On Fri, Sep 28, 2001 at 06:55:36PM -0500, Jay Maynard wrote:
> I've got a 5000/240 running 1.4 (yeah, but I haven't had to mess with it in
> two years) that's recently started complaining:
> 
> CPU memory read ECC error at 0x01d7806c
>    ECC 0xb48b9d00
> 
> I assume there's a bad SIMM in the box. Does the address it's
> complaining about (the same address every time, BTW) map to a specific
> slot? How can I figure out which it is, if so?

the same thing happened to my /240 not so long ago.  ;)

http://mail-index.netbsd.org/port-pmax/1999/08/15/0002.html

memory starts at zero, so if you have 8MB modules in there, 0x01d7806c
is about 3.7 * 8MB...  so I guess the fourth module is the one giving
you troubles.

I haven't had any memory errors in a long time -- did Simon's module
identifier routines ever make it back into a later 1.4 release?  (I'm
running 1.4.3A on my /240...)

-- 
  Aaron J. Grier | "Not your ordinary poofy goof." | agrier@poofygoof.com
  "If you are wondering why I stopped after ten chapters, it is because
           Chapter 11 was never considered for a book on Alpha!"
    --  Dileep P. Bhandarkar, _Alpha Implementations and Architecture_