Subject: Re: unhappy decstation 5000/240
To: Matt Thomas <matt@3am-software.com>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: port-pmax
Date: 01/25/1998 14:53:24
Matt Thomas writes:

>At 12:21 AM 1/25/98 , Aaron J. Grier wrote:
>On Sat, 24 Jan 1998, CyberPeasant wrote:
>
>> Bad memory? 
>
>>>Must be... I got 1.2g booted on it with just the 32MB card in it. 
>>>Methinks one of the 8MB cards is dead.  (It's got 32+8+8.)  I also
>>>reseated the processor module, and that seems to have helped.

>You can't do that.  All memory cards must be the same type (except
>if you have 32mb cards, the last one can be a 8mb card -- this leaves
>a 24mb memory hole but since it's at the end it's ok).  

Uh, you can't do it on Ultrix or NetBSD (yet).  However, I added
noncontig memory support to the Stanford V-kernel.  The hardware
supports it just fine.  With MACHINE_NONCONTIG we might even support
it in NetBSD.

Hm, OK, maybe the PROM tests don't grok it properly; I've never
tried. But if memory serves, the PROM will display the config
correctly.