mm/slub.c: fix random_seq offset destruction
authorSean Rees <sean@erifax.org>
Wed, 8 Feb 2017 22:30:59 +0000 (14:30 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 8 Feb 2017 23:41:43 +0000 (15:41 -0800)
Commit 210e7a43fa90 ("mm: SLUB freelist randomization") broke USB hub
initialisation as described in

  https://bugzilla.kernel.org/show_bug.cgi?id=177551.

Bail out early from init_cache_random_seq if s->random_seq is already
initialised.  This prevents destroying the previously computed
random_seq offsets later in the function.

If the offsets are destroyed, then shuffle_freelist will truncate
page->freelist to just the first object (orphaning the rest).

Fixes: 210e7a43fa90 ("mm: SLUB freelist randomization")
Link: http://lkml.kernel.org/r/20170207140707.20824-1-sean@erifax.org
Signed-off-by: Sean Rees <sean@erifax.org>
Reported-by: <userwithuid@gmail.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Thomas Garnier <thgarnie@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/slub.c

index 7aa6f433f4de554d308e774d9e9b40507c6ab48a..7ec0a965c6a3a366550b2500f9880f65e5025138 100644 (file)
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -1422,6 +1422,10 @@ static int init_cache_random_seq(struct kmem_cache *s)
        int err;
        unsigned long i, count = oo_objects(s->oo);
 
+       /* Bailout if already initialised */
+       if (s->random_seq)
+               return 0;
+
        err = cache_random_seq_create(s, count, GFP_KERNEL);
        if (err) {
                pr_err("SLUB: Unable to initialize free list for %s\n",