Subject: Re: RiscPC & Xserver testing
To: None <port-arm32@netbsd.org>
From: Andy Wingate <xyzzy@ox.compsoc.net>
List: port-arm32
Date: 06/26/1999 13:59:59
In message <Pine.NEB.3.96.990603151652.11174A-100000@pain.devlab.nc.com>
          Mark Brinicombe <mark@causality.com> wrote:

> On Thu, 3 Jun 1999, Dave wrote:

[snip]

> > However, when trying to change to 16 bpp, 'setdisplay -x 800 -y
> > 600 -c 16 -m /etc/monitors.conf' , it complains about (from
> > memory) 'multiple open of /dev/vidcvideo0 by <process no>'. Any
> > ideas anyone?
> 
> Ok the multiple open is actually a warning that can be ignored. That
> after than you should find that when Xvidc launches it is using a
> 16bpp mode.

There I was, about to ask how to use setdisplay since looking at the
script it wasn't too obvious and I had tried certain arrangements of
the paramters each time getting errors, when I see this message
sitting in my mailing-list folder. Does /etc/monitors.conf need to be
of a certain format, or will it be already OK from an ordinary 1.4
install?

BTW, I did get 1.4 installed (on my RiscPC with Rev J StrongARM) using
a 19981001 install kernel initially (since it happily partitioned my
disk) then doing a MAKEDEV all afterwards to get a mouse driver and
other things that had gone missing for whatever reason through my
method of installation.

I am using the Xvidc-1 binary at present and it does seem fairly
reliable. I've had WindowMaker, Afterstep, fvwm1 and 2, and now a
freshly built Blackbox as window managers. The Mozilla 5.0a binary
seems happy to crash wonderfully, but that's probably to be expected
whereas Arena seems happy. I've had Netscapes 4.04 and 4.6 running
from other hosts and piped through ssh for a number of hours and also
Amaya. I'm looking forward to getting it all running in 16bpp since
then hopefully some of the colours will appear as you would expect
rather than some bizarre other shade occasionally. I only really
noticed colour problems when configuring background colours, and
otherwise it's been very good - I've only had one real hang.

I did seem to get (when using my etherh card) "ring buffer memory
overrun on device ne0"-type errors at various times - is the etherh
driver still a little unreliable?

Out of interest, what sort of value for maxproc is sensible? I did
have it set to 64, but at one stage kept running out of processes
while building three packages simultaneously and trying to get on with
something else in X.

Cheers,
Andy.
-- 
Andy Wingate <xyzzy@ox.compsoc.net>
3rd year Chemist, Oxford University
http://www.ox.compsoc.net/~xyzzy/
Pain is nothing that a downpour won't erase