From owner-freebsd-hardware  Thu Jan 30 01:42:06 1997
Return-Path: <owner-hardware>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id BAA01402
          for hardware-outgoing; Thu, 30 Jan 1997 01:42:06 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id BAA01397
          for <freebsd-hardware@freebsd.org>; Thu, 30 Jan 1997 01:42:02 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.3/8.6.9) id UAA05534; Thu, 30 Jan 1997 20:40:20 +1100
Date: Thu, 30 Jan 1997 20:40:20 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199701300940.UAA05534@godzilla.zeta.org.au>
To: freebsd-hardware@freebsd.org, plm@xs4all.nl
Subject: Re: Fast 16550A compatible serial card??
Sender: owner-hardware@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>As cheap alternative to a Hayes ESP someone recommended me a serial
>card called T9000+. It contains a 16550A compatible, yet runs at
>max. speed of 460k8.
>
>Is this possible? And is it possible under FreeBSD?

16550A's are specified to work up to 1.5Mbps.  Supporting 460.8Kbps
properly would require a few changes (it might work modulo some
buffer overflows at 460.8Kbps when the software thinks the speed is
115.2Kbps).  Expect an ISA bus overhead of 15-20% per port at 46080
cps bidirectional saturated (corresponding to 4-5% at 11520 cps).
At 150000 cps, the ISA bus overhead would be 50-65% per port (too
large for normal operation) and the fifo trigger level would certainly
have to be reduced to avoid silo overflows, and you would need a fast
cpu faster than a 486 just to complete the i/o processing in the
remaining 50-35% of the time.

Bruce