Today I attempted to a variety of truely simple things.
Namely, copy in all of the source distribution.

I was prevented from happily completing this task by
numerous:

biodone alreay...

In several cases the DB ps and callout commands had
the following (abreviated since I didn't have a way
to record all of the text output).

ps output:

rm
sh         wait
sh         wait
pagedaemon third_sleep
init       wait
swapper    scheduler

callout output:

1   1   _roundrobin
2   1   _pffasttimeout
20  18  _pfslowtimeout
49  29  _schedcpu
40  0   _if_slowtimeout

registers:

d0 -- _acct_process

The rest of the registers looked like 'unintersting' stuff.
At the moment can't judge what is useful or not.
---------

I don't know if the fact that I'm using a Mac II, which is
'slow' is part of the reason I see this crash so fequently
or what.

An alternate idea is that I have a noisy SCSI cable and I'm
either getting a SCSI parity error and the driver isn't doing
the right thing, or is doing a retry, but doesn't do the
right thing on the biodone bit.

Alternatively, the slowness could affect how fast I can get
to an interupt, say the ADB interupt, or while doing SCSI
stuff the ADB stuff affects the situation.

However, in most of the failures I was not 'using' ADB, all
that was happening was that the screen was 'scrolling'.
I did see the failure when I had no screen output, so I'm
not at all convinced that it is something dealing with
ADB/Screen operations, unless it's like vertical blank level.


John Clark
clark@synergy.encinitas.ca.us