Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 29 Jan 1996 01:47:56 +0100 (MET)
From:      J Wunsch <j@uriah.heep.sax.de>
To:        freebsd-scsi@FreeBSD.ORG
Subject:   Re: scsi cleanup
Message-ID:  <199601290047.BAA05392@uriah.heep.sax.de>
In-Reply-To: <199601282202.OAA02211@ref.tfs.com> from "Julian Elischer" at Jan 28, 96 02:02:30 pm

next in thread | previous in thread | raw e-mail | index | archive | help
As Julian Elischer wrote:
> 
> > [1995/03/28] kern/275      qic-02 streamer won't work
> not really SCSI? 

Not SCSI.  But i've got a donation of two drives (Wangtek and
Archive), so i'll pick this up some time later.  This looks like an
``extended weekend project''. :)

> > [1995/06/18] misc/530      Failed install from SCSI tape
> install error I think

Nevertheless, our sysinstall should get an option to allow some basic
settings that mt(1) does.  I think of blocksize and perhaps density
codes.  Just for those oddball weird tape drives (or old firmware
revs) where the drive misidentifies itself.  Things are fine for a
running system (you've got an rc.local there), but with sysinstall,
you're sorta outta luck.

> > get the SCSI reprobe code working properly on all host adapters.
> 
> agreed. This requires that interrupts are fully set up
> at the time of probing.. I'm not convinced of this..

Nope.  I've once started to rewrite the code, but eventually gave up.
It's necessary to pass the information down through all the layers
whether the actual probe must be handled in polled mode (system
initialization, dumping) or in interrupt mode.  This can be done, but
is a lot of work for the current architecture.

> > Be sure you can work with the SCSI system even if only a host
> > adapter was found at boot time.

> I think this basically requires the existenc in a non-optional way
> of the 'super-scsi' device or some equivalent.

I think this would be fine.

> > Get the CD-ROM writer code working.

> joerg seems to have this under control.

I'm really interested in a discussion of the approach i've been
taking.  I hope it's extensible enough.

I don't really have a clue yet how to handle the case that a CD-R
drive includes all the CD-ROM functionality.

> > Fix it so you can use a 1542A.
> 
> anyone have one?

I've got a 1540A, but it works fine.  It looks like only very early
firmware rev's are affected.  (I know of at least one other 1542A that
works quite well, in this case with FreeBSD 2.1R, even at a customer
of us. :)

> I want to write more documentation about how it all goes together.
> The code needs to have more comments put in it.

Well, i'm more in favour of documenting the generic scsi functions in
the newly-created section 9 of the manual.  I don't think we need to
write man pages for every oddball function in a device-specific
driver, but it might be useful for the ``recyclable'' generic code.

After all: Julian, though one needs a few sheets of paper to have
hardcopies from some files, it has been possible at all to dig through
the code, even in its current state.  One needs hardcopies of the
SCSI-2 specs, too, but that's probably the least the must be expected
from a SCSI driver developer.  I'd explicitly thank you and Peter for
all your good work!  My efforts with the worm driver wouldn't have
been possible otherwise.

p.s.: I think we are almost the first who can claim full support for
a CD-R under multiuser conditions. :)  (I could even fire up xdm and
log into an X session while the burning was in progress.)

-- 
cheers, J"org

joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601290047.BAA05392>