Skip to contentSkip to conversation

mm,hugetlb: take hugetlb_lock before decrementing h->resv_huge_pages

PATCH Applied Rik van Riel 2 participants 2 messages Oct 18, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm,hugetlb: take hugetlb_lock before decrementing h->resv_huge_pages

The h->*_huge_pages counters are protected by the hugetlb_lock, but alloc_huge_page has a corner case where it can decrement the counter outside of the lock.

This could lead to a corrupted value of h->resv_huge_pages, which we have observed on our systems.

Take the hugetlb_lock before decrementing h->resv_huge_pages to avoid a potential race.

Fixes a88c76954804 ("mm: hugetlb: fix hugepage memory leak caused by wrong reserve count")
Cc Naoya Horiguchi <[email protected]>
Cc Glen McCready <[email protected]>
Cc Mike Kravetz <[email protected]>
Cc Muchun Song <[email protected]>
Cc Andrew Morton <[email protected]>
---
 mm/hugetlb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
mm/hugetlb.c +1 -1
index b586cdd75930..dede0337c07c 100644
@@ -2924,11 +2924,11 @@ struct page *alloc_huge_page(struct vm_area_struct *vma,
 		page = alloc_buddy_huge_page_with_mpol(h, vma, addr);
 		if (!page)
 			goto out_uncharge_cgroup;
+		spin_lock_irq(&hugetlb_lock);
 		if (!avoid_reserve && vma_has_reserves(vma, gbl_chg)) {
 			SetHPageRestoreReserve(page);
 			h->resv_huge_pages--;
 		}
-		spin_lock_irq(&hugetlb_lock);
 		list_add(&page->lru, &h->hugepage_activelist);
 		set_page_refcounted(page);
 		/* Fall through */
Signature · 1 line
2.37.2
Reviewed-by on v1 1/1
#1 lore raw
Quoted 18 lines · Rik van Riel

The h->*_huge_pages counters are protected by the hugetlb_lock, but alloc_huge_page has a corner case where it can decrement the counter outside of the lock.

This could lead to a corrupted value of h->resv_huge_pages, which we have observed on our systems.

Take the hugetlb_lock before decrementing h->resv_huge_pages to avoid a potential race.

Fixes a88c76954804 ("mm: hugetlb: fix hugepage memory leak caused by wrong reserve count")
Cc Naoya Horiguchi <[email protected]>
Cc Glen McCready <[email protected]>
Cc Mike Kravetz <[email protected]>
Cc Muchun Song <[email protected]>
Cc Andrew Morton <[email protected]>
Quoted 3 lines
---
 mm/hugetlb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Thanks Rik! That case did slip through the cracks.

Signature · 1 line
Mike Kravetz
Quoted 22 lines
mm/hugetlb.c +1 -1
index b586cdd75930..dede0337c07c 100644
@@ -2924,11 +2924,11 @@ struct page *alloc_huge_page(struct vm_area_struct *vma,
 		page = alloc_buddy_huge_page_with_mpol(h, vma, addr);
 		if (!page)
 			goto out_uncharge_cgroup;
+		spin_lock_irq(&hugetlb_lock);
 		if (!avoid_reserve && vma_has_reserves(vma, gbl_chg)) {
 			SetHPageRestoreReserve(page);
 			h->resv_huge_pages--;
 		}
-		spin_lock_irq(&hugetlb_lock);
 		list_add(&page->lru, &h->hugepage_activelist);
 		set_page_refcounted(page);
 		/* Fall through */
Signature · 1 line
2.37.2