From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kirill Korotaev Subject: Re: [Devel] Re: [PATCH] Copy mac_len in skb_clone() as well Date: Thu, 15 Mar 2007 13:33:12 +0300 Message-ID: <45F920E8.4020304@sw.ru> References: <20070314130711.GA6812@localhost.sw.ru> <20070315.030235.79052031.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: adobriyan@sw.ru, netdev@vger.kernel.org To: devel@openvz.org Return-path: Received: from mailhub.sw.ru ([195.214.233.200]:29944 "EHLO relay.sw.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751360AbXCOKSz (ORCPT ); Thu, 15 Mar 2007 06:18:55 -0400 In-Reply-To: <20070315.030235.79052031.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org David Miller wrote: > From: Alexey Dobriyan > Date: Wed, 14 Mar 2007 16:07:11 +0300 > > >>ANK says: "It is rarely used, that's wy it was not noticed. >>But in the places, where it is used, it should be disaster." >> >>Signed-off-by: Alexey Dobriyan > > > Applied. > > What bug triggered that helped you discover this? Or is it > merely from a code audit? Ohhh, it is a fairy-tale to tell the truth :) We had some unexplainable problems with java application in OpenVZ kernel. It didn't work sometimes, but worked fine (!) with CONFIG_SLAB_DEBUG. Alexey blamed java :), but ... Then we found that poising one of the bits in slab cache was curing it. After that we found that the problem is related to fclone cache. And then we found that not all the fields are initialized during cloning. The bug was related to our own skb->field we introduced, but we analyzed the code and found this as well. Thanks, Kirill