Re: [PATCH] mm/zswap: remove SWP_SYNCHRONOUS_IO swapcache bypass workaround

[Chengming Zhou <chengming.zhou@linux.dev>]

On 2026/2/2 01:47, Kairui Song wrote:
> From: Kairui Song <kasong@tencent.com>
> 
> Since commit f1879e8a0c60 ("mm, swap: never bypass the swap cache even
> for SWP_SYNCHRONOUS_IO"), all swap-in operations go through the swap
> cache, including those from SWP_SYNCHRONOUS_IO devices like
> zram. Which means the workaround for swap cache bypassing
> introduced by commit 25cd241408a2 ("mm: zswap: fix data loss on
> SWP_SYNCHRONOUS_IO devices") is no longer needed. Remove it, but
> keep the comments that are still helpful.
> 
> Suggested-by: Yosry Ahmed <yosry.ahmed@linux.dev>
> Signed-off-by: Kairui Song <kasong@tencent.com>

Reviewed-by: Chengming Zhou <chengming.zhou@linux.dev>

Thanks!

> ---
>   mm/zswap.c | 19 ++++++-------------
>   1 file changed, 6 insertions(+), 13 deletions(-)
> 
> diff --git a/mm/zswap.c b/mm/zswap.c
> index 3d2d59ac3f9c..8cd61603ff79 100644
> --- a/mm/zswap.c
> +++ b/mm/zswap.c
> @@ -1589,11 +1589,11 @@ int zswap_load(struct folio *folio)
>   {
>   	swp_entry_t swp = folio->swap;
>   	pgoff_t offset = swp_offset(swp);
> -	bool swapcache = folio_test_swapcache(folio);
>   	struct xarray *tree = swap_zswap_tree(swp);
>   	struct zswap_entry *entry;
>   
>   	VM_WARN_ON_ONCE(!folio_test_locked(folio));
> +	VM_WARN_ON_ONCE(!folio_test_swapcache(folio));
>   
>   	if (zswap_never_enabled())
>   		return -ENOENT;
> @@ -1624,22 +1624,15 @@ int zswap_load(struct folio *folio)
>   		count_objcg_events(entry->objcg, ZSWPIN, 1);
>   
>   	/*
> -	 * When reading into the swapcache, invalidate our entry. The
> -	 * swapcache can be the authoritative owner of the page and
> +	 * We are reading into the swapcache, invalidate zswap entry.
> +	 * The swapcache is the authoritative owner of the page and
>   	 * its mappings, and the pressure that results from having two
>   	 * in-memory copies outweighs any benefits of caching the
>   	 * compression work.
> -	 *
> -	 * (Most swapins go through the swapcache. The notable
> -	 * exception is the singleton fault on SWP_SYNCHRONOUS_IO
> -	 * files, which reads into a private page and may free it if
> -	 * the fault fails. We remain the primary owner of the entry.)
>   	 */
> -	if (swapcache) {
> -		folio_mark_dirty(folio);
> -		xa_erase(tree, offset);
> -		zswap_entry_free(entry);
> -	}
> +	folio_mark_dirty(folio);
> +	xa_erase(tree, offset);
> +	zswap_entry_free(entry);
>   
>   	folio_unlock(folio);
>   	return 0;
> 
> ---
> base-commit: 2c263046cbe6d9d5fce3dfeba063f199f7e6298f
> change-id: 20251226-zswap-syncio-cleanup-a05b7fc6180f
> 
> Best regards,