[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1322566405.1834.11.camel@tellur>
Date: Tue, 29 Nov 2011 12:33:25 +0100
From: Lucas Stach <dev@...xeye.de>
To: Borislav Petkov <bp@...en8.de>
Cc: LKML <linux-kernel@...r.kernel.org>,
Andreas Herrmann <andreas.herrmann3@....com>
Subject: Re: Hangs with Linux 3.2.0-rc3
Am Dienstag, den 29.11.2011, 09:06 +0100 schrieb Borislav Petkov:
> Hi,
>
> On Tue, Nov 29, 2011 at 08:36:58AM +0100, Lucas Stach wrote:
> > [please cc me on every answer, I'm not on the list]
> >
> > Hi,
> >
> > since a few days I get lockups every few hours. I think it's related to
> > me installing a the 3.2.0-rc3 kernel. I can't say for sure if the
> > problem existed in earlier rc's as it only happens after a few hours of
> > work on this machine and I haven't had time to retest the other rc's.
>
> yeah, that might help as a sort of a coarse-grained bisection. I haven't
> seen any issues with 3.1 so far.
>
> But it looks like this is a Bulldozer in a Gigabyte board, correct? Can
> you send full dmesg from the boot along with /proc/cpuinfo please?
>
Yep, you are right, this is a FX-8150 on a Gigabyte Ga-970A-UD3 with
latest (F5a) BIOS. Now that I think about it I remember I updated the
BIOS some days ago, but only change according to Gigabyte website is an
update to a newer AGESA version. Don't know if this could be related.
I will change back to rc2 kernel today to see if the problem persists. I
don't know if the problem could be workload dependent, but I had done
some burn-in tests with this system with older BIOS (F4) and 3.2.0-rc2
kernel where I had 100% sustained cpu load on all cores where I didn't
encounter any lockups in more than 48h, now I use the system under low
load with occasional spikes and have lockups every few hours.
You will find the full dmesg of a fresh boot and cpuinfo attached to
this mail.
Sorry I have so much variables in here, but this is a relatively new
system and therefore it's still a bit in flux.
Thanks,
Lucas
> Thanks.
>
> Leaving the rest in for Andreas.
>
> > If I can be of any help tracking this down, I gladly test patches.
> >
> > Backtrace as found in syslog:
> >
[snip]
View attachment "dmesg" of type "text/plain" (94327 bytes)
View attachment "cpuinfo" of type "text/plain" (8569 bytes)
Powered by blists - more mailing lists