First I must point out that everything works alright when I don't use
DRI (more specifically, when I enable xinerama to disable DRI as a side
effect). So the card's fine and X is fine and my installation isn't
completely broken and my configuration must be right.
But with DRI I wind up with a non-usable X. The first couple X-related
commands (xsetroot, etc.) in my .xinitrc execute and their results are
visible (the root window no longer drives me batty), but it doesn't take
long for the display to freeze (I don't see wmaker start). Along the
top of the screen I have a bar of garbage several hundred pixels thick.
But stranger things are in evidence... X is taking 100% CPU. My mouse
continues to work, but my keyboard is sometimes stuck (no LEDs, no
Ctrl+Alt+...). And if I ssh in remotely and do the following:
andy@blender $ ssh chainsaw
andy@chainsaw$ DISPLAY=:0 xkbbell
the X server beeps, so it's not totally frozen. I can't kill X, not
even with SIGKILL, but I can take out xinit and .xinitrc. If I try to
start a second X server while the broken one is running (which means any
time between invoking the initial startx and me rebooting the machine),
it blocks until I give up and kill it.
andy@chainsaw$ /usr/sbin/lspci
00:09.0 VGA compatible controller:
ATI Technologies Inc Radeon RV200 QW [Radeon 7500]
andy@chainsaw$ X -version
XFree86 Version 4.4.0
andy@chainsaw$ uname -a
Linux chainsaw 2.4.22 #4 Wed Sep 24 00:25:40 CDT 2003 i686 unknown
andy@chainsaw$ cat /etc/X11/XF86Config | dwim --excerpt
Section "ServerLayout"
Identifier "3d"
Screen 0 "Alpha" 0 0
InputDevice "Logitech" "CorePointer"
InputDevice "Sejin" "CoreKeyboard"
Option "xinerama" "off"
EndSection
Section "Module"
Load "dri"
Load "glx"
EndSection
Section "DRI"
Mode 0666
EndSection
Section "Device"
Identifier "Radeon.DVI"
Driver "radeon"
VendorName "ATI Technologies Inc"
BoardName "Radeon RV200 QW [Radeon 7500]"
BusID "PCI:0:9:0"
Screen 0
Option "MonitorLayout" "CRT, CRT"
EndSection
Of course I can provide more information, including the full XF86Config
and XFree86.0.log files. I can also text out patches and CVS versions
and newer kernel versions (2.4 or 2.6).
NoAccel doesn't help matters any.
Ah, yes, I also had trouble with XF86Config and XFree86.0.log getting
corrupted on some rare occasions. XF86Config sometimes was overwritten
with something like the following:
serial 0xffffffff 0xffffffff 0xffffffff 0xffffffff 0x00070300 0x00ffff00 0x00000000
bttv 0x0000109e 0x00000350 0xffffffff 0xffffffff 0x00000000 0x00000000 0x00000000
bttv 0x0000109e 0x00000351 0xffffffff 0xffffffff 0x00000000 0x00000000 0x00000000
bttv 0x0000109e 0x0000036e 0xffffffff 0xffffffff 0x00000000 0x00000000 0x00000000
bttv 0x0000109e 0x0000036f 0xffffffff 0xffffffff 0x00000000 0x00000000 0x00000000
sis900 0x00001039 0x00000900 0xffffffff 0xffffffff 0x00000000 0x00000000 0x00000000
And XFree86.0.log grew a whole bunch of NULs at the end and other random
bits elsewhere (I don't have an example handy, and I forget the details).
When I first had this problem (back in May when I built this XFree86
release), I thought it was some kind of weird conflict with the S3 card
I had in the same system. Now that I removed the card yet still have
the same problem I'm ready to ask for help.
Thanks in advance.
--
Andy Goth + unununium@xxxxxxxxxxxxx + http://ioioio.net/
_______________________________________________
XFree86 mailing list
XFree86@xxxxxxxxxxx
http://XFree86.Org/mailman/listinfo/xfree86
