From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:56146) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1acmEe-0006zx-HI for qemu-devel@nongnu.org; Sun, 06 Mar 2016 22:53:34 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1acmEc-000065-SX for qemu-devel@nongnu.org; Sun, 06 Mar 2016 22:53:32 -0500 Date: Mon, 7 Mar 2016 14:36:55 +1100 From: David Gibson Message-ID: <20160307033655.GD22546@voom.fritz.box> References: <1457074461-14285-1-git-send-email-bharata@linux.vnet.ibm.com> <1457074461-14285-6-git-send-email-bharata@linux.vnet.ibm.com> <20160304113845.667c19ad@nial.brq.redhat.com> <20160304110253.GB5054@in.ibm.com> <20160304190720.4d64abc4@nial.brq.redhat.com> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="KdquIMZPjGJQvRdI" Content-Disposition: inline In-Reply-To: <20160304190720.4d64abc4@nial.brq.redhat.com> Subject: Re: [Qemu-devel] [RFC PATCH v1 05/10] cpu: Abstract CPU core type List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Igor Mammedov , g@voom.fritz.box Cc: mjrosato@linux.vnet.ibm.com, thuth@redhat.com, pkrempa@redhat.com, ehabkost@redhat.com, aik@ozlabs.ru, qemu-devel@nongnu.org, agraf@suse.de, armbru@redhat.com, borntraeger@de.ibm.com, qemu-ppc@nongnu.org, Bharata B Rao , pbonzini@redhat.com, mdroth@linux.vnet.ibm.com, afaerber@suse.de --KdquIMZPjGJQvRdI Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, Mar 04, 2016 at 07:07:20PM +0100, Igor Mammedov wrote: > On Fri, 4 Mar 2016 16:32:53 +0530 > Bharata B Rao wrote: >=20 > > On Fri, Mar 04, 2016 at 11:38:45AM +0100, Igor Mammedov wrote: > > > On Fri, 4 Mar 2016 12:24:16 +0530 > > > Bharata B Rao wrote: > > > =20 > > > > Add an abstract CPU core type that could be used by machines that w= ant > > > > to define and hotplug CPUs in core granularity. > > > >=20 > > > > Signed-off-by: Bharata B Rao > > > > --- > > > > hw/cpu/Makefile.objs | 1 + > > > > hw/cpu/core.c | 44 +++++++++++++++++++++++++++++++++++++++= +++++ > > > > include/hw/cpu/core.h | 30 ++++++++++++++++++++++++++++++ > > > > 3 files changed, 75 insertions(+) > > > > create mode 100644 hw/cpu/core.c > > > > create mode 100644 include/hw/cpu/core.h > > > >=20 > > > > diff --git a/hw/cpu/Makefile.objs b/hw/cpu/Makefile.objs > > > > index 0954a18..942a4bb 100644 > > > > --- a/hw/cpu/Makefile.objs > > > > +++ b/hw/cpu/Makefile.objs > > > > @@ -2,4 +2,5 @@ obj-$(CONFIG_ARM11MPCORE) +=3D arm11mpcore.o > > > > obj-$(CONFIG_REALVIEW) +=3D realview_mpcore.o > > > > obj-$(CONFIG_A9MPCORE) +=3D a9mpcore.o > > > > obj-$(CONFIG_A15MPCORE) +=3D a15mpcore.o > > > > +obj-y +=3D core.o > > > > =20 > > > > diff --git a/hw/cpu/core.c b/hw/cpu/core.c > > > > new file mode 100644 > > > > index 0000000..d8caf37 > > > > --- /dev/null > > > > +++ b/hw/cpu/core.c > > > > @@ -0,0 +1,44 @@ > > > > +/* > > > > + * CPU core abstract device > > > > + * > > > > + * Copyright (C) 2016 Bharata B Rao > > > > + * > > > > + * This work is licensed under the terms of the GNU GPL, version 2= or later. > > > > + * See the COPYING file in the top-level directory. > > > > + */ > > > > +#include "hw/cpu/core.h" > > > > + > > > > +static char *core_prop_get_slot(Object *obj, Error **errp) > > > > +{ > > > > + CPUCore *core =3D CPU_CORE(obj); > > > > + > > > > + return g_strdup(core->slot); > > > > +} > > > > + > > > > +static void core_prop_set_slot(Object *obj, const char *val, Error= **errp) > > > > +{ > > > > + CPUCore *core =3D CPU_CORE(obj); > > > > + > > > > + core->slot =3D g_strdup(val); > > > > +} > > > > + > > > > +static void cpu_core_instance_init(Object *obj) > > > > +{ > > > > + object_property_add_str(obj, "slot", core_prop_get_slot, core_= prop_set_slot, > > > > + NULL); > > > > +} > > > > + > > > > +static const TypeInfo cpu_core_type_info =3D { > > > > + .name =3D TYPE_CPU_CORE, > > > > + .parent =3D TYPE_DEVICE, > > > > + .abstract =3D true, > > > > + .instance_size =3D sizeof(CPUCore), > > > > + .instance_init =3D cpu_core_instance_init, > > > > +}; > > > > + > > > > +static void cpu_core_register_types(void) > > > > +{ > > > > + type_register_static(&cpu_core_type_info); > > > > +} > > > > + > > > > +type_init(cpu_core_register_types) > > > > diff --git a/include/hw/cpu/core.h b/include/hw/cpu/core.h > > > > new file mode 100644 > > > > index 0000000..2daa724 > > > > --- /dev/null > > > > +++ b/include/hw/cpu/core.h > > > > @@ -0,0 +1,30 @@ > > > > +/* > > > > + * CPU core abstract device > > > > + * > > > > + * Copyright (C) 2016 Bharata B Rao > > > > + * > > > > + * This work is licensed under the terms of the