Skip to contentSkip to conversation

mm: reinstate unconditional writeback start in balance_dirty_pages()

PATCH Applied Joanne Koong 5 participants 6 messages Mar 26 – Mar 27

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v2 cover mm: reinstate unconditional writeback start in balance_dirty_pages()

Changelog
--------
v1: https://lore.kernel.org/linux-fsdevel/[email protected]/
v1 -> v2:
* fix it by bringing back the removed logic instead of the strictlimit wb check (Jan and Christoph)

Joanne Koong (1):
mm: reinstate unconditional writeback start in balance_dirty_pages()

 mm/page-writeback.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
Signature · 1 line
2.52.0
#1 lore raw

PATCH v2 1/1 mm: reinstate unconditional writeback start in balance_dirty_pages()

Commit 64dd89ae01f2 ("mm/block/fs: remove laptop_mode") removed this unconditional writeback start from balance_dirty_pages():

       if (unlikely(!writeback_in_progress(wb)))
	       wb_start_background_writeback(wb);

This logic needs to be reinstated to prevent performance regressions for strictlimited BDIs and memcg setups. The problem occurs because:

a) For strictlimited BDIs, throttling is calculated using per-wb thresholds. The per-wb threshold can be exceeded even when the global dirty threshold was not exceeded (nr_dirty < gdtc->bg_thresh)

b) For memcg-based throttling, memcg uses its own dirty count / thresholds and can trigger throttling even when the global threshold isn't exceeded

Without the unconditional writeback start, IO is throttled as it waits for dirty pages to be written back but there is no writeback running. This leads to severe stalls. On fuse, buffered write performance dropped from 1400 MiB/s to 2000 KiB/s.

Reinstate the unconditional writeback start so that writeback is guaranteed to be running whenever IO needs to be throttled.

Fixes 64dd89ae01f2 ("mm/block/fs: remove laptop_mode")
---
 mm/page-writeback.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
mm/page-writeback.c +21 -0
index 601a5e048d12..c1a4b32af1a7 100644
@@ -1858,6 +1858,27 @@ static int balance_dirty_pages(struct bdi_writeback *wb,
 			break;
 		}
 
+		/*
+		 * Unconditionally start background writeback if it's not
+		 * already in progress. We need to do this because the global
+		 * dirty threshold check above (nr_dirty > gdtc->bg_thresh)
+		 * doesn't account for these cases:
+		 *
+		 * a) strictlimit BDIs: throttling is calculated using per-wb
+		 * thresholds. The per-wb threshold can be exceeded even when
+		 * nr_dirty < gdtc->bg_thresh
+		 *
+		 * b) memcg-based throttling: memcg uses its own dirty count and
+		 * thresholds and can trigger throttling even when global
+		 * nr_dirty < gdtc->bg_thresh
+		 *
+		 * Writeback needs to be started else the writer stalls in the
+		 * throttle loop waiting for dirty pages to be written back
+		 * while no writeback is running.
+		 */
+		if (unlikely(!writeback_in_progress(wb)))
+			wb_start_background_writeback(wb);
+
 		mem_cgroup_flush_foreign(wb);
 
 		/*
Signature · 1 line
2.52.0
maintainer
#2 lore raw
Quoted 17 lines · On Thu, 26 Mar 2026 14:51:27 -0700 Joanne Koong <[email protected]> wrote:

Commit 64dd89ae01f2 ("mm/block/fs: remove laptop_mode") removed this unconditional writeback start from balance_dirty_pages():

       if (unlikely(!writeback_in_progress(wb)))
	       wb_start_background_writeback(wb);

This logic needs to be reinstated to prevent performance regressions for strictlimited BDIs and memcg setups. The problem occurs because:

a) For strictlimited BDIs, throttling is calculated using per-wb thresholds. The per-wb threshold can be exceeded even when the global dirty threshold was not exceeded (nr_dirty < gdtc->bg_thresh)

b) For memcg-based throttling, memcg uses its own dirty count / thresholds and can trigger throttling even when the global threshold isn't exceeded

Quoted 7 lines

Without the unconditional writeback start, IO is throttled as it waits for dirty pages to be written back but there is no writeback running. This leads to severe stalls. On fuse, buffered write performance dropped from 1400 MiB/s to 2000 KiB/s.

