I've been using xorg-x11 for a while on my PowerPC system with a Radeon 9100. GLX/DRI is much more stable than with xfree (that would crash after about an hour if I loaded dri) but a new problem started showing up suddenly. I did not experience this in the beginning, it started showing up about week ago and I have not recompiled xorg-x11. After running X11 for a while (duration varies) the screen goes black, sometimes it is enough to switch to a console and back and sometimes I have to reboot. running dmesg shows: "[drm] Loading R200 Microcode" The only warning I can find in the Xorg log that might be relevant is this: (WW) ****INVALID IO ALLOCATION**** b: 0x1000 e: 0x10ff correcting^G But I had that with xfree as well and xorg-x11 did work perfectly for the first couple of weeks I used it.
Try disabling DPMS in your conf file, if it's enabled.
Removing blocker, otherwise every xorg bug will start depending on it.
This did not help. :(
This a dup of bug #47922?
No, I don't think so. It's not just that the screen blanks, the entire computer hangs.
Is it a hard or soft hang? e.g., can you SSH in remotely and kill X / ping it / etc, or is the machine totally unresponsive to anything?
I'm not entirely sure, if I'm playing music at the same time it seems to get stuck in a loop, simply playing what's in the hardware buffer. The next time it happens I'm going to find another box and ping/ssh here. I'll also see if I can set up a serial debugger and check for any kernel output. Right now I'm trying out a newer kernel with some bugfixes for my USB-controller, I've heard that it could cause X11 to hang if you ran an unpatched kernel.
It seems to happen less frequent, but it did happen again today. This time I tried ssh:ing and ping:ing the box, and it seems totally unresponsive. Do you have any recommendations on how I could debug this? I have the ability to set up a serial debug console if that helps.
There's a version of gdb around patched to support X. One location is ftp://people.redhat.com/mharris/gdb-xfree86/. You'll need to run it remotely from another box.
I think I have the same problem -- come back to my box to find it black and unresponsive. I compiled the magic sysrq key into the kernel adn it still responds to to an old Alt-SysRq S,U,B
To add... i'm on an athlon system with a RADEON 9000.
I have also observed this behavior on the same chipset. My machine is also an Athlon (AthlonXP 1600+ (1400MHz)), in this case with a VIA-based motherboard. When I can recover the system, as the screen blanks the monitor shows the 'out of range' error code, and switching resolutions (or to a text console and back) solves the problem. When it locks hard, the monitor actually switches off. I would be willing to attach a debugger as well if necessary.
I've had that bug on a FireGL 8800 (R200) and after some fiddling I found that updating the 2D driver from CVS corrected the bug. For x86 people, go to http://www.freedesktop.org/~dri/snapshots/ and install a snapshot ( 20041021 worked for me, i needed both common* and r200*) . For PPC people, I believe you need to use CVS. Be sure to update (the included install.sh script does that for you) 1/ the kernel driver, radeon.ko under 2.6 2/ the 2D X.org driver, radeon.o + ati_drv.o 3/ the 3D DRI driver, r200_dri.so
detailed instructions on how to build X.org and DRI are at : http://dri.sourceforge.net/cgi-bin/moin.cgi/Building?action=highlight&value=ToDo sorry for not having said that the 1st time.
See this on Dell PE2850 with ATI Radeon 7000/VE. An "/etc/init.d/xdm start" will start the "greeter", but as soon as I log in, the box actually locks up solid, with the Dell LCD screen showing Error E07F0. There is no output to X logs. I have tried all the BS on the web - setting 'Option "PciOsConfig" "1"' in the ServerFlags, as well as both 'Option "VBERestore"' and 'Option "NoINT10"' I have also tried with and without the radeon/drm kernel modules. It doesn't help. RH bugzilla indicates that this is tied to some "feature" in xorg-x11 that transpired between 6.8.1.1 and 6.8.2.x. However, a currently sync'd Gentoo system doesn't have access to anything prior to 6.8.2-r4 . . . and attempted to patch in a 6.0.8-r2 ebuild only to have it fail on a patch for radeon monitoring. This bug is severe . . . as it actually causes hardware failure - please change severity to critical.
Could you try with 6.8.99.15-r4 please? If the problem still exists, let's move this bug upstream.
I am out of town; I apologize for the delay, but should be able to get that tried by next Tuesday. Just to let you know I did receive the request, and will try.
So, did you manage to get it working correctly?
Please reopen if this remains an issue with modular X.