From owner-freebsd-hackers  Mon Sep  9 02:46:14 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id CAA10905
          for hackers-outgoing; Mon, 9 Sep 1996 02:46:14 -0700 (PDT)
Received: from relay.philips.nl (ns.philips.nl [130.144.65.1])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id CAA10895
          for <freebsd-hackers@FreeBSD.org>; Mon, 9 Sep 1996 02:46:09 -0700 (PDT)
Received: (from smap@localhost) by relay.philips.nl (8.6.9/8.6.9-950414) id LAA06825; Mon, 9 Sep 1996 11:41:49 +0200
Received: from unknown(192.26.173.32) by ns.philips.nl via smap (V1.3+ESMTP) with ESMTP
	id sma006686; Mon Sep  9 11:40:47 1996
Received: from spooky.lss.cp.philips.com (spooky.lss.cp.philips.com [130.144.199.105]) by smtp.nl.cis.philips.com (8.6.10/8.6.10-0.9z-02May95) with ESMTP id LAA04431; Mon, 9 Sep 1996 11:43:37 +0200
Received: (from guido@localhost) 
	by spooky.lss.cp.philips.com (8.6.10/8.6.10-0.991c-08Nov95) id LAA12024; Mon, 9 Sep 1996 11:40:43 +0200
From: Guido van Rooij <Guido.vanRooij@nl.cis.philips.com>
Message-Id: <199609090940.LAA12024@spooky.lss.cp.philips.com>
Subject: Re: vx device broken in 21.1.5?
To: dfr@render.com (Doug Rabson)
Date: Mon, 9 Sep 1996 11:40:43 +0200 (MET DST)
Cc: Guido.vanRooij@nl.cis.philips.com, freebsd-hackers@FreeBSD.org
Reply-To: Guido.vanRooij@nl.cis.philips.com
In-Reply-To: <Pine.BSF.3.95.960906114003.4396F-100000@minnow.render.com> from Doug Rabson at "Sep 6, 96 11:45:31 am"
X-Mailer: ELM [version 2.4ME+ PL19 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

Doug Rabson wrote:
> On Fri, 6 Sep 1996, Doug Rabson wrote:
> 
> > On Fri, 6 Sep 1996, Guido van Rooij wrote:
> > 
> > > I am getting a very poor performance with a 3c590 card in FreeBSd 2.1.5.
> > > When Ftp-ing I get like 600K/sec put performance, but only 10K/sec 
> > > when getting data. Are there some overrun problems?
> > 
> > There are some problems but I don't know if they are overruns.  I have
> > been experimenting with some patches based on examining the OpenBSD driver
> > with some confusing but promising results.  I'll try to come up with a
> > patch today.
> 
> Try this patch.  I pulled in some code from OpenBSD which detects a
> particular overrun hang in this card.  This improves the throughput of the
> device enourmously.  Oddly, the overrun hang which is code detects has
> never happened for me.  Part of the test is to read the VX_W4_FIFO_DIAG
> register; if I reduce the patch to simply reading this register and
> discarding the result, I get a similar performance improvement.  Since I
> have no documentation for the card, I have no idea why this should be the
> case.

I tried the patch but it doesn't solve part of my problem. It makes things
worse.

Basically, now I am getting buffer overflows when *putting* out a 
file with ftp. I checked what happened with a ping -f -s 1000 somenearbyhost
and noticed that both the original 2.1.5 driver as well as the one with
your patch give `no buffer space available' after some time. But with your
patch I am getting them earlier.


-Guido