]> git.ipfire.org Git - thirdparty/kernel/stable.git/commitdiff
mm/migrate: putback split folios when numa hint migration fails
authorPeter Xu <peterx@redhat.com>
Mon, 8 Jul 2024 21:55:37 +0000 (17:55 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 11 Aug 2024 10:57:47 +0000 (12:57 +0200)
[ Upstream commit 6e49019db5f7a09a9c0e8ac4d108e656c3f8e583 ]

This issue is not from any report yet, but by code observation only.

This is yet another fix besides Hugh's patch [1] but on relevant code
path, where eager split of folio can happen if the folio is already on
deferred list during a folio migration.

Here the issue is NUMA path (migrate_misplaced_folio()) may start to
encounter such folio split now even with MR_NUMA_MISPLACED hint applied.
Then when migrate_pages() didn't migrate all the folios, it's possible the
split small folios be put onto the list instead of the original folio.
Then putting back only the head page won't be enough.

Fix it by putting back all the folios on the list.

[1] https://lore.kernel.org/all/46c948b4-4dd8-6e03-4c7b-ce4e81cfa536@google.com/

[akpm@linux-foundation.org: remove now unused local `nr_pages']
Link: https://lkml.kernel.org/r/20240708215537.2630610-1-peterx@redhat.com
Fixes: 7262f208ca68 ("mm/migrate: split source folio if it is on deferred split list")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Yang Shi <shy828301@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Huang Ying <ying.huang@intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
mm/migrate.c

index 6b5affe49cf9198bfbb19f83ed7335b54fe34d95..9dabeb90f772d861d3cabf9a1a58215833a2641c 100644 (file)
@@ -2634,20 +2634,13 @@ int migrate_misplaced_folio(struct folio *folio, struct vm_area_struct *vma,
        int nr_remaining;
        unsigned int nr_succeeded;
        LIST_HEAD(migratepages);
-       int nr_pages = folio_nr_pages(folio);
 
        list_add(&folio->lru, &migratepages);
        nr_remaining = migrate_pages(&migratepages, alloc_misplaced_dst_folio,
                                     NULL, node, MIGRATE_ASYNC,
                                     MR_NUMA_MISPLACED, &nr_succeeded);
-       if (nr_remaining) {
-               if (!list_empty(&migratepages)) {
-                       list_del(&folio->lru);
-                       node_stat_mod_folio(folio, NR_ISOLATED_ANON +
-                                       folio_is_file_lru(folio), -nr_pages);
-                       folio_putback_lru(folio);
-               }
-       }
+       if (nr_remaining && !list_empty(&migratepages))
+               putback_movable_pages(&migratepages);
        if (nr_succeeded) {
                count_vm_numa_events(NUMA_PAGE_MIGRATE, nr_succeeded);
                if (!node_is_toptier(folio_nid(folio)) && node_is_toptier(node))