From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael Chan" Subject: Re: [PATCH 1/1] tg3: Avoid Send BD corruption Date: Thu, 11 Sep 2008 15:55:58 -0700 Message-ID: <1221173758.12785.32.camel@HP1> References: <20080911051244.GA27384@xw6200.broadcom.net> <20080910.224618.113269063.davem@davemloft.net> <20080911165913.GA27725@xw6200.broadcom.net> <20080911.145154.112457696.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: "Matthew Carlson" , "netdev@vger.kernel.org" , "andy@greyhouse.net" To: "David Miller" Return-path: Received: from mms2.broadcom.com ([216.31.210.18]:3717 "EHLO mms2.broadcom.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752819AbYIKW5Y (ORCPT ); Thu, 11 Sep 2008 18:57:24 -0400 In-Reply-To: <20080911.145154.112457696.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, 2008-09-11 at 14:51 -0700, David Miller wrote: > From: "Matt Carlson" > Date: Thu, 11 Sep 2008 09:59:13 -0700 > > > On Wed, Sep 10, 2008 at 10:46:18PM -0700, David Miller wrote: > > > Are users really hitting this or was this discovered by your own > > > internal testing and verification? > > > > The bug was discovered by our own internal testing, but this bug is > > severe enough to warrant submission to the net-2.6 kernel. I'll be > > sending a patch to stable as well. > > > > These chips are relatively new so there isn't a whole lot of opportunity > > for users to experience the problem yet. As the chips are distributed, > > the likelyhood that users will encounter this problem is pretty high. > > That doesn't matter. That isn't the criteria for inclusion outside of > the merge window, or into -stable. Send BD corruption is quite serious as it can cause tg3_tx() to crash. We've seen a number of similar crashes over the years caused by re-ordered IOs and the nr_frags getting modified by HTB. Since regression/security/oops fixes are allowed, shouldn't this qualify since it prevents a crash in tg3_tx()? > > You should read postings such as the following one and the other ones > in it's thread so that you can become familiar with the criteria: > > http://marc.info/?l=linux-netdev&m=122048757705315&w=2 > > And by those descriptions, this patch is not appropriate and I'm > therefore not going to try and submit it for 2.6.27 and risk getting > my head chopped off by Linus again in public. > > > > >