Bug#841289: [drm/i915] GPU HANG: ecode 9:0:0xfffffffe, in Xorg [2507], reason: Engine(s) hung, action: reset

[Raphaël Halimi]

Le 31/10/2017 à 21:53, Jonas Meurer a écrit :
> I finally discovered some weeks ago that tlp seems to be responsible
> here. As soon as I disable TLP in /etc/default/tlp, I don't run into the
> bug anymore.

Hi,

Which version of TLP do you use ? I believe it may have been fixed in
1.0, please try with the version from backports (or backports-sloppy if
you use jessie) and let me know.

You may have to tweak some settings, please read :

http://linrunner.de/en/tlp/docs/tlp-configuration.html
http://linrunner.de/en/tlp/docs/tlp-faq.html
http://linrunner.de/en/tlp/docs/tlp-troubleshooting.html

Regards,

-- 
Raphaël Halimi



signature.asc
Description: OpenPGP digital signature

Bug#841289: [drm/i915] GPU HANG: ecode 9:0:0xfffffffe, in Xorg [2507], reason: Engine(s) hung, action: reset

[Jonas Meurer]

reassign 841289 tlp
thanks

On Mon, 14 Nov 2016 23:49:14 +0100 Jonas Meurer 
wrote:
> Am 19.10.2016 um 13:53 schrieb Jonas Meurer:
> > recently, something on my system related to drm/i915 broke. Since no
> > xserver/drm packages have been upgraded, I strongly suspect the latest
> > linux kernel upgrade to be the culprit.
> > 
> > Since several days, the X server on my Lenovo Thinkpad T460 sometimes
> > freezes, apparently when gnome locks and turns of the screen after some
> > time of inactivity. This renders the system completely unusable, it
> > seems to not react at all anymore. Also changing to a text consoly
> > (tty) doesn't work.
> > 
> > Below you find a copy of the relevant syslog entries. Unfortunately,
> > since the system is rendered unusable, I'm not able to save the crash
> > details from /sys/class/drm/card0/error. After reboot, it obviously is
> > empty again.
> 
> Eventually I managed to save the content of /sys/class/drm/card0/error
> after the GPU HANG by logging into the system over SSH. You find a copy
> of the file attached to this mail.
> 
> Apparently, the bug occurs only when I'm in battery mode. I'm not
> absolutely sure, but I believe that I never had the issue with my laptop
> being connected to power.

I finally discovered some weeks ago that tlp seems to be responsible
here. As soon as I disable TLP in /etc/default/tlp, I don't run into the
bug anymore.

Cheers
 jonas




signature.asc
Description: OpenPGP digital signature

Bug#841289: [drm/i915] GPU HANG: ecode 9:0:0xfffffffe, in Xorg [2507], reason: Engine(s) hung, action: reset

[Jonas Meurer]

Control: forwarded -1 https://bugs.freedesktop.org/show_bug.cgi?id=98332

Am 19.10.2016 um 14:35 schrieb Ben Hutchings:
> On Wed, 2016-10-19 at 13:53 +0200, Jonas Meurer wrote:
>> Package: src:linux
>> Version: 4.7.6-1
>> Severity: grave
>> Justification: renders package unusable
>>
>> Hello,
>>
>> recently, something on my system related to drm/i915 broke. Since no
>> xserver/drm packages have been upgraded, I strongly suspect the
>> latest
>> linux kernel upgrade to be the culprit.
> 
> Please open a bug upstream as recommended:

Thanks for the prompt reply.

> Then let us know the bug number or URL.

It can be found here: https://bugs.freedesktop.org/show_bug.cgi?id=98332

Cheers,
 jonas




signature.asc
Description: OpenPGP digital signature

Bug#841289: [drm/i915] GPU HANG: ecode 9:0:0xfffffffe, in Xorg [2507], reason: Engine(s) hung, action: reset

[Ben Hutchings]

On Wed, 2016-10-19 at 13:53 +0200, Jonas Meurer wrote:
> Package: src:linux
> Version: 4.7.6-1
> Severity: grave
> Justification: renders package unusable
> 
> Hello,
> 
> recently, something on my system related to drm/i915 broke. Since no
> xserver/drm packages have been upgraded, I strongly suspect the
> latest
> linux kernel upgrade to be the culprit.

Please open a bug upstream as recommended:

[...]
> Oct 12 21:44:40 calvin2 kernel: [11318.754599] [drm] GPU hangs can indicate a 
> bug anywhere in the entire 
> gfx stack, including userspace.
> Oct 12 21:44:40 calvin2 kernel: [11318.754603] [drm] Please file a _new_ bug 
> report on bugs.freedesktop.o
> rg against DRI -> DRM/Intel
[...]

Then let us know the bug number or URL.

Ben.

-- 
Ben Hutchings
I'm always amazed by the number of people who take up solipsism because
they heard someone else explain it. - E*Borg on alt.fan.pratchett



signature.asc
Description: This is a digitally signed message part

Bug#841289: [drm/i915] GPU HANG: ecode 9:0:0xfffffffe, in Xorg [2507], reason: Engine(s) hung, action: reset

[Jonas Meurer]

Package: src:linux
Version: 4.7.6-1
Severity: grave
Justification: renders package unusable

Hello,

recently, something on my system related to drm/i915 broke. Since no
xserver/drm packages have been upgraded, I strongly suspect the latest
linux kernel upgrade to be the culprit.

