Re: need a bit of early boot help

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

>> ...if it had been 45 years ago you could have bug-reported the
>> diagnostics to DEC for pointing out the wrong hardware :-)
> Hahaa, yes!  And it's not like it was that complex of a problem
> either.  Just got sent on a wild goose chase by bad diagnostics.  It
> was booting NetBSD that gave us the hint that we needed.

Reminds me of an incident back in the '80s.  I was working on a project
using KA630s for robotics: one under mtXinu 4.3+NFS running the upper
layer, with an auxiliary CPU running the lower layer (under a *very*
cut-down custom kernel).

But it used interprocessor doorbell interrupts heavily.  And a few
percent of them went missing.

My colleague (the principal researcher on the project) finally built a
standalone test program which, running overnight, quantified the
percentage that went missing.  He then reported it to DEC.

It was a hardware bug: an etch run crossed the whole board, and the
resulting capacitance, with the pullup resistance, meant it never
should have worked; it worked at all only because of DEC overdesign.
As long as there was still a DEC service department, you could get the
fix to KA630s - a hardware patch to reduce the pullup's resistance - if
you knew what ECO to ask for.

Simple enough...but it was a right pain to track down.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B