbkVXkHwz3NwO6L
 3SeTnLWqnADbjl2qsuyF0uYHqURygVz7g+r4X7AO5D1IRyCCkmtDOuwumxERiQ3p
 3vZMQrWh+y3SgRiF8brDG5KTshhxcinKdHEYXrwq3XgaHZg4mtLI4XjOyZlJruoX
 MGWhZjga6+RGysH0RKjZbHaMr/f4m3X00SHa/5Ibcp6Q21TIx6M=
 =8iJB
 -----END PGP SIGNATURE-----

Merge tag 'xfs-5.7-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs updates from Darrick Wong:
 "There's a lot going on this cycle with cleanups in the log code, the
  btree code, and the xattr code.

  We're tightening of metadata validation and online fsck checking, and
  introducing a common btree rebuilding library so that we can refactor
  xfs_repair and introduce online repair in a future cycle.

  We also fixed a few visible bugs -- most notably there's one in
  getdents that we introduced in 5.6; and a fix for hangs when disabling
  quotas.

  This series has been running fstests & other QA in the background for
  over a week and looks good so far.

  I anticipate sending a second pull request next week. That batch will
  change how xfs interacts with memory reclaim; how the log batches and
  throttles log items; how hard writes near ENOSPC will try to squeeze
  more space out of the filesystem; and hopefully fix the last of the
  umount hangs after a catastrophic failure. That should ease a lot of
  problems when running at the limits, but for now I'm leaving that in
  for-next for another week to make sure we got all the subtleties
  right.

  Summary:

   - Fix a hard to trigger race between iclog error checking and log
     shutdown.

   - Strengthen the AGF verifier.

   - Ratelimit some of the more spammy error messages.

   - Remove the icdinode uid/gid members and just use the ones in the
     vfs inode.

   - Hold ILOCK across insert/collapse range.

   - Clean up the extended attribute interfaces.

   - Clean up the attr flags mess.

   - Restore PF_MEMALLOC after exiting xfsaild thread to avoid
     triggering warnings in the process accounting code.

   - Remove the flexibly-sized array from struct xfs_agfl to eliminate
     compiler warnings about unaligned pointers and packed structures.

   - Various macro and typedef removals.

   - Stale metadata buffers if we decide they're corrupt outside of a
     verifier.

   - Check directory data/block/free block owners.

   - Fix a UAF when aborting inactivation of a corrupt xattr fork.

   - Teach online scrub to report failed directory and attr name lookups
     as a metadata corruption instead of a runtime error.

   - Avoid potential buffer overflows in sysfs files by using scnprintf.

   - Fix a regression in getdents lookups due to a mistake in pointer
     arithmetic.

   - Refactor btree cursor private data structures to use anonymous
     unions.

   - Cleanups in the log unmounting code.

   - Fix a potential mishandling of ENOMEM errors on multi-block
     directory buffer lookups.

   - Fix an incorrect test in the block allocation code.

   - Cleanups and name prefix shortening in the scrub code.

   - Introduce btree bulk loading code for online repair and scrub.

   - Fix a quotaoff log item leak (and hang) when the fs goes down
     midway through a quotaoff operation.

   - Remove di_version from the incore inode.

   - Refactor some of the log shutdown checking code.

   - Record the forcing of the log unmount records in the log force
     counters.

   - Fix a longstanding bug where quotacheck would purge the
     administrator's default quota grace interval and warning limits.

   - Reduce memory usage when scrubbing directory and xattr trees.

   - Don't let fsfreeze race with GETFSMAP or online scrub.

   - Handle bio_add_page failures more gracefully in xlog_write_iclog"

* tag 'xfs-5.7-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (108 commits)
  xfs: prohibit fs freezing when using empty transactions
  xfs: shutdown on failure to add page to log bio
  xfs: directory bestfree check should release buffers
  xfs: drop all altpath buffers at the end of the sibling check
  xfs: preserve default grace interval during quotacheck
  xfs: remove xlog_state_want_sync
  xfs: move the ioerror check out of xlog_state_clean_iclog
  xfs: refactor xlog_state_clean_iclog
  xfs: remove the aborted parameter to xlog_state_done_syncing
  xfs: simplify log shutdown checking in xfs_log_release_iclog
  xfs: simplify the xfs_log_release_iclog calling convention
  xfs: factor out a xlog_wait_on_iclog helper
  xfs: merge xlog_cil_push into xlog_cil_push_work
  xfs: remove the di_version field from struct icdinode
  xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
  xfs: simplify di_flags2 inheritance in xfs_ialloc
  xfs: only check the superblock version for dinode size calculation
  xfs: add a new xfs_sb_version_has_v3inode helper
  xfs: fix unmount hang and memory leak on shutdown during quotaoff
  xfs: factor out quotaoff intent AIL removal and memory free
  ...