Skip to contentSkip to conversation

mm: restore SHRINKER_NONSLAB on THP and zswap shrinkers

PATCH In review Qinyun Tan 6 participants 9 messages Sep 4 – Sep 9

Subsystems (3)
lore
Share & tools
lore

Qinyun Tan

author
#0 lore raw

cover mm: restore SHRINKER_NONSLAB on THP and zswap shrinkers

While reviewing the patch "mm/list_lru: don't copy stale shrinker id from non-memcg-aware shrinkers" (now in mm-unstable), Michal noticed that the two non-slab memcg-aware shrinkers, the THP deferred split shrinker and the zswap shrinker, are registered without SHRINKER_NONSLAB [1].

Without that flag, booting with cgroup.memory=nokmem silently demotes both shrinkers to non-memcg-aware ones: memcg (limit-induced) reclaim neither splits partially unmapped THPs nor writes back zswapped pages of the cgroup under pressure.

Patch 1 restores the flag on the deferred split shrinker; it was lost when commit fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru") re-registered the shrinker.

Patch 2 adds the flag to the zswap shrinker, which has never carried it since its introduction in commit b5ba474f3f51 ("zswap: shrink zswap pool based on memory pressure").

The series is based on mm-new, on top of the list_lru patch above, but the changes are independent of it.

[1] https://lore.kernel.org/lkml/[email protected]/

Qinyun Tan (2):
mm: thp: restore SHRINKER_NONSLAB on the deferred split shrinker
mm: zswap: mark the zswap shrinker SHRINKER_NONSLAB

 mm/huge_memory.c | 3 ++-
 mm/zswap.c       | 4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)
Signature · 1 line
2.43.7
author
#1 lore raw

PATCH 1/2 mm: thp: restore SHRINKER_NONSLAB on the deferred split shrinker

On a system booted with cgroup.memory=nokmem, the deferred split shrinker is quietly demoted to a non-memcg-aware one. As a result, partially unmapped THPs are only split under global reclaim; memcg (limit-induced) reclaim never splits them, so a cgroup under memory pressure keeps its underused THPs intact.

This is a regression from commit fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru"), which re-registered the shrinker without SHRINKER_NONSLAB. The shrinker had carried this flag since commit 0a432dcbeb32 ("mm: shrinker: make shrinker not depend on memcg kmem") precisely so it would keep working with kmem accounting disabled. Without the flag, shrinker_memcg_alloc() fails with -ENOSYS under nokmem and the shrinker loses its memcg awareness.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Restore the flag.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru")
---
 mm/huge_memory.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
mm/huge_memory.c +2 -1
index 644d6905b49cc..16f9aa5b2d077 100644
@@ -1022,7 +1022,8 @@ int folio_memcg_alloc_deferred(struct folio *folio)
 static int __init thp_shrinker_init(void)
 {
 	deferred_split_shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE |
-						 SHRINKER_MEMCG_AWARE,
+						 SHRINKER_MEMCG_AWARE |
+						 SHRINKER_NONSLAB,
 						 "thp-deferred_split");
 	if (!deferred_split_shrinker)
 		return -ENOMEM;
Signature · 1 line
2.43.7
reviewer Acked-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] mm: thp: restore SHRINKER_NONSLAB on the deferred split shrinker

Quoted 19 lines · Qinyun Tan

On a system booted with cgroup.memory=nokmem, the deferred split shrinker is quietly demoted to a non-memcg-aware one. As a result, partially unmapped THPs are only split under global reclaim; memcg (limit-induced) reclaim never splits them, so a cgroup under memory pressure keeps its underused THPs intact.

This is a regression from commit fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru"), which re-registered the shrinker without SHRINKER_NONSLAB. The shrinker had carried this flag since commit 0a432dcbeb32 ("mm: shrinker: make shrinker not depend on memcg kmem") precisely so it would keep working with kmem accounting disabled. Without the flag, shrinker_memcg_alloc() fails with -ENOSYS under nokmem and the shrinker loses its memcg awareness.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Restore the flag.

Quoted 9 lines

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru")
---
 mm/huge_memory.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Makes sense.

