NVIDIA driver (nvidia-kernel-source) :-(
Sridhar:
>This may not be relevant but i believe you also should have a
>
> Load "glx"
That is okay.
>> When I 'startx' I get a blank screen with a square cursor at the top left.
>
>I'm not sure what is happening at this point. When you get this screen,
>can you switch to another console terminal? If so, can you provide the
>output of "ps axf > /tmp/ps.log" at the time? It might help to so what
>is running (or not) at the time.
I can switch to a virtual console. Also, I've notice what I wrote about the
square cursor isn't 100% true. When X starts I get a blank screen (no blinking
square cursor). If I switch to a virtual console (e.g. F3-Ctrl-Alt) AND THEN
switch back (F7-Ctrl-Alt) I get the blank screen WITH the blinking square cursor.
I don't see any significance in that, but I thought I'd mention it 'cause someone
else might see something in that.
I did the ps & redirect.
$ ps axf > ~/ps.log # full log at end
A grep of that file for 'X11' ( $ grep X11 ps.log ) yields this result:
---
244 ? S 0:00 /usr/bin/X11/xfs -daemon
308 ? S 0:00 /usr/X11R6/bin/Xprt -ac -pn -nolisten tcp -audit 4 -fp /usr/X11R6/lib/X11/fonts/Type1,/usr/X11R6/lib/X11/fonts/100dpi,/usr/X11R6/lib/X11/fonts/75dpi,/usr/X11R6/lib/X11/fonts/misc :64
---
It is obvious things aren't running properly if one compares it to a 'ps'
when X is working the the open driver (nv):
---
$ ps axf |grep X11
758 tty4 S 0:00 \_ /bin/sh /usr/bin/X11/startx
769 tty4 S 0:00 \_ xinit /usr/X11R6/lib/X11/xinit/xinitrc -- /usr/X11R6/lib/X11/xinit/xserverrc
770 ? S< 0:10 \_ /usr/bin/X11/X -dpi 100 -nolisten tcp
---
I saw a few warnings in 'XFree86.0.log' file:
$ grep "(WW)" XFree86.0.log
---
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(WW) The directory "/usr/lib/X11/fonts/CID" does not exist.
(WW) The directory "/usr/lib/X11/fonts/cyrillic" does not exist.
(WW) Open APM failed (/dev/apm_bios) (No such file or directory)
---
I don't know if they have any significance to the problems I'm having.
I beginning to wonder whether the 'XF86Config-4' has problems. The way I
installed Debian "Sid": (1) installed "Woody" (2) installed X (configured X)
(3) upgraded to "Sid."
I installed Debian a couple of times... interesting is that if you install "Woody"
then upgrade and then install X the configure routinue (for X) that runs asks
you different questions. I remember this 'cause when I did it this way I was
asked a few more questions and I wasn't sure how to answer some of 'em. What
I ended-up doing is abandoning the second install (and doing it a third time
after wiping the partitions) 'cause I had no apparent X problems when installing
X in "Woody" before doing the upgrade to "Sid."
(from an earlier post)
>Do you have your kernel's header or source?
Kernel headers. I'm not sure where to get the source... currently don't care either
if things can be taken care of with the headers.
>To require nvidia-glx when you installed
>nvidia-kernel-source would be incorrect as nvidia-glx is not needed by
>the nvidia-kernel-source package. However, when the resultant
>nvidia-kernel binary package was created it should have required
>nvidia-glx.
There is something I don't quite follow, because I tried the install with
the 'nvidia-kernel-source' package and it did NOT appear to have installed
the nvidia driver (nvidia-glx) properly (in a previous post I explained how
'lsmod' showed the 'nvidia' module as "unused").
The way I fixed that is:
# apt-get install nvidia-glx
You're suggesting that is wrong?
(It seemed to be doing good things. The 'nvidia' module was no longer listed
as 'unused.')
Anyway, I went on that hunch, 'cause I can't think of any other ideas.
I removed nvidia-glx (# dpkg -r nvidia-glx) and changed the 'XF86Config-4'
(so it would look for the 'nvidia' driver). Then I dropped out of X
(with Ctrl-Alt-Bkspc) and *BOOM* my machine locks-up---no consoles... no
killing X... I had to three-finger-salute. It doesn't seem like that
is progress. I'm really close to fed-up.
This seems like a long shot:
Is it possible that because I didn't have 'nvidia-kernel-common' installed
when I built the dpkg (# debian/rules binary_modules) it screwed-up the
compile? Does the order of the install matter?
process list
==========
PID TTY STAT TIME COMMAND
6 ? SW 0:15 [kupdated]
5 ? SW 0:00 [bdflush]
4 ? SW 0:00 [kswapd]
3 ? SWN 0:01 [ksoftirqd_CPU0]
1 ? S 0:04 init [2]
2 ? SW 0:00 [keventd]
7 ? SW 0:00 [i2oevtd]
9 ? SW 0:00 [kreiserfsd]
52 ? SW 0:00 [khubd]
59 ? SW 0:00 [kapmd]
84 ? SW 0:00 [kjournald]
85 ? SW 0:00 [kjournald]
86 ? SW 0:00 [kjournald]
118 ? SW 0:00 [eth0]
128 ? S 0:00 dhclient -pf /var/run/dhclient.eth0.pid eth0
216 ? S 0:00 /sbin/syslogd
219 ? S 0:00 /sbin/klogd
228 ? S 0:00 /usr/sbin/inetd
235 ? S 0:00 /usr/sbin/sshd
244 ? S 0:00 /usr/bin/X11/xfs -daemon
303 ? S 0:00 /bin/bash /etc/rc2.d/S20xprint posix_sh_forced start
306 ? S 0:00 \_ tee -a /dev/null
307 ? S 0:00 \_ logger -p lpr.notice -t Xprt_64
308 ? S 0:00 /usr/X11R6/bin/Xprt -ac -pn -nolisten tcp -audit 4 -fp /usr/X11R6/lib/X11/fonts/Type1,/usr/X11R6/lib/X11/fonts/100dpi,/usr/X11R6/lib/X11/fonts/75dpi,/usr/X11R6/lib/X11/fonts/misc :64
326 ? S 0:00 /usr/sbin/atd
329 ? S 0:00 /usr/sbin/cron
358 tty1 S 0:00 /sbin/getty 38400 tty1
359 tty2 S 0:00 /sbin/getty 38400 tty2
360 tty3 S 0:00 /sbin/getty 38400 tty3
361 tty4 S 0:00 -bash
739 tty4 R 0:00 \_ ps axf
362 tty5 S 0:00 /sbin/getty 38400 tty5
364 tty6 S 0:00 /sbin/getty 38400 tty6
439 ? S 0:00 kdeinit: Running...
455 ? S 0:02 \_ /usr/bin/artsd -F 10 -S 4096 -s 60 -m artsmessage -l 3 -f
551 ? S 0:00 \_ kdeinit: kio_file file /tmp/ksocket-michael/klauncherKi8qJb.slave-socket /tmp/ksocket-michael/kateOXMTOa.slave-socket
442 ? S 0:00 kdeinit: dcopserver --nosid
445 ? S 0:00 kdeinit: klauncher
==========
Reply to: