From owner-freebsd-hackers@FreeBSD.ORG  Wed Aug 31 22:53:11 2005
Return-Path: <owner-freebsd-hackers@FreeBSD.ORG>
X-Original-To: freebsd-hackers@freebsd.org
Delivered-To: freebsd-hackers@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id A184A16A41F;
	Wed, 31 Aug 2005 22:53:11 +0000 (GMT)
	(envelope-from hselasky@c2i.net)
Received: from swip.net (mailfe14.swipnet.se [212.247.155.161])
	by mx1.FreeBSD.org (Postfix) with ESMTP id E459143D45;
	Wed, 31 Aug 2005 22:53:10 +0000 (GMT)
	(envelope-from hselasky@c2i.net)
X-T2-Posting-ID: Y1QAsIk9O44SO+J/q9KNyQ==
Received: from mp-217-132-64.daxnet.no ([193.217.132.64] verified)
	by mailfe14.swip.net (CommuniGate Pro SMTP 4.3.4)
	with ESMTP id 13974270; Thu, 01 Sep 2005 00:53:09 +0200
From: Hans Petter Selasky <hselasky@c2i.net>
To: freebsd-hackers@freebsd.org
Date: Thu, 1 Sep 2005 00:54:01 +0200
User-Agent: KMail/1.7
References: <200508302009.aa99975@nowhere.iedowse.com>
	<200508312239.04897.hselasky@c2i.net> <43161F60.8010408@samsco.org>
In-Reply-To: <43161F60.8010408@samsco.org>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200509010054.02793.hselasky@c2i.net>
Cc: Scott Long <scottl@samsco.org>, Eugene Grosbein <eugen@kuzbass.ru>,
	Ian Dowse <iedowse@iedowse.com>, freebsd-usb@freebsd.org,
	"Eygene A. Ryabinkin" <freebsd@rea.mbslab.kiae.ru>
Subject: Re: Low umass performance with USB 2.0 ports
X-BeenThere: freebsd-hackers@freebsd.org
X-Mailman-Version: 2.1.5
Precedence: list
Reply-To: hselasky@c2i.net
List-Id: Technical Discussions relating to FreeBSD
	<freebsd-hackers.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-hackers>, 
	<mailto:freebsd-hackers-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-hackers>
List-Post: <mailto:freebsd-hackers@freebsd.org>
List-Help: <mailto:freebsd-hackers-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-hackers>,
	<mailto:freebsd-hackers-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Wed, 31 Aug 2005 22:53:11 -0000

On Wednesday 31 August 2005 23:21, Scott Long wrote:
> >>
> >>Actually, I just peeked inside the Linux EHCI code and it does a dummy
> >>read immediately after writing to the status register:
> >>
> >>         /* clear (just) interrupts */
> >>         writel (status, &ehci->regs->status);
> >>         readl (&ehci->regs->command);   /* unblock posted write */
> >>
> >>I wonder if that's the whole trick here.  Would someone be willing to
> >>try the attached patch instead of the one that Ian posted?
> >>
> >>Scott
> >
> > This is not documented in the EHCI chip specification.
>
> Flushing posted writes is something that all programmers of PCI devices
> should understand, so it usually isn't documented in device manuals.
>
> > There exists the
> > doorbell to ensure that the EHCI controller is finished with data
> > structures. Also I have noticed that the existing EHCI driver does not
> > always dequeue structures from the controller before accessing them.
>
> Can you point to an example here?

In the official USB system, when an endpoint is opened, a QH structure is  
allocated and inserted into the USB controller's schedule. After that the 
EHCI driver will just write new values to that structure, while it is still 
scheduled. In my opinion you should unqueue this structure and wait for 
doorbell, before touching it.

>
> > If Scott's patch doesn't work, could you have tried to install the
> > following (compiles on FreeBSD 5/6/7):
>
> Yeah, looks like my guess was wrong.
>
> Scott

--HPS