i915 driver hangs up, GPU wedged.
AaroneusTheGreat
Status: New User - Welcome
Joined: 11 Jun 2013
Posts: 3
Location: North Carolina, USA
Reply Quote
So I've been getting this odd problem where during certain graphically intensive situations, X completely goes down and I can't get it back up by restarting it by dropping into the terminal. Today I finally got a message come through on dmesg and I captured it finally!

I am looking for help on who to talk to and/or possible solutions for what I'm experiencing. To help whomever is nice enough to read this and look into it I'm including my inxi output as well as the dmesg of the error.

Inxi output :

:: Code ::

System:    Host: mordecaipc Kernel: 3.9-5.dmz.1-liquorix-amd64 x86_64 (64 bit, gcc: 4.7.3)
           Desktop: Gnome dm: gdm3 Distro: Debian GNU/Linux 7.0
Machine:   System: TOSHIBA (portable) product: Satellite Pro C650 version: PSC13U-00W01E serial: ZA018389Q
           Mobo: TOSHIBA model: Portable PC Bios: INSYDE version: 1.60 date: 09/02/2010
CPU:       Dual core Intel Core i3 CPU M 370 (-HT-MCP-) clocked at 2399.00 MHz
Graphics:  Card: Intel Core Processor Integrated Graphics Controller bus-ID: 00:02.0 chip-ID: 8086:0046
           X.org: 1.12.4 drivers: intel (unloaded: fbdev,vesa) tty size: 80x24 Advanced Data: N/A for root
Network:   Card-1: Atheros AR9285 Wireless Network Adapter (PCI-Express) driver: ath9k bus-ID: 02:00.0 chip-ID: 168c:002b
           Card-2: Atheros AR8152 v1.1 Fast Ethernet
           driver: atl1c ver: 1.0.1.1-NAPI port: 3000 bus-ID: 01:00.0 chip-ID: 1969:2060
Drives:    HDD Total Size: 320.1GB (-)
Info:      Processes: 169 Uptime: 18 min Memory: 1025.9/3753.1MB Runlevel: 2 Gcc sys: 4.7.2
           Client: Shell (bash 4.2.37 running in gnome-terminal) inxi: 1.9.7


and here's the dmesg error log:

:: Code ::

wlan0: associated
IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
telepathy-idle[4401]: segfault at 0 ip 00007f430d197240 sp 00007fff877b7b98 error 4 in libgio-2.0.so.0.3200.4[7f430d137000+14c000]
[drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung
[drm] capturing error event; look for more information in/sys/kernel/debug/dri/0/i915_error_state
[drm:init_ring_common] *ERROR* failed to set render ring head to zero ctl 00000000 head b9c17e9c tail 00000000 start 00003000
[drm:init_ring_common] *ERROR* render ring initialization failed ctl 0001f001 head b9c17e9c tail 00000000 start 00003000
[drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung
[drm:i915_reset] *ERROR* GPU hanging too fast, declaring wedged!
[drm:i915_reset] *ERROR* Failed to reset chip.


I don't think the bit about the ipv6 error has anything to do with it but I included it for completeness.

Thank you in advance to anyone who spends any time on this. I really do appreciate it. I've scoured the Internet for several hours about this and I haven't found much that has been useful yet.
Back to top
techAdmin
Status: Site Admin
Joined: 26 Sep 2003
Posts: 4124
Location: East Coast, West Coast? I know it's one of them.
Reply Quote
good issue report, I'd send this upstream to the intel driver group of Xorg if I were you, or the Debian packager first in case there's already a fix upstream.

Either case it's not something that liquorix or anyone here can fix for you, unless it turns out there's a kms patch for kernels that is related to the issue.

It's always annoying to be an early bug detector, then there just is not enough information online, but with the code you pasted in here for error, that should help a bit in terms of starting the ball rolling.

I suggest posting bug first to debian bugs, see if you get any results there first.
Back to top
Thanks.
AaroneusTheGreat
Status: New User - Welcome
Joined: 11 Jun 2013
Posts: 3
Location: North Carolina, USA
Reply Quote
Thanks for helping me figure out where to direct this. I'll go and see what I can find out. In the interest of helping others, would you like me to post any solutions I find here so others may find them easily?
Back to top
techAdmin
Status: Site Admin
Joined: 26 Sep 2003
Posts: 4124
Location: East Coast, West Coast? I know it's one of them.
Reply Quote
Oh, yes, please post solutions, or, if a new version of xorg driver fixes it, post which release does that.

It's always nice for problem threads to contain the solutions as well,.
Back to top
techAdmin
Status: Site Admin
Joined: 26 Sep 2003
Posts: 4124
Location: East Coast, West Coast? I know it's one of them.
Reply Quote
:: Code ::
apt-cache show xserver-xorg-video-intel


will show dependencies, particularly the drm stuff, the maintainer emails etc, and debian bugs: www.debian.org/Bugs/

is where you search for bug / issue reports and post your own if you don't find anything.

Generally with packaged drivers you want to go to the distro first because it's possible the issue has already been fixed upstream. Not necessarily, but that's the general way to do it.

You could also post directly to Xorg just to see if it's been fixed yet, give them the same basic data you gave here, that's enough to get them started.

Also show this:

apt-cache policy xserver-xorg-video-intel

so they know what version you are running.
Back to top
I found a workaround.
AaroneusTheGreat
Status: New User - Welcome
Joined: 11 Jun 2013
Posts: 3
Location: North Carolina, USA
Reply Quote
Thanks for the tips, I have passed the report on along to the maintainers of the driver, and hopefully it'll help them find a fix for this.

I did manage to hunt down a work around, it's not a fix but it makes the system run like it should. I'm not entirely sure how it works to be honest, but after scouring the Internet for a few days I'm not really sure I care, because it does work.

Add this to /etc/default/grub

:: Code ::
i915.i915_enable_rc6=0


to the line

:: Code ::
GRUB_CMDLINE_LINUX_DEFAULT


and it'll stop the hangs from the i915 driver. Now from what I do understand about this it changes a mode for the graphics chip, so it might change how the chip functions during normal operation, possibly affecting performance or power usage, but I don't know the details.
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 1117
Reply Quote
Can you try 3.9.6 that's now in the repository? There's a few i915 and DRM bug fixes, one that specifically mentions wedged intel gpus.
Back to top
Display posts from previous:   

All times are GMT - 8 Hours