Subject: Re: NetBSD and RISC OS 4.27
To: None <port-arm32@netbsd.org>
From: Dave Daniels <dave_daniels@argonet.co.uk>
List: port-arm32
Date: 02/04/2002 20:13:25
In article <E16XOCm-0006Ps-0Y@anchor-post-34.mail.demon.net>,
   Chris Gilbert <chris@paradox.demon.co.uk> wrote:
> Yep, that's a known issue.  Requires work on both ends IE the bootloader has 
> to pass in more sections and the kernel has to make use of them.

>
I have been aware of the four section limit since the Kinetic card
came out. I thought that this was one of the reasons why Reinoud
Zandijk wrote his new bootloader?

Regarding what I tested, I only tried !BtRiscBSD. !BtNetBSD is a
nicer version of the front end, but it still calls the old
bootloader proper. My concern was that something would break in
that.

> Sounds like one of your simms just has 1 bank on it (I think)  try
> swapping the simms over see if it then reports 4 sections.

I do not think that swapping the two SIMM strips will have any
effect. The one SIMM strip is at physical address 18000000. IIRC
the other one was at 10000000, that is, they occupied two
separate address ranges.

>> 2) Not so important this one. When you reboot RISC OS, RISC OS
>>    moans that the machine was not shut down cleanly.

> Hmm, wonder how they do that?  Must be some flag in a file
> somewhere that's set on shutdown.

Gavan Fantom spotted it. There is a flag set in the CMOS which RISC
OS checks when it boots now. I guess we could either zap the flag
when the bootloader runs or find out if there is a way to set it
legally. Should I open a PR for this?

I think I had better join the acorn32 mailing list for the rest of
this thread rather than keep posting in arm32.

Dave Daniels