When running KVM on POWER, we allow the user to pass "-cpu POWERx" instead
of "-cpu host". This is achieved by patching the ppc_cpu_aliases[] array
so that "POWERx" points to the CPU class with the same PVR as the host CPU.
This causes CPUs to be instantiated from this CPU class instead of the
TYPE_HOST_POWERPC_CPU class which is used with "-cpu host". These CPUs thus
miss all the KVM specific tuning from kvmppc_host_cpu_class_init().
This currently causes QEMU with "-cpu POWER9" to fail when running KVM on a
POWER9 DD1 host:
qemu-system-ppc64: Register sync failed... If you're using kvm-hv.ko, only
"-cpu host" is possible
kvm_init_vcpu failed: Invalid argument
Let's have the "POWERx" alias to point to TYPE_HOST_POWERPC_CPU directly,
so that "-cpu POWERx" instantiates CPUs from the same class as "-cpu host".
Signed-off-by: Greg Kurz <groug@kaod.org>
---
target/ppc/kvm.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/target/ppc/kvm.c b/target/ppc/kvm.c
index f2f7c531bc7b..f7a7ea5858a3 100644
--- a/target/ppc/kvm.c
+++ b/target/ppc/kvm.c
@@ -2445,6 +2445,7 @@ static int kvm_ppc_register_host_cpu_type(void)
.class_init = kvmppc_host_cpu_class_init,
};
PowerPCCPUClass *pvr_pcc;
+ ObjectClass *oc;
DeviceClass *dc;
int i;
@@ -2455,6 +2456,9 @@ static int kvm_ppc_register_host_cpu_type(void)
type_info.parent = object_class_get_name(OBJECT_CLASS(pvr_pcc));
type_register(&type_info);
+ oc = object_class_by_name(type_info.name);
+ g_assert(oc);
+
#if defined(TARGET_PPC64)
type_info.name = g_strdup_printf("%s-"TYPE_SPAPR_CPU_CORE, "host");
type_info.parent = TYPE_SPAPR_CPU_CORE,
@@ -2474,7 +2478,6 @@ static int kvm_ppc_register_host_cpu_type(void)
dc = DEVICE_CLASS(ppc_cpu_get_family_class(pvr_pcc));
for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) {
if (strcmp(ppc_cpu_aliases[i].alias, dc->desc) == 0) {
- ObjectClass *oc = OBJECT_CLASS(pvr_pcc);
char *suffix;
ppc_cpu_aliases[i].model = g_strdup(object_class_get_name(oc));
On 05/07/2017 10:49, Greg Kurz wrote: > diff --git a/target/ppc/kvm.c b/target/ppc/kvm.c > index f2f7c531bc7b..f7a7ea5858a3 100644 > --- a/target/ppc/kvm.c > +++ b/target/ppc/kvm.c > @@ -2445,6 +2445,7 @@ static int kvm_ppc_register_host_cpu_type(void) > .class_init = kvmppc_host_cpu_class_init, > }; > PowerPCCPUClass *pvr_pcc; > + ObjectClass *oc; > DeviceClass *dc; > int i; > > @@ -2455,6 +2456,9 @@ static int kvm_ppc_register_host_cpu_type(void) > type_info.parent = object_class_get_name(OBJECT_CLASS(pvr_pcc)); > type_register(&type_info); > > + oc = object_class_by_name(type_info.name); > + g_assert(oc); > + > #if defined(TARGET_PPC64) > type_info.name = g_strdup_printf("%s-"TYPE_SPAPR_CPU_CORE, "host"); > type_info.parent = TYPE_SPAPR_CPU_CORE, > @@ -2474,7 +2478,6 @@ static int kvm_ppc_register_host_cpu_type(void) > dc = DEVICE_CLASS(ppc_cpu_get_family_class(pvr_pcc)); > for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) { > if (strcmp(ppc_cpu_aliases[i].alias, dc->desc) == 0) { > - ObjectClass *oc = OBJECT_CLASS(pvr_pcc); > char *suffix; > > ppc_cpu_aliases[i].model = g_strdup(object_class_get_name(oc)); > A far better solution :) Tested-by: Laurent Vivier <lvivier@redhat.com> Reviewed-by: Laurent Vivier <lvivier@redhat.com>
On 05.07.2017 10:49, Greg Kurz wrote: > When running KVM on POWER, we allow the user to pass "-cpu POWERx" instead > of "-cpu host". This is achieved by patching the ppc_cpu_aliases[] array > so that "POWERx" points to the CPU class with the same PVR as the host CPU. > This causes CPUs to be instantiated from this CPU class instead of the > TYPE_HOST_POWERPC_CPU class which is used with "-cpu host". These CPUs thus > miss all the KVM specific tuning from kvmppc_host_cpu_class_init(). > > This currently causes QEMU with "-cpu POWER9" to fail when running KVM on a > POWER9 DD1 host: > > qemu-system-ppc64: Register sync failed... If you're using kvm-hv.ko, only > "-cpu host" is possible > kvm_init_vcpu failed: Invalid argument > > Let's have the "POWERx" alias to point to TYPE_HOST_POWERPC_CPU directly, > so that "-cpu POWERx" instantiates CPUs from the same class as "-cpu host". > > Signed-off-by: Greg Kurz <groug@kaod.org> > --- > target/ppc/kvm.c | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/target/ppc/kvm.c b/target/ppc/kvm.c > index f2f7c531bc7b..f7a7ea5858a3 100644 > --- a/target/ppc/kvm.c > +++ b/target/ppc/kvm.c > @@ -2445,6 +2445,7 @@ static int kvm_ppc_register_host_cpu_type(void) > .class_init = kvmppc_host_cpu_class_init, > }; > PowerPCCPUClass *pvr_pcc; > + ObjectClass *oc; > DeviceClass *dc; > int i; > > @@ -2455,6 +2456,9 @@ static int kvm_ppc_register_host_cpu_type(void) > type_info.parent = object_class_get_name(OBJECT_CLASS(pvr_pcc)); > type_register(&type_info); > > + oc = object_class_by_name(type_info.name); > + g_assert(oc); > + > #if defined(TARGET_PPC64) > type_info.name = g_strdup_printf("%s-"TYPE_SPAPR_CPU_CORE, "host"); > type_info.parent = TYPE_SPAPR_CPU_CORE, > @@ -2474,7 +2478,6 @@ static int kvm_ppc_register_host_cpu_type(void) > dc = DEVICE_CLASS(ppc_cpu_get_family_class(pvr_pcc)); > for (i = 0; ppc_cpu_aliases[i].alias != NULL; i++) { > if (strcmp(ppc_cpu_aliases[i].alias, dc->desc) == 0) { > - ObjectClass *oc = OBJECT_CLASS(pvr_pcc); > char *suffix; > > ppc_cpu_aliases[i].model = g_strdup(object_class_get_name(oc)); > Yes, I think that should be fine, too. Reviewed-by: Thomas Huth <thuth@redhat.com>
On Wed, Jul 05, 2017 at 10:49:52AM +0200, Greg Kurz wrote: > When running KVM on POWER, we allow the user to pass "-cpu POWERx" instead > of "-cpu host". This is achieved by patching the ppc_cpu_aliases[] array > so that "POWERx" points to the CPU class with the same PVR as the host CPU. > This causes CPUs to be instantiated from this CPU class instead of the > TYPE_HOST_POWERPC_CPU class which is used with "-cpu host". These CPUs thus > miss all the KVM specific tuning from kvmppc_host_cpu_class_init(). > > This currently causes QEMU with "-cpu POWER9" to fail when running KVM on a > POWER9 DD1 host: > > qemu-system-ppc64: Register sync failed... If you're using kvm-hv.ko, only > "-cpu host" is possible > kvm_init_vcpu failed: Invalid argument > > Let's have the "POWERx" alias to point to TYPE_HOST_POWERPC_CPU directly, > so that "-cpu POWERx" instantiates CPUs from the same class as "-cpu host". > > Signed-off-by: Greg Kurz <groug@kaod.org> Nice. I had mistakenly thought it already did that. I've applied this to ppc-for-2.10. Just to make sure I've understood, this obsoletes Laurent's earlier patches moving DD1 workarounds about, yes? -- 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
On Thu, 6 Jul 2017 17:32:33 +1000 David Gibson <david@gibson.dropbear.id.au> wrote: > On Wed, Jul 05, 2017 at 10:49:52AM +0200, Greg Kurz wrote: > > When running KVM on POWER, we allow the user to pass "-cpu POWERx" instead > > of "-cpu host". This is achieved by patching the ppc_cpu_aliases[] array > > so that "POWERx" points to the CPU class with the same PVR as the host CPU. > > This causes CPUs to be instantiated from this CPU class instead of the > > TYPE_HOST_POWERPC_CPU class which is used with "-cpu host". These CPUs thus > > miss all the KVM specific tuning from kvmppc_host_cpu_class_init(). > > > > This currently causes QEMU with "-cpu POWER9" to fail when running KVM on a > > POWER9 DD1 host: > > > > qemu-system-ppc64: Register sync failed... If you're using kvm-hv.ko, only > > "-cpu host" is possible > > kvm_init_vcpu failed: Invalid argument > > > > Let's have the "POWERx" alias to point to TYPE_HOST_POWERPC_CPU directly, > > so that "-cpu POWERx" instantiates CPUs from the same class as "-cpu host". > > > > Signed-off-by: Greg Kurz <groug@kaod.org> > > Nice. I had mistakenly thought it already did that. > > I've applied this to ppc-for-2.10. > > Just to make sure I've understood, this obsoletes Laurent's earlier > patches moving DD1 workarounds about, yes? > Yes it does. Cheers, -- Greg
© 2016 - 2024 Red Hat, Inc.