| Summary: | MGA will crash sometimes the whole system | ||
|---|---|---|---|
| Product: | DRI | Reporter: | Daniel Stone <daniel> |
| Component: | DRM/other | Assignee: | Ian Romanick <idr> |
| Status: | RESOLVED INVALID | QA Contact: | |
| Severity: | critical | ||
| Priority: | high | CC: | matt |
| Version: | unspecified | ||
| Hardware: | x86 (IA32) | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
FreeDesktop Bugzilla Database Corruption Fix User
2005-07-04 09:57:44 UTC
I missed a thing: I am using a Matrox MGA G550 graphics card. Cheers, Timo Which xscreensaver module is it where the lock-up occurs? Is it always the same one? Are you able to SSH into the computer when the hang occurs? If so, can you try and attach gdb to the offending xscreensaver module and get a backtrace? Are you able to reproduce this bug with a more recent DRM and client-side DRI module? What kernel version are you running? (In reply to comment #2) > Which xscreensaver module is it where the lock-up occurs? Is it always the same > one? No, it's not always the same one. They will be chosen random. Screensavers, that have already locked the system are for example BlockTube, Bouncing Cow, Juggle, Pong and Interference. > > Are you able to SSH into the computer when the hang occurs? If so, can you try > and attach gdb to the offending xscreensaver module and get a backtrace? > I'm not sure, but I will try this. > Are you able to reproduce this bug with a more recent DRM and client-side DRI Does that mean I need to build a whole new kernel? - Because it is a development workstation in a company, I think it's not possible, or in other words I'm not allowed :-(, to build a new kernel for it. > module? What kernel version are you running? > uname -r says 2.6.10-5-386 Cheers, Timo I also saw this a few days after upgrading a machine; it's now running the same Ubunto release that Timo is, with a MGA G400 AGP graphics card. I can ssh into the system and it seems happy other than the wedged X subsystem; sorry I started killing stuff before finding this bug so I didn't try attaching gdb to anything. Note that even ctrl-alt-backspace didn't work because X was too dead to even talk to the keyboard; I killed -9 the Xorg process and that didn't help; I killed -9 gdm and then the dead screensaver disappeared and I had a blank screen with a working mouse cursor and the keyboard started responding to things like Caps Lock; then I hit ctrl-alt-backspace and it looked like it was trying to restart gdm but no luck (no gdm running, and further ctrl-alt-backspace doesn't help); this did provoke a few new messages in syslog from agpgart "Found an AGP2.0 compliant device ; Putting AGPV2 device at 0000:00:00.0 into 1x mode". This doesn't seem that new; see for example this post: http://lkml.org/lkml/2003/5/6/390 Prior to upgrading to Hoary last week though, I was running RH8 (with XFree86 4.2) on the same hardware for a couple years without seeing this problem. Though I don't know for sure whether I was using DRI. Hi all, sorry for losing sight on that. Sadly I'm not able to have a SSH connection on this machine. Nearly a month I use XOrg without the DRI extension and since that time the problem is gone. Cheers, Timo I had a similar problem with xorg 6.8.2-49 using ubuntu. I disabled the AGPMode which seems to not be avaiable any more( http://www.x.org/X11R6.8.2/doc/mga.4.html ) This solved the problem in my case. Can you post your xorg.conf? And try if it still happens if you disable this option? Bye chris. closing old bug with no response for a while, and the reporter's account lost |
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.