| Summary: | Intel GPU hangs on resume from suspend - screen freezes, must hard reboot | ||
|---|---|---|---|
| Product: | DRI | Reporter: | Rahul <rahulk> |
| Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
| Status: | CLOSED WONTFIX | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
| Severity: | major | ||
| Priority: | medium | CC: | david.weinehall, intel-gfx-bugs |
| Version: | XOrg git | ||
| Hardware: | Other | ||
| OS: | other | ||
| Whiteboard: | |||
| i915 platform: | BDW | i915 features: | power/suspend-resume |
|
Description
Rahul
2015-05-24 21:22:14 UTC
> sudo cat /sys/class/drm/card0/error
[sudo] password for taprah:
no error state collected
(In reply to Rahul from comment #1) > > sudo cat /sys/class/drm/card0/error > [sudo] password for taprah: > no error state collected Did you try this after the message appeared in the log and before a reboot? Here's what I do, I suspend the session, I resume it by double-clicking my mouse, and I get a blank white screen where the mouse is still moving but nothing else can be done other than reboot. I can only look at the log *after* reboot. (In reply to Rahul from comment #3) > Here's what I do, I suspend the session, I resume it by double-clicking my > mouse, and I get a blank white screen where the mouse is still moving but > nothing else can be done other than reboot. I can only look at the log > *after* reboot. So SSH doesn't work before a reboot? Hi Ander - the screen is frozen on wakeup. Maybe ssh works. I have no other computer on the network to ssh into it from if that's what you are suggesting. I can't do anything on the screen - it's just plain white and blank. Maybe applications are still running - but I can't tell if they are. I can't bring up an xterm or anything to get more information. So if I understand you correctly, upon reboot /sys/class/drm/card0/error gets overwritten? Can I set up a daemon to watch the file and copy it to another location? Maybe that way I can get you a stacktrace? (In reply to Rahul from comment #5) > Hi Ander - the screen is frozen on wakeup. Maybe ssh works. I have no other > computer on the network to ssh into it from if that's what you are > suggesting. I can't do anything on the screen - it's just plain white and > blank. Maybe applications are still running - but I can't tell if they are. > I can't bring up an xterm or anything to get more information. So if I > understand you correctly, upon reboot /sys/class/drm/card0/error gets > overwritten? In a way yes. That file is actually an interface to i915 code that outputs the error state recorded in the driver. So when the driver is loaded after the reboot, that error state is clear. Can I set up a daemon to watch the file and copy it to another > location? Maybe that way I can get you a stacktrace? That could work, though I'm not sure inotify would work for that particular file. You might have to constantly poll the contents of the file. It is pretty much impossible to debug the GPU hang without that error state. tail -f /sys/class/drm/card0/error | tee ~/somewhere_else.log should work? I'll try it tonight and update here, thanks. Please reproduce with a recent upstream kernel, and attach dmesg. Can you still reproduce this on a recent kernel? This bug has been in NEEDINFO state for 1.5 years now; if there's no further update I'll close it. closing bug due to inactivity by the submitter, thanks David for also looking into this :) |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.