Nouveau - Jul 2014 - [Bug 81072] New: GPU lockup after "read fault at 0x0000039000 [PAGE_NOT_PRESENT]"

https://bugs.freedesktop.org/show_bug.cgi?id=81072

          Priority: medium
            Bug ID: 81072
          Assignee: nouveau at lists.freedesktop.org
           Summary: GPU lockup after "read fault at 0x0000039000
                    [PAGE_NOT_PRESENT]"
        QA Contact: xorg-team at lists.x.org
          Severity: normal
    Classification: Unclassified
                OS: Linux (All)
          Reporter: brian.campbell at editshare.com
          Hardware: Other
            Status: NEW
           Version: 7.7 (2011)
         Component: Driver/nouveau
           Product: xorg

Created attachment 102454
  --> https://bugs.freedesktop.org/attachment.cgi?id=102454&action=edit
Xorg log with crash

About once every couple of weeks, my display locks up. Sometimes I am able to
switch virtual terminals and reboot; sometimes after switching virtual
terminals, something tries to restart X and the display becomes completely
garbled and unusable. I can't reproduce this reliably, but it generally
seems
to happen when I have a lot of windows open and am in the middle of doing
something like opening a new window or switching to the Gnome Shell overview.

I filed a bug with Debian a few months ago, haven't heard anything:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=744166. Here are a few lines
excerpted from dmesg that seem to indicate when the problem started on that
crash (I forgot to save dmesg output from the most recent crash before
rebooting); the rest is included in the Debian bug report (sadly, appears not
to be complete; I'll increase my log buffer size and get a full log the next
time this happens). I've attached my Xorg.0.log.old from my most recent
crash,
and the Xorg.log from the older crash is attached to the Debian bug report.

[112444.556301] nouveau E[   PFIFO][0000:01:00.0] read fault at 0x0000039000
[PAGE_NOT_PRESENT] from PCOPY0/PCOPY0 on channel 0x001fdb2000 [DRM]
[112475.742287] nouveau E[     DRM] GPU lockup - switching to software fbcon
[112490.737414] nouveau E[Xorg[1800]] failed to idle channel 0xcccc0001
[Xorg[1800]]
[112505.727471] nouveau E[Xorg[1800]] failed to idle channel 0xcccc0001
[Xorg[1800]]
[112507.726348] nouveau E[   PFIFO][0000:01:00.0] playlist update failed
[112522.716209] nouveau E[Xorg[1800]] failed to idle channel 0xcccc0000
[Xorg[1800]]
[112537.706272] nouveau E[Xorg[1800]] failed to idle channel 0xcccc0000
[Xorg[1800]]
[112539.707186] nouveau E[   PFIFO][0000:01:00.0] channel 2 [Xorg[1800]] kick
timeout
[112539.708006] nouveau W[   PFIFO][0000:01:00.0] INTR 0x00000100: 0x0000000d
[112541.706666] nouveau E[   PFIFO][0000:01:00.0] playlist update failed

Here are my current package versions:

libdrm-nouveau2: 2.4.54-1
xserver-xorg-video-nouveau: 1:1.0.10-1+b1
xorg: 1:7.7+7
linux-image-3.14-1-amd64: 3.14.7-1

If there's any other relevant information I can provide, let me know. I
can't
reproduce this on demand, and I don't have a known good version to bisect
from;
I switched from Mandriva running the proprietary drivers (which seemed to have
some problems of their own, causing substantial slowness), to Debian running
nouveau, and the problem has been happening occasionally ever since the switch.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20140709/691ab32d/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

Ilia Mirkin <imirkin at alum.mit.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|GPU lockup after "read      |[NVC1] GPU lockup after
                   |fault at 0x0000039000       |"read fault at
0x0000039000
                   |[PAGE_NOT_PRESENT]"         |[PAGE_NOT_PRESENT] from
                   |                            |PCOPY0"

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20140709/ca2f2e77/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

Brian Campbell <brian.campbell at editshare.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #102454|0                           |1
        is obsolete|                            |

--- Comment #1 from Brian Campbell <brian.campbell at editshare.com> ---
Created attachment 102944
  --> https://bugs.freedesktop.org/attachment.cgi?id=102944&action=edit
Another Xorg.0.log with crash

