From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: RE: bnx2 FTQ issues on 2.6.32 + xen 4.0.1; Date: Sat, 30 Jul 2011 11:37:11 -0400 Message-ID: <20110730153711.GA6541@dumpdata.com> References: <20110729141613.GC5458@dumpdata.com> <20110730141539.GA7801@dumpdata.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: MaoXiaoyun Cc: xen devel , mchan@broadcom.com List-Id: xen-devel@lists.xenproject.org On Sat, Jul 30, 2011 at 11:27:24PM +0800, MaoXiaoyun wrote: > > I have below configuration in grub.conf, did u mean change 115200 to sth? > > iommu=off x2apic=off console=com1,vga com1=115200,8n1 Hm, that is what I had in mind. That is not enough to get the data before your machine crashes? When it crashes - you mean the watchdog timer hits - so your machine is still functioninig right? Then you should be able to see the output of the driver in /var/log/messages,syslog I think? You are still able to recover when the bnx drivers goes bonk, right? > > thanks. > > > > > Date: Sat, 30 Jul 2011 10:15:39 -0400 > > From: konrad.wilk@oracle.com > > To: tinnycloud@hotmail.com > > CC: mchan@broadcom.com; xen-devel@lists.xensource.com > > Subject: Re: [Xen-devel] RE: bnx2 FTQ issues on 2.6.32 + xen 4.0.1; > > > > > > > I've spent more time on this. > > > > > It's easy to reproduce. I have two physical host, on top of which, totally start 15VMS, doing migrating over > > > > > and over again. Since xen live migration invovle large mount of memory copy, so I can see that the network is > > > > > saturated. About 16 hours later, the FTQ dump will show up. > > > > > > > > > > I've debugged more. > > > > > Below debug code is added when FTQ happens. > > > > > Does anyone know what is going on? > > > > > > > > No idea. Did Michael Chan send you a private email with some ideas? > > > > > > > > > > Yep, Michael give me some debug patch. But unfortunately, when this bug show up, there is no enough time for the > > > kernel to print all debugged log before it crashes. > > > > Can you use a serial console at a high baud rate? >