GNU GPL, version 2= or later. > > > > + * See the COPYING file in the top-level directory. > > > > + */ > > > > +#ifndef HW_CPU_CORE_H > > > > +#define HW_CPU_CORE_H > > > > + > > > > +#include "qemu/osdep.h" > > > > +#include "hw/qdev.h" > > > > + > > > > +#define TYPE_CPU_CORE "cpu-core" > > > > + > > > > +#define CPU_CORE(obj) \ > > > > + OBJECT_CHECK(CPUCore, (obj), TYPE_CPU_CORE) > > > > + > > > > +typedef struct CPUCore { > > > > + /*< private >*/ > > > > + DeviceState parent_obj; > > > > + > > > > + /*< public >*/ > > > > + char *slot; > > > > +} CPUCore; > > > > + > > > > +#define CPU_CORE_SLOT_PROP "slot" =20 > > > as it's generic property I'd rename to 'core' so it would fit all use= rs =20 > >=20 > > Ok. Also note that this is a string property which is associated with t= he > > link name (string) that we created from machine object to this core. I = think > > it would be ideal if this becomes an interger property in which case it > > becomes easier to feed the core location into your CPUSlotProperties.co= re. > agreed, it should be core number. The slot stuff is continuing to confuse me a bit. I see that we need some kind of "address" value, but how best to do it is not clear to me. Changing this to an integer sounds like it's probably a good idea. I'm a bit wary of just calling it "core" though. Do all platforms even necessarily have a core id? I'm wondering if the addressing is something that needs to move the the platform specific subtypes, while some other stuff can move to the generic base type. > > > on top of that I'd add numeric 'threads' property to base class so > > > all derived cores would inherit it. > > >=20 > > > Then as easy integration with -smp threads=3Dx, a machine could push > > > a global variable 'cpu-core.threads=3D[smp_threads]' which would > > > make every created cpu-core object to have threads set > > > at instance_init() time (device_init). > > >=20 > > > That way user won't have to specify 'threads=3Dy' for every > > > device_add spapr-core,core=3Dx > > > as it will be taken from global property 'cpu-core.threads' > > > but if user wishes he/she still could override global by explicitly > > > providing thread property at device_add time: > > > device_add spapr-core,core=3Dx,threads=3Dy > > >=20 > > > wrt this series it would mean, instead of creating threads in property > > > setter, delaying threads creation to core.realize() time, > > > but since realize is allowed to fail it should be fine do so. =20 > >=20 > > Ok that would suit us as there are two properties on which thread creat= ion > > is dependent upon: nr_threads and cpu_model. If thread objects can be > > created at core realize time, then we don't have to resort to the uglin= ess > > of creating the threads from either of the property setters. I always > > assumed that we shouldn't be creating objects from realize, but if that > > is fine, it is good. > since realize is allowed to fail, it should be safe from hotplug pov > to create internal objects there, as far as proper cleanups are done > for failure path. Right, moving the "nr_threads" property to the base type seems like a good idea to me. Even for subtypes which have a fixed number of threads we should be able to handle by just verifying and rejecting in the subtype's realize. And even for those core types, being able to read the number of threads in a generic way is a good thing. I'm not clear from the above if you're also intending to move at least the adding of the threads as child properties is supposed to go into the base type, but that also sounds like a good idea, again for consistency. --=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 --KdquIMZPjGJQvRdI Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAEBAgAGBQJW3PdXAAoJEGw4ysog2bOSNhoQAKQVcgeAnUC0ReD/vTY8OHbO S2HFWyA/1fLvBOI7Eipv+byQ5dpJnzb95PBVfa/0KbBdQxD3gqlN9FcGSLZlisOJ dm9Pw076YssvqapwOZBLaZDr+B2ThssBQx8dOo/8PeP29akKxR4lu05G2R8ldYgs rb9uoLXWR/4tPcMtrtk2td7HDKbo4cgK9Pa0mL95QsfDLVJIjGfnNdU83K3SuyZX eZeyRCtlMHGEoX+JRIZur6Xd6KpTh3aUoMIMyHUTjlFT7oaCNV/TyFxzX74LE63b 1TkOfqvrKlsGnm4NmhwSSb3sKcm9x+4yoygBeLBb/9+AZUXG8877aS6ZLEMU08Nd wXxPjUDRdy+w1BzgQ8m7h5cGMKloNhdwf96vkFOF9ut35HFEwJXGYZ4cM8wV+czU kvZl3ENk/HbXiaooKFQloZ385naHeRwag+iRdGLjbv7d47bBndvW1tYMy9tiqv9X hF7GJV7MwngnggsrBA5rtPR2jMGvh54wF3f9Aag8JSrai+ZMkzE0s/cR0apD490H JdBS57dblvpqPResIE3D2LjpNWCuVRnUxNiS5jkM2fzfU0rGC8euI0o7SMm1eQpy Em/2thDM+SxZZArOWikijbFFn8HqM3SSt5NdcPY489s+63s14zy7N6b11OwIIqIN QUeK0Q7iEJUQ9aomVZg6 =VlV7 -----END PGP SIGNATURE----- --KdquIMZPjGJQvRdI--