[GIT PULL] Btrfs fixes for 7.2-rc6

[GIT PULL] Btrfs fixes for 7.2-rc6

[David Sterba]

Hi,

please pull the following btrfs fixes. They're one-liners or short
focused fixes (I've removed other less serious ones).

A heads up for one more fix, for a problem that resurfaced last week. In
short we have to bring back the "fixup worker", detect and handle pages
set dirty without filesystem notification. There's a reliable reproducer
for at least one such case. The fix heavily builds on the removed code
so we have a solid base, the new changes are wrt folios and subpage mode.

The v3 [1] seems to be the final one, we'll give it a test, estimated
time of the pull request is next week.

[1] https://lore.kernel.org/linux-btrfs/69d0043e0f6a3d17048dfde857127ab0bf331154.1785190866.git.boris@bur.io/

This was not expected and we hoped to leave the fixup worker in the
past, but it's hard to argue with a reproducer. The remaining time
before release should be sufficient to get the fix finalized.

Thanks.


- zoned mode:
  - fix assertion and handle case of finished zone and truncated extent
  - fix zone metadata write pointer on actual zone reset
  - fix deadlock caused metadata writeback and transaction commit
  - fix return value reuse leading to confusion about chunk reservations

- raid56 scrub
  - fix tracking of sector checksums when there are not checksums found
  - fix inverted logic when submitting parity read bio

- mount/remount fixes
  - fix leaking 'remount in progress' state which can break other
    operations to work (qgroup rescan, autodefrag, reclaim)
  - adjust using global block reserve after read-only mount when using
    rescue= option
  - handle missing raid stripe tree when mounted with 'ignorebadroots'

- fix -Wmaybe-uninitialized warning in GET_CSUMS ioctl

----------------------------------------------------------------
The following changes since commit 75a41e3e51eb0d6f7a4969c5056298c2aa3e85f2:

  btrfs: fix GET_SUBVOL_INFO after compat refactor (2026-07-14 07:05:10 +0200)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux.git tags/for-7.2-rc5-tag

for you to fetch changes up to c4c0673e4cb15b0c127e6d00732a2427bdd12c11:

  btrfs: raid56: fix scrub read assembly submitting no reads (2026-07-21 06:44:08 +0200)

----------------------------------------------------------------
Dongjiang Zhu (2):
      btrfs: skip global block reserve accounting for rescue mounts
      btrfs: report missing raid stripe tree root during lookup

Guanghui Yang (1):
      btrfs: zoned: fix missing chunk metadata reservation

Johannes Thumshirn (3):
      btrfs: zoned: fix deadlock between metadata writeback and transaction commit
      btrfs: zoned: reset meta_write_pointer on zone reset
      btrfs: zoned: skip fully truncated ordered extents at zone finish

Mykola Lysenko (1):
      btrfs: raid56: fix scrub read assembly submitting no reads

Paul E. McKenney (1):
      btrfs: initialize 'args' to avoid compiler warning in btrfs_ioctl_get_csums()

Qu Wenruo (2):
      btrfs: fix leaking BTRFS_FS_STATE_REMOUNTING flag
      btrfs: raid56: fix an incorrect csum skip during scrub

 fs/btrfs/block-group.c      | 34 +++++++++++++++++++---------------
 fs/btrfs/block-rsv.c        | 19 +++++++++++++++++--
 fs/btrfs/disk-io.c          | 11 +----------
 fs/btrfs/fs.h               |  9 +++++++++
 fs/btrfs/ioctl.c            |  2 +-
 fs/btrfs/raid-stripe-tree.c |  6 ++++++
 fs/btrfs/raid56.c           | 15 ++++++++-------
 fs/btrfs/super.c            |  8 +++++---
 fs/btrfs/zoned.c            | 27 ++++++++++++++++++++++++++-
 9 files changed, 92 insertions(+), 39 deletions(-)

Re: [GIT PULL] Btrfs fixes for 7.2-rc6

[Linus Torvalds]

On Tue, 28 Jul 2026 at 02:25, David Sterba <dsterba@suse.com> wrote:
>
> A heads up for one more fix, for a problem that resurfaced last week. In
> short we have to bring back the "fixup worker", detect and handle pages
> set dirty without filesystem notification. There's a reliable reproducer
> for at least one such case. The fix heavily builds on the removed code
> so we have a solid base, the new changes are wrt folios and subpage mode.

Ok, not optimal timing, but I guess if it's mostly resurrecting old
tested code it's all good.

          Linus