On Fri 23-02-18 16:06:15, Jeff Moyer wrote:
> Prior to commit d47992f86b307 ("mm: change invalidatepage prototype to
> accept length"), an offset of 0 meant that the full page was being
> invalidated. After that commit, we need to instead check the length.
>
> Fixes: d47992f86b307 ("mm: change invalidatepage prototype to accept length")
> Signed-off-by: Jeff Moyer <jmoyer@xxxxxxxxxx>
>
> ---
> Note that I found this via code inspection. I'm not sure if this causes
> issues in the wild.
The only possible issue is that try_to_release_page() was called more often
than necessary. Otherwise the issue is harmless but still it's good to have
this fixed. The patch looks good. You can add:
Reviewed-by: Jan Kara <jack@xxxxxxx>
Honza
>
> diff --git a/fs/buffer.c b/fs/buffer.c
> index 9a73924db22f..6b98609356ed 100644
> --- a/fs/buffer.c
> +++ b/fs/buffer.c
> @@ -1511,7 +1511,7 @@ void block_invalidatepage(struct page *page, unsigned int offset,
> * The get_block cached value has been unconditionally invalidated,
> * so real IO is not possible anymore.
> */
> - if (offset == 0)
> + if (length == PAGE_SIZE)
> try_to_release_page(page, 0);
> out:
> return;
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
