Subject: Fun with 1.6
To: None <port-pmax@netbsd.org>
From: Christopher Tribo <t1345@hopi.dtcc.edu>
List: port-pmax
Date: 11/16/2002 03:53:12
	Well I downloaded the 1.6 ISO from the NetBSD ftp site. Booted it 
up, a few seconds into PAX writing out the etc.tgz set (fresh install) to 
the hard disk, sysinst dumped core with a segmentation fault. So I 
finished the install off manually. Rebooted, and my machine is stuck at

>> boot 3/rz1/netbsd -a

	This isn't the first time this has happened, back with 1.4 I sent 
a PR that was in effect, every time I halted any of my PMAXen with a 
Quantum hard disk, I had to power cycle the disk after halting or it 
locked up the entire SCSI chain. cnfg 3 won't even return anything, it 
just hangs until I power cycle the disk. At the time I was using some old 
Quantum Pro Drive 250 and 500 MB Apple ROM'd hard disks. Now I'm using a 
DEC DSP3105 and I'm seeing the same thing. (The DSP3105 *is* a quantum 
drive) From the limited research I've done, I suspect that tagged quening 
is what is causing this behavior. Did the 4.4 BSD derived driver have 
tagged quening enabled? I know the dmesg from 1.4[.x] didn't mention 
anything about it. After I come up with a few more 1 GB seagate hawks I 
might try using one of those in place of the DSP 3105 and stripe a few 
together for enough space to download the source tree and build a kernel 
with tagged queneing disabled. Any thoughts?


	Chris