Bug#578912: linux-image-2.6.32-4-686: soft lockups with radeon KMS
tags 578912 moreinfo
thanks
On Fri, Apr 23, 2010 at 03:40:16PM +0200, Zsolt Rizsanyi wrote:
> Package: linux-2.6
> Version: 2.6.32-11
> Severity: important
>
> When Radeon KMS is enabled with my Radeon X1300 the system gets into
> soft lockups. The lockups happen after 1-8 hours of use (altough it
> might be related to some event eg. DPMS or anything else, not sure about
> that).
>
> Each lockup last from about 1-60 seconds, but the majority is about 5-15
> seconds. When the lockup starts to happen it happens quite frequently.
> with 5-60 second breaks between them.
> During the lockup the screen is frozen, the mouse pointer does
> not move. (Interestingly if I type it gets processed after the lockup.
> But if I move the pointer during the lockup nothing happens from it.)
>
> After the lockup is finished everything seems to be working as if
> nothing has happend.
>
> When the screen locks up some very strange effect can be seen on the
> screen I have never seen or heard about. It looks like the effects when
> you log out from a graphical session and the screen gets dimmed with
> applying a pattern to it. Except in this case the dimming is not
> completely uniform, but looks like it is spreading accross the screen.
> It is constantly changing during the lockup getting dimmer and dimmer.
> When the lockup start almost nothing can be seen (sometimes only a small
> part of the screen is smudged). But it quickly gets darker (in 1-2
> seconds).
>
> I have tried to take pictures of the screen when the lockup happens with
> my digital camera. The pictures are not the best because of the
> reflections, but this is the best I can make. I will be ataching them to
> the bug report.
> In the pictures you can see how the screen gets darker by time in one of
> the longer lockups. The last picture is only for reference which was
> taken after the lockup (so that you can see how light is the screen
> normally).
>
> This problem has always started appearing when X was running, but this
> is no surprise since I'm usually running X. Butonce when I was shutting
> down the machine I have seen the lockup occuring outside of X too.
>
> If I reboot with a different kernel image (2.6.32-3-686 version 2.6.32-9
> in particular) this does not happen at all (even after many days of use,
> with suspend/resumes).
>
> I have tried the 2.6.33 kernel from experimental and this happens there
> too. (Which is no surprise, since they have the same KMS code.)
>
> When these lockups happen I don't see any errors printed in any of the
> log files (checked dmesg, syslog, X log).
>
> BTW I would consider this bug as release critical, but since nobody has
> reported this yet I assume that this affects only a small fraction of
> users. So I have set it to important severity :)
>
> I have added Alex Deucher to the CC, because I have seen that he was
> active on all the radeon KMS reports I have checked (to see if this
> issue is already reported or not).
>
> If there is any more information you need please ask.
Please test if the problem persists with the 2.6.34 from experimental.
If so, please report this bug upstream at the upstream bugzilla
(http://bugzilla.kernel.org), so that it can be investigated/
fixed upstream and backported for Squeeze.
(Product: Drivers, Component: Video(DRI - non Intel))
Once done, please add the bug number to this bug.
(We're asking you to file the bug report yourself, since the
kernel.org developers will have specific inquiries to your
hardware/setup)
Cheers,
Moritz
Reply to: