]> www.infradead.org Git - users/jedix/linux-maple.git/commit
mm/compaction: fix bug in hugetlb handling pathway
authorVishal Moola (Oracle) <vishal.moola@gmail.com>
Tue, 1 Apr 2025 02:10:24 +0000 (19:10 -0700)
committerAndrew Morton <akpm@linux-foundation.org>
Sat, 12 Apr 2025 00:32:36 +0000 (17:32 -0700)
commita84edd52f0a0fa193f0f685769939cf84510755b
treee2fb77bd9a6dedd531416f81ab64fc84df12719f
parent770c8d55c42868239c748a3ebc57c9e37755f842
mm/compaction: fix bug in hugetlb handling pathway

The compaction code doesn't take references on pages until we're certain
we should attempt to handle it.

In the hugetlb case, isolate_or_dissolve_huge_page() may return -EBUSY
without taking a reference to the folio associated with our pfn.  If our
folio's refcount drops to 0, compound_nr() becomes unpredictable, making
low_pfn and nr_scanned unreliable.  The user-visible effect is minimal -
this should rarely happen (if ever).

Fix this by storing the folio statistics earlier on the stack (just like
the THP and Buddy cases).

Also revert commit 66fe1cf7f581 ("mm: compaction: use helper compound_nr
in isolate_migratepages_block") to make backporting easier.

Link: https://lkml.kernel.org/r/20250401021025.637333-1-vishal.moola@gmail.com
Fixes: 369fa227c219 ("mm: make alloc_contig_range handle free hugetlb pages")
Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Acked-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
mm/compaction.c