On Mon, Dec 12, 2011 at 04:55:52PM -0500, Xi Wang wrote: > Commit fa8b18ed didn't prevent the integer overflow and possible > memory corruption. "count" can go negative and bypass the check. > > Signed-off-by: Xi Wang <xi.wang at gmail.com> Looks good. Nice catch! Reviewed-by: Ben Myers <bpm at sgi.com>