From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:55445) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bMpux-0002GJ-4O for qemu-devel@nongnu.org; Tue, 12 Jul 2016 01:07:36 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1bMpuu-00021b-Ga for qemu-devel@nongnu.org; Tue, 12 Jul 2016 01:07:34 -0400 Date: Tue, 12 Jul 2016 14:41:53 +1000 From: David Gibson Message-ID: <20160712044153.GS16355@voom.fritz.box> References: <1467903025-13383-1-git-send-email-bharata@linux.vnet.ibm.com> <1467903025-13383-4-git-send-email-bharata@linux.vnet.ibm.com> <20160708052413.GM14675@voom.fritz.box> <20160708064112.GK25522@in.ibm.com> <20160708073952.GQ14675@voom.fritz.box> <20160708125959.2a5c35bb@nial.brq.redhat.com> <20160711032601.GL16355@voom.fritz.box> <20160711101557.29789b68@nial.brq.redhat.com> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="WfjJTYUClXDJCnRl" Content-Disposition: inline In-Reply-To: <20160711101557.29789b68@nial.brq.redhat.com> Subject: Re: [Qemu-devel] [RFC PATCH v2 3/5] spapr: Set stable_cpu_id for threads of CPU cores List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Igor Mammedov Cc: nikunj@linux.vnet.ibm.com, groug@kaod.org, qemu-devel@nongnu.org, qemu-ppc@nongnu.org, Bharata B Rao , pbonzini@redhat.com --WfjJTYUClXDJCnRl Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, Jul 11, 2016 at 10:15:57AM +0200, Igor Mammedov wrote: > On Mon, 11 Jul 2016 13:26:01 +1000 > David Gibson wrote: >=20 > > On Fri, Jul 08, 2016 at 12:59:59PM +0200, Igor Mammedov wrote: > > > On Fri, 8 Jul 2016 17:39:52 +1000 > > > David Gibson wrote: > > > =20 > > > > On Fri, Jul 08, 2016 at 12:11:12PM +0530, Bharata B Rao wrote: =20 > > > > > On Fri, Jul 08, 2016 at 03:24:13PM +1000, David Gibson wrote: = =20 > > > > > > On Thu, Jul 07, 2016 at 08:20:23PM +0530, Bharata B Rao wrote: = =20 > > > > > > > Conditonally set stable_cpu_id for CPU threads that are creat= ed as part > > > > > > > of spapr CPU cores. The use of stable_cpu_id is enabled for p= series-2.7 > > > > > > > onwards. > > > > > > >=20 > > > > > > > Signed-off-by: Bharata B Rao > > > > > > > --- > > > > > > > hw/ppc/spapr_cpu_core.c | 7 +++++++ > > > > > > > 1 file changed, 7 insertions(+) > > > > > > >=20 > > > > > > > diff --git a/hw/ppc/spapr_cpu_core.c b/hw/ppc/spapr_cpu_core.c > > > > > > > index b104778..0ec3513 100644 > > > > > > > --- a/hw/ppc/spapr_cpu_core.c > > > > > > > +++ b/hw/ppc/spapr_cpu_core.c > > > > > > > @@ -293,8 +293,15 @@ static void spapr_cpu_core_realize(Devic= eState *dev, Error **errp) > > > > > > > for (i =3D 0; i < cc->nr_threads; i++) { > > > > > > > char id[32]; > > > > > > > obj =3D sc->threads + i * size; > > > > > > > + CPUState *cs; > > > > > > > =20 > > > > > > > object_initialize(obj, size, typename); > > > > > > > + cs =3D CPU(obj); > > > > > > > + > > > > > > > + /* Use core_id (which is actually cpu_dt_id) as stab= le CPU id */ > > > > > > > + if (cs->has_stable_cpu_id) { > > > > > > > + cs->stable_cpu_id =3D cc->core_id + i; > > > > > > > + } =20 > > > > > >=20 > > > > > > Testing cs->has_stable_cpu_id here in machine type specific cod= e seems > > > > > > really weird. It's the machine type that knows whether it has a > > > > > > stable ID to give to the CPU or not, rather than the other way = around. > > > > > >=20 > > > > > > Since we haven't yet had a release with cpu cores, I think the = right > > > > > > thing is for cpu_core to unconditionally set the stable ID (and= set > > > > > > has_stable_id to true). =20 > > > > >=20 > > > > > Right, we can set cpu_stable_id unconditionally here since this c= ode path > > > > > (core realize) will be taken only for pseries-2.7 onwards. has_st= able_id > > > > > will get set as part of the property we defined in SPAPR_COMPAT_2= _7. =20 > > > >=20 > > > > Hrm, that's true. But when you describe it like that it sounds lik= e a > > > > really non-obvious and fragile dependency between different compone= nts. =20 > > > that's how compat stuff is typically done for devices, > > > CPUs shouldn't be an exception.