Re: Maybe I need a better video card? Editing xorg.conf?
On Thu, Sep 29, 2011 at 02:13:15AM -0700, Ming-Ching Tiew wrote:
> --- On Thu, 9/29/11, Ming-Ching Tiew <email@example.com> wrote:
> > Today there is one case of dead graphics which I get to
> > text mode and then 'init 2' & 'init 5', compiz can't be
> > started but metacity is fallbacked to.
> Sent out prematurely due window manager changed. :)
> Here is some log :-
> $ dmesg |grep drm
> [ 0.592351] [drm] Initialized drm 1.1.0 20060810
> [ 0.608369] [drm] MTRR allocation failed. Graphics performance may suffer.
> [ 0.609174] [drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
> [ 0.609176] [drm] Driver supports precise vblank timestamp query.
> [ 0.984756] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state
> [ 0.985004] [drm:i915_report_and_clear_eir] *ERROR* EIR stuck: 0x00000010, masking
> [ 1.160423] fbcon: inteldrmfb (fb0) is primary device
> [ 1.244720] fb0: inteldrmfb frame buffer device
> [ 1.244722] drm: registered panic notifier
> [ 1.269949] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
> [40784.632039] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
> [40784.632061] [drm:i915_wait_request] *ERROR* i915_wait_request returns -11 (awaiting 5653091 at 5653084, next 5653092)
> [40785.134030] [drm:i915_reset] *ERROR* Failed to reset chip.
Looks like the "firmware" of the card got stuck.
If that happens again, could you please try
echo "base=0xd0000000 size=0x10000000 type=write-combining" > /proc/mtrr
from the console, before restarting the X server with init 5?
(base=0xd0000000 size=0x10000000 are correct for my intel card).
This may reallocate at least the MTRR registers that got lost.
But, I wonder what the Xorg server is SUPPOSED to do when the GPU hangs?