Skip to contentSkip to conversation

drm/i915: protect force_wake_(get|put) with the gt_lock

PATCH Nicolas Kalkhof 2 participants 2 messages Nov 7, 2011

lore
Share & tools
lore
#0 lore raw

Hi Daniel,

confirmed. Linus's linux-next.git shows the same behaviour. Until now I've used a patched 3.1-rc-6+ Kernel from Dave Airlie's branch (git://people.freedesktop.org/~airlied/linux' from Oct 26th.). dmesg/syslog shows nothing special, neither does i915_error_state.

Any advice?

Nic

-----Ursprüngliche Nachricht-----
Von: "Daniel Vetter" <[email protected]>
Gesendet: Nov 7, 2011 5:05:12 PM
An: "Nicolas Kalkhof" <[email protected]>
Betreff: Re: [Intel-gfx] [PATCH] drm/i915: protect force_wake_(get|put) with the gt_lock

Quoted 24 lines · Nicolas Kalkhof
Quoted 11 lines

Hi Daniel,

thanks for the git branch. Strange thing is with the patched kernel the cpu load goes up after working with my desktop with no obvious reason. top shows a high user and system load but no process using up all the cpu cores. The cpufreq-daemon clocks the cpu cores down to the minimal frequency (800mhz) when idle while top still shows full cpu load but no process using up the cpu. The cpu load drops to normal after I kill X and immediately goes up again if I restart X.

I'm kinda lost here :( Any Ideas?

Step one is to check whether -linus master has the same issues. My branch is based upon

commit 37be944a0270402f9cda291a930b0286f6dc92f5
Merge: ca836a2 1717c0e
Author: Linus Torvalds <[email protected]>
Date: Fri Oct 28 05:54:23 2011 -0700

Merge branch 'drm-core-next' of
git://people.freedesktop.org/~airlied/linux'

Quoted 8 lines

If that doesn't show the same issue, can you please run a bisect (only 13 commits) to find out which patch introduced this?

Thanks, Daniel
--
Daniel Vetter
Mail: [email protected]
Mobile: +41 (0)79 365 57 48

___________________________________________________________
SMS schreiben mit WEB.DE FreeMail - einfach, schnell und kostenguenstig. Jetzt gleich testen! http://f.web.de/?mc=021192
_______________________________________________
Intel-gfx mailing list [email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#1 lore raw
Quoted 8 lines · Nicolas Kalkhof

Hi Daniel,

confirmed. Linus's linux-next.git shows the same behaviour. Until now I've used a patched 3.1-rc-6+ Kernel from Dave Airlie's branch
(git://people.freedesktop.org/~airlied/linux' from Oct 26th.). dmesg/syslog shows nothing special, neither does i915_error_state.

Any advice?

If the behaviour-difference is clear, you could try to bisect this. The alternative is to use your favourite system profiler and see where the cycles get wasted (either sysprof or perf). I suggest you'll try whatever approach your more familiar with first and then switch over to the other if the first one doesn't yield any clear results. If the difference is really clear, I'd start with the bisect (aside: don't restrict the bisect to any subdir, the change causing the regression could equally likely be somewhere in the core kernel).

Yours, Daniel

Signature · 3 lines
Daniel Vetter
Mail: [email protected]
Mobile: +41 (0)79 365 57 48