From mboxrd@z Thu Jan  1 00:00:00 1970
From: George Dunlap <george.dunlap@eu.citrix.com>
Subject: Re: [PATCH RESEND 05/12] xen: numa-sched: make space
 for per-vcpu node-affinity
Date: Wed, 6 Nov 2013 14:47:21 +0000
Message-ID: <527A5679.9000301@eu.citrix.com>
References: <20131105142844.30446.78671.stgit@Solace>	
	<20131105143500.30446.9976.stgit@Solace>	
	<5279143702000078000FFB15@nat28.tlf.novell.com>	
	<527908B2.5090208@eu.citrix.com> <52790A93.4020903@eu.citrix.com>
	<1383738090.9207.113.camel@Solace>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; Format="flowed"
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xen.org>
Received: from mail6.bemta14.messagelabs.com ([193.109.254.103])
	by lists.xen.org with esmtp (Exim 4.72)
	(envelope-from <George.Dunlap@citrix.com>) id 1Ve4OG-0002uz-Qo
	for xen-devel@lists.xenproject.org; Wed, 06 Nov 2013 14:47:29 +0000
In-Reply-To: <1383738090.9207.113.camel@Solace>
List-Unsubscribe: <http://lists.xen.org/cgi-bin/mailman/options/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xen.org>
List-Help: <mailto:xen-devel-request@lists.xen.org?subject=help>
List-Subscribe: <http://lists.xen.org/cgi-bin/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=subscribe>
Sender: xen-devel-bounces@lists.xen.org
Errors-To: xen-devel-bounces@lists.xen.org
To: Dario Faggioli <dario.faggioli@citrix.com>
Cc: Keir Fraser <keir@xen.org>, Ian Campbell <Ian.Campbell@citrix.com>, Li Yechen <lccycc123@gmail.com>, Andrew Cooper <Andrew.Cooper3@citrix.com>, Ian Jackson <Ian.Jackson@eu.citrix.com>, Jan Beulich <JBeulich@suse.com>, Matt Wilson <msw@amazon.com>, xen-devel <xen-devel@lists.xenproject.org>, Elena Ufimtseva <ufimtseva@gmail.com>
List-Id: xen-devel@lists.xenproject.org

On 06/11/13 11:41, Dario Faggioli wrote:
> There's of course the option of renaming XEN_DOMCTL_setvcpuaffinity to
> XEN_DOMCTL_setvcpuhardaffinity (and, while there, put some '_' in it!
> )... But is that really an option? And even if it is an option for the
> hypervisor, and perhaps, libxc, I'm quite sure it's not for libxl... Am
> I wrong?

I think vcpuaffinity is a domctl, which means it's a interface to libxc, 
and thus malleable.  So I *think* we could actually just add a parameter 
to vcpuaffinity to say "hard" or "soft".  (Correct me if I'm wrong, Jan.)

Where we need to consider backwards compatibility is in the libxl 
interface.  I'm not sure whether the best thing to do there.  I think we 
can't break the existing API, so we need to keep 
libxl_set_vcpuaffinity() available for old callers.  Options include:
* Just add libxl_set_vcpuaffinity_soft(), and leave the first one as it is
* Add libxl_set_vcpu_affinity_[something]() which takes a soft/hard 
flag, and deprecate the existing one
* Redefine libxl_set_vcpuaffinity() to include a soft/hard flag, bumping 
the API version and having a work-around for old callers.

IanJ / IanC, any opinions?

  -George