Re: [PATCHSET v3 1/2] xfs: LLM-inspired bug fixes, part 1

On Mon, 13 Jul 2026 23:03:57 -0700, Darrick J. Wong wrote:
> Someone gave me a pretty generous license for a large language model, so
> I pointed it at every single file under fs/xfs and told it to find bugs.
> After personally filtering out the errors and garbage, here are fixes
> for the real bugs that I think it found.
> 
> Most of these fixes target the online fsck code because I'm most
> familiar with it, but the LLM will read other parts of the xfs codebase
> so there are miscellaneous bug fixes.
> 
> [...]

Applied to for-next, thanks!

[1/8] xfs: don't replace the wrong part of the cow fork
      commit: a1caeeadbf57ff86dfc3454398c46de86056a74e
[2/8] xfs: make cow repair somewhat flaky when debugging knob enabled
      commit: bcb0621204e6999bb8f3a6c7deb5c6abfe0f197b
[3/8] xfs: move cow_replace_mapping to xfs_bmap_util.c
      commit: 60a1dde9d28948a88ed5a06cf0b6179b1d487f85
[4/8] xfs: don't wrap around quota ids in dqiterate
      commit: d766e4e5e85d829629c3ba503802fe1303d7b591
[5/8] xfs: use rtrefcount btree cursor in xchk_xref_is_rt_cow_staging
      commit: ee248157da501f0c02688fb64e5359f2832b0b01
[6/8] xfs: use the rt version of the cow staging checker
      commit: 881f2eb0fcdea2f46a50fa6a892fe9e3ccf19e01
[7/8] xfs: write the rg superblock when fixing it
      commit: 9af789fa274bc14b907f811f79fa988a6fc6d4e7
[8/8] xfs: grab rtrmap btree when checking rgsuper
      commit: ea6e2d9de25d2095845e0cdf2274e581fd40d636

Best regards,
-- 
Carlos Maiolino <[email protected]>