Re: [PATCH 1/6] mm/page_io: convert write completion handlers to folios

On Thu, Aug 06, 2026 at 06:55:33PM -0400, Tal Zussman wrote:
> Convert swap_write_end() and swap_fs_write_complete() to operate on
> folios directly instead of going through the folio-compat page APIs.
> This removes calls to end_page_writeback() and set_page_dirty(), and the
> last caller of ClearPageReclaim(), saving two calls to compound_head()
> per folio on the write error path.
>
> Signed-off-by: Tal Zussman <[email protected]>

Very lovely, thanks! :)

I see you remove PageReclaim in the subsequent commit which makes sense as a
follow-up.

Reviewed-by: Lorenzo Stoakes (ARM) <[email protected]>

> ---
>  mm/page_io.c | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/mm/page_io.c b/mm/page_io.c
> index e4fa7ffffe8b..43b4383a4729 100644
> --- a/mm/page_io.c
> +++ b/mm/page_io.c
> @@ -496,13 +496,13 @@ static void swap_write_end(struct swap_iocb *sio, bool failed)
>  	int p;
>
>  	for (p = 0; p < sio->nr_bvecs; p++) {
> -		struct page *page = sio->bvecs[p].bv_page;
> +		struct folio *folio = bvec_folio(&sio->bvecs[p]);
>
>  		if (failed) {
> -			set_page_dirty(page);
> -			ClearPageReclaim(page);
> +			folio_mark_dirty(folio);
> +			folio_clear_reclaim(folio);
>  		}
> -		end_page_writeback(page);
> +		folio_end_writeback(folio);
>  	}
>  	mempool_free(sio, sio_pool);
>  }
> @@ -513,16 +513,16 @@ static void swap_fs_write_complete(struct kiocb *iocb, long ret)
>  	bool failed = ret != sio->len;
>
>  	if (failed) {
> -		struct page *page = sio->bvecs[0].bv_page;
> +		struct folio *folio = bvec_folio(&sio->bvecs[0]);
>
>  		/*
>  		 * In the case of swap-over-nfs, this can be a temporary failure
>  		 * if the system has limited memory for allocating transmit
> -		 * buffers.  Mark the page dirty and avoid
> +		 * buffers.  Mark the folio dirty and avoid
>  		 * folio_rotate_reclaimable but rate-limit the messages.
>  		 */
>  		pr_err_ratelimited("Write error %ld on dio swapfile (%llu)\n",
> -				   ret, swap_dev_pos(page_swap_entry(page)));
> +				   ret, swap_dev_pos(folio->swap));
>  	}
>
>  	swap_write_end(sio, failed);
>
> --
> 2.39.5
>

--
Cheers, Lorenzo