Subject: Re: status of xy/xd drivers
To: None <davided@sco.com>
From: Chuck Cranor <chuck@dworkin.wustl.edu>
List: port-sun3
Date: 02/20/1996 13:39:52
>what's the current status of the xy/xd drivers ?  are they:
>	- in a usable state,
>	- not finished, but someone's on it
>	- not finished, currently ignored
>or what ?  i haven't got any scsi for my sun3, and diskless seems awfully slow
>if there is useful work to be done in this respect, i'd love to help.

hi-

   both the xd and xy drivers are in a usable state under the sparc
port.   i believe charles has an "xd" running under the sun3.   the
xy under sun3 is untested as far as i know.    [the issues WRT to
sparc vs sun3 are dvma related].

   the "xd" driver is more stable than the "xy" driver.   [the "xd"
hardware is also faster than the "xy" hardware.]   the main problem 
with the "xy" driver is that every now and then it seems to loose an 
interrupt or something like that.   this causes disk I/O to stop until 
the software watchdog timer goes off.   at that point the driver 
recovers from the problem and everything is fine [i.e.  you don't 
loose any data].   also, i haven't had the chance to test the bad144 
code in either driver that much (in order for this code to get used you 
need a track with more than one error on it so that sector slipping 
doesn't fix all the problems).

   note that access to the sunos format program is very useful if
you want to use SMD disks since NetBSD does not have a format program.

   my future plans for these drivers involve [a] updating the sparc dvma
stuff in xd to fit with the new sun4m stuff, [b] sync'ing the sparc and
sun3 dvma APIs, and [c] merging the sun3 and sparc versions of the drivers
together into a generic sun version.

   if you are interested in contributing to this effort, things to
consider working on are [1] getting the sun3 versions of the drivers
fully working (charles or jason can help here?), [2] writing a format
program, or [3] trying to track down the lost interrupt problem in xy.c.
both Gordon and I have docs on these xylogics controllers, and the
source code has lots of comments.

chuck
-- 
Chuck Cranor, Graduate Student
Computer and Communications Research Center
Washington University, St. Louis MO  USA
E-Mail: chuck@maria.wustl.edu / cranor@udel.edu