linux-kernel - Re: Kernel panic with intel

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <11892.1318465118@turing-police.cc.vt.edu>
Date:	Wed, 12 Oct 2011 20:18:38 -0400
From:	Valdis.Kletnieks@...edu
To:	Dave Hansen <dave@...ux.vnet.ibm.com>
Cc:	"Udo A. Steinberg" <udo@...ervisor.org>,
	Dave Airlie <airlied@...il.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	iommu@...ts.linux-foundation.org,
	David Woodhouse <dwmw2@...radead.org>
Subject: Re: Kernel panic with intel_iommu=on

On Wed, 12 Oct 2011 10:19:55 PDT, Dave Hansen said:
> I'm seeing this all of a sudden on 3.1 kernels.  It might be because of
> the firefox upgrade to 5.0 that others were talking about.  But, it's
> quite easy to trigger.  Seems to happen in the first 10 minutes of using
> the system.  I was only able to capture this with netconsole.

I was seeing DMAR issues as well, but I figured it was self-inflicted because
it was due to the NVidia binary driver. Rebooted with 'intel_iommu=off' and it
behaved itself.  Basically, I got a flood of messages like these:

Oct 12 17:20:09 turing-police kernel: [  301.401157] DRHD: handling fault status reg 2
Oct 12 17:20:09 turing-police kernel: [  301.401164] DMAR:[DMA Read] Request device [01:00.0] fault addr d5b76000
Oct 12 17:20:09 turing-police kernel: [  301.401166] DMAR:[fault reason 01] Present bit in root entry is clear

Oct 12 17:20:13 turing-police kernel: [  305.615039] DRHD: handling fault status reg 2
Oct 12 17:20:13 turing-police kernel: [  305.615047] DMAR:[DMA Read] Request device [01:00.0] fault addr 110de9000
Oct 12 17:20:13 turing-police kernel: [  305.615048] DMAR:[fault reason 01] Present bit in root entry is clear

Oct 12 17:20:14 turing-police kernel: [  306.056109] DRHD: handling fault status reg 2
Oct 12 17:20:14 turing-police kernel: [  306.056116] DMAR:[DMA Read] Request device [01:00.0] fault addr ca861000
Oct 12 17:20:14 turing-police kernel: [  306.056117] DMAR:[fault reason 01] Present bit in root entry is clear

Oct 12 17:20:14 turing-police kernel: [  306.059950] DRHD: handling fault status reg 2
Oct 12 17:20:14 turing-police kernel: [  306.059959] DMAR:[DMA Read] Request device [01:00.0] fault addr ca861000
Oct 12 17:20:14 turing-police kernel: [  306.059960] DMAR:[fault reason 01] Present bit in root entry is clear

Oct 12 17:20:14 turing-police kernel: [  306.064171] DRHD: handling fault status reg 2
Oct 12 17:20:14 turing-police kernel: [  306.064179] DMAR:[DMA Read] Request device [01:00.0] fault addr ca861000
Oct 12 17:20:14 turing-police kernel: [  306.064180] DMAR:[fault reason 01] Present bit in root entry is clear

If that tells the iommu people something, fine.  If it doesn't, that's OK too, I'll
take it up with NVidia.


Content of type "application/pgp-signature" skipped