Subject: Re: Problems with memory (was: Unsuccessful attempt...)
To: None <port-arm32@NetBSD.ORG>
From: Kjetil Bernhard Thomassen <thomassk@oslo.geco-prakla.slb.com>
List: port-arm32
Date: 08/31/1996 18:09:58
> From thomassk Sat Aug 31 02:06:40 1996
> Subject: Re: Unsuccessful attempt on 1.2 beta scratch-installation
> 
>> Date: Tue, 13 Aug 1996 18:05:13 +0200 (MESZ)
>> From: Markus Baeurle <emw4maba@rghx50.gp.fht-esslingen.de>
>> 
>> On Sun, 11 Aug 1996, Kjetil Bernhard Thomassen wrote:
>> 
>>> I have now finally managed to install RiscBSD to my SCSI-disk.
>>> 
>>> The problem I reported seems to be in the 1.2 kernels. None of
>>> them are able to copy the floppy to the RAM disk.
>>> 
>>> The solution was to use the old bootloader and the 4197 kernel.
>> 
>> I can't confirm this. I just tried it yesterday night with the new 
>> bootloader from the 1.2-beta directory and kernel bsd-4444 and the boot 
>> floppy was loaded without any problem. OK it was the 1.1 boot floppy, but 
>> I don't see why this should make any difference.
> 
> No, the 1.1 boot floppy makes no difference, but the RAM configuration
> does.
> 
> I tried the same installation procedure on my Friends Risc PC. It is
> almost identical to mine except that he has 8+8 MB DRAM.
> 
> I started off with 32+16 MB, and now I have 32+32 MB.
> 
> With 32+16 MB I was able to boot with bsd-4444 once I had the
> root file system ready.
> 
> Now that I have more memory, I am unable to boot at all with bsd-4444.
> The boot process proceeds to the point where it checks the clock
> and then panics with the following error message:
> kmem_suballoc: bad status return of 3.
> panic: kmem_suballoc
> Stopped at _Debugger+0x10: ldmdb r11, {r11, r13, r15}
> db>
> 
> I then type co<cr>
> 
> and is returned to the RiscBSD debug/monitor shell
> 
> I guess I have to try a later kernel even though I have an
> EtherH card.

I have now tried the latest kernel bsd-4554, and I still have the
same problem.

Now it dumps the memory when I do a 'co' on the db> prompt:
  dumping to dev 1801, offset 63238
  dump 0 1 2 ... 31 pmap_enter: Vf1000000 P14f41000 in pmap Vf013a720
     prot=00000001, wired=0
  32 33 34 ... 63 succeded

So, where has it dumped this?
Specifically, what is device 1801 and is the offset in blocks?

As said, I have no problems booting with bsd-4197, so this is a
problem with later kernels.

An interesting thing happened though when I now rebooted with this
kernel: It started to give me error messages like this:
savecore: can't find device 0/0

I hope that this problem can be solved so that I will be able to
use newer kernels.

I am quite happy to try any kernel handed over me in order to solve
this problem.

I believe that this will become more and more of a problem as people
are adding more memory to their Risc PC's.

Kjetil B.