mm: fix BUG in __split_huge_page_pmd

commit 750e8165f5e87b6a142be953640eabb13a9d350a upstream.

Occasionally we hit the BUG_ON(pmd_trans_huge(*pmd)) at the end of
__split_huge_page_pmd(): seen when doing madvise(,,MADV_DONTNEED).

It's invalid: we don't always have down_write of mmap_sem there: a racing
do_huge_pmd_wp_page() might have copied-on-write to another huge page
before our split_huge_page() got the anon_vma lock.

Forget the BUG_ON, just go back and try again if this happens.

Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: David Rientjes <rientjes@google.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>

1 files changed, 9 insertions, 1 deletions

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index b92d0ce428b1..0164b09c1e99 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2699,6 +2699,7 @@ void __split_huge_page_pmd(struct vm_area_struct *vma, unsigned long address,
 
         mmun_start = haddr;
         mmun_end   = haddr + HPAGE_PMD_SIZE;
+again:
         mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end);
         spin_lock(&mm->page_table_lock);
         if (unlikely(!pmd_trans_huge(*pmd))) {
@@ -2721,7 +2722,14 @@ void __split_huge_page_pmd(struct vm_area_struct *vma, unsigned long address,
         split_huge_page(page);
 
         put_page(page);
-        BUG_ON(pmd_trans_huge(*pmd));
+
+        /*
+         * We don't always have down_write of mmap_sem here: a racing
+         * do_huge_pmd_wp_page() might have copied-on-write to another
+         * huge page before our split_huge_page() got the anon_vma lock.
+         */
+        if (unlikely(pmd_trans_huge(*pmd)))
+                goto again;
 }
 
 void split_huge_page_pmd_mm(struct mm_struct *mm, unsigned long address,