[PATCH] mm: vmscan: remove the redundant FOLIOREF_RECLAIM_CLEAN logic

[PATCH] mm: vmscan: remove the redundant FOLIOREF_RECLAIM_CLEAN logic

[Baolin Wang]

folio_check_references() will return FOLIOREF_RECLAIM_CLEAN for referenced
file folios, indicating that we can proceed to reclaim clean file folios
or keep them if they are dirty file folios. However, after commit 6b0dfabb3555
("fs: Remove aops->writepage"), we no longer attempt to write back filesystem
folios through reclaim. Instead, we always activate dirty file folios and wakeup
the flush workers to write them back. As a result, the FOLIOREF_RECLAIM_CLEAN
logic is now redundant: for dirty file folios, we will no longer reach the
'references == FOLIOREF_RECLAIM_CLEAN' branch in shrink_folio_list().

Additionally, lazyfree folios are also placed on the file LRU list, but if a
lazyfree folio becomes dirty, try_to_unmap() will fail and thus prevent reclaim
of the re-dirtied lazyfree folios.

Therefore, we can drop the FOLIOREF_RECLAIM_CLEAN-related logic.

Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>
---
 mm/vmscan.c | 8 --------
 1 file changed, 8 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 299b5d9e8836..2f7e5878faa5 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -823,7 +823,6 @@ void folio_putback_lru(struct folio *folio)
 
 enum folio_references {
 	FOLIOREF_RECLAIM,
-	FOLIOREF_RECLAIM_CLEAN,
 	FOLIOREF_KEEP,
 	FOLIOREF_ACTIVATE,
 };
@@ -920,10 +919,6 @@ static enum folio_references folio_check_references(struct folio *folio,
 		return FOLIOREF_KEEP;
 	}
 
-	/* Reclaim if clean, defer dirty folios to writeback */
-	if (referenced_folio && folio_is_file_lru(folio))
-		return FOLIOREF_RECLAIM_CLEAN;
-
 	return FOLIOREF_RECLAIM;
 }
 
@@ -1235,7 +1230,6 @@ static unsigned int shrink_folio_list(struct list_head *folio_list,
 			stat->nr_ref_keep += nr_pages;
 			goto keep_locked;
 		case FOLIOREF_RECLAIM:
-		case FOLIOREF_RECLAIM_CLEAN:
 			; /* try to reclaim the folio below */
 		}
 
@@ -1381,8 +1375,6 @@ static unsigned int shrink_folio_list(struct list_head *folio_list,
 				goto activate_locked;
 			}
 
-			if (references == FOLIOREF_RECLAIM_CLEAN)
-				goto keep_locked;
 			if (!may_enter_fs(folio, sc->gfp_mask))
 				goto keep_locked;
 			if (!sc->may_writepage)
-- 
2.47.3

Re: [PATCH] mm: vmscan: remove the redundant FOLIOREF_RECLAIM_CLEAN logic

[Johannes Weiner]

On Mon, Jun 29, 2026 at 04:04:06PM +0800, Baolin Wang wrote:
> folio_check_references() will return FOLIOREF_RECLAIM_CLEAN for referenced
> file folios, indicating that we can proceed to reclaim clean file folios
> or keep them if they are dirty file folios. However, after commit 6b0dfabb3555
> ("fs: Remove aops->writepage"), we no longer attempt to write back filesystem
> folios through reclaim. Instead, we always activate dirty file folios and wakeup
> the flush workers to write them back. As a result, the FOLIOREF_RECLAIM_CLEAN
> logic is now redundant: for dirty file folios, we will no longer reach the
> 'references == FOLIOREF_RECLAIM_CLEAN' branch in shrink_folio_list().
> 
> Additionally, lazyfree folios are also placed on the file LRU list, but if a
> lazyfree folio becomes dirty, try_to_unmap() will fail and thus prevent reclaim
> of the re-dirtied lazyfree folios.
> 
> Therefore, we can drop the FOLIOREF_RECLAIM_CLEAN-related logic.
> 
> Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>

Good catch and nice cleanup :)

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

Re: [PATCH] mm: vmscan: remove the redundant FOLIOREF_RECLAIM_CLEAN logic

[Shakeel Butt]

On Mon, Jun 29, 2026 at 04:04:06PM +0800, Baolin Wang wrote:
> folio_check_references() will return FOLIOREF_RECLAIM_CLEAN for referenced
> file folios, indicating that we can proceed to reclaim clean file folios
> or keep them if they are dirty file folios. However, after commit 6b0dfabb3555
> ("fs: Remove aops->writepage"), we no longer attempt to write back filesystem
> folios through reclaim. Instead, we always activate dirty file folios and wakeup
> the flush workers to write them back. As a result, the FOLIOREF_RECLAIM_CLEAN
> logic is now redundant: for dirty file folios, we will no longer reach the
> 'references == FOLIOREF_RECLAIM_CLEAN' branch in shrink_folio_list().
> 
> Additionally, lazyfree folios are also placed on the file LRU list, but if a
> lazyfree folio becomes dirty, try_to_unmap() will fail and thus prevent reclaim
> of the re-dirtied lazyfree folios.
> 
> Therefore, we can drop the FOLIOREF_RECLAIM_CLEAN-related logic.
> 
> Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com>

Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>