I have a Dell Optiplex GX620 with a Radeon X600 card installed in the PEG slot: 01:00.0 VGA compatible controller: ATI Technologies Inc RV370 5B62 [Radeon X600 (PCIE)] It has an Intel EM64T CPU and 2GB of DDR2 PC2-5300 memory (4 512MB DIMMs). I am running Fedore Core 4 (Linux) on it, using kernel 2.6.15-1.1831_FC4 (the latest) and xorg-x11-6.8.2-37.FC4.49.2. About 5-10 seconds after starting X, the entire machine locks, hard. And when I say "lock", I mean "the machine even stops responding to ICMP ECHO requests". There's no errors and nothing unusual in any log files. Since I'm using Fedora Core 4, I reported this bug via Red Hat's Bugzilla: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=172335 Another person with the same hardware (a Dell Optiplex GX260 with a Radeon X600) reported the same problem: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=176826 The second reporter was able to get some console messages with netconsole: Uhhuh. NMI received. Dazed and confused, but trying to continue Uhhuh. NMI received. Dazed and confused, but trying to continue You probably have a hardware problem with your RAM chips You probably have a hardware problem with your RAM chips I have not been able to confirm if I get the same console messages. The second reporter also reports that the proprietary ATI driver doesn't hang the machine, but I haven't been able to test that, either. I notice that the RV370 5B62 card wasn't supported until fairly recently: https://bugs.freedesktop.org/show_bug.cgi?id=3602 Based on the available data, I strongly suspect that there's something special about this particular Radeon card that the xorg radeon driver doesn't [yet] know how to handle. If you need additional information (other than what's available in this report and in the reports in Red Hat's Bugzilla), please let me know.
I'm raising the severity to critical, as this bug causes a hang/crash. I would be happy to help debug this problem in any way I can.
If you comment out the Load "dri" line in xorg.conf and reboot and restart X, is the problem still present? If that works around the problem, uncomment the line and move the radeon.ko kernel module into /root, and reboot and test again. Does the problem still occur? If disabling the dri X server module prevents the problem and/or removing the radeon kernel module, then this is probably a combination kernel DRM bug and X bug. In X11R7, various random R300+ hardware totally hangs if the X server's DRI module is loaded, even if DRI is not supported on the chip. The X server should not load the DRM in the first place if DRI is unsupported on a given chip, but it seems to do so anyway. The kernel DRM however should not hang the chip either. In FC5 development we have just disabled R300+ kernel DRM support to work around similar problems on X300, X800, and other R300 based hardware. Please update the report with the results of your testing.
Unfortunately, the problem persists regardless of whether the Load "dri" line in xorg.conf is commented out. Just for completeness, I commented out the Load "dri" line, commented out the Section "DRI" section, moved the radeon.ko kernel module into /root, rebooted, and then fired up X. The machine wedged in exactly the same way. My next testing step was to load this machine with FC5test3, but at this point, I'm just going to wait until FC5 final is released. If you have any other suggestions to try in the meantime, I'm all ears, but otherwise, I'll report back once FC5 is out and I can test it.
In comment #3, that should have read, "my next testing step was going to be to load..."
What if you try the radeon driver from CVS ?
I tried to backport the ati driver tree from CVS into 6.8.2 (what Fedora Core 4 has), but failed. :( I plan to wait until FC5 is released and try X11R7. If that doesn't work, then at least building the ati driver from CVS should be a straightforward process. Either way, I'll report back.
Hello everybody. I have a r350 and I am having the same problems... I tried different patches / versions of the radeon driver, but unfortunately the old ones didn't load... The patched one (a variation from #1912, Dynamic Clock disabled) didn't solve the problem and the newest one I couldn't backport so it compiles correctly. I tried deleting libdrm, deleting libgl-dri... So that doesn't get accidently loaded by the X server, but that didn't change anything. I now switched to the fglrx driver, because my machine hardlooked in 3 days approx. 50 times... This killed my file systems twice, so it's a very severe problem. I would gladly test out stuff for you guys, because running fglrx just for widescreen support (the only thing I need, and 2d) is way overkill. :-) TIA
James: Did FC5 resolve this issue for you?
xf86-video-ati 6.5.8.0 fixed the issue for me... I don't know if I can set it like this.
I apologize for the delay, but: YES, the FC5 radeon driver fixed this problem. I've been using it for several months now, and have not experienced a single lock-up.
Changing resolution to FIXED.
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.