Re: [drm:i915_gem_object_unbind] *ERROR* Attempting to unbind pinned buffer

On Sun, Aug 30, 2015 at 3:06 PM, Adrian Chadd <adrian_at_freebsd.org> wrote:

> Hi,
>
> has anyone asked dumbbell directly about it?
>
>
>
> -a
>
>
> On 30 August 2015 at 14:25, Slawa Olhovchenkov <slw_at_zxy.spb.ru> wrote:
> > On Sun, Aug 30, 2015 at 11:59:26PM +0300, Ruslan Makhmatkhanov wrote:
> >
> >> Slawa Olhovchenkov wrote on 08/30/2015 22:17:
> >> > On Sun, Aug 30, 2015 at 09:58:31PM +0300, Ruslan Makhmatkhanov wrote:
> >>
> >> >> No doubt that this is not the root cause, but frankly I haven't that
> >> >> "GPU hung" messages in my system. I have others like this one
> triggered
> >> >> on shutdown:
> >> >> error: [drm:pid1041:intel_lvds_enable] *ERROR* timed out waiting for
> >> >> panel to power off
> >> >>
> >> >> And this one spamming almost with the same frequency as "pinned
> buffer":
> >> >> error: [drm:pid1016:gen6_sanitize_pm] *ERROR* Power management
> >> >> discrepancy: GEN6_RP_INTERRUPT_LIMITS expected 000d0000, was 180d0000
> >> >>
> >> >> But I had not investigated that yet and not sure they are related.
> >> >> It's on r287029 head.
> >> >
> >> > All of this related to import new DRI/DRM code and such code in Linux
> >> > have same problems.
> >> > r282141 in stable related to r279599 and r275209 in current.
> >> > Can you try to revert r279599?
> >>
> >> You are right. After reverting r279599 two of this messages ("timed out
> >> waiting for panel to power off" and "unbind pinned buffer") disappeared,
> >> while "Power management discrepancy" is still there. Should I try to
> >> revert r275209 too?
> >
> > I think r275209 is not relevant here.
>
>
This thread has been quiet of late and seems to be an annoyance that a real
problem.

I would like to suggest that it might belong on x11_at_ rather than current_at_.
I see the same issue on stable and first saw it  in late March after the
MFC of  r280369 which incorporated r277487, r277959, r278146-278148,
r278152 and r278159.These are all older than the commits suggested. r277487
was the primary commit to head.

I reported this to x11_at_ back on April 1, but when I went a day with no
errors, I sent a note that it was transient and had disappeared on April 9.
Clearly I was too quick on the draw as the messages reappeared in a day or
two, but I never got around to following up.
--
Kevin Oberman, Network Engineer, Retired
E-mail: rkoberman_at_gmail.com
PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683