=20 > > > (consistency with other devices helps here in long run) > > > =20 > > > > > > The backup path that does thread-based cpu > > > > > > init, can set has_stable_id to false (if that's not the default= ). =20 > > > > >=20 > > > > > Default is off, but turning it on for 2.7 will be inherited by 2.6 > > > > > and others below. Hence I have code to explicitly disable this pr= op > > > > > for 2.6 and below via SPAPR_COMPAT_2_6. =20 > > > >=20 > > > > This is all seeming terribly awkward. =20 > > > Typically default is set the way so new machine type doesn't have > > > to enable it explicitly. > > >=20 > > > However the way it's done here helps not to touch/check every user > > > that uses cpu_index, limiting series impact only to code that > > > asks for it, it look a lot safer to got this rout for now. > > >=20 > > > =20 > > > > Can we try investigating a > > > > different approach: > > > >=20 > > > > 1. Rename cpu_index to cpu_id, but it's still used in the same > > > > places it's used. > > > >=20 > > > > 2. Remove assumptions that cpu_id values are contiguous or > > > > dense > > > > =20 > > > > 3. Machine type decides whether it wants to populate the cpu_id > > > > values explicitly, or leave it to generic code to calculate > > > > them as cpu_index is calculated now. > > > >=20 > > > > 4. Ideally, generic code enforces that the machine type populat= es > > > > either all or none of the cpu_id values. > > > >=20 > > > > Does that seem workable? =20 > > > Ideally we will get there some day (and may be get rid of cpu_index a= ltogether), > > > but for now it seems too invasive with a lot of chances to introduce = non obvious > > > regression. =20 > >=20 > > Yes, that's a risk. But I'm basically no longer convinced that it's > > any higher than the risk of the same thing with the current approach. > to me it looks much less risky as it affects only migration path and > could be done in time for 2.7 >=20 > if we go to cpu_index re-factoring path then I'm afraid it quite a bit > more complex and it's not a 2.7 material. Yeah, ok. Having looked in a bit more depth, I agree with you. >=20 >=20 > >=20 > > > So I'd keep approach used in this series. =20 > >=20 >=20 --=20 David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson --WfjJTYUClXDJCnRl Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAEBAgAGBQJXhHURAAoJEGw4ysog2bOS73wP/Rw5CRPJiBN65UvzG2jVOwAU Ztj/qOW+reTCRxHypeXiEJPN4azYaTuBqXDeQ2A4gO6O82emDv/ODTMLAouDFUeI wl0qpT47OXAwp0ZA046Ad9ZhMYPul12aSXgfXiF4RFjg7PJEiRyTGRAwz4fNkMCs 49aG5MdJMbabCvnoxQ+iUldezJgAob4Pf3eIM7teg1SyDuloCQgRUyeLHRkBws0H mJpbvILvc89b7E1p6dKS62ajnrhFtvGwsnitIyvuXOCd3xfOod+9rMqkniDBaYyF 7H8Tbrd7RD0iM++k74yF6hdun0PksVVl4uokATRoeJddwN0Alshgyria3Hz+0NVy 7frFKimoIECcY/QvnSiUknuOAdkHQfwcD8l0b7TSxWLKnRiSSG65hetEWT47lkLF SGVgOJRQ94ISd4INNvR0sxOnDBGcvbabSJahyLTe3KONHn3+BiwdhofzRRBsaLFV F7DwI9lpESvLOIwLP2MwehETR4Sn5RWIi6eipK5GrjJBXAjT/a4B2+MAbOdhLUiV 3HRPwJJw7OYIT4HSee2CPmgD8jYSkJ6ii5s5in45mcBweHpKA/GZcCfNKHuJVFpL cwUPljPZ68fhUCQAFGLULESVm0LwM+4g6apFGwdNfIBT3A+S8qUjVSTCK3Esg3yg 6WyzCB6c8ENwA4BoeNCe =kn8l -----END PGP SIGNATURE----- --WfjJTYUClXDJCnRl--