Before I file more bug reports, I thought I'd see whether other
people are having the same issues I'm having.
The system in question is a PowerComputing PowerCenter 132
(Macintosh clone with 604e processor running at 132 MHz), with 80 MB
of RAM. I'm using the built-in display adapter (Platinum, with 4 MB
of VRAM) with a Portrait Display Labs Pivot 1700 monitor.
Currently running kernel 2.2.12.
1. gdm locks up the keyboard when started at boot time.
X starts, gdm comes up, the mouse works, but you can't type.
Which also means you can't switch to a virtual terminal to stop
and restart gdm. (Luckily we have three Unix machines in our
house, so it's only annoying to have to telnet or ssh in and
fix the problem.)
2. Frame buffer console annoyances.
When my machine starts up, the text is red on a black screen.
We run fbset from a script in /etc/rc.boot to correct this
problem. The text turns to white on black (hooray), but the
cursor is invisible. Other virtual terminals display in
various funky colors. X starts (with gdm) at the wrong
resolution.
#!/bin/sh
# /etc/rc.boot/0fbdev
#
# Sets resolution and refresh rate for framebuffer.
# 11 Jul 1999 by MEO
#
FBSET=/usr/sbin/fbset
# If the serial executable has been removed abort the configuration
[ -x ${FBSET} ] || exit 0
echo -n "Configuring frame buffer..."
# fbset --all 1024x768-75
${FBSET} -a -x -depth 16 1152x864-80
echo "done."
If I quit gdm (and the X server) and run the same fbset command
line from the /etc/rc.boot/0fbdev script by hand, then restart
gdm, I can sometimes get X to come up in the right resolution
(sometimes I need to run fbset by hand while X is running, then
quit and restart gdm). Virtual terminals are white text on
black background, but no matter what I do, the cursor remains
invisible.
Oh, and the resolution actually used seems to be 1152x870-75.
(Or so fbset reports when run with no arguments.) Note that X
misbehaves no matter what fbset reports until after the magical
manual fbsets.
3. fdutils don't work on PowerMacs?
I wanted to make some boot floppies. fdformat doesn't work --
it claims to be obsolete and recommends superformat.
superformat, however, doesn't appear to recognize Mac floppy
drives. Running 'superformat /dev/fd0 hd' results in the error
``get drive characteristics: Unknown error 515'' and upsets the
kernel enough that it decides that the floppy drive is ``BUSY''
from that point on, preventing you from doing anything else
with the floppy drive (such as ejecting the disk) until the
system is restarted.
4. ~/.Xmodmap file for pointer isn't read.
I have a .Xmodmap file in my home directory to swap the second
and third buttons (on my Kensington TurboMouse track ball,
which has four real buttons, and a chord (making five, all of
which are detected by Linux).
~/.Xmodmap:
pointer = 1 3 2 4 5
When X starts (from gdm) it reads this file, but doesn't seem
to find anything to work with. When I load the file manually
(using xmodmap ~/.Xmodmap), or run 'xmodmap -e "pointer = 1 3 2
4 5"', it works fine. Any idea why it isn't loaded properly
during the startup process? (And, yes, you have to specify all
five buttons.)
.gnome-errors:
[...]
/etc/gdm/Sessions//Gnome: Reading /etc/X11/Xmodmap...
! dump of work queue
keycode 0x3b = BackSpace
keycode 0x7d = Delete
/etc/gdm/Sessions//Gnome: checking /etc/X11/Xsession.options
for option allow-user-resources
Found allow-user-resources in /etc/X11/Xsession.options
We supposedly have allow-user-resources...
Now we test to see if there's a file called /home/claire/.Xresources...
/etc/gdm/Sessions//Gnome: There was! So we read
/home/claire/.Xresources...
/etc/gdm/Sessions//Gnome: checking /etc/X11/Xsession.options
for option allow-user-modmap
Found allow-user-modmap in /etc/X11/Xsession.options
/etc/gdm/Sessions//Gnome: Reading /home/claire/.Xmodmap...
! dump of work queue
Starting /usr/bin/gnome-session --purge-delay=2500 without ssh...
[...]
Also, does anyone have any ideas on how I could use the fourth
button (and maybe the chord) for anything useful?
5. Does gpm work on PowerMacs? If so, how should it be
configured?
I have gpm running on my system, as installed from the base
install (I assume) and with the arguments it came with
device=/dev/adbmouse
responsiveness=
type=ms
append="-l \"a-zA-Z0-9_.:~/\300-\326\330-\366\370-\377\""
It appears to have no effect in console applications that seem
like they might use its features, however.
6. Recent kernels don't work.
2.2.13 and 2.2.14 (any prerelease patch level -- all from
standard kernel sources (i.e., *not* vger kernel source -- and
where is that, these days, anyway?) don't produce functional
kernels. They cause a machine check in kernel mode, ``probably
due to mm fault with mmu off''.
Versions of packages mentioned in this message (all are the latest
binary-powerpc versions as of January 5, 2000):
Name Version Description
===============-==============-============================================
fbset 2.1-5 Framebuffer device maintenance program.
fdutils 5.3-3 Linux floppy utilities
gdm 2.0-0.beta4.5 GNOME Display Manager
gpm 1.17.8-11 General Purpose Mouse Interface
powerpc-utils 1.1.3-1 Various utilities for Linux/PowerPC
xbase-clients 3.3.5-1.1 miscellaneous X clients
xserver-common 3.3.5-1.1 files and utilities common to all X servers
xserver-fbdev 3.3.5-1.1 X server for framebuffer-based graphics driv
Any help would be greatly appreciated -- I would be glad to hear
I've just misread a man page or dropped a semicolon in a
configuration file. Like everyone else, I have things I should be
doing that I'm not getting done, at least in part because I'm
wrestling with some of these problems.
Thanks!
C.
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Behind the counter a boy with a shaven head stared vacantly into space,
a dozen spikes of microsoft protruding from the socket behind his ear.
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
C.M. Connelly [EMAIL PROTECTED] SHC, DS
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
