OSDN Git Service

mm/memory-failure.c: recheck PageHuge() after hugetlb page migrate successfully
authorJianguo Wu <wujianguo@huawei.com>
Thu, 19 Dec 2013 01:08:54 +0000 (17:08 -0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 25 Jan 2014 16:27:11 +0000 (08:27 -0800)
commit a49ecbcd7b0d5a1cda7d60e03df402dd0ef76ac8 upstream.

After a successful hugetlb page migration by soft offline, the source
page will either be freed into hugepage_freelists or buddy(over-commit
page).  If page is in buddy, page_hstate(page) will be NULL.  It will
hit a NULL pointer dereference in dequeue_hwpoisoned_huge_page().

  BUG: unable to handle kernel NULL pointer dereference at 0000000000000058
  IP: [<ffffffff81163761>] dequeue_hwpoisoned_huge_page+0x131/0x1d0
  PGD c23762067 PUD c24be2067 PMD 0
  Oops: 0000 [#1] SMP

So check PageHuge(page) after call migrate_pages() successfully.

[wujg: backport to 3.10:
 - adjust context]

Signed-off-by: Jianguo Wu <wujianguo@huawei.com>
Tested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mm/memory-failure.c

index 6a7f9ca..7e3601c 100644 (file)
@@ -1499,10 +1499,16 @@ static int soft_offline_huge_page(struct page *page, int flags)
                pr_info("soft offline: %#lx: migration failed %d, type %lx\n",
                        pfn, ret, page->flags);
        } else {
-               set_page_hwpoison_huge_page(hpage);
-               dequeue_hwpoisoned_huge_page(hpage);
-               atomic_long_add(1 << compound_trans_order(hpage),
-                               &num_poisoned_pages);
+               /* overcommit hugetlb page will be freed to buddy */
+               if (PageHuge(page)) {
+                       set_page_hwpoison_huge_page(hpage);
+                       dequeue_hwpoisoned_huge_page(hpage);
+                       atomic_long_add(1 << compound_order(hpage),
+                                       &num_poisoned_pages);
+               } else {
+                       SetPageHWPoison(page);
+                       atomic_long_inc(&num_poisoned_pages);
+               }
        }
        /* keep elevated page count for bad page */
        return ret;