tools, slub: Fix off-by-one buffer corruption after readlink() call
authorThomas Jarosch <thomas.jarosch@intra2net.com>
Mon, 17 Oct 2011 14:48:10 +0000 (16:48 +0200)
committerPekka Enberg <penberg@kernel.org>
Tue, 18 Oct 2011 16:57:59 +0000 (19:57 +0300)
readlink() never zero terminates the provided buffer.
Therefore we already do

    buffer[count] = 0;

This leads to an off-by-one buffer corruption as readlink()
might return the full size of the buffer.

The common technique is to reduce the buffer size by one.
Another fix would be to check

  if (count < 0 || count == sizeof(buffer))
      fatal();

Reducing the buffer size by one is easier IMHO.

Signed-off-by: Thomas Jarosch <thomas.jarosch@intra2net.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Christoph Lameter <cl@gentwo.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
tools/slub/slabinfo.c

index 868cc93..cc1a378 100644 (file)
@@ -1145,7 +1145,7 @@ static void read_slab_dir(void)
                switch (de->d_type) {
                   case DT_LNK:
                        alias->name = strdup(de->d_name);
-                       count = readlink(de->d_name, buffer, sizeof(buffer));
+                       count = readlink(de->d_name, buffer, sizeof(buffer)-1);
 
                        if (count < 0)
                                fatal("Cannot read symlink %s\n", de->d_name);