Since several days, the X server on my Lenovo Thinkpad T460 sometimes
freezes, apparently when gnome locks and turns of the screen after some
time of inactivity. This renders the system completely unusable, it
seems to not react at all anymore. Also changing to a text consoly
(tty) doesn't work.

Below you find a copy of the relevant syslog entries. Unfortunately,
since the system is rendered unusable, I'm not able to save the crash
details from /sys/class/drm/card0/error. After reboot, it obviously is
empty again.

Please tell me if I can do anything more to debug this bug. It's very
annoying as it sometimes implies loosing things that you just worked on.

Unfortunately, I don't know another way to reproduce the bug apart from
waiting for the automatic screen lock. Then you have to be (un)lucky
because it doesn't freeze every time.

Cheers,
 jonas

(Assumed) relevant logs from /var/log/syslog:

[...]
Oct 12 21:43:01 calvin2 kernel: [11219.762878] [drm] RC6 on
Oct 12 21:43:25 calvin2 kernel: [11243.760753] [drm] RC6 on
Oct 12 21:43:54 calvin2 kernel: [11272.758240] [drm] RC6 on
Oct 12 21:44:23 calvin2 kernel: [11301.754923] [drm] RC6 on
Oct 12 21:44:40 calvin2 kernel: [11318.753426] [drm] stuck on render ring
Oct 12 21:44:40 calvin2 kernel: [11318.754593] [drm] GPU HANG: ecode 
9:0:0xfffe, in Xorg [2507], reas
on: Engine(s) hung, action: reset
Oct 12 21:44:40 calvin2 kernel: [11318.754599] [drm] GPU hangs can indicate a 
bug anywhere in the entire 
gfx stack, including userspace.
Oct 12 21:44:40 calvin2 kernel: [11318.754603] [drm] Please file a _new_ bug 
report on bugs.freedesktop.o
rg against DRI -> DRM/Intel
Oct 12 21:44:40 calvin2 kernel: [11318.754607] [drm] drm/i915 developers can 
then reassign to the right c
omponent if it's not a kernel issue.
Oct 12 21:44:40 calvin2 kernel: [11318.754610] [drm] The gpu crash dump is 
required to analyze gpu hangs,
 so please always attach it.
Oct 12 21:44:40 calvin2 kernel: [11318.754614] [drm] GPU crash dump saved to 
/sys/class/drm/card0/error
Oct 12 21:44:40 calvin2 kernel: [11318.756979] drm/i915: Resetting chip after 
gpu hang
Oct 12 21:44:41 calvin2 kernel: [11319.761430] [drm] RC6 on
Oct 12 21:44:50 calvin2 kernel: [11328.752757] [drm] stuck on render ring
Oct 12 21:44:50 calvin2 kernel: [11328.753891] [drm] GPU HANG: ecode 
9:0:0xfffe, in Xorg [2507], reason: Engine(s) hung, action: reset
Oct 12 21:44:50 calvin2 kernel: [11328.756633] drm/i915: Resetting chip after 
gpu hang
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
EDID vendor "LGD", prod id 1188
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
Printing DDC gathered Modelines:
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
Modeline "1920x1080"x0.0  138.70  1920 1968 2000 2080  1080 1083 1088  
+hsync -vsync (66.7 kHz eP)
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
EDID vendor "LGD", prod id 1188
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
Printing DDC gathered Modelines:
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
Modeline "1920x1080"x0.0  138.70  1920 1968 2000 2080  1080 1083 1088  
+hsync -vsync (66.7 kHz eP)
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
EDID vendor "LGD", prod id 1188
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
Printing DDC gathered Modelines:
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: (II) modeset(0): 
Modeline "1920x1080"x0.0  138.70  1920 1968 2000 2080  1080 1083 1088  
+hsync -vsync (66.7 kHz eP)
Oct 12 21:44:51 calvin2 kernel: [11329.760399] [drm] RC6 on
Oct 12 21:44:51 calvin2 /usr/lib/gdm3/gdm-x-session[2505]: 
intel_do_flush_locked failed: Input/output error
Oct 12 21:44:51 calvin2 firefox-esr.desktop[3113]: firefox-esr: Fatal IO error 
11 (Die Ressource ist zur Zeit nicht verfügbar) on X server :0.
Oct 12 21:44:51 calvin2 pidgin.desktop[2749]: Pidgin: Fatal IO error 11 (Die 
Ressource ist zur Zeit nicht verfügbar) on X server :0.
[...]
Oct 12 21:44:51 calvin2 kernel: [11329.834184] Qt bearer threa[2850]: segfault 
at 0 ip 7fd4d6fbe9e5 sp 7fd4b7b8e560 error 4 in 
libQt5DBus.so.5.6.1[7fd4d6f5c000+87000]
Oct 12 21:44:51 calvin2 nautilus-autostart.desktop[2779]: Server response: 
STATUS:OK:/home/user
Oct 12 21:44:51 calvin2 org.a11y.atspi.Registry[2589]: XIO:  fatal IO error 11 
(Resource temporarily unavailable) on X server ":0"
Oct 12 21:44:51 calvin2 org.a11y.atspi.Registry[2589]:   after 17393 
requests (17393 known processed) with 0 events remaining.
Oct 12 21:44:51 calvin2