Subject: Re: Problems installing NetBSD-acorn32 v2.0
To: None <port-acorn32@NetBSD.ORG>
From: None <netbsd@wynded.co.uk>
List: port-acorn32
Date: 02/07/2005 21:55:38
gavan@coolfactor.org wrote:

[re setting WimpSlot to 8000k]
> It shouldn't make a difference, but it may be worth trying to increase the 
> wimpslot as far as it will go.
I tried it with 20000k and it made no difference at all.

> Have you tried other (non-INSTALL) images?
I have now - the generic kernel that's on the FTP site. I get exactly the same behaviour with that.

> Also, this may seem like a silly question, but what's your memory 
> configuration? Is it a single 64MB SIMM, or two 32MB SIMMS? If the former, 
> is the SIMM in slot 0?
Two 32MB SIMMs. I think they're of a similar age; possibly even the same
manufacturer. ISTR buying them as a pair.

> Hmm. 64MB. That means no VRAM. Have you tried a different screen mode?
Oops. I forgot to mention the 1MB VRAM that the machine has.
It's running in 1024x768x256, but I've also tried it at 800x600x256 with the same results.
[Sometimes the bootloader appears to be writing over the screen memory, and this seems to occur more often at 800x600]

[Re the two installation kernels being different sizes]
> How different are the sizes? Does a comparison of the files reveal 
> anything interesting? (ie how far into the files do they differ, etc)
*diffbin kernel.netbsd-INST misc.BtNetBSD.InstKern
(i.e. diffing the decompressed kernels)
returns:
  At offsets      32 and      32
  Skipping        49 and      81 bytes

  Unable to synchronise - skipping remainder of files.
  At offsets 3774971 and 3775003
  Skipping       649 and  148828 bytes

i.e. the files are mostly identical, but there's a block (of varying size) that's different just after the file header, and the tail differs.

> That's an oversight. Could you please file a PR for the missing checksum.
Will do - once I get my machine working again. :)

However, significant progress today. I've switched to using the old bootloader
[BtNetBSD rather than boot32]. This boots both installation kernels and now appears to be happily initializing the disk.

So it seems that boot32 doesn't like something about my machine.
From a pragmatic point of view, I'm happy to live with that knowledge. However,
if I can give any input that others avoid a rather unpleasant 'out of the box'
experience, I'm happy to do so.

Thanks for your suggestions.

Neil