Subject: Re: Lombard PBG3 + kernel 2.2.12
From: Hugh Caley (hcaley@loomer.com)
Date: Thu Sep 09 1999 - 23:23:38 MDT
Peter Foster wrote:
> Gives video acceleration + USB mouse + trackpad
>
> Heaps of help from and thanks to Hugh Caley, and his page at
> http://www.loomer.com/linuxppc/#lombard
> I believe this is a slightly different strain of 2.2.12 than the
> www.kernel.org one.
Thanks!
It is a build of Paul Mackerras' 2.2.12 stable tree, specifically for my
Lombard laptop, although it may work with other machines.
> Put Hugh's /etc/X11/XF86Config where it belongs, there. Adjust as he
> suggests, commenting out in the Section "Pointer"
> Protocol "IMPS/2"
> Device "/dev/usbmouse"
> and adding
> Protocol "MouseSystems"
> Device "/dev/gpmdata"
> Or maybe he's fixed the 'other' file, so you don't need to even do that.
>
There should be a fixed version out there now with the proper settings for
using two mice. gpm and/or X will fail to start if you don't have both mice
attached!
>
> In BootX, use kernel args
> adb_buttons=103,111 video=atyfb:vmode:16,cmode:32,mclk:100
> and of course the "No Video Driver" is unchecked.
>
>
> To use the mouse and the trackpad, do the gpm together.
> Note that YDCS1.1 starts gpm, so it needs to be killed first, or not
> started.
> I put these in my rc.local file. I assume it could be done more elegantly
> by not starting gpm in the first place, but I have not quite figured out
> how to control the booup process yet. I bet its really easy.
>
I'm not familiar with YD, but it is usually started in /etc/rc.d/rc3.d, I
believe. Simply remove the S**gpm link from this folder.
>
> # restart (-k) gpm with 2 mice, first is the trackpad, second is the usb
> mouse
> echo "killing and restarting gpm"
> gpm -k
> gpm -t bm -m /dev/mouse -M -t ps2 -m /dev/usbmouse
>
> I don't remember mucking with my /dev/mouse or /dev/usbmouse files. Or if
> I did, I put them back the way I found them.
>
I didn't have to create them.
>
> The two mice should both work outside of Xwindows.
> Start x with
> startx -- -bpp 16
>
> (When I shut down X, I get a lot of
> mach64CreateGC: unsupported depth: 16. Using cfbCreateGC instead.)
>
There are some X applications that will not work in 32-bit color using
XF68_FBDev (the XFree86 X server), notably Xfmail and Lyx. Starting in 16 bpp
allows these apps to be used. And using Xpmac will allow the use of these
apps at 32 bpp, if you need it.
>
> When I start it with -bpp 32, it gives me video artifacts, those diagonal
> lines that I vaguely remember people complaining about.
>
They are annoying but they don't seem to interfere with the machines'
processes.
>
> ppp does not work for me. The kernel tells me that it does not support
> ppp. Dang.
>
I am compiling PPP support into these kernels, but I haven't been able to get
it to work. I suspect that Paul, who is the PPP maintainer, uses his own
trees as testbeds for ppp code and they are a bit spotty yet. Sorry about
that.
>
> Odd stuff in the /usr/var/messages:
> Sep 9 22:02:11 clintonite modprobe: can't locate module lo:0
> Sep 9 22:02:11 clintonite modprobe: can't locate module lo:1
> Sep 9 22:02:11 clintonite modprobe: can't locate module lo:2
> Sep 9 22:02:11 clintonite modprobe: can't locate module lo:3
> Sep 9 22:02:11 clintonite modprobe: can't locate module lo:4
> Sep 9 22:02:11 clintonite modprobe: can't locate module lo:5
>
> and so on. Same stuff for eth0:n. But in the messages it says that both
> interfaces come up. And ifconfig confirms, I think. Odd.
>
These I don't know about.
>
> Also in the messages, it says, on bootup,
> rpc.nfsd: nfssvc: function not implemented
> and then tells me that Starting keytable [FAILED] on the bootup screen,
> but in the messages file it says "Starting keytable succeeded.", as
>
> keytable: Loading keymap:
> keytable: Loading /usr/lib/kbd/keymaps/mac/mac-us-std.kmap.gz
> keytable: Loading system font:
> rc: Starting keytable succeeded
>
I don't know much about these, other than I doubt that the rpc.nfsd: nfssvc
error is related to keytable ; )
>
> On shutdown, on the shutdown screen, it complains that NFS quotas or
> daemon (too fast to see) is failing to be shut down.
I'm glad this is helping people.
Hugh
-- Hugh Caley, Unix Administrator Babcock & Brown, San Francisco 510-524-1672 hughc@babcockbrown.com
This archive was generated by hypermail 2a24 : Fri Oct 01 1999 - 16:13:43 MDT