Linux - KernelThis forum is for all discussion relating to the Linux kernel.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
Hi there. My graphics card is giving me troubles.
I have RedHat ES4, with ATI's latest drivers on my Asus P4C800 motherboard, and an ATI 9500 card.
Yesterday I upgraded my kernel to the latest stable (2.6.16.9?), and I installed ATI's drivers as well, to get DRI working (I never had DRI working before.)
However, ever since then, I get random lockups, mainly while "flying" in flightgear. The mouse moves but that's about it.
I can ssh to the machine and find that the simulator's process, fgfs is using 99.9% of the CPU. Interestingly, I am unable to reboot the machine from that ssh session - shutdown starts and hangs.
This morning it also got stuck while I was just emailing from that machine.
I did a lot of googling but was unable to find a clear-cut answer. Someone suggested to disabled Agp Fastwrites, but my motherboard has no such feature in its BIOS.
Did anyone experience this problem? If so, were you able to solve it?
Unfortunately, you did two things at once. Of the two, I'd say that ATI's drivers are the most likely culprit. I would start there ... to make very, very sure that you configured them with the right options (or whatever) for exactly the type of card you actually have, the actual amount of memory and so-on. Verify that the driver is compatible with the BIOS-level that you actually have. If there are any video diagnostic programs, run them.
Check also the /var/log/messages file, and/or other files in that directory, to see if any messages have been logged.
Next time it hangs, it would be good to try kill -9 processid on the game-process. This command is an unconditional order to the process to die immediately. See if it does so. If it doesn't, then it's almost-certainly some kind of driver (or hardware) problem, because only a low-level driver routine can put a process into such an "un-killable" state.
At that point, it's time to surf ATI's site for reports of similar issues.
Many thanks! I will check the options that it gives in xorg.conf. My MB bios is the latest production, but I will about the ATI9500's BIOS.
There were no messages in either Xorg.0.log or /var/log/messages.
Kill -9 just made the machine "more stuck" - even the ssh session I was on, got stuck. It really felt like low-level driver crash. I doubt it was the kernel, too.
I reported this to ATI, and they gave me a generic response of three URL's, "these drivers are provided as-is, see the following URL's for help with linux."
It definitely sounds like a low-level hang of some kind to me, as in a driver problem.
(1) Just to be sure, do the entire "make distclean" route to get a brand-new kernel. {Naturally, back-up your .config file first, but you do that all the time anyway, right? } Force everything to be recompiled from scratch, including all modules.
(2) Probably the best place to begin searching is right here.
(3) Sigh... never ask any tech-support goon anywhere anything about Linux. That's not in the goon manual.
Isn't it like the Nvidia drivers where you have to remove the "DRI" line from your xorg.conf.
I had a similar problem with a RH WS4 machine and removing the DRI line fixed it.
This may be of no help, but I experienced random crashes/reboots while playing games with an nvidia FX5700. Of course, I assumed there was some problem with the graphics card/driver, since that's the only time it happended. After several weeks of upgrading drivers and other fiddling about, I finally figured out that I'd left the memory clock settings in the bios to spd. The bios was taking the memory clock settings from the ram, which is PC3200 (i.e. 200/400MHz). However, I have an Athlon XP 2800+ which has a 166/33MHz FSB. So, while the board was happy running the memory at 200MHz, there clearly was some timing problem between the memory and the processor, which was causing these random crashes. The problem didn't arise until the processor was pushed flat out running the game(s), so it didn't show up in everyday use. I set the memory clock manually to 166/333 and it has been fine ever since. May have no relation to your problem but it's worth verifying the your memory and processor are in sync.
Isn't it like the Nvidia drivers where you have to remove the "DRI" line from your xorg.conf.
I had a similar problem with a RH WS4 machine and removing the DRI line fixed it.
It worked for me on Ubuntu 6.06 with the included drivers. It says
Section "DRI"
Mode 0666
delete it all
Edit: It makes it so you have to have root access to use DRI
It worked for me on Ubuntu 6.06 with the included drivers. It says
Section "DRI"
Mode 0666
delete it all
Edit: It makes it so you have to have root access to use DRI
Thanks so much. I commented out (rather than deleted) all the dri and DRI references and the "hanging" seems to have stopped. After making the changes I hit CTRL ALT Backspace and restarted the Gui and it seems to be a done deal.
I have nVidia GeForce2 MX400.
One last thing: I set Ubuntu up to allow ssh logins. When the machine hung, I was able to use putty from a Windows box to get in the back door and check the process status and kill processes. The console was hung up solid with the Xorg process consuming all the CPU 99.8 to 99.9% but I had no problem coming in via ssh.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.