mm: skip dirty file folios during isolation

[PATCH] mm: skip dirty file folios during isolation

Posted by zhaoyang.huang 2 weeks, 5 days ago

From: Zhaoyang Huang <zhaoyang.huang@unisoc.com>

Since dirty file folios are no longer writeout in reclaiming after
'commit 84798514db50 ("mm: Remove swap_writepage() and
shmem_writepage()")', there is no need to isolate them from LRU which
could help to improve the scan efficiency.

Signed-off-by: Zhaoyang Huang <zhaoyang.huang@unisoc.com>
---
 mm/vmscan.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 10f1e7d716ca..3a029dede5ce 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -4507,6 +4507,7 @@ static bool sort_folio(struct lruvec *lruvec, struct folio *folio, struct scan_c
 static bool isolate_folio(struct lruvec *lruvec, struct folio *folio, struct scan_control *sc)
 {
 	bool success;
+	struct address_space *mapping = folio_mapping(folio);
 
 	/* swap constrained */
 	if (!(sc->gfp_mask & __GFP_IO) &&
@@ -4514,6 +4515,11 @@ static bool isolate_folio(struct lruvec *lruvec, struct folio *folio, struct sca
 	     (folio_test_anon(folio) && !folio_test_swapcache(folio))))
 		return false;
 
+	/* skip dirty file pages since they can't be writeout */
+	if (mapping && !shmem_mapping(mapping) &&
+		!folio_test_anon(folio) && folio_test_dirty(folio))
+		return false;
+
 	/* raced with release_pages() */
 	if (!folio_try_get(folio))
 		return false;
-- 
2.25.1

Re: [PATCH] mm: skip dirty file folios during isolation

Posted by Johannes Weiner 2 weeks, 5 days ago

On Wed, Mar 18, 2026 at 05:06:56PM +0800, zhaoyang.huang wrote:
> From: Zhaoyang Huang <zhaoyang.huang@unisoc.com>
> 
> Since dirty file folios are no longer writeout in reclaiming after
> 'commit 84798514db50 ("mm: Remove swap_writepage() and
> shmem_writepage()")', there is no need to isolate them from LRU which
> could help to improve the scan efficiency.

The reclaim scanner needs to sample dirty state to coordinate if early
writeback is necessary:

try_to_shrink_lruvec():
	/*
	 * If too many file cache in the coldest generation can't be evicted
	 * due to being dirty, wake up the flusher.
	 */
	if (sc->nr.unqueued_dirty && sc->nr.unqueued_dirty == sc->nr.file_taken)
		wakeup_flusher_threads(WB_REASON_VMSCAN);

Re: [PATCH] mm: skip dirty file folios during isolation

Posted by Zhaoyang Huang 2 weeks, 4 days ago

On Wed, Mar 18, 2026 at 9:50 PM Johannes Weiner <hannes@cmpxchg.org> wrote:
>
> On Wed, Mar 18, 2026 at 05:06:56PM +0800, zhaoyang.huang wrote:
> > From: Zhaoyang Huang <zhaoyang.huang@unisoc.com>
> >
> > Since dirty file folios are no longer writeout in reclaiming after
> > 'commit 84798514db50 ("mm: Remove swap_writepage() and
> > shmem_writepage()")', there is no need to isolate them from LRU which
> > could help to improve the scan efficiency.
>
> The reclaim scanner needs to sample dirty state to coordinate if early
> writeback is necessary:
>
> try_to_shrink_lruvec():
>         /*
>          * If too many file cache in the coldest generation can't be evicted
>          * due to being dirty, wake up the flusher.
>          */
>         if (sc->nr.unqueued_dirty && sc->nr.unqueued_dirty == sc->nr.file_taken)
>                 wakeup_flusher_threads(WB_REASON_VMSCAN);
Thanks for heads up. Just find out that dirty file pages have been
promoted to younger gen under MGLRU, will update the code in patchv2