Re: Booting on RiscPC

[Stephen Borrill <netbsd%precedence.co.uk@localhost>]

On Tue, 26 Mar 2019, Mike Pumford wrote:
> On 25/03/2019 21:22, Mike Pumford wrote:
> > On 25/03/2019 19:09, Mike Pumford wrote:
> > > I'm in the process of setting up a more up to date current tree for myself 
> > > where I'll apply these patches and see what happens.
> > Well I got a kernel with patches booting and got the same division by zero 
> > panic you did. However I had a working keyboard so was able to get a 
> > backtrace.
> >
> > Division by zero is in statclock (called from hardclock().
> >
> > There also seems to be something really bad in the podule detection code as 
> > it didn't pick up any of the podules properly. :(
> >
> > I'll keep on diggind and see if I can work it out.
> Well the good new is that I've got a booting kernel.
>
> You need an updated version of the timerfix.patch (attached). Division by 
> zero was in setstatclockrate() and has been there forever.

I must admit to potentially wasting some of your time with my most recent 
reports. It turns out I was booting with -d (drop to ddb) which I'd added 
during a previous round of debugging, so I was notseeing a panic at this point.
However, I was originally seeing the panic you've fixed I believe.

I've built a -current kernel with timerfix.patch from this mail and 
irqfix.patch from 
http://mail-index.netbsd.org/port-acorn32/2019/03/25/msg000079.html

I can confirm this boots fine, but the keyboard is still messed up. For 
example:
z = [[
x = gg
c = ff
s = ]]
b = mm
1 = uu
n = nn

For each keypress, I get two incorrect characters.

--
Stephen