Re: [PATCH] f2fs: skip further SIT updates for already invalidated blocks

[Chao Yu <chao@kernel.org>]

Hi Seongjae,

On 11/12/26 18:36, Seongjae Jeong wrote:
>> I don't see how commit a66cdd985532 ("f2fs: introduce discard_map for f2fs_trim_fs")
>> introduce the bug? from the timeline, at that point there is no error handling in
>> update_sit_entries() if sit bitmap is inconsistent, right?
>>
> 
> Hi Chao,
> 
> Commit a66cdd985532 added the following two lines:
> 
>     if (!f2fs_test_and_clear_bit(offset, se->cur_valid_map))
>         f2fs_bug_on(sbi, 1);

Previously, we will trigger f2fs_bug_on() and there is no further error handling
at that point, so that we will update SIT_I(sbi)->written_valid_blocks as well
incorrectly, so I don't think this is a bug of a66cdd985532.

Thanks,

> + if (f2fs_test_and_clear_bit(offset, se->discard_map))
> +     sbi->discard_blks++;
> 
> Even if the cur_valid_map bit was already clear, the added code
> could still clear the discard bit and increment discard_blks if the
> discard bit was set.
> 
> Since my patch skips that update for an already invalidated block,
> I thought a66cdd985532 was an appropriate Fixes tag for the discard
> bitmap part of the change.
> 
> Thanks,
> Seongjae