Reinstate the unconditional writeback start so that writeback is guaranteed to be running whenever IO needs to be throttled.

Thanks, I queued this in mm.git's mm-hotfixes-unstable branch for test and review. This starts it on the track to mainline before the 7.0 release.

I assume we want to be prompt with this because the fuse regression sounds pretty horrid.

maintainer Reviewed-by on v2 1/1
#4 lore raw
In reply to Joanne Koong
Quoted 22 lines · Joanne Koong

Commit 64dd89ae01f2 ("mm/block/fs: remove laptop_mode") removed this unconditional writeback start from balance_dirty_pages():

       if (unlikely(!writeback_in_progress(wb)))
	       wb_start_background_writeback(wb);

This logic needs to be reinstated to prevent performance regressions for strictlimited BDIs and memcg setups. The problem occurs because:

a) For strictlimited BDIs, throttling is calculated using per-wb thresholds. The per-wb threshold can be exceeded even when the global dirty threshold was not exceeded (nr_dirty < gdtc->bg_thresh)

b) For memcg-based throttling, memcg uses its own dirty count / thresholds and can trigger throttling even when the global threshold isn't exceeded

Without the unconditional writeback start, IO is throttled as it waits for dirty pages to be written back but there is no writeback running. This leads to severe stalls. On fuse, buffered write performance dropped from 1400 MiB/s to 2000 KiB/s.

Quoted 5 lines

Reinstate the unconditional writeback start so that writeback is guaranteed to be running whenever IO needs to be throttled.

Fixes 64dd89ae01f2 ("mm/block/fs: remove laptop_mode")

Looks good. Feel free to add:

Honza

Quoted 39 lines
---
 mm/page-writeback.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
mm/page-writeback.c +21 -0
index 601a5e048d12..c1a4b32af1a7 100644
@@ -1858,6 +1858,27 @@ static int balance_dirty_pages(struct bdi_writeback *wb,
 			break;
 		}
 
+		/*
+		 * Unconditionally start background writeback if it's not
+		 * already in progress. We need to do this because the global
+		 * dirty threshold check above (nr_dirty > gdtc->bg_thresh)
+		 * doesn't account for these cases:
+		 *
+		 * a) strictlimit BDIs: throttling is calculated using per-wb
+		 * thresholds. The per-wb threshold can be exceeded even when
+		 * nr_dirty < gdtc->bg_thresh
+		 *
+		 * b) memcg-based throttling: memcg uses its own dirty count and
+		 * thresholds and can trigger throttling even when global
+		 * nr_dirty < gdtc->bg_thresh
+		 *
+		 * Writeback needs to be started else the writer stalls in the
+		 * throttle loop waiting for dirty pages to be written back
+		 * while no writeback is running.
+		 */
+		if (unlikely(!writeback_in_progress(wb)))
+			wb_start_background_writeback(wb);
+
 		mem_cgroup_flush_foreign(wb);
 
 		/*
Signature · 1 line
2.52.0
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
Acked-by on v2 1/1
#5 lore raw
In reply to Joanne Koong
Quoted 22 lines · Joanne Koong

Commit 64dd89ae01f2 ("mm/block/fs: remove laptop_mode") removed this unconditional writeback start from balance_dirty_pages():

       if (unlikely(!writeback_in_progress(wb)))
	       wb_start_background_writeback(wb);

This logic needs to be reinstated to prevent performance regressions for strictlimited BDIs and memcg setups. The problem occurs because:

a) For strictlimited BDIs, throttling is calculated using per-wb thresholds. The per-wb threshold can be exceeded even when the global dirty threshold was not exceeded (nr_dirty < gdtc->bg_thresh)

b) For memcg-based throttling, memcg uses its own dirty count / thresholds and can trigger throttling even when the global threshold isn't exceeded

Without the unconditional writeback start, IO is throttled as it waits for dirty pages to be written back but there is no writeback running. This leads to severe stalls. On fuse, buffered write performance dropped from 1400 MiB/s to 2000 KiB/s.

Quoted 5 lines

Reinstate the unconditional writeback start so that writeback is guaranteed to be running whenever IO needs to be throttled.

Fixes 64dd89ae01f2 ("mm/block/fs: remove laptop_mode")

Thanks for the fix. This was a combination of the old comment and my overly hasty read of the existing dirty > bg_thresh && !writeback_in_progress() block. The patch looks good to me.