On Wed 22-01-25 19:41:28, libaokun@huaweicloud.com wrote:
> From: Baokun Li <libaokun1@huawei.com>
>
> And after commit 95257987a638 ("ext4: drop EXT4_MF_FS_ABORTED flag") in
> v6.6-rc1, the EXT4_FLAGS_SHUTDOWN bit is set in ext4_handle_error() under
> errors=remount-ro mode. This causes the read to fail even when the error
> is triggered in errors=remount-ro mode.
>
> To correct the behavior under errors=remount-ro, EXT4_FLAGS_SHUTDOWN is
> replaced by the newly introduced EXT4_FLAGS_EMERGENCY_RO. This new flag
> only prevents writes, matching the previous behavior with SB_RDONLY.
>
> Fixes: 95257987a638 ("ext4: drop EXT4_MF_FS_ABORTED flag")
> Closes: https://lore.kernel.org/all/22d652f6-cb3c-43f5-b2fe-0a4bb6516a04@huawei.com/
> Suggested-by: Jan Kara <jack@suse.cz>
> Signed-off-by: Baokun Li <libaokun1@huawei.com>
> Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Looks good. Feel free to add:
Reviewed-by: Jan Kara <jack@suse.cz>
Honza
> ---
> fs/ext4/super.c | 19 ++++++++-----------
> 1 file changed, 8 insertions(+), 11 deletions(-)
>
> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
> index d8116c9c2bd0..098e62727aec 100644
> --- a/fs/ext4/super.c
> +++ b/fs/ext4/super.c
> @@ -708,11 +708,8 @@ static void ext4_handle_error(struct super_block *sb, bool force_ro, int error,
> if (test_opt(sb, WARN_ON_ERROR))
> WARN_ON_ONCE(1);
>
> - if (!continue_fs && !sb_rdonly(sb)) {
> - set_bit(EXT4_FLAGS_SHUTDOWN, &EXT4_SB(sb)->s_ext4_flags);
> - if (journal)
> - jbd2_journal_abort(journal, -EIO);
> - }
> + if (!continue_fs && !ext4_emergency_ro(sb) && journal)
> + jbd2_journal_abort(journal, -EIO);
>
> if (!bdev_read_only(sb->s_bdev)) {
> save_error_info(sb, error, ino, block, func, line);
> @@ -738,17 +735,17 @@ static void ext4_handle_error(struct super_block *sb, bool force_ro, int error,
> sb->s_id);
> }
>
> - if (sb_rdonly(sb) || continue_fs)
> + if (ext4_emergency_ro(sb) || continue_fs)
> return;
>
> ext4_msg(sb, KERN_CRIT, "Remounting filesystem read-only");
> /*
> - * EXT4_FLAGS_SHUTDOWN was set which stops all filesystem
> - * modifications. We don't set SB_RDONLY because that requires
> - * sb->s_umount semaphore and setting it without proper remount
> - * procedure is confusing code such as freeze_super() leading to
> - * deadlocks and other problems.
> + * We don't set SB_RDONLY because that requires sb->s_umount
> + * semaphore and setting it without proper remount procedure is
> + * confusing code such as freeze_super() leading to deadlocks
> + * and other problems.
> */
> + set_bit(EXT4_FLAGS_EMERGENCY_RO, &EXT4_SB(sb)->s_ext4_flags);
> }
>
> static void update_super_work(struct work_struct *work)
> --
> 2.39.2
>
--
Jan Kara <jack@suse.com>
SUSE Labs, CR