Re: [Bug #11382] e1000e: 2.6.27-rc1 corrupts EEPROM/NVM
[Posted September 24, 2008 by corbet]
| From: |
| Jiri Kosina <jkosina-AT-suse.cz> |
| To: |
| Jonathan Corbet <corbet-AT-lwn.net> |
| Subject: |
| Re: [Bug #11382] e1000e: 2.6.27-rc1 corrupts EEPROM/NVM |
| Date: |
| Wed, 24 Sep 2008 18:56:01 +0200 (CEST) |
| Message-ID: |
| <alpine.LNX.1.10.0809241852230.4671@pegasus.suse.cz> |
| Cc: |
| David Miller <davem-AT-davemloft.net>, airlied-AT-gmail.com,
jeffrey.t.kirsher-AT-intel.com, david.vrabel-AT-csr.com, rjw-AT-sisk.pl,
linux-kernel-AT-vger.kernel.org, kernel-testers-AT-vger.kernel.org,
chrisl-AT-vmware.com |
On Wed, 24 Sep 2008, Jonathan Corbet wrote:
> A data point, just in case it helps... I've not had time to update my
> desktop system, so this all-Intel, ICH9, e1000e-based box has been stuck
> at 2.6.27-rc3. It has rawhide as of shortly after the floodgates
> reopened (but with my own kernel); that means X server 1.5.0 and i810
> 2.4.2-3.
> It's happy as a clam. I'm not sure how often this problem bites, but it
> hasn't gotten me.
Thanks for the information.
Seems like it quite often triggers during the very first probing of
graphics card during the initial X startup. Karsten is currently writing a
tool that will safely restore the EEPROM contents to the card. When this
gets done, testing will get much easier and hopefully we'll be able to
isolate whether it is e1000e driver (I currently don't think so), DRM
kernel code, or xorg 7.4 causing this.
Thanks,
--
Jiri Kosina
SUSE Labs
