mm: thp: avoid false positive VM_BUG_ON_PAGE in page_move_anon_rmap()
authorKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Fri, 27 May 2016 21:27:38 +0000 (14:27 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 27 May 2016 21:49:37 +0000 (14:49 -0700)
If page_move_anon_rmap() is refiling a pmd-splitted THP mapped in a tail
page from a pte, the "address" must be THP aligned in order for the
page->index bugcheck to pass in the CONFIG_DEBUG_VM=y builds.

Link: http://lkml.kernel.org/r/1464253620-106404-1-git-send-email-kirill.shutemov@linux.intel.com
Fixes: 6d0a07edd17c ("mm: thp: calculate the mapcount correctly for THP pages during WP faults")
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Tested-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: <stable@vger.kernel.org> [4.5]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/rmap.c

index 8a83993..0ea5d90 100644 (file)
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -1098,6 +1098,8 @@ void page_move_anon_rmap(struct page *page,
 
        VM_BUG_ON_PAGE(!PageLocked(page), page);
        VM_BUG_ON_VMA(!anon_vma, vma);
+       if (IS_ENABLED(CONFIG_DEBUG_VM) && PageTransHuge(page))
+               address &= HPAGE_PMD_MASK;
        VM_BUG_ON_PAGE(page->index != linear_page_index(vma, address), page);
 
        anon_vma = (void *) anon_vma + PAGE_MAPPING_ANON;