[comp.unix.microport] Tatu Ylonen's SCSI driver

ylo@sauna.HUT.FI (Tatu Yl|nen) (06/19/89)

> I would like to know whether this code is likely to work with
> other SCSI adapters rather than the Seagate ST-01 - i.e. the Adaptec SCSI
> adapter, etc.

I would consider it very unlikely that it would work with any other
adapter.  The driver needs intimate information on the i/o ports,
memory areas and dma channels used by the adapter.  These are likely
to differ greatly between different adapters.  I would consider it unlikely
that anyone would clone the ST-01.

What I hve heard, the Adaptec adapters use dma.  The driver does not
support dma, as the ST-01 did not (and I did not have any documentation
on how to use the dma channels in sysV/386).  I have also heard that
technical documentation of the Adaptec adapter would be hard to obtain;
I haven't tried though.  (Using a dma-driven controller would probably
improve performance by several hundred percent from that obtained
with the ST-01 (under msdos even it performs quite well, though).

If there are no other criteria for selecting the controller and you
can't get any other controllers to work, the ST-01 costs something
like $50 (I don't know the price in the US).

    Tatu Ylonen    ylo@hupu.hut.fi

hwh@cup.portal.com (Harold W Hankins) (06/20/89)

Wolf N. Paul (wnp@killer.dallas.tx.us) asks:

> I would like to know whether this code is likely to work with other
> SCSI adapters rather than the Seagate ST-01 -- i.e. the Adaptec SCSI
> adapter, etc.

The only other one I know of that it has a good chance of working on
is the one by Future Domain, from the pictures I've seen, it's almost
identical to the ST-01.  In fact, Future Domain did the MSDOS BIOS ROM
on the ST-01 (it has their copyright on it).

The ST-01 card is a small (what do we call them, quarter cards?) PC
card with only 3 chips, 5 small capacitors and 3 jumpers. 
The main chip (the scsi controller) is made by Texas Instruments.
(p/n 11746-501 CF61891FN N 34213).  It's completely memory mapped,
it uses no I/O ports.  All of its registers and its buffer ram are
mapped into memory in the space the second half of the rom should
be.

The strangest thing about it is it's price, around $32.00 (not a typo,
thats thirty-two dollars).

Hank Hankins		hwh@cup.portal.com
Point of Sales Systems, Inc
Camarillo, CA

buck@siswat.UUCP (A. Lester Buck) (06/20/89)

In article <YLO.89Jun19143718@sauna.HUT.FI>, ylo@sauna.HUT.FI (Tatu Yl|nen) writes:
> 
> > I would like to know whether this code is likely to work with
> > other SCSI adapters rather than the Seagate ST-01 - i.e. the Adaptec SCSI
> > adapter, etc.
> 
> I would consider it very unlikely that it would work with any other
> adapter.  The driver needs intimate information on the i/o ports,
> memory areas and dma channels used by the adapter.  These are likely
> to differ greatly between different adapters.  I would consider it unlikely
> that anyone would clone the ST-01.

Well, the ST-01 is a clone itself.  It is just Seagate's repackaging
of the Future Domain TMC-841, at a somewhat reduced price.

> What I hve heard, the Adaptec adapters use dma.  The driver does not
> support dma, as the ST-01 did not (and I did not have any documentation
> on how to use the dma channels in sysV/386).  I have also heard that
> technical documentation of the Adaptec adapter would be hard to obtain;
> I haven't tried though.  (Using a dma-driven controller would probably
> improve performance by several hundred percent from that obtained
> with the ST-01 (under msdos even it performs quite well, though).

I had no problem getting the technical manual for the 1540/1542 adapter
from Adaptec. It costs $5.00 but must be ordered separately (it doesn't
come with the adapter itself).

You don't quote transfer rates for your driver, but I wouldn't expect
an improvement of such magnitude.  When connected and transferring
asynchronously (Adaptec can also do synch transfers), both adapters
operate at about 1.5 MB/sec.  The ST-01 cannot be doing anything
else (i.e., other computing) but it can have other SCSI devices
disconnected.  The Adaptec could overlap computing.  Remember that
all the stock AT controllers do not use DMA either, so DMA is nice
but not essential.  (Also, the Adaptec can transfer host adapter->memory
at a much higher rate (up to 10 MB/sec) than across the SCSI bus.)

-- 
A. Lester Buck		...!texbell!moray!siswat!buck

wnp@killer.DALLAS.TX.US (Wolf Paul) (07/22/89)

Has anyone with some hardware understanding looked at the recently
posted SCSI driver for System V/386.

I would like to know whether this code is likely to work with other
SCSI adapters rather than the Seagate ST-01 -- i.e. the Adaptec SCSI
adapter, etc.

Thanks -- if I receive any significant responses I will summarize.

Wolf Paul
-- 
Wolf N. Paul * 3387 Sam Rayburn Run * Carrollton TX 75007 * (214) 306-9101
UUCP:   {texbell, killer, dalsqnt}!dcs!wnp
DOMAIN: wnp@killer.dallas.tx.us or wnp%dcs@texbell.swbt.com

laurie@ucscb.UCSC.EDU (60648000) (08/02/89)

In article <8400@killer.DALLAS.TX.US> wnp@killer.Dallas.TX.US (Wolf Paul) writes:
>
>Has anyone with some hardware understanding looked at the recently
>posted SCSI driver for System V/386.
>
>I would like to know whether this code is likely to work with other
>SCSI adapters rather than the Seagate ST-01 -- i.e. the Adaptec SCSI
>adapter, etc.

The hardware interface for the Adaptec controllers is as different as can be 
from the the Sleezegate. Microport does have an Adaptec 154x device driver
that didn't quite make it out to the masses though. It went thru beta and it
(according to the beta testers and uport internally) was very solid and an
excellent performer. I would suggest that you try to get ahold of Peter at
uport and squeeze a driver out of him. 

Ken Chapin

zeeff@b-tech.ann-arbor.mi.us (Jon Zeeff) (08/05/89)

>from the the Sleezegate. Microport does have an Adaptec 154x device driver
>that didn't quite make it out to the masses though. It went thru beta and it
>(according to the beta testers and uport internally) was very solid and an
>excellent performer. I would suggest that you try to get ahold of Peter at
>uport and squeeze a driver out of him. 

Please let us know if you get this driver.  I'll believe uport is alive if
you do.






-- 
Branch Technology            |  zeeff@b-tech.ann-arbor.mi.us
                             |  Ann Arbor, MI