From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Lameter Subject: Re: [Bugme-new] [Bug 33502] New: Caught 64-bit read from uninitialized memory in __alloc_skb Date: Tue, 19 Apr 2011 12:09:30 -0500 (CDT) Message-ID: References: <20110418153852.153d3ed3.akpm@linux-foundation.org> <1303181466.4152.39.camel@edumazet-laptop> <1303182557.4152.48.camel@edumazet-laptop> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: Andrew Morton , netdev@vger.kernel.org, bugzilla-daemon@bugzilla.kernel.org, bugme-daemon@bugzilla.kernel.org, casteyde.christian@free.fr, Vegard Nossum , Pekka Enberg To: Eric Dumazet Return-path: Received: from smtp102.prem.mail.ac4.yahoo.com ([76.13.13.41]:37398 "HELO smtp102.prem.mail.ac4.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1752223Ab1DSRJi (ORCPT ); Tue, 19 Apr 2011 13:09:38 -0400 In-Reply-To: <1303182557.4152.48.camel@edumazet-laptop> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 19 Apr 2011, Eric Dumazet wrote: > get_freepointer(s, object) can access to freed memory and kmemcheck > triggers the fault, while this_cpu_cmpxchg_double() would presumably > detect a change of tid and would not perform the freelist/tid change. Sounds right. The new lockless patchset for slub that uses a locked cmpxchg16b will make this behavior even more common since it will do more speculative accesses.