From owner-freebsd-questions  Fri Mar 17 20:31:53 1995
Return-Path: questions-owner
Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id UAA10929 for questions-outgoing; Fri, 17 Mar 1995 20:31:53 -0800
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id UAA10921; Fri, 17 Mar 1995 20:31:38 -0800
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id OAA07994; Sat, 18 Mar 1995 14:30:39 +1000
Date: Sat, 18 Mar 1995 14:30:39 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199503180430.OAA07994@godzilla.zeta.org.au>
To: dleeds@eagle.ais.net, rgrimes@gndrsh.aac.dev.com
Subject: Re: ultrastor 34f
Cc: hackers@FreeBSD.org, questions@FreeBSD.org
Sender: questions-owner@FreeBSD.org
Precedence: bulk

>> I want to know if the ultrastor 34f controller (scsi-2) VL will work with 
>> freebsd, I heard it does, but want to know if it is stable and works
>> well before I buy it.

>The Ultrastor 34F does work, it is stable, it does have the nasty habit
>of causeing high interrupt latency as it seems to grab the bus for long
>periods of time.

``long'' is apparently about 250 usec.  For the BT445C the latency is
< 16 usec.  (I can't measure the overhead directly but I can measure
the maximum time that it takes for several instructions.  A U34F hogs
the bus in such a way that 10 inb() instructions take 250 usec longer,
while with a BT445C they only take 16 usec longer.)

There are some other problems:

1) media detection is not completely implemented in the driver.
2) there are some performance problems.  Command overhead is higher.
   The driver or the controller doesn't seem to handle scatter/gather
   well.  The speed of reading /dev/rsd0d drops by 40% when memory
   gets fragmented.

>Utrastor is also out of business, so if you are buying one get a real
>good deal on it.

>If any one wants one of these I will gladly sell mine for $150.00 as

$50 would be a good deal :-).

Bruce