From owner-freebsd-hackers  Wed Nov 26 04:43:06 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id EAA13650
          for hackers-outgoing; Wed, 26 Nov 1997 04:43:06 -0800 (PST)
          (envelope-from owner-freebsd-hackers)
Received: from mail.award.de (mail.award.de [195.30.16.205])
          by hub.freebsd.org (8.8.7/8.8.7) with SMTP id EAA13638
          for <hackers@freebsd.org>; Wed, 26 Nov 1997 04:43:01 -0800 (PST)
          (envelope-from timog@mail.award.de)
Received: (qmail 3616 invoked by uid 1015); 26 Nov 1997 12:36:02 -0000
Message-ID: <19971126133602.62657@mail.award.de>
Date: Wed, 26 Nov 1997 13:36:02 +0100
From: tg@award.de
To: sos@FreeBSD.dk
Cc: hackers@freebsd.org
Subject: Re: PIIX configuration
References: <19971126100027.53229@mail.award.de> <199711261031.LAA05158@sos.freebsd.dk>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Mailer: Mutt 0.88e
In-Reply-To: =?iso-8859-1?Q?=3C199711261031=2ELAA05158=40sos=2Efreebsd=2Edk=3E=3B_fro?=
 =?iso-8859-1?Q?m_S=F8ren_Schmidt_on_Wed=2C_Nov_26=2C_1997_at_11=3A31=3A1?=
 =?iso-8859-1?Q?3AM_+0100?=
Reply_To: <tg@award.de>
Sender: owner-freebsd-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

On Wed, Nov 26, 1997 at 11:31:13AM +0100, Søren Schmidt wrote:
> In reply to tg@award.de who wrote:
> > Hello all,
> > while looking through the source in /sys/pci yesterday, I noticed that the
> > detection code for the Intel 82371xx IDE chip (PIIX) was present in 2 places.
> > 
> > The detection code in pcisupport.c will find the chip first, claim it and
> > display that it found one, but will not do anything with the chip.
> > 
> > This effectively blocks the detection code in wd82371.c, which means that the
> > more sophisticated code there does not get executed at all, so no DMA setup.
> > Quite a pity if your IDE disk supports DMA.
> 
> You must be looking at a 2.2.x system.
> For more up to date (E)IDE  driver system look in curent, there is
> suppport for DMA etc modes.

Err... Sh*t. That's why I always ask MY customers which version they are using.
The system in question is a 2.2.5-Stable cvsupped 11/17. Sorry.

Is there any chance to take the code from current and mix it into 2,2,x without
too much work ?
 
> Most of the speed difference comes from the fact that linux has delayed
> write on as default, try mounting your disks with the async flag, and
> you will see semilar performance (and less security against corrupted
> disks on power outages etc).

I don't think that the major speed difference comes from the sync vs. async
mounting. The speed difference was measured using bonnie with files > 100M
and that's about twice the available RAM. Besides, I usually mount anything
sync if possible. Old electrical installation, you know :-).  

Regards,

Timo