Signature · 2 lines
Best Regards,
Yan, Zi
reviewer Reviewed-by on v1 1/2
#3 lore raw

Re: [PATCH 1/2] mm: thp: restore SHRINKER_NONSLAB on the deferred split shrinker

In reply to Qinyun Tan
Quoted 17 lines · Qinyun Tan

On a system booted with cgroup.memory=nokmem, the deferred split shrinker is quietly demoted to a non-memcg-aware one. As a result, partially unmapped THPs are only split under global reclaim; memcg (limit-induced) reclaim never splits them, so a cgroup under memory pressure keeps its underused THPs intact.

This is a regression from commit fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru"), which re-registered the shrinker without SHRINKER_NONSLAB. The shrinker had carried this flag since commit 0a432dcbeb32 ("mm: shrinker: make shrinker not depend on memcg kmem") precisely so it would keep working with kmem accounting disabled. Without the flag, shrinker_memcg_alloc() fails with -ENOSYS under nokmem and the shrinker loses its memcg awareness.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Quoted 8 lines

Restore the flag.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru")

---

LGTM.

maintainer Acked-by on v1 1/2
#4 lore raw

Re: [PATCH 1/2] mm: thp: restore SHRINKER_NONSLAB on the deferred split shrinker

In reply to Qinyun Tan
Quoted 17 lines · Qinyun Tan

On a system booted with cgroup.memory=nokmem, the deferred split shrinker is quietly demoted to a non-memcg-aware one. As a result, partially unmapped THPs are only split under global reclaim; memcg (limit-induced) reclaim never splits them, so a cgroup under memory pressure keeps its underused THPs intact.

This is a regression from commit fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru"), which re-registered the shrinker without SHRINKER_NONSLAB. The shrinker had carried this flag since commit 0a432dcbeb32 ("mm: shrinker: make shrinker not depend on memcg kmem") precisely so it would keep working with kmem accounting disabled. Without the flag, shrinker_memcg_alloc() fails with -ENOSYS under nokmem and the shrinker loses its memcg awareness.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Quoted 8 lines

Restore the flag.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru")

---

Signature · 3 lines
Cheers,

David
reviewer Acked-by on v1 1/2
#5 lore raw

Re: [PATCH 1/2] mm: thp: restore SHRINKER_NONSLAB on the deferred split shrinker

In reply to Qinyun Tan
Quoted 33 lines · Qinyun Tan

On a system booted with cgroup.memory=nokmem, the deferred split shrinker is quietly demoted to a non-memcg-aware one. As a result, partially unmapped THPs are only split under global reclaim; memcg (limit-induced) reclaim never splits them, so a cgroup under memory pressure keeps its underused THPs intact.

This is a regression from commit fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru"), which re-registered the shrinker without SHRINKER_NONSLAB. The shrinker had carried this flag since commit 0a432dcbeb32 ("mm: shrinker: make shrinker not depend on memcg kmem") precisely so it would keep working with kmem accounting disabled. Without the flag, shrinker_memcg_alloc() fails with -ENOSYS under nokmem and the shrinker loses its memcg awareness.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Restore the flag.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes fafaeceb89a5 ("mm: switch deferred split shrinker to list_lru")
---
 mm/huge_memory.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
mm/huge_memory.c +0 -0
index 644d6905b49cc..16f9aa5b2d077 100644

mm/huge_memory.c int folio_memcg_alloc_deferred(struct folio *folio) · line 1029

Quoted 9 lines
mm/huge_memory.c +2 -1
@@ -1022,7 +1022,8 @@ int folio_memcg_alloc_deferred(struct folio *folio)
 static int __init thp_shrinker_init(void)
 {
 	deferred_split_shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE |
-						 SHRINKER_MEMCG_AWARE,
+						 SHRINKER_MEMCG_AWARE |
+						 SHRINKER_NONSLAB,
 						 "thp-deferred_split");
 	if (!deferred_split_shrinker)
 		return -ENOMEM;
author
#6 lore raw

PATCH 2/2 mm: zswap: mark the zswap shrinker SHRINKER_NONSLAB

In reply to Qinyun Tan

With kmem accounting disabled (cgroup.memory=nokmem), memcg reclaim never invokes the zswap shrinker: a cgroup under memory pressure does not write back its own zswapped pages to make room; they are only written back by global reclaim or once the global pool limit kicks in.

