s390/mm: Check for valid vma before zapping in gmap_discard
authorJanosch Frank <frankja@linux.ibm.com>
Thu, 16 Aug 2018 08:02:31 +0000 (09:02 +0100)
committerJanosch Frank <frankja@linux.ibm.com>
Wed, 12 Sep 2018 12:46:37 +0000 (14:46 +0200)
Userspace could have munmapped the area before doing unmapping from
the gmap. This would leave us with a valid vmaddr, but an invalid vma
from which we would try to zap memory.

Let's check before using the vma.

Fixes: 1e133ab296f3 ("s390/mm: split arch/s390/mm/pgtable.c")
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Message-Id: <20180816082432.78828-1-frankja@linux.ibm.com>
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
arch/s390/mm/gmap.c

index bb44990..911c7de 100644 (file)
@@ -708,11 +708,13 @@ void gmap_discard(struct gmap *gmap, unsigned long from, unsigned long to)
                vmaddr |= gaddr & ~PMD_MASK;
                /* Find vma in the parent mm */
                vma = find_vma(gmap->mm, vmaddr);
+               if (!vma)
+                       continue;
                /*
                 * We do not discard pages that are backed by
                 * hugetlbfs, so we don't have to refault them.
                 */
-               if (vma && is_vm_hugetlb_page(vma))
+               if (is_vm_hugetlb_page(vma))
                        continue;
                size = min(to - gaddr, PMD_SIZE - (gaddr & ~PMD_MASK));
                zap_page_range(vma, vmaddr, size);