On Fri 16 Apr, Gilbert, Christopher wrote:
> > Should a single 64MB SIMM work in a RiscPC with a 1.3G kernel?
>
> I vaguely remember something about this from a long long time ago, I think
> that at 64MB or more you had to add a boot option of nbuf=500 or something
> like that, this might not be necessary anymore, I can't remember.

The code to read the value of nbuf= is commented out in machdep.c, so I
suppose it was felt to be superfluous.

> Other possibility is to try putting the simm in the other slot and seeing
> if it works then. 
                                                                           
I believe the people concerned have tried that, both with and without
the original 16MB SIMM.

>Does it work in RISC-OS ?

It reports the right amount of RAM, but it hasn't been extensively checked.

Has any of this pmap-type bag of worms been addressed for 1.4? IMHO, the
apparent limit on RAM (128MB stops a 1.3F CATS booting) and the very real
limit on processes (see my postings last year; RiscPCs and CATS proved to
frequently stiff as you approached max_processes; max_processes itself
is limited to 255) severely limit the use of NetBSD/arm32 besides as a
research interest.

-- 
Dr. Stephen Borrill