Re: [PATCH 0/4] btrfs: squota lifetime fixes

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

在 2026/5/13 05:04, Boris Burkov 写道:
> A recent report of read-only filesystems since enabling squotas:
> 
> https://lore.kernel.org/linux-btrfs/adnBhWfJQ1n3hZC8@merlins.org/
> 
> led me to investigate a handful of issues in squotas and find some trivial and
> not so trivial bugs in their lifetime management. The aim of the series is to
> strengthen enforcement of the invariant: "a squota qgroup must outlast extents
> that are accounted to it".
> 
> This series consists of two fixes to squota lifetime bugs and two patches that
> reduce the consequences of errors in lifetime management from full fs goes
> read-only abort to "warning and inaccurate squotas which can be deleted
> eventually"
> 
> Additional improvements to btrfs check and fstests to follow.

Only minor debug related comments in the last two patches.

Otherwise looks good to me.

Reviewed-by: Qu Wenruo <wqu@suse.com>

Thanks,
Qu

> 
> Boris Burkov (4):
>    btrfs: check for subvolume before deleting squota qgroup
>    btrfs: fix squota accounting during enable generation
>    btrfs: clamp to avoid squota underflow
>    btrfs: swallow btrfs_record_squota_delta() enoent
> 
>   fs/btrfs/fs.h     |  1 +
>   fs/btrfs/qgroup.c | 99 ++++++++++++++++++++++++++++++++---------------
>   2 files changed, 68 insertions(+), 32 deletions(-)
>