mm, hugetlb: do not use a page in page cache for cow optimization
authorJoonsoo Kim <iamjoonsoo.kim@lge.com>
Wed, 11 Sep 2013 21:21:04 +0000 (14:21 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 11 Sep 2013 22:57:27 +0000 (15:57 -0700)
Currently, we use a page with mapped count 1 in page cache for cow
optimization.  If we find this condition, we don't allocate a new page and
copy contents.  Instead, we map this page directly.  This may introduce a
problem that writting to private mapping overwrite hugetlb file directly.
You can find this situation with following code.

        size = 20 * MB;
        flag = MAP_SHARED;
        p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0);
        if (p == MAP_FAILED) {
                fprintf(stderr, "mmap() failed: %s\n", strerror(errno));
                return -1;
        }
        p[0] = 's';
        fprintf(stdout, "BEFORE STEAL PRIVATE WRITE: %c\n", p[0]);
        munmap(p, size);

        flag = MAP_PRIVATE;
        p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0);
        if (p == MAP_FAILED) {
                fprintf(stderr, "mmap() failed: %s\n", strerror(errno));
        }
        p[0] = 'c';
        munmap(p, size);

        flag = MAP_SHARED;
        p = mmap(NULL, size, PROT_READ|PROT_WRITE, flag, fd, 0);
        if (p == MAP_FAILED) {
                fprintf(stderr, "mmap() failed: %s\n", strerror(errno));
                return -1;
        }
        fprintf(stdout, "AFTER STEAL PRIVATE WRITE: %c\n", p[0]);
        munmap(p, size);

We can see that "AFTER STEAL PRIVATE WRITE: c", not "AFTER STEAL PRIVATE
WRITE: s".  If we turn off this optimization to a page in page cache, the
problem is disappeared.

So, I change the trigger condition of optimization.  If this page is not
AnonPage, we don't do optimization.  This makes this optimization turning
off for a page cache.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Acked-by: Hillf Danton <dhillf@gmail.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Davidlohr Bueso <davidlohr.bueso@hp.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/hugetlb.c

index a13be48b818becf82f5f27df2add172f37c23301..da027a3307af000d757f759f50d3ee706f9be391 100644 (file)
@@ -2528,7 +2528,6 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma,
 {
        struct hstate *h = hstate_vma(vma);
        struct page *old_page, *new_page;
-       int avoidcopy;
        int outside_reserve = 0;
        unsigned long mmun_start;       /* For mmu_notifiers */
        unsigned long mmun_end;         /* For mmu_notifiers */
@@ -2538,10 +2537,8 @@ static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma,
 retry_avoidcopy:
        /* If no-one else is actually using this page, avoid the copy
         * and just make the page writable */
-       avoidcopy = (page_mapcount(old_page) == 1);
-       if (avoidcopy) {
-               if (PageAnon(old_page))
-                       page_move_anon_rmap(old_page, vma, address);
+       if (page_mapcount(old_page) == 1 && PageAnon(old_page)) {
+               page_move_anon_rmap(old_page, vma, address);
                set_huge_ptep_writable(vma, address, ptep);
                return 0;
        }