Re: [PATCH v1 8/9] iomap: use find_next_bit() for dirty bitmap scanning

[Christoph Hellwig <hch@infradead.org>]

On Thu, Oct 09, 2025 at 03:56:10PM -0700, Joanne Koong wrote:
> Use find_next_bit()/find_next_zero_bit() for iomap dirty bitmap
> scanning. This uses __ffs() internally and is more efficient for
> finding the next dirty or clean bit than manually iterating through the
> bitmap range testing every bit.
> 
> Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
> Suggested-by: Christoph Hellwig <hch@infradead.org>
> ---
>  fs/iomap/buffered-io.c | 73 ++++++++++++++++++++++++++++++------------
>  1 file changed, 52 insertions(+), 21 deletions(-)
> 
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index 66c47404787f..37d2b76ca230 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -76,15 +76,49 @@ static void iomap_set_range_uptodate(struct folio *folio, size_t off,
>  		folio_mark_uptodate(folio);
>  }
>  
> -static inline bool ifs_block_is_dirty(struct folio *folio,
> -		struct iomap_folio_state *ifs, int block)
> +/**
> +* ifs_next_dirty_block - find the next dirty block in the folio
> +* @folio: The folio
> +* @start_blk: Block number to begin searching at
> +* @end_blk: Last block number (inclusive) to search
> +*
> +* If no dirty block is found, this will return end_blk + 1.
> +*/

I'm not a huge fan of using the very verbose kerneldoc comments for
static functions where this isn't even turned into generated
documentation.  Can you shorten the comments into normal non-structured
ones?

> +	if (start_blk > end_blk)
> +		return 0;
> +	else if (start_blk == end_blk)

No need for an else after a return.

> +		nblks = ifs_next_clean_block(folio, start_blk + 1, end_blk)
> +				- start_blk;

Kernel style keeps the operator before the line break.

> +	for_each_clean_block(folio, first_blk, last_blk)

Given that this is the only user of this macro in the entire series,
what is the point of it?  Just open coding the search would be simpler.