From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:57852) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fahzU-0001kt-8w for qemu-devel@nongnu.org; Wed, 04 Jul 2018 09:38:41 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fahzQ-0001Ja-7T for qemu-devel@nongnu.org; Wed, 04 Jul 2018 09:38:40 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:42086 helo=mx1.redhat.com) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fahzQ-0001J1-1y for qemu-devel@nongnu.org; Wed, 04 Jul 2018 09:38:36 -0400 References: <1530709942-87947-1-git-send-email-jingqi.liu@intel.com> From: Paolo Bonzini Message-ID: Date: Wed, 4 Jul 2018 15:38:32 +0200 MIME-Version: 1.0 In-Reply-To: <1530709942-87947-1-git-send-email-jingqi.liu@intel.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH] kvm: x86: Add support for -machine split-lock-ac List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jingqi Liu Cc: rth@twiddle.net, ehabkost@redhat.com, mtosatti@redhat.com, qemu-devel@nongnu.org, wei.w.wang@intel.com, "Hu, Robert" On 04/07/2018 15:12, Jingqi Liu wrote: > Add the option split-lock-ac to control whether the #AC > exception is generated for split locked accesses, which > is introduced for the machine, there is an example to enable it: > -machine split-lock-ac=3Don > It is disabled on default. >=20 > When bit 29 of the MSR (33H) is set, the processor > causes an #AC exception to be issued instead of suppressing LOCK on > bus(during split lock access). This should be a CPU feature, not a machine feature. As mentioned in the review of the kernel patch, please work with Robert to use the same infrastructure for both MSR_TEST_CTL and MSR_IA32_ARCH_CAPABILITIES. (Robert, does IceLake have this feature? If so, we cannot create the CPU model until everything is in place). BTW, why is the availability of the feature not exposed with a CPUID bit? It doesn't make much sense. Paolo > Signed-off-by: Jingqi Liu > --- > accel/kvm/kvm-all.c | 10 ++++++++++ > hw/core/machine.c | 19 +++++++++++++++++++ > include/hw/boards.h | 1 + > linux-headers/linux/kvm.h | 1 + > qemu-options.hx | 3 +++ > 5 files changed, 34 insertions(+) >=20 > diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c > index eb7db92..e960a4d 100644 > --- a/accel/kvm/kvm-all.c > +++ b/accel/kvm/kvm-all.c > @@ -1699,6 +1699,16 @@ static int kvm_init(MachineState *ms) > =20 > s->sync_mmu =3D !!kvm_vm_check_extension(kvm_state, KVM_CAP_SYNC_M= MU); > =20 > + if (ms->split_lock_ac && > + kvm_check_extension(s, KVM_CAP_X86_SPLIT_LOCK_AC)) { > + ret =3D kvm_vm_enable_cap(s, KVM_CAP_X86_SPLIT_LOCK_AC, 0); > + if (ret) { > + fprintf(stderr, "Could not enable the feature that #AC exc= eption" > + "will be generated for split lock accesses!\n"); > + goto err; > + } > + } > + > return 0; > =20 > err: > diff --git a/hw/core/machine.c b/hw/core/machine.c > index 617e5f8..e054119 100644 > --- a/hw/core/machine.c > +++ b/hw/core/machine.c > @@ -350,6 +350,22 @@ static void machine_set_memory_encryption(Object *= obj, const char *value, > ms->memory_encryption =3D g_strdup(value); > } > =20 > +static void machine_set_split_lock_ac(Object *obj, bool value, > + Error **errp) > +{ > + MachineState *ms =3D MACHINE(obj); > + > + ms->split_lock_ac =3D value; > +} > + > +static bool machine_get_split_lock_ac(Object *obj, Error **errp) > +{ > + MachineState *ms =3D MACHINE(obj); > + > + return ms->split_lock_ac; > +} > + > + > void machine_class_allow_dynamic_sysbus_dev(MachineClass *mc, const ch= ar *type) > { > strList *item =3D g_new0(strList, 1); > @@ -634,6 +650,9 @@ static void machine_class_init(ObjectClass *oc, voi= d *data) > &error_abort); > object_class_property_set_description(oc, "memory-encryption", > "Set memory encyption object to use", &error_abort); > + object_class_property_add_bool(oc, "split-lock-ac", > + machine_get_split_lock_ac, machine_set_split_lock_ac, > + &error_abort); > } > =20 > static void machine_class_base_init(ObjectClass *oc, void *data) > diff --git a/include/hw/boards.h b/include/hw/boards.h > index 79069dd..9c1e2c0 100644 > --- a/include/hw/boards.h > +++ b/include/hw/boards.h > @@ -256,6 +256,7 @@ struct MachineState { > bool suppress_vmdesc; > bool enforce_config_section; > bool enable_graphics; > + bool split_lock_ac; > char *memory_encryption; > DeviceMemoryState *device_memory; > =20 > diff --git a/linux-headers/linux/kvm.h b/linux-headers/linux/kvm.h > index 98f389a..4bfa879 100644 > --- a/linux-headers/linux/kvm.h > +++ b/linux-headers/linux/kvm.h > @@ -949,6 +949,7 @@ struct kvm_ppc_resize_hpt { > #define KVM_CAP_GET_MSR_FEATURES 153 > #define KVM_CAP_HYPERV_EVENTFD 154 > #define KVM_CAP_HYPERV_TLBFLUSH 155 > +#define KVM_CAP_X86_SPLIT_LOCK_AC 156 > =20 > #ifdef KVM_CAP_IRQ_ROUTING > =20 > diff --git a/qemu-options.hx b/qemu-options.hx > index 81b1e99..935cefe 100644 > --- a/qemu-options.hx > +++ b/qemu-options.hx > @@ -44,6 +44,7 @@ DEF("machine", HAS_ARG, QEMU_OPTION_machine, \ > " nvdimm=3Don|off controls NVDIMM support (default=3D= off)\n" > " enforce-config-section=3Don|off enforce configura= tion section migration (default=3Doff)\n" > " s390-squash-mcss=3Don|off (deprecated) controls s= upport for squashing into default css (default=3Doff)\n" > + " split-lock-ac=3Don|off controls support #AC excep= tion for split locked accesses (default=3Doff)\n" > " memory-encryption=3D@var{} memory encryption obje= ct to use (default=3Dnone)\n", > QEMU_ARCH_ALL) > STEXI > @@ -113,6 +114,8 @@ NOTE: this parameter is deprecated. Please use @opt= ion{-global} > @option{migration.send-configuration}=3D@var{on|off} instead. > @item memory-encryption=3D@var{} > Memory encryption object to use. The default is none. > +@item split-lock-ac=3Don|off > +Enables or disables #AC exception for split locked accesses. > @end table > ETEXI > =20 >=20