Crash happened again, here's another Xorg.log.0.old.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20140716/33cac8d5/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

--- Comment #2 from Brian Campbell <brian.campbell at editshare.com> ---
Created attachment 102945
  --> https://bugs.freedesktop.org/attachment.cgi?id=102945&action=edit
dmesg from most recent crash

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20140716/acea342b/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

--- Comment #3 from Andrew <andrew at fritztech.com> ---
I have the same issue (same symptoms and apparent cause). I can usually SSH
into the box, however a hard reset is required before the GPU will function
again. A soft reboot (using reboot or halt) leaves the screen on and with
whatever image was there when he GPU locked up. Generally the lock up occurs
when interacting with firefox but that might be coincidental.

Here is the tail end of dmesg last time it occurred:

[631973.764168] scsi 21:0:0:0: Direct-Access     Generic  Ultra HS-SD/MMC  1.82
PQ: 0 ANSI: 0
[631973.764503] sd 21:0:0:0: Attached scsi generic sg8 type 0
[631973.781306] sd 21:0:0:0: [sdh] Attached SCSI removable disk
[670004.997024] type=1400 audit(1415799667.273:94): apparmor="STATUS"
operation="profile_replace" profile="unconfined"
name="/usr/lib/cups/backend/cups-pdf" pid=24612
comm="apparmor_parser"
[670004.997030] type=1400 audit(1415799667.273:95): apparmor="STATUS"
operation="profile_replace" profile="unconfined"
name="/usr/sbin/cupsd"
pid=24612 comm="apparmor_parser"
[670004.997300] type=1400 audit(1415799667.273:96): apparmor="STATUS"
operation="profile_replace" profile="unconfined"
name="/usr/sbin/cupsd"
pid=24612 comm="apparmor_parser"
[757535.322498] type=1400 audit(1415887157.175:97): apparmor="STATUS"
operation="profile_replace" profile="unconfined"
name="/usr/lib/cups/backend/cups-pdf" pid=3423
comm="apparmor_parser"
[757535.322505] type=1400 audit(1415887157.175:98): apparmor="STATUS"
operation="profile_replace" profile="unconfined"
name="/usr/sbin/cupsd"
pid=3423 comm="apparmor_parser"
[757535.322776] type=1400 audit(1415887157.175:99): apparmor="STATUS"
operation="profile_replace" profile="unconfined"
name="/usr/sbin/cupsd"
pid=3423 comm="apparmor_parser"
[757612.584375] nouveau E[   PFIFO][0000:01:00.0] read fault at 0x0000039000
[PAGE_NOT_PRESENT] from PCOPY0/PCOPY0 on channel 0x001fdb2000 [DRM]

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20141113/c2e616ef/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

--- Comment #4 from Patryk Zawadzki <patrys at pld-linux.org> ---
I am getting a slightly different error and am not sure whether it's the
same
issue:

nouveau E[   PFIFO][0000:01:00.0] read fault at 0x0011980000 [PAGE_NOT_PRESENT]
from PGRAPH/GPC0/PROP on channel 0x00bf739000 [Xorg[2305]]
nouveau E[   PFIFO][0000:01:00.0] PGRAPH engine fault on channel 10,
recovering...

After that the screen freezes and I need to do a hard reset. Applications
continue running and I can SSH in but restarting Xorg and trying to soft reboot
do not help.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20151005/3a4b85e0/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

--- Comment #5 from Patryk Zawadzki <patrys at pld-linux.org> ---
Just got some more:

nouveau E[Xorg[2265]] failed to idle channel 0xcccc0000 [Xorg[2265]]
nouveau E[Xorg[2265]] failed to idle channel 0xcccc0000 [Xorg[2265]]

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20151005/24f45e22/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=81072

Martin Peres <martin.peres at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |MOVED

--- Comment #6 from Martin Peres <martin.peres at free.fr> ---
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been
closed from further activity.

You can subscribe and participate further through the new bug through this link
to our GitLab instance:
https://gitlab.freedesktop.org/xorg/driver/xf86-video-nouveau/issues/119.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://lists.freedesktop.org/archives/nouveau/attachments/20191204/3b49582f/attachment.html>