On Tue, 24 Oct 2023 22:27:06 +0800 Zhongkun He <hezhongkun.hzk@xxxxxxxxxxxxx> wrote:
> The zswap_writeback_entry() will add a page to the swap cache, decompress
> the entry data into the page, and issue a bio write to write the page back
> to the swap device. Move the page to the tail of lru list through
> SetPageReclaim(page) and folio_rotate_reclaimable().
>
> Currently, about half of the pages will fail to move to the tail of lru
> list because there is no LRU flag in page which is not in the LRU list but
> the cpu_fbatches. So fix it.
Some reviewer input on this change would be helpful, please.
Also, is there any measurable change in overall performance?
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -1139,6 +1139,11 @@ static int zswap_writeback_entry(struct zswap_entry *entry,
> /* move it to the tail of the inactive list after end_writeback */
> SetPageReclaim(page);
>
> + if (!PageLRU(page)) {
> + /* drain lru cache to help folio_rotate_reclaimable() */
> + lru_add_drain();
> + }
> +
> /* start writeback */
> __swap_writepage(page, &wbc);
> put_page(page);
> --
> 2.25.1
>
