On Tue, 24 Oct 2000, Mike Pumford wrote:
> > 	There was quite a lot of interest in NetBSD/arm32, including quite
> > 	a few people who had used earlier versions (1.2, 1.3), also tried
> > 	ArmLinux, and were interested in trying out the latest version.
>
> Hopefully they will find things a little nicer. IMO NetBSD 1.5 is a
> massive improvement over 1.2, 1.3 and 1.4.

Certainly seems nicer than when I last did anything with NetBSD (about 2 yrs 
back, so that's about 1.2-1.3)

> > 	    - Patch to avoid RiscOS 4 NVRAM corruption (PR:8759)
> > 		Dave Daniels submitted a patch to enable NetBSD to
> > 		update the NVRAM checksum (required for RiscOS 4).
> > 		If anyone out is running RiscOS 3.{5,6,7} can they please
> > 		test this patch to confirm it does not affect them?
> > 		If we get rapid enough response we should be able to get
> > 		this into the 1.5 release.
>
> I can test this at the weekend. Is there anything specific that needs
> to be checked?

if you type reboot, it doesn't trash your cmos settings (that's basically it, 
RO4 actually checks the checksum on a reboot, one thought I've just had is 
how safe it is with the clock, (the clock updates the time in the cmos at 
some point)  Anyway this is a contact risc-os ltd I think (didn't have time 
at the show to quiz them.)

> > 	    - Packages
> > 		- Chris built around 70 binary packages for 1.5, and is
> > 		  continuing to build more. The initial set is now up
> > 		  at: ftp://ftp.netbsd.org/pub/NetBSD/packages/1.5/arm32
> > 	        - We hope to get chimera, kfm, and some other graphical
> > 		  browsers uploaded in a while.
>
> I have a 1.5ALPHA2 RiscPC which could be left building a chunk of
> packages if someone drops me a list. I keep prodding at mozilla from
> time to time. I have modified the ARMLinux assembler stubs required for
> xptcall but I have never been totally satisfied that I have got them
> right. If someone wants to have a shot I'm happy to pass them on.

Yep, mozilla didn't compile for me either.

> > 	    - !BtRiscBSD
> > 		- Can people confirm if the RiscOS4 fixed version of !BtRiscBSD
> > 		  also works under 3.5, 3.6, and 3.7?
>
> Again I can try this at the weekend.
>
> > 		- We need to find out if we can distribute memfix inside
> > 		  !BtRiscBSD for RiscOS 4 users.
> > 		- !BtRiscBSD defaults to 1024x768 - users with monitors unable
> > 		  to display that resolution need to exit a config file to
> > 		  run - that should be fixed.
> > 		- The number of supported resolutions should be expanded.
> > 		- Need to determine if !BtRiscBSD can boot gzipped kernels.
> > 		- Is !BtRiscBSD 'max proc' configure option still needed?
>
> IIRC it is still picked up by the kernel. I suspect it could be
> removed. The default for !BtRiscBSD should be 640x480@60Hz this will be
> safe for almost any RiscPC/A7000 class machine which were shipped with
> SVGA monitors. The mode support is actually compiled into the kernel at
> boot time.

It crossed my mind that we should increase the number of modes available at 
boot time.

> > 	    - hardware/software support
> > 		- We need people to test their ide, scsi, and other cards
> > 		  under 1.5 and report back what works and does not.
>
> I have already submitted a PR (with a patch) to make the Acorn SCSI
> card work in 1.5. IIRC the patch has not been applied in current or on
> the 1.5 branch yet.
>
> The machine also has a Simtec IDE interface and i-cubed Etherlan 600
> Network card which both work just fine.

Great :)

> >		- We should approach more hardware vendors for samples and
> >	 	  documentation.
>
> Indeed. It would be nice to get in touch with the people producing
> newer RISC OS machines so that NetBSD can be ported to them as and when
> they are released.

There are some interesting machines that seem to be popping up, companies 
approached:
Riscstation: Willing to lend a riscstation to a developer to get it working 
(I believe it's very much like A7000 series hardware (but has fp on board:)
Castle: Willing to give info out on cards (including kinetic).  however 
they've requested that any really interesting parts are kept obscure in the 
code.
Cumana:  Said they've given tech docs on scsi card to scott stevens about 3 
yrs ago, but willing to do so again.

One company I didn't get time to speak to was millepede (or is it 
microdigital)  anyway they're the ones doing the omega, but they also mention 
a dual processor (including xscale chip) pci based system.  Could be worth 
approaching them via email.  Anyway I've a bunch of email addrs to mail and 
talk to people.

Chris