[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#761182: xserver-xorg-video-nouveau: NV94: screen black after resume from S3, floods of kernel errors



On 2014-09-11 15:34 +0200, Zack Weinberg wrote:

> Package: xserver-xorg-video-nouveau
> Version: 1:1.0.11-1
> Severity: important
>
> After resume from S3, the monitor does not turn back on and there are
> floods of "PGRAPH" error messages in the kernel log.  The OS is otherwise
> functional (one can ssh in, for instance) but an attempt to reboot hangs
> indefinitely because the X server has become unkillable.
>
> Sample error messages:
>
> Sep 10 19:06:25 nouveau  [    VOLT][0000:02:00.0] GPU voltage: 1000000uv
> Sep 10 19:06:25 nouveau  [  PTHERM][0000:02:00.0] fan management: automatic
> Sep 10 19:06:25 nouveau  [     CLK][0000:02:00.0] --: core 500 MHz shader 1250 MHz memory 499 MHz
> Sep 10 19:06:25 nouveau  [     DRM] resuming client object trees...
> Sep 10 19:06:25 nouveau  [     DRM] resuming display...
> Sep 10 19:06:27 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH TLB flush idle timeout fail
> Sep 10 19:06:27 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_STATUS  : 0x01000001 BUSY ROP
> Sep 10 19:06:27 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_VSTATUS0: 0x00000000
> Sep 10 19:06:27 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_VSTATUS1: 0x00000000
> Sep 10 19:06:27 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_VSTATUS2: 0x00200000 ROP
> Sep 10 19:06:29 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH TLB flush idle timeout fail
> Sep 10 19:06:29 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_STATUS  : 0x01000001 BUSY ROP
> Sep 10 19:06:29 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_VSTATUS0: 0x00000000
> Sep 10 19:06:29 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_VSTATUS1: 0x00000000
> Sep 10 19:06:29 nouveau E[  PGRAPH][0000:02:00.0] PGRAPH_VSTATUS2: 0x00200000 ROP
>
> The block from "PGRAPH TLB flush idle timeout fail" repeats indefinitely
> until the machine is forcibly powered off, with occasional other
> errors like
>
> Sep 10 19:09:24 nouveau E[     PFB][0000:02:00.0] trapped read at 0x0020012010 on channel 0x0001fbe9 [unknown] SEMAPHORE_BG/PFIFO_READ/00 reason: PAGE_NOT_PRESENT
>
> and
>
> Sep 10 19:10:36 nouveau E[Xorg[835]] failed to idle channel 0xcccc0001 [Xorg[835]]
>
> This is a regression from driver version 1.0.10 and/or kernel 3.14, both were
> upgraded simultaneously.

It's a kernel problem, most likely the same as 
https://bugs.freedesktop.org/show_bug.cgi?id=81136.  You should also see
a GPU lockup if you suspend/resume without X running.

Try booting with nouveau.nofbaccel=1, for me that helped.

Cheers,
       Sven


Reply to: