Created attachment 101318 [details] /sys/class/drm/card0/error [ 85.435367] [drm] stuck on render ring [ 85.436107] [drm] GPU HANG: ecode 0:0x85dffffd, in X [593], reason: Ring hung, action: reset [ 85.436111] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 85.436113] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 85.436114] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 85.436116] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 85.436118] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 87.437121] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off [ 91.367204] [drm] stuck on render ring [ 91.368537] [drm] GPU HANG: ecode 0:0x85dffffd, in X [593], reason: Ring hung, action: reset [ 91.368699] [drm:i915_context_is_banned] *ERROR* gpu hanging too fast, banning! [ 93.369002] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off [ 230.516279] [drm] stuck on render ring [ 230.517059] [drm] GPU HANG: ecode 0:0x85dffffd, in glxspheres64 [3870], reason: Ring hung, action: reset [ 232.517981] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off [ 236.528072] [drm] stuck on render ring [ 236.529124] [drm] GPU HANG: ecode 0:0x85dfbffb, in glxspheres64 [3870], reason: Ring hung, action: reset [ 238.529891] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off My distribution is ArchLinux with stock kernel 3.15.1-1. I do not see any graphical corruption, the visual display freezes for a few seconds. The last pre-compiled kernel that does not express this hang is 3.12.9-2.
That's different. It died after the second CA glyph pass. The batch is free of the most recent errors and doesn't have any corruption. Is this a regression (from 3.14)? Would you be able to bisect?
(In reply to comment #1) > That's different. It died after the second CA glyph pass. The batch is free > of the most recent errors and doesn't have any corruption. Is this a > regression (from 3.14)? Would you be able to bisect? I don't know if this is a recent regression or not. I just updated my kernel from 3.12.9 to 3.15.1 because I encountered similar crashes with the intermediary kernels. I can bisect, but I do not know what range I should bisect over.
Is it OK to attach my card0/error here too? Or should I open a new bug report? I'm seeing this error since a few weeks now, probably since I installed Ubuntu 14.04 (which has kernel 3.10.x). In the meantime I installed 3.14 and 3.15 but the bug keeps popping up. Here is a syslog from a few moments ago. When this happens I see some windows show funny characters, and some other artefacts. And the system is inresponsive for a few seconds. Jul 13 21:15:40 rapper kernel: [258170.239604] [drm] stuck on render ring Jul 13 21:15:40 rapper kernel: [258170.240267] [drm] GPU HANG: ecode 0:0xf5fffffe, in Xorg [1854], reason: Ring hung, action: reset Jul 13 21:15:40 rapper kernel: [258170.240269] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Jul 13 21:15:40 rapper kernel: [258170.240270] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel Jul 13 21:15:40 rapper kernel: [258170.240270] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. Jul 13 21:15:40 rapper kernel: [258170.240271] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. Jul 13 21:15:40 rapper kernel: [258170.240272] [drm] GPU crash dump saved to /sys/class/drm/card0/error Jul 13 21:15:42 rapper kernel: [258172.242543] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
Don't mean to hijack this bugreport in any way, but I wanted to mention that I am also seeing the same/similar behaviour on my Ivy Bridge / HD4000 (i3-3225) system, using Ubuntu's mainline kernel 3.15.6-031506-generic Posting here as I didn't want to create a duplicate bug report. (let me know if I actually should do just that.) [Mon Jul 28 21:36:47 2014] [drm] stuck on render ring [Mon Jul 28 21:36:47 2014] [drm] GPU HANG: ecode 0:0xf5f7fffe, in cinnamon [2586], reason: Ring hung, action: reset [Mon Jul 28 21:36:47 2014] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [Mon Jul 28 21:36:47 2014] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [Mon Jul 28 21:36:47 2014] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [Mon Jul 28 21:36:47 2014] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [Mon Jul 28 21:36:47 2014] [drm] GPU crash dump saved to /sys/class/drm/card0/error [Mon Jul 28 21:36:49 2014] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off [Tue Jul 29 00:32:59 2014] [drm] stuck on render ring [Tue Jul 29 00:32:59 2014] [drm] GPU HANG: ecode 0:0xf5f7fffe, in cinnamon [2586], reason: Ring hung, action: reset [Tue Jul 29 00:33:01 2014] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off [Tue Jul 29 07:53:25 2014] [drm] stuck on render ring [Tue Jul 29 07:53:25 2014] [drm] GPU HANG: ecode 0:0xf5f7fffe, in cinnamon [2586], reason: Ring hung, action: reset [Tue Jul 29 07:53:27 2014] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off [Tue Jul 29 09:21:32 2014] [drm] stuck on render ring [Tue Jul 29 09:21:32 2014] [drm] GPU HANG: ecode 0:0xf5f7fffe, in cinnamon [2586], reason: Ring hung, action: reset [Tue Jul 29 09:21:34 2014] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off [...] [Tue Jul 29 19:38:10 2014] [drm] stuck on render ring [Tue Jul 29 19:38:10 2014] [drm] GPU HANG: ecode 0:0xf5f7fffe, in cinnamon [2586], reason: Ring hung, action: reset [Tue Jul 29 19:38:12 2014] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off [Tue Jul 29 21:06:16 2014] [drm] stuck on render ring [Tue Jul 29 21:06:16 2014] [drm] GPU HANG: ecode 0:0xf5f7fffe, in cinnamon [2586], reason: Ring hung, action: reset [Tue Jul 29 21:06:18 2014] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
Created attachment 103684 [details] /sys/class/drm/card0/error #2 GPU crash dump associated with the previous dmesg entries
(In reply to comment #4) > Don't mean to hijack this bugreport in any way, but I wanted to mention that > I am also seeing the same/similar behaviour on my Ivy Bridge / HD4000 > (i3-3225) system, using Ubuntu's mainline kernel 3.15.6-031506-generic > > Posting here as I didn't want to create a duplicate bug report. > (let me know if I actually should do just that.) In future, please do create new bug reports as deciding whether it is a duplicate is fairly tricky. In this case, you actually have bug 77104.
Is there any chance you can retest with the latest kernel or bisect the change that introduced the hangs between 3.12 and 3.15?
On the current mainline kernel, 3.19-rc4 as of 2015-01-11, I still get the gpu hang through a few different ways: [ 48.398180] [drm] stuck on render ring [ 48.398984] [drm] GPU HANG: ecode 7:0:0x85dffffd, in Xorg.bin [354], reason: Ring hung, action: reset [ 48.398988] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 48.398989] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 48.398991] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 48.398993] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 48.398994] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 48.399059] [drm:intel_pipe_set_base [i915]] *ERROR* pin & fence failed [ 48.404834] drm/i915: Resetting chip after gpu hang [ 54.389676] [drm] stuck on render ring [ 54.390431] [drm] GPU HANG: ecode 7:0:0x85dffffd, in Xorg.bin [354], reason: Ring hung, action: reset [ 54.390500] [drm:intel_pipe_set_base [i915]] *ERROR* pin & fence failed [ 54.390513] [drm:i915_set_reset_status [i915]] *ERROR* gpu hanging too fast, banning! [ 54.396341] drm/i915: Resetting chip after gpu hang I can trigger this error only once after rebooting and then running the command: xrandr --output HDMI1 --off --output VIRTUAL1 --off --output eDP1 --mode 1920x1080 --pos 0x0 --rotate normal --output VGA1 --off Starting chromium without the flag --disable-gpu always produces the error: [ 561.799135] [drm] stuck on render ring [ 561.799939] [drm] GPU HANG: ecode 7:0:0x85dfbff8, in chromium [917], reason: Ring hung, action: reset [ 561.805793] drm/i915: Resetting chip after gpu hang
Created attachment 112251 [details] /sys/class/drm/card0/error.gz I'd had to gzip the file because it was over the file size limit.
Timeout, closing. Please reopen if the problem persists with latest kernels.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct.