Re: On https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=775c17386a6fd

On 5/26/26 7:09 PM, Dmitry Antipov wrote:
> Note that [1] opens the door for the things like [2] and [3] again. This happens
> when ocfs2_validate_dx_root() detects bad signature and ocfs2_error() remounts
> the filesystem readonly; since read operations are still allowed after that,
> ocfs2_dx_dir_lookup_rec() may be issued for a dx root block with invalid extent
> records (in particular, out-of-range l_count and l_next_free_rec values).
> 
> Dmitry
> 
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=775c17386a6fd
> [2] https://syzkaller.appspot.com/bug?extid=caacd220635a9cc3bac9
> [3] https://syzkaller.appspot.com/bug?extid=30b53487d00b4f7f0922

Don't understand in which case we will see a corrupted dr_list in
ocfs2_dx_dir_lookup_rec().

When ocfs2_validate_dx_root() fails, it will jump to read_failure logic
and buffer is not set to uptodate.

Could you please list a detail flow for it?

Thanks,
Joseph