If __unmap_hugepage_range() tries to unmap the address range over which
hugepage migration is on the way, we get the wrong page because pte_page()
doesn't work for migration entries. This patch simply clears the pte for
migration entries as we do for hwpoison entries.
Signed-off-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx> # [2.6.36+]
---
ChangeLog v4:
- check hwpoisoned hugepage and migrating hugepage with the same check
instead of doing separately
---
mm/hugetlb.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git mmotm-2014-09-09-14-42.orig/mm/hugetlb.c mmotm-2014-09-09-14-42/mm/hugetlb.c
index 461ee1e10067..1ecb625bc498 100644
--- mmotm-2014-09-09-14-42.orig/mm/hugetlb.c
+++ mmotm-2014-09-09-14-42/mm/hugetlb.c
@@ -2653,9 +2653,10 @@ void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct *vma,
goto unlock;
/*
- * HWPoisoned hugepage is already unmapped and dropped reference
+ * Migrating hugepage or HWPoisoned hugepage is already
+ * unmapped and its refcount is dropped, so just clear pte here.
*/
- if (unlikely(is_hugetlb_entry_hwpoisoned(pte))) {
+ if (unlikely(!pte_present(pte))) {
huge_pte_clear(mm, address, ptep);
goto unlock;
}
--
1.9.3
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>