[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#438736: intel driver produces kernel crash in a variety of situations

On Mon, Aug 20, 2007 at 12:21:12AM +0200, Cesare Leonardi wrote:
> Robert Millan wrote:
> >I've recently upgraded my hardware (motherboard + cpu + ram) and am now
> >getting Linux crashes in a variety of situations.  I've detected them
> >when any of the following conditions are met:
> >
> >  - When starting a second X session.
> >  - When time goes backwards (e.g. due to ntpdate) while X is running (but
> >    NOT when X isn't running)
> >  - When stopping X.
> [...]
> >  - X driver is xserver-xorg-video-i810.  Card model:
> Are you sure this is a kernel bug?

Well, since X is a userland process, I wouldn't expect it to be able to
crash the kernel, even if it plays with /dev/agpgart in a bad way.  It
can even be a security issue.

> I see that the constant is that crashes happen when X is running (or 
> when is finishing running).
> You are using xserver-xorg-video-i810, but since some months this is a 
> transitional package that points to the newer (and less stable) 
> xserver-xorg-video-intel.
> Since this transition i had noticed numerous crash (and various problems 
> using video applications, but these seems to be resolved), and reading 
> some bug reports seems that this driver still has different problems on 
> which Intel is working on.
> I suggest you to look at:
> http://bugs.debian.org/xserver-xorg-video-intel

Yes, I've seen those reports.  That's why I holded on backporting the driver
from testing/sid to my etch system.

> In the meanwhile try to use the generic vesa driver to see if the 
> crashes still happens. If they will not occur anymore, you can try with 
> the intel driver from unstable or experimental.
> If they'll persists... this could be a kernel bug.  ;-)

Do you know if Linux had improvements in this area from 2.6.18 to 2.6.22 ?

Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)

Reply to: