Skip to contentSkip to conversation

workingset: fix confusion around eviction vs refault container

PATCH Applied Nhat Pham 1 participant 2 messages Jan 4, 2023

Subsystems (3)
lore
Share & tools
lore

Nhat Pham

author
#0 lore raw

PATCH 1/1 workingset: fix confusion around eviction vs refault container

From: Johannes Weiner <[email protected]>

Refault decisions are made based on the lruvec where the page was evicted, as that determined its LRU order while it was alive. Stats and workingset aging must then occur on the lruvec of the new page, as that's the node and cgroup that experience the refault and that's the lruvec whose nonresident info ages out by a new resident page. Those lruvecs could be different when a page is shared between cgroups, or the refaulting page is allocated on a different node.

There are currently two mix-ups:

1. When swap is available, the resident anon set must be considered when comparing the refault distance. The comparison is made against the right anon set, but the check for swap is not. When pages get evicted from a cgroup with swap, and refault in one without, this can incorrectly consider a hot refault as cold - and vice versa. Fix that by using the eviction cgroup for the swap check.

2. The stats and workingset age are updated against the wrong lruvec
altogether: the right cgroup but the wrong NUMA node. When a page refaults on a different NUMA node, this will have confusing stats and distort the workingset age on a different lruvec - again possibly resulting in hot/cold misclassifications down the line.

Fix the swap check and the refault pgdat to address both concerns.

This was found during code review. It hasn't caused notable issues in production, suggesting that those refault-migrations are relatively rare in practice.

---
 mm/workingset.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
mm/workingset.c +2 -1
index ae7e984b23c6..79585d55c45d 100644
@@ -457,6 +457,7 @@ void workingset_refault(struct folio *folio, void *shadow)
 	 */
 	nr = folio_nr_pages(folio);
 	memcg = folio_memcg(folio);
+	pgdat = folio_pgdat(folio);
 	lruvec = mem_cgroup_lruvec(memcg, pgdat);
 
 	mod_lruvec_state(lruvec, WORKINGSET_REFAULT_BASE + file, nr);
@@ -474,7 +475,7 @@ void workingset_refault(struct folio *folio, void *shadow)
 		workingset_size += lruvec_page_state(eviction_lruvec,
 						     NR_INACTIVE_FILE);
 	}
-	if (mem_cgroup_get_nr_swap_pages(memcg) > 0) {
+	if (mem_cgroup_get_nr_swap_pages(eviction_memcg) > 0) {
 		workingset_size += lruvec_page_state(eviction_lruvec,
 						     NR_ACTIVE_ANON);
 		if (file) {
Signature · 1 line
2.30.2
author
#1 lore raw
Quoted 61 lines · Nhat Pham

From: Johannes Weiner <[email protected]>

Refault decisions are made based on the lruvec where the page was evicted, as that determined its LRU order while it was alive. Stats and workingset aging must then occur on the lruvec of the new page, as that's the node and cgroup that experience the refault and that's the lruvec whose nonresident info ages out by a new resident page. Those lruvecs could be different when a page is shared between cgroups, or the refaulting page is allocated on a different node.

There are currently two mix-ups:

1. When swap is available, the resident anon set must be considered when comparing the refault distance. The comparison is made against the right anon set, but the check for swap is not. When pages get evicted from a cgroup with swap, and refault in one without, this can incorrectly consider a hot refault as cold - and vice versa. Fix that by using the eviction cgroup for the swap check.

2. The stats and workingset age are updated against the wrong lruvec
altogether: the right cgroup but the wrong NUMA node. When a page refaults on a different NUMA node, this will have confusing stats and distort the workingset age on a different lruvec - again possibly resulting in hot/cold misclassifications down the line.

Fix the swap check and the refault pgdat to address both concerns.

This was found during code review. It hasn't caused notable issues in production, suggesting that those refault-migrations are relatively rare in practice.

---
 mm/workingset.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
mm/workingset.c +2 -2
index ae7e984b23c6..79585d55c45d 100644
@@ -457,6 +457,7 @@ void workingset_refault(struct folio *folio, void *shadow)
         */
        nr = folio_nr_pages(folio);
        memcg = folio_memcg(folio);
+       pgdat = folio_pgdat(folio);
        lruvec = mem_cgroup_lruvec(memcg, pgdat);
 
        mod_lruvec_state(lruvec, WORKINGSET_REFAULT_BASE + file, nr);
@@ -474,7 +475,7 @@ void workingset_refault(struct folio *folio, void *shadow)
                workingset_size += lruvec_page_state(eviction_lruvec,
                                                     NR_INACTIVE_FILE);
        }
-       if (mem_cgroup_get_nr_swap_pages(memcg) > 0) {
+       if (mem_cgroup_get_nr_swap_pages(eviction_memcg) > 0) {
                workingset_size += lruvec_page_state(eviction_lruvec,
                                                     NR_ACTIVE_ANON);
                if (file) {
--
Quoted 1 line

2.30.2

Oh this one is sent out twice too... Something is wrong with my pipeline... Anyway, please disregard the first email and only review this one!