From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+willy=40w.ods.org-S932194AbWAPDWv@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S932194AbWAPDWv (ORCPT <rfc822;willy@w.ods.org>);
	Sun, 15 Jan 2006 22:22:51 -0500
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932195AbWAPDWv
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Sun, 15 Jan 2006 22:22:51 -0500
Received: from mx1.rowland.org ([192.131.102.7]:54546 "HELO mx1.rowland.org")
	by vger.kernel.org with SMTP id S932192AbWAPDWu (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Sun, 15 Jan 2006 22:22:50 -0500
Date: Sun, 15 Jan 2006 22:22:49 -0500 (EST)
From: Alan Stern <stern@rowland.harvard.edu>
X-X-Sender: stern@netrider.rowland.org
To: Reuben Farrelly <reuben-lkml@reub.net>
cc: Andrew Morton <akpm@osdl.org>, <linux-kernel@vger.kernel.org>,
       <jgarzik@pobox.com>, Greg KH <greg@kroah.com>,
       <linux-usb-devel@lists.sourceforge.net>,
       Neil Brown <neilb@cse.unsw.edu.au>, <linux-acpi@vger.kernel.org>
Subject: Re: [linux-usb-devel] Re: 2.6.15-mm3 [USB lost interrupt bug]
In-Reply-To: <43CAD1BB.60301@reub.net>
Message-ID: <Pine.LNX.4.44L0.0601152212340.1929-100000@netrider.rowland.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

On Mon, 16 Jan 2006, Reuben Farrelly wrote:

> On 13/01/2006 4:53 a.m., Alan Stern wrote:
> > On Thu, 12 Jan 2006, Reuben Farrelly wrote:
> > 
> >>>> Initializing USB Mass Storage driver...
> >>>> irq 193: nobody cared (try booting with the "irqpoll" option)
> > 
> >>>> handlers:
> >>>> [<c027017e>] (usb_hcd_irq+0x0/0x56)
> >>>> Disabling IRQ #193
> >>> USB lost its interrupt.  Could be USB, more likely ACPI.
> >> I've seen this one happen nearly every boot since then including bootups that 
> >> are otherwise OK (no oopses), so it's probably worth more looking into rather 
> >> than being written off as a 'once off':
> >>
> >> uhci_hcd 0000:00:1d.3: Unlink after no-IRQ?  Controller is probably using the 
> >> wrong IRQ.

Note the PCI ID is 1d.3 and the IRQ is 193.

> Hi Alan,
> 
> If it's any use, here's some simply and easy-to-get information which may even 
> be what you are looking for:
> 
> [root@tornado dovecot]# uname -a
> Linux tornado.reub.net 2.6.15-mm1 #1 SMP Sun Jan 8 03:42:25 NZDT 2006 i686 i686 
> i386 GNU/Linux
> [root@tornado ~]# cat /proc/interrupts
>             CPU0       CPU1
>    0:   21638510          0    IO-APIC-edge  timer
>    4:        356          0    IO-APIC-edge  serial
>    8:          1          0    IO-APIC-edge  rtc
>    9:          0          0   IO-APIC-level  acpi
>   14:          1          0    IO-APIC-edge  ide0
>   50:          3          0   IO-APIC-level  ehci_hcd:usb1, uhci_hcd:usb2
> 169:        120          0   IO-APIC-level  uhci_hcd:usb5
> 177:    2837992          0   IO-APIC-level  sky2
> 185:      61450          0   IO-APIC-level  uhci_hcd:usb4, serial
> 193:    4722447          0   IO-APIC-level  libata, uhci_hcd:usb3

Note that in the earlier kernel, IRQ 193 is assigned to usb3.  That's the 
second UHCI controller, since usb1 is EHCI.

> [root@tornado ~]# lspci

> 00:1d.0 USB Controller: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) 
> USB UHCI #1 (rev 03)
> 00:1d.1 USB Controller: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) 
> USB UHCI #2 (rev 03)
> 00:1d.2 USB Controller: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) 
> USB UHCI #3 (rev 03)
> 00:1d.3 USB Controller: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) 
> USB UHCI #4 (rev 03)
> 00:1d.7 USB Controller: Intel Corporation 82801FB/FBM/FR/FW/FRW (ICH6 Family) 
> USB2 EHCI Controller (rev 03)

Note that 1d.3 is the fourth UHCI controller; the second is 1d.1.

> I guess this looks like it was assigned the same IRQ ?

I don't think so.  To be certain you'd have to check the boot-up log and
verify that 1d.1 is usb3 and 1d.3 is usb5.

>>From the information presented here, it looks like -mm1 correctly routes
the 1d.1 controller to IRQ 193 and the 1d.3 controller to IRQ 169, whereas
-mm3 incorrectly routes the 1d.3 controller to IRQ 193.  That would make 
it an ACPI problem.

Alan Stern