Subject: Re: [PATCH] Configure IDE probe delays
From: Alan Cox <alan@lxorguk.ukuu.org.uk>
To: Lee Revell <rlrevell@joe-job.com>
Cc: Mark Lord <lkml@rtr.ca>, Jeff Garzik <jgarzik@pobox.com>,
       bzolnier@milosz.na.pl, Greg Stark <gsstark@mit.edu>,
       Denis Vlasenko <vda@port.imtp.ilyichevsk.odessa.ua>,
       Todd Poynor <tpoynor@mvista.com>,
       Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
       tim.bird@am.sony.com, dsingleton@mvista.com
In-Reply-To: <1094067412.1970.48.camel@krustophenia.net>
References: <20040730191100.GA22201@slurryseal.ddns.mvista.com>
	 <200408272005.08407.bzolnier@elka.pw.edu.pl>
	 <1093630121.837.39.camel@krustophenia.net>
	 <200408272059.51779.bzolnier@elka.pw.edu.pl> <4135CC9E.5060905@rtr.ca>
	 <4135E017.1000901@pobox.com>  <4135EC84.6070407@rtr.ca>
	 <1094067412.1970.48.camel@krustophenia.net>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Message-Id: <1094064126.3100.1.camel@localhost.localdomain>
Mime-Version: 1.0
Date: Wed, 01 Sep 2004 19:42:09 +0100
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 780
Lines: 15

On Mer, 2004-09-01 at 20:36, Lee Revell wrote:
> The effect can be measured using a recent version of the voluntary
> preemption patches, and disabling hardirq preemption.  In this situation
> the IDE I/O completion is by far the longest non-preemptible code path,
> so can be easily profiled from the latency traces.

A lot of IDE controllers hold off the CPU for long times when you do I/O
cycles. The other factor is PIO which defaults to IRQ masking for safety
on old controllers. For PCI we should probably default the other way.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/