| Summary: | EQ overflowing - recurring total X crash | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | lockheed <qwrules> |
| Component: | Drivers/Gallium/r600 | Assignee: | Default DRI bug account <dri-devel> |
| Status: | RESOLVED DUPLICATE | QA Contact: | |
| Severity: | critical | ||
| Priority: | medium | CC: | andyrtr |
| Version: | git | ||
| Hardware: | x86-64 (AMD64) | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: |
Xorg.0.log
dmesg Updated log with more ERROR info |
||
|
Description
lockheed
2014-02-02 22:43:10 UTC
Please attach your Xorg.0.log file and the output of dmesg. Created attachment 93265 [details]
Xorg.0.log
Xorg log up until the manual reboot after the crash.
@Michel Dänzer, did you mean output of dmesg directly after crash, or just a general one? It may be difficult to get it with screwed up console, but I will try to get it next time it crashes (In reply to comment #3) > @Michel Dänzer, did you mean output of dmesg directly after crash, or just a > general one? Preferably the former, but the latter will do in the meantime. Does the problem also occur without Option "EXAVSync"? Created attachment 93454 [details]
dmesg
Here is dmesg directly after crash.
Does that help?
>>>>> Does the problem also occur without Option "EXAVSync"?
If anything, turning it off makes it worse. With it, I get the crash anywhere between few hours, to 2 days. Without it, it crashed in less than an hour.
I think I found a way how to replicate it fairly regularly. When I connect to a remote VNC server with VNC client (I use remmina) and then open some graphically intensive website in www browser on the remote host, and scroll it vigorously up and down, the crash occurs fairly quickly. I'd like to know if anyone is doing anything about it. Were you able to identify the faulty code? I just lost 3 hours of audio work because of another crash. This time autosave did not help me. I am forced to go back to catalyst, which require me to downgrade xorg server, too. I would also like to add, that I tried kernels up to 3.14 RC1 with the same result. (In reply to comment #8) > Were you able to identify the faulty code? It's not that simple. Something is causing the GPU to hang, but it can be tricky to isolate what exactly that is. Please try eliminating some variables, e.g.: Does it also happen with radeon.dpm=0? Does it also happen if you disable / enable compositing? ... *** Bug 74511 has been marked as a duplicate of this bug. *** (In reply to comment #10) > Please try eliminating some variables, e.g.: Does it also happen with > radeon.dpm=0? Does it also happen if you disable / enable compositing? ... I will try radeon.dpm=0 when I get a chance. Can I use dynpm then, or better not to avoid skewing the results? But what do you mean by "disable / enable compositing"? I am using openbox + compton. (In reply to comment #12) > (In reply to comment #10) > > Please try eliminating some variables, e.g.: Does it also happen with > > radeon.dpm=0? Does it also happen if you disable / enable compositing? ... > > > > I will try radeon.dpm=0 when I get a chance. Can I use dynpm then, or better > not to avoid skewing the results? Just disable dpm for now. I have a similar problem, the screen freezes ( keyboard and mouse are frozen too ) for one minute, then X crashes with an automatic reboot of X, after the reboot all seems Ok but 2 minutes later the same problem occurs again, I use archlinux 64 bits, radeon driver ( I have an ati radeon HD4650 pcie ), mesa 10.0.3 and kernel 3.12.9, with dpm mode enabled, but this bug is very rare and difficult to reproduce, and I forgot to copy the Xorg log file ( because at each reboot the Xorg log is automaticaly erased ), but for what I saw with my configuration the bug seems to be related with vdpau driver which is part of mesa package, and also related with the linux kernel because vdpau relies on linux kernel, a possible pattern to trigger the bug with my configuration : - use KDE 4.12 with opengl 2.0 enabled ( kwin ) - enable DPM feature for radeon driver - use a software who uses vdpau features ( for example "kaffeine" who is a dvb-t software, mplayer ) - open multiple GUI programs like firefox, notepad, dolphin ( whatever you want ) soon or later ( it can take hours or even days ) something will go wrong, the first step is kaffeine who will show a black screen instead of a video image, then the second step is a total freeze of KDE 4 and keyboard/mouse, the freeze will last 30 seconds to 1 minute, then X will crash and automaticaly reboot, what I am sure : if I disable "vdpau" feature in the configuration of Kaffeine the bug will never occur, "vdpau" seems to be the key element in my configuration for this bug (In reply to comment #12) > But what do you mean by "disable / enable compositing"? I am using openbox + > compton. Does the problem also happen without compton, or with a different compton backend? (In reply to comment #14) > I have a similar problem, [...] Please file your on report. Update: 1. radeon.dpm=1, compton backed: both "xrender" and "glx" Crash occurs. 2. radeon.dpm=0, "glx" compton backend (did not feel the need to test xrender) I have been working with a VNC session over few hours quite heavily (which usually causes the crash very quickly) and crash has not occurred. So it looks like it is DPM-related, and my guess is it occurs when the frequency stepping is changed. However, with radeon.dpm=0 the GPU temp is about 15-20 C higher, even with dynpm on. *** This bug has been marked as a duplicate of bug 66963 *** I am getting more detailed logs in the recent version. I hope those can help to solve the issue. New log attached. Created attachment 95412 [details]
Updated log with more ERROR info
(In reply to comment #19) > I am getting more detailed logs in the recent version. I hope those can help > to solve the issue. Unfortunately not, the backtraces in the log file just show where the driver hangs due to the hanging GPU, they don't say anything about what caused the GPU to hang. Unless you can reproduce this problem with DPM disabled, please follow up on bug 66963 instead of reopening this one. *** This bug has been marked as a duplicate of bug 66963 *** |
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.