Commit efeda7a41e09efce506a68c3549b60b16dd7dedd
Committed by
Linus Torvalds
Linus Torvalds
1 parent
b16957c643
Exists in
master
and in
4 other branches
thp: fix splitting of hwpoisoned hugepages
The poisoned THP is now split with split_huge_page() in
collect_procs_anon(). If kmalloc() is failed in collect_procs(),
split_huge_page() could not be called. And the work after
split_huge_page() for collecting the processes using poisoned page will
not be done, too. So the processes using the poisoned page could not be
killed.
The condition becomes worse when CONFIG_DEBUG_VM == "Y". Because the
poisoned THP could not be split, system panic will be caused by
VM_BUG_ON(PageTransHuge(page)) in try_to_unmap().
This patch does:
1. move split_huge_page() to the place before collect_procs().
This can be sure the failure of splitting THP is caused by itself.
2. when splitting THP is failed, stop the operations after it.
This can avoid unexpected system panic or non sense works.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Jin Dongming <jin.dongming@np.css.fujitsu.com>
Reviewed-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Showing 1 changed file with 28 additions and 2 deletions Side-by-side Diff
mm/memory-failure.c
| ... | ... | @@ -386,8 +386,6 @@ |
| 386 | 386 | struct task_struct *tsk; |
| 387 | 387 | struct anon_vma *av; |
| 388 | 388 | |
| 389 | - if (!PageHuge(page) && unlikely(split_huge_page(page))) | |
| 390 | - return; | |
| 391 | 389 | read_lock(&tasklist_lock); |
| 392 | 390 | av = page_lock_anon_vma(page); |
| 393 | 391 | if (av == NULL) /* Not actually mapped anymore */ |
| ... | ... | @@ -893,6 +891,34 @@ |
| 893 | 891 | printk(KERN_INFO |
| 894 | 892 | "MCE %#lx: corrupted page was clean: dropped without side effects\n", |
| 895 | 893 | pfn); |
| 894 | + } | |
| 895 | + } | |
| 896 | + | |
| 897 | + if (PageTransHuge(hpage)) { | |
| 898 | + /* | |
| 899 | + * Verify that this isn't a hugetlbfs head page, the check for | |
| 900 | + * PageAnon is just for avoid tripping a split_huge_page | |
| 901 | + * internal debug check, as split_huge_page refuses to deal with | |
| 902 | + * anything that isn't an anon page. PageAnon can't go away fro | |
| 903 | + * under us because we hold a refcount on the hpage, without a | |
| 904 | + * refcount on the hpage. split_huge_page can't be safely called | |
| 905 | + * in the first place, having a refcount on the tail isn't | |
| 906 | + * enough * to be safe. | |
| 907 | + */ | |
| 908 | + if (!PageHuge(hpage) && PageAnon(hpage)) { | |
| 909 | + if (unlikely(split_huge_page(hpage))) { | |
| 910 | + /* | |
| 911 | + * FIXME: if splitting THP is failed, it is | |
| 912 | + * better to stop the following operation rather | |
| 913 | + * than causing panic by unmapping. System might | |
| 914 | + * survive if the page is freed later. | |
| 915 | + */ | |
| 916 | + printk(KERN_INFO | |
| 917 | + "MCE %#lx: failed to split THP\n", pfn); | |
| 918 | + | |
| 919 | + BUG_ON(!PageHWPoison(p)); | |
| 920 | + return SWAP_FAIL; | |
| 921 | + } | |
| 896 | 922 | } |
| 897 | 923 | } |
| 898 | 924 |