When a process which used a hwpoisoned hugepage tries to exit() or
munmap(), the kernel can print out "bad pmd" message because page
table walker in free_pgtables() encounters 'hwpoisoned entry' on pmd.
This is because currently we fail to clear the hwpoisoned entry in
__unmap_hugepage_range(), so this patch simply does it.
Signed-off-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
---
mm/hugetlb.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git v3.7-rc8.orig/mm/hugetlb.c v3.7-rc8/mm/hugetlb.c
index e61a749..fe7c2a7 100644
--- v3.7-rc8.orig/mm/hugetlb.c
+++ v3.7-rc8/mm/hugetlb.c
@@ -2387,8 +2387,10 @@ void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct *vma,
/*
* HWPoisoned hugepage is already unmapped and dropped reference
*/
- if (unlikely(is_hugetlb_entry_hwpoisoned(pte)))
+ if (unlikely(is_hugetlb_entry_hwpoisoned(pte))) {
+ pte_clear(mm, address, ptep);
continue;
+ }
page = pte_page(pte);
/*
--
1.7.11.7
--
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>