thp, dax: do not try to withdraw pgtable from non-anon VMA
authorKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Wed, 17 Feb 2016 21:11:12 +0000 (13:11 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 19 Feb 2016 00:23:24 +0000 (16:23 -0800)
DAX doesn't deposit pgtables when it maps huge pages: nothing to
withdraw. It can lead to crash.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Matthew Wilcox <willy@linux.intel.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/huge_memory.c

index 08fc0ba..722546d 100644 (file)
@@ -1700,7 +1700,8 @@ bool move_huge_pmd(struct vm_area_struct *vma, struct vm_area_struct *new_vma,
                pmd = pmdp_huge_get_and_clear(mm, old_addr, old_pmd);
                VM_BUG_ON(!pmd_none(*new_pmd));
 
-               if (pmd_move_must_withdraw(new_ptl, old_ptl)) {
+               if (pmd_move_must_withdraw(new_ptl, old_ptl) &&
+                               vma_is_anonymous(vma)) {
                        pgtable_t pgtable;
                        pgtable = pgtable_trans_huge_withdraw(mm, old_pmd);
                        pgtable_trans_huge_deposit(mm, new_pmd, pgtable);