Subject: Re: Shell on 3100 hung in thread_block
To: Donn Cave <donn@u.washington.edu>
From: Per Fogelstrom <pefo@enea.se>
List: port-pmax
Date: 03/21/1995 19:57:14
> 
> Quoth rnillson,
> | ... So much for the silver lining, when I started
> | to do some more stuff while at the same time doing some tape I/O
> | (from another terminal) I experienced the same kind hang-up situation
> | as when I run on my 8MB DS.
> 
> NetBSD's non-disconnectable SCSI implementation could be a factor in this.

Correct me if im wrong but both the 'asc' and 'sii' driver does disconnect
if it can be negotiated with the device. At least from what i could judge
from the source code.

Tape devices had a flaw in the design (at least some of them and some time
ago) that made them malfunction together with hosts that could not do
'odd byte' disconnects. To solve this problem, disconnect was disabled.
Some newer tape devices can be configured to do disconnects at some even
byte boundary to avoid the problem.

Well more could have happened, haven't been struggling with these for a while.