[6.1,00/26] xfs backports to catch 6.1.y up to 6.6

Message

Leah Rumancik Sept. 24, 2024, 6:38 p.m. UTC

Hello again,

Here is the next set of XFS backports, this set is for 6.1.y and I will
be following up with a set for 5.15.y later. There were some good
suggestions made at LSF to survey test coverage to cut back on
testing but I've been a bit swamped and a backport set was overdue.
So for this set, I have run the auto group 3 x 8 configs with no
regressions seen. Let me know if you spot any issues.

This set has already been ack'd on the XFS list.

Thanks,
Leah

https://lkml.iu.edu/hypermail/linux/kernel/2212.0/04860.html
52f31ed22821
[1/1] xfs: dquot shrinker doesn't check for XFS_DQFLAG_FREEING

https://lore.kernel.org/linux-xfs/[e-mail]/
4da112513c01
[1/1] xfs: Fix deadlock on xfs_inodegc_worker

https://www.spinics.net/lists/linux-xfs/msg68547.html
601a27ea09a3
[1/1] xfs: fix extent busy updating

https://www.spinics.net/lists/linux-xfs/msg67254.html
c85007e2e394
[1/1] xfs: don't use BMBT btree split workers for IO completion

fixes from start of the series
https://lore.kernel.org/linux-xfs/[e-mail]/
[00/42] xfs: per-ag centric allocation alogrithms

1dd0510f6d4b
[01/42] xfs: fix low space alloc deadlock

f08f984c63e9
[02/42] xfs: prefer free inodes at ENOSPC over chunk allocation

d5753847b216
[03/42] xfs: block reservation too large for minleft allocation

https://lore.kernel.org/linux-xfs/Y+Z7TZ9o+KgXLcV8@magnolia/
60b730a40c43
[1/1] xfs: fix uninitialized variable access

https://lore.kernel.org/linux-xfs/[e-mail]/
0c7273e494dd
[1/1] xfs: quotacheck failure can race with background inode inactivation

https://lore.kernel.org/linux-xfs/20230412024907.GP360889@frogsfrogsfrogs/
8ee81ed581ff
[1/1] xfs: fix BUG_ON in xfs_getbmap()

https://www.spinics.net/lists/linux-xfs/msg71062.html
[0/4] xfs: bug fixes for 6.4-rc1

[1/4] xfs: don't unconditionally null args->pag in xfs_bmap_btalloc_at_eof
    skip, fix for a commit from 6.3

8e698ee72c4e
[2/4] xfs: set bnobt/cntbt numrecs correctly when formatting new AGs

[3/4] xfs: flush dirty data and drain directios before scrubbing cow fork
    skip, scrub

[4/4] xfs: don't allocate into the data fork for an unshare request
    skip, more of an optimization than a fix

1bba82fe1afa
[5/4] xfs: fix negative array access in xfs_getbmap
  (fix of 8ee81ed)

https://lore.kernel.org/linux-xfs/[e-mail]/
[0/4] xfs: bug fixes for 6.4-rcX

89a4bf0dc385
[1/4] xfs: buffer pins need to hold a buffer reference

[2/4] xfs: restore allocation trylock iteration
  skip, for issue introduced in 6.3

cb042117488d
[3/4] xfs: defered work could create precommits
  (dependency for patch 4)

82842fee6e59
[4/4] xfs: fix AGF vs inode cluster buffer deadlock

https://lore.kernel.org/linux-xfs/[e-mail]/
348a1983cf4c
(fix for 82842fee6e5)
[1/1] xfs: fix unlink vs cluster buffer instantiation race

https://lore.kernel.org/linux-xfs/[e-mail]/
d4d12c02b
[1/1] xfs: collect errors from inodegc for unlinked inode recovery

https://lore.kernel.org/linux-xfs/20230524121041.GA4128075@ceph-admin/
c3b880acadc9
[1/1] xfs: fix ag count overflow during growfs

4b827b3f305d
[1/1] xfs: remove WARN when dquot cache insertion fails
  requested on list to reduce bot noise

https://www.spinics.net/lists/linux-xfs/msg73214.html
5cf32f63b0f4
[1/2] xfs: fix the calculation for "end" and "length"

[2/2] introduces new feature, skipping

https://lore.kernel.org/linux-xfs/[e-mail]/
3c90c01e4934
[1/1] xfs: correct calculation for agend and blockcount
  (fixes 5cf32)

https://lore.kernel.org/all/20230901160020.GT28186@frogsfrogsfrogs/
68b957f64fca
[1/1] xfs: load uncached unlinked inodes into memory on demand

https://www.spinics.net/lists/linux-xfs/msg74960.html
[0/3] xfs: reload entire iunlink lists

f12b96683d69
[1/3] xfs: use i_prev_unlinked to distinguish inodes that are not on the unlinked list

83771c50e42b
[2/3] xfs: reload entire unlinked bucket lists
  (dependency of 49813a21ed)