The zswap shrinker is registered memcg-aware but without SHRINKER_NONSLAB, so under nokmem it is treated as a slab shrinker:
before commit 03375203e1da ("mm: do not allocate shrinker info with cgroup.memory=nokmem") it was skipped by the !memcg_kmem_online() check in shrink_slab_memcg(); since that commit it is demoted to non-memcg-aware at registration. But the zswap shrinker is not a slab
shrinker: it tracks zswap entries and resolves the owning memcg from the folio's objcg, independent of kmem accounting.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Mark it SHRINKER_NONSLAB so it keeps its memcg awareness and runs under memcg reclaim with nokmem.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes b5ba474f3f51 ("zswap: shrink zswap pool based on memory pressure")
---
 mm/zswap.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
mm/zswap.c +2 -2
index 761cd699e0a3e..8bcc2997d7719 100644
@@ -1263,8 +1263,8 @@ static struct shrinker *zswap_alloc_shrinker(void)
 {
 	struct shrinker *shrinker;
 
-	shrinker =
-		shrinker_alloc(SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE, "mm-zswap");
+	shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE |
+				  SHRINKER_NONSLAB, "mm-zswap");
 	if (!shrinker)
 		return NULL;
 
Signature · 1 line
2.43.7
reviewer Acked-by on v1 2/2
#7 lore raw

Re: [PATCH 2/2] mm: zswap: mark the zswap shrinker SHRINKER_NONSLAB

Quoted 34 lines · Qinyun Tan

With kmem accounting disabled (cgroup.memory=nokmem), memcg reclaim never invokes the zswap shrinker: a cgroup under memory pressure does not write back its own zswapped pages to make room; they are only written back by global reclaim or once the global pool limit kicks in.

The zswap shrinker is registered memcg-aware but without SHRINKER_NONSLAB, so under nokmem it is treated as a slab shrinker:
before commit 03375203e1da ("mm: do not allocate shrinker info with cgroup.memory=nokmem") it was skipped by the !memcg_kmem_online() check in shrink_slab_memcg(); since that commit it is demoted to non-memcg-aware at registration. But the zswap shrinker is not a slab
shrinker: it tracks zswap entries and resolves the owning memcg from the folio's objcg, independent of kmem accounting.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Mark it SHRINKER_NONSLAB so it keeps its memcg awareness and runs under memcg reclaim with nokmem.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes b5ba474f3f51 ("zswap: shrink zswap pool based on memory pressure")
---
 mm/zswap.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
mm/zswap.c +0 -0
index 761cd699e0a3e..8bcc2997d7719 100644

mm/zswap.c static struct shrinker *zswap_alloc_shrinker(void) · line 1270

Quoted 10 lines
mm/zswap.c +2 -2
@@ -1263,8 +1263,8 @@ static struct shrinker *zswap_alloc_shrinker(void)
 {
 	struct shrinker *shrinker;
 
-	shrinker =
-		shrinker_alloc(SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE, "mm-zswap");
+	shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE |
+				  SHRINKER_NONSLAB, "mm-zswap");
 	if (!shrinker)
 		return NULL;
 
maintainer Acked-by on v1 2/2
#8 lore raw

Re: [PATCH 2/2] mm: zswap: mark the zswap shrinker SHRINKER_NONSLAB

In reply to Qinyun Tan
Quoted 18 lines · Qinyun Tan

With kmem accounting disabled (cgroup.memory=nokmem), memcg reclaim never invokes the zswap shrinker: a cgroup under memory pressure does not write back its own zswapped pages to make room; they are only written back by global reclaim or once the global pool limit kicks in.

The zswap shrinker is registered memcg-aware but without SHRINKER_NONSLAB, so under nokmem it is treated as a slab shrinker:
before commit 03375203e1da ("mm: do not allocate shrinker info with cgroup.memory=nokmem") it was skipped by the !memcg_kmem_online() check in shrink_slab_memcg(); since that commit it is demoted to non-memcg-aware at registration. But the zswap shrinker is not a slab
shrinker: it tracks zswap entries and resolves the owning memcg from the folio's objcg, independent of kmem accounting.

This was noticed by Michal during review of the patch "mm/list_lru:
don't copy stale shrinker id from non-memcg-aware shrinkers" [1].

Quoted 8 lines

Mark it SHRINKER_NONSLAB so it keeps its memcg awareness and runs under memcg reclaim with nokmem.

[1] https://lore.kernel.org/lkml/[email protected]/

Fixes b5ba474f3f51 ("zswap: shrink zswap pool based on memory pressure")