From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+w=401wt.eu-S1763244AbXERSDf@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1763244AbXERSDf (ORCPT <rfc822;w@1wt.eu>);
	Fri, 18 May 2007 14:03:35 -0400
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1761958AbXERSDA
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Fri, 18 May 2007 14:03:00 -0400
Received: from ebiederm.dsl.xmission.com ([166.70.28.69]:50202 "EHLO
	ebiederm.dsl.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1761580AbXERSC7 (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Fri, 18 May 2007 14:02:59 -0400
From: ebiederm@xmission.com (Eric W. Biederman)
To: "Yinghai Lu" <yhlu.kernel@gmail.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>, mingo@elte.hu, ak@suse.de,
       akpm@linux-foundation.org, linux-kernel@vger.kernel.org,
       "Zou, Nanhai" <nanhai.zou@intel.com>,
       "Mallick, Asit K" <asit.k.mallick@intel.com>,
       "Packard, Keith" <keith.packard@intel.com>
Subject: Re: [patch] x86_64, irq: check remote IRR bit before migrating level triggered irq
References: <20070517230324.GB8089@linux-os.sc.intel.com>
	<m1ejlf0y96.fsf@ebiederm.dsl.xmission.com>
	<20070518010119.GB12088@linux-os.sc.intel.com>
	<m1veeqjitm.fsf@ebiederm.dsl.xmission.com>
	<86802c440705181030m6d72edf6w87749275c46c7c93@mail.gmail.com>
Date: Fri, 18 May 2007 12:01:37 -0600
In-Reply-To: <86802c440705181030m6d72edf6w87749275c46c7c93@mail.gmail.com>
	(Yinghai Lu's message of "Fri, 18 May 2007 10:30:08 -0700")
Message-ID: <m1abw2105a.fsf@ebiederm.dsl.xmission.com>
User-Agent: Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux)
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

"Yinghai Lu" <yhlu.kernel@gmail.com> writes:

> Eric,
>
> ioapic_level irq is limited, So if we keep vector number not changed
> when imgration to other cpus.  It that could help.

We can solve the problem without doing that, and keeping the same
vector number during migration keeps x86 from scaling.  Personally
I would prefer to disallow irq migration.

> it will need modify a little with assign_irq_vector and
> irq_complete_move/smp_irq_move_cleanup_interrupt. because it assume
> vector must be changed.

Yes it does.

Eric