From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dario Faggioli Subject: Re: [PATCH RESEND 05/12] xen: numa-sched: make space for per-vcpu node-affinity Date: Wed, 6 Nov 2013 00:08:01 +0100 Message-ID: <1383692881.15457.59.camel@Abyss> 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> <52791B8702000078000FFBC4@nat28.tlf.novell.com> <5279114B.9080405@eu.citrix.com> <52792326.4050206@eu.citrix.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============5526405490166494502==" Return-path: Received: from mail6.bemta14.messagelabs.com ([193.109.254.103]) by lists.xen.org with esmtp (Exim 4.72) (envelope-from ) id 1VdpjD-00027S-W1 for xen-devel@lists.xenproject.org; Tue, 05 Nov 2013 23:08:08 +0000 In-Reply-To: <52792326.4050206@eu.citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: George Dunlap Cc: MarcusGranado , Justin Weaver , Jan Beulich , Li Yechen , Andrew Cooper , Juergen Gross , Ian Jackson , Matt Wilson , xen-devel , Daniel De Graaf , KeirFraser , Elena Ufimtseva , Ian Campbell List-Id: xen-devel@lists.xenproject.org --===============5526405490166494502== Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-GzRBIIePmevmbpToNjqY" --=-GzRBIIePmevmbpToNjqY Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On mar, 2013-11-05 at 16:56 +0000, George Dunlap wrote: > [Coming back after going through the whole series] >=20 > This is basically the main architectural question that needs to be=20 > sorted out with the series: Do we bake in that the "soft affinity" is=20 > specifically for NUMA-ness, or not? >=20 Damn... I was really hoping to be almost done with this series! Unfortunately, this point you're raising is a very good one and, even more unfortunately, I like the way in which you propose to solve it, so I've got to re-implement this thing, or --I'm sure of that-- I won't be able to get any sleep at night for months! :-P :-P :-P > The patch the way it is does make this connection, and that has several= =20 > implications: > * There is no more concept of a separate "domain numa affinity" (Patch= =20 > 06); the domain numa affinity is just a pre-calculated union of the vcpu= =20 > affinities. > * The interface to this "soft affinity" is a bitmask of numa nodes, not= =20 > a bitmask of cpus. >=20 > If we're OK with that direction, then I think this patch series looks=20 > pretty good. >=20 We're not... As said, now that you put it this way, it really sound so much wrong, that I have to fix it! :-) > Release-wise, I think as long as we're OK with libxl providing a=20 > "set_vcpu_numa_affinity", then we can always come back and change the=20 > implementation later if we want to maintain that distinction internally. >=20 That's true too. Let's see how far and how quick can I get to implementing what you propose and then decide. BTW, if I can ask (and sorry if it's my fault for having missed it), what's the updated release schedule? Thanks and Regards, Dario --=20 <> (Raistlin Majere) ----------------------------------------------------------------- Dario Faggioli, Ph.D, http://about.me/dario.faggioli Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK) --=-GzRBIIePmevmbpToNjqY Content-Type: application/pgp-signature; name="signature.asc" Content-Description: This is a digitally signed message part Content-Transfer-Encoding: 7bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.15 (GNU/Linux) iEYEABECAAYFAlJ5elEACgkQk4XaBE3IOsQ6ZACglCjJ7txMz4MfTQRcj4nb8B48 AtkAoKmuvfYn/SbOY3IKoQP5/EQRXZhJ =T5c+ -----END PGP SIGNATURE----- --=-GzRBIIePmevmbpToNjqY-- --===============5526405490166494502== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org http://lists.xen.org/xen-devel --===============5526405490166494502==--