drm error logs

jmz

New Member


Messages: 3

Hi

Upgraded my 10-Stable system yesterday and now my logs and console screen are being flooded with drm error messages. There is no loss of functionality though; i'm not sure what's happening but I'd quite like to fix it or at least least suppress these messages:

Code:
error: [drm:pid6:i915_gem_object_unbind] *ERROR* Attempting to unbind pinned buffer
I've searched loads, only found a few notes on a bug report a couple months back but the user there upgraded his system to a release of Stable older mine (r286868M) which resolved the issue. Nothing further was added to the bug report after that.

If anyone can advise I'd be very happy. Thanks very much.

Jamie
 

protocelt

Daemon

Reaction score: 411
Messages: 1,253

Unfortunately I don't have an answer to your problem but there is a PR 194766 filed that does look very similar or the same as this issue. One user mentions this was still an issue with 10.2-PRERELEASE(r285858).
 

ljboiler

Well-Known Member

Reaction score: 120
Messages: 375

I've been seeing this for quite some time with the Intel driver. Haven't seen the GPU hang mentioned in that PR, however, so I've just been ignoring it.
 
OP
OP
jmz

jmz

New Member


Messages: 3

Ok guys, thanks for your replies. Looks like I'll just have to live with it for now. I upgraded the system yesterday and it still persists. Luckily it does not cause any loss of functionality.

Jamie
 

fnoyanisi

Active Member

Reaction score: 101
Messages: 219

On my laptop, the only issue with drm I have noticed after 10.2-RELEASE upgrade was my battery life problem.

There is a thread here for further reference.
 
Last edited by a moderator:

rhsbsd

Member

Reaction score: 12
Messages: 69

Need a little help here.
Code:
[drm:pid1137:gen6_sanitize_pm] *ERROR* Power management discrepancy: GEN6_RP_INTERRUPT_LIMITS expected 1a000000, was 12060000
This is the exact error. So the way I read this: DRM is calling pid1137 which is /usr/bin/x which is calling function 'gen6_sanitize_pm'. On this machine there is only one reference to this function:
Code:
root@home:/ # dtrace -l | grep 'gen6'
52432  fbt  i915kms  i915_gen6_forcewake_count_info entry
52433  fbt  i915kms  i915_gen6_forcewake_count_info return
52530  fbt  i915kms  gen6_pm_rps_work_func entry
52531  fbt  i915kms  gen6_pm_rps_work_func return
52532  fbt  i915kms  gen6_queue_rps_work entry
52637  fbt  i915kms  gen6_fdi_link_train entry
52638  fbt  i915kms  gen6_fdi_link_train return
52719  fbt  i915kms  intel_gen6_queue_flip entry
52720  fbt  i915kms  intel_gen6_queue_flip return
52926  fbt  i915kms  gen6_init_clock_gating entry
52927  fbt  i915kms  gen6_init_clock_gating return
52928  fbt  i915kms  gen6_sanitize_pm entry<<<<<
52989  fbt  i915kms  gen6_add_request entry
52990  fbt  i915kms  gen6_add_request return
52991  fbt  i915kms  gen6_bsd_ring_write_tail entry
52992  fbt  i915kms  gen6_render_ring_flush entry
52993  fbt  i915kms  gen6_render_ring_flush return
52994  fbt  i915kms  gen6_ring_dispatch_execbuffer entry
52995  fbt  i915kms  gen6_ring_dispatch_execbuffer return
52996  fbt  i915kms  gen6_ring_flush entry
52997  fbt  i915kms  gen6_ring_flush return
52998  fbt  i915kms  gen6_ring_get_irq entry
52999  fbt  i915kms  gen6_ring_get_irq return
53000  fbt  i915kms  gen6_ring_get_seqno entry
53001  fbt  i915kms  gen6_ring_put_irq entry
53002  fbt  i915kms  gen6_ring_sync entry
53003  fbt  i915kms  gen6_ring_sync return
53187  fbt  i915kms  gen6_update_ring_freq entry
53188  fbt  i915kms  gen6_update_ring_freq return
53367  fbt  i915kms  gen6_gt_force_wake_get entry
53372  fbt  i915kms  gen6_enable_rps entry
53432  fbt  i915kms  gen6_disable_rps entry
53474  fbt  i915kms  gen6_set_rps entry
53615  fbt  i915kms  gen6_gt_force_wake_put entry
As you can see this is in located in module /boot/kernel/i915kms.ko. My question is anyone know how to initialize the i915 frame buffer from a running system. Then with a little dtracing maybe this could be remedied. It seems as if there are two distinct 'i915' modules involves here and although it looks as if DRM could be at fault, it just may not be either.
 
Top