Subject: Re: Conner CTD 8000 DDS-2 DAT Drive
To: Joseph R. Rach <nomad@udel.edu>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: port-sparc
Date: 07/02/1996 10:22:43
On Sun, 30 Jun 96 23:48:39 EDT 
 "Joseph R. Rach" <nomad@udel.edu> wrote:

 >     I'm a new user to netbsd/sparc! I'm currently using NetBSD 1.1 on a
 >     sparc1 clone. Having just gotten the system loaded and running,
 >     including compiling X11R6.1, I'm ready to load my personal files
 >     back on. The problem is I backuped up my files on dds-2 drive, and
 >     it doesn't seem to be supported (directly). The only real problem
 >     is using mt. Here is one error message I'm getting:
 > sd0(esp0:0:0): timed out (ecb 0xf8659c98 (flags 0x12, dleft 2000), state 3,
 >    phase 257, msgpriq 0, msgout 0) AGAIN
 > esp0: SCSI bus reset

Hmm, the esp driver has gone though some changes.  You might consider 
upgrading to the latest NetBSD/sparc 1.2_ALPHA (is there a BETA snapshot 
for the SPARC yet?) snapshot, available at:

	ftp://ftp.NetBSD.ORG/pub/NetBSD/arch/sparc/

There are snapshot and prerelease style sets there.

 >     Taking a quick look in /sys/scsi it appears that dds-dc is listed,
 >     but not dds-2. It looks like I can put a hook in for my drive in
 >     st.c(?). Has anyone gotten this drive to work?

Actually, the error message is coming from your SCSI disk ... looks like 
the tape might be in the middle of executing some command, a disconnect 
botched, or something, and the bus wedged.  The driver then resets the 
bus, and you are apparently able to continue on normally.

 >     Having never been in the position of having kernel source (100% src
 >     for the whole system actually) I'm going to have quite some fun!
 >     Thanks to the NetBSD community! I've been bitten by the BSD
 >     Devil...

Yah, it's pretty cool, huh?  :-)

 >     For others having a similiar problem, implementing mt -f /dev/nrst0
 >     fsf # is not hard. Just do multiple tar -tf /dev/nrst0. I'm sure
 >     there is a more creative solution, but I'll leave that exercise to
 >     someone more qualified.

Ok ... it defintely looks like a disconnect/reselect issue with your tape 
drive (note that the "tar" command issues reads which advance the tape, 
whereas the "mt fsf" command actually issues a SCSI command to skip file 
marks).  The bus is tied up while the command is in-progress, unless the 
device disconnects.  Note, not all devices _like_ to disconnect; your 
tape drive might not...

 >     O yeah, almost forgot. Is there a tmpfs or like in BSD? I used to
 >     use:swap       /tmp          tmp  rw 0 0
 >     in /etc/fstab with 4.1.4.

Yes, the MFS (memory file system) ... I use an /etc/fstab entry like this:

/dev/sd0b       /tmp            mfs     rw,-s=32768     0 0

...where sd0b is my primary swap partition.  You can also use the word 
"swap" if you like, a'la:

swap            /tmp            mfs     rw,-s=32768     0 0

Ciao.

 -- save the ancient forests - http://www.bayarea.net/~thorpej/forest/ -- 
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939