49813a21ed57
[3/3] xfs: make inode unlinked bucket recovery work with quotacheck
  (dependency for 537c013)

https://lore.kernel.org/all/169565629026.1982077.12646061547002741492.stgit@frogsfrogsfrogs/
537c013b140d
[1/1] xfs: fix reloading entire unlinked bucket lists
  (fix of 68b957f64fca)


Darrick J. Wong (8):
  xfs: fix uninitialized variable access
  xfs: load uncached unlinked inodes into memory on demand
  xfs: fix negative array access in xfs_getbmap
  xfs: use i_prev_unlinked to distinguish inodes that are not on the
    unlinked list
  xfs: reload entire unlinked bucket lists
  xfs: make inode unlinked bucket recovery work with quotacheck
  xfs: fix reloading entire unlinked bucket lists
  xfs: set bnobt/cntbt numrecs correctly when formatting new AGs

Dave Chinner (12):
  xfs: dquot shrinker doesn't check for XFS_DQFLAG_FREEING
  xfs: don't use BMBT btree split workers for IO completion
  xfs: fix low space alloc deadlock
  xfs: prefer free inodes at ENOSPC over chunk allocation
  xfs: block reservation too large for minleft allocation
  xfs: quotacheck failure can race with background inode inactivation
  xfs: buffer pins need to hold a buffer reference
  xfs: defered work could create precommits
  xfs: fix AGF vs inode cluster buffer deadlock
  xfs: collect errors from inodegc for unlinked inode recovery
  xfs: remove WARN when dquot cache insertion fails
  xfs: fix unlink vs cluster buffer instantiation race

Long Li (1):
  xfs: fix ag count overflow during growfs

Shiyang Ruan (2):
  xfs: fix the calculation for "end" and "length"
  xfs: correct calculation for agend and blockcount

Wengang Wang (1):
  xfs: fix extent busy updating

Wu Guanghao (1):
  xfs: Fix deadlock on xfs_inodegc_worker

Ye Bin (1):
  xfs: fix BUG_ON in xfs_getbmap()

 fs/xfs/libxfs/xfs_ag.c          |  19 ++-
 fs/xfs/libxfs/xfs_alloc.c       |  69 +++++++--
 fs/xfs/libxfs/xfs_bmap.c        |  16 +-
 fs/xfs/libxfs/xfs_bmap.h        |   2 +
 fs/xfs/libxfs/xfs_bmap_btree.c  |  19 ++-
 fs/xfs/libxfs/xfs_btree.c       |  18 ++-
 fs/xfs/libxfs/xfs_fs.h          |   2 +
 fs/xfs/libxfs/xfs_ialloc.c      |  17 +++
 fs/xfs/libxfs/xfs_log_format.h  |   9 +-
 fs/xfs/libxfs/xfs_trans_inode.c | 113 +-------------
 fs/xfs/xfs_attr_inactive.c      |   1 -
 fs/xfs/xfs_bmap_util.c          |  18 +--
 fs/xfs/xfs_buf_item.c           |  88 ++++++++---
 fs/xfs/xfs_dquot.c              |   1 -
 fs/xfs/xfs_export.c             |  14 ++
 fs/xfs/xfs_extent_busy.c        |   1 +
 fs/xfs/xfs_fsmap.c              |   1 +
 fs/xfs/xfs_fsops.c              |  13 +-
 fs/xfs/xfs_icache.c             |  58 +++++--
 fs/xfs/xfs_icache.h             |   4 +-
 fs/xfs/xfs_inode.c              | 260 ++++++++++++++++++++++++++++----
 fs/xfs/xfs_inode.h              |  36 ++++-
 fs/xfs/xfs_inode_item.c         | 149 ++++++++++++++++++
 fs/xfs/xfs_inode_item.h         |   1 +
 fs/xfs/xfs_itable.c             |  11 ++
 fs/xfs/xfs_log_recover.c        |  19 ++-
 fs/xfs/xfs_mount.h              |  11 +-
 fs/xfs/xfs_notify_failure.c     |  15 +-
 fs/xfs/xfs_qm.c                 |  72 ++++++---
 fs/xfs/xfs_super.c              |   1 +
 fs/xfs/xfs_trace.h              |  46 ++++++
 fs/xfs/xfs_trans.c              |   9 +-
 32 files changed, 841 insertions(+), 272 deletions(-)

Comments

Dave Chinner Sept. 24, 2024, 10:03 p.m. UTC | #1

On Tue, Sep 24, 2024 at 11:38:25AM -0700, Leah Rumancik wrote:
> Hello again,
> 
> Here is the next set of XFS backports, this set is for 6.1.y and I will
> be following up with a set for 5.15.y later. There were some good
> suggestions made at LSF to survey test coverage to cut back on
> testing but I've been a bit swamped and a backport set was overdue.
> So for this set, I have run the auto group 3 x 8 configs with no
> regressions seen. Let me know if you spot any issues.
> 
> This set has already been ack'd on the XFS list.

Hi Leah, can you pick up this recently requested fix for the series,
too?

https://lore.kernel.org/linux-xfs/[e-mail]/T/

-Dave.