From: Reiji Watanabe <[email protected]>
The following patches will use the number of counters information
from the arm_pmu and use this to set the PMCR.N for the guest
during vCPU reset. However, since the guest is not associated
with any arm_pmu until userspace configures the vPMU device
attributes, and a reset can happen before this event, assign a
default PMU to the guest just before doing the reset.
Signed-off-by: Reiji Watanabe <[email protected]>
Signed-off-by: Raghavendra Rao Ananta <[email protected]>
---
arch/arm64/kvm/arm.c | 20 ++++++++++++++++++++
arch/arm64/kvm/pmu-emul.c | 12 ++----------
include/kvm/arm_pmu.h | 6 ++++++
3 files changed, 28 insertions(+), 10 deletions(-)
diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
index 78b0970eb8e6..708a53b70a7b 100644
--- a/arch/arm64/kvm/arm.c
+++ b/arch/arm64/kvm/arm.c
@@ -1313,6 +1313,23 @@ static bool kvm_vcpu_init_changed(struct kvm_vcpu *vcpu,
KVM_VCPU_MAX_FEATURES);
}
+static int kvm_vcpu_set_pmu(struct kvm_vcpu *vcpu)
+{
+ struct kvm *kvm = vcpu->kvm;
+
+ if (!kvm_arm_support_pmu_v3())
+ return -EINVAL;
+
+ /*
+ * When the vCPU has a PMU, but no PMU is set for the guest
+ * yet, set the default one.
+ */
+ if (unlikely(!kvm->arch.arm_pmu))
+ return kvm_arm_set_default_pmu(kvm);
+
+ return 0;
+}
+
static int __kvm_vcpu_set_target(struct kvm_vcpu *vcpu,
const struct kvm_vcpu_init *init)
{
@@ -1328,6 +1345,9 @@ static int __kvm_vcpu_set_target(struct kvm_vcpu *vcpu,
bitmap_copy(kvm->arch.vcpu_features, &features, KVM_VCPU_MAX_FEATURES);
+ if (kvm_vcpu_has_pmu(vcpu) && kvm_vcpu_set_pmu(vcpu))
+ goto out_unlock;
+
/* Now we know what it is, we can reset it. */
kvm_reset_vcpu(vcpu);
diff --git a/arch/arm64/kvm/pmu-emul.c b/arch/arm64/kvm/pmu-emul.c
index eb5dcb12dafe..cc30c246c010 100644
--- a/arch/arm64/kvm/pmu-emul.c
+++ b/arch/arm64/kvm/pmu-emul.c
@@ -717,8 +717,7 @@ static struct arm_pmu *kvm_pmu_probe_armpmu(void)
* It is still necessary to get a valid cpu, though, to probe for the
* default PMU instance as userspace is not required to specify a PMU
* type. In order to uphold the preexisting behavior KVM selects the
- * PMU instance for the core where the first call to the
- * KVM_ARM_VCPU_PMU_V3_CTRL attribute group occurs. A dependent use case
+ * PMU instance for the core during the vcpu reset. A dependent use case
* would be a user with disdain of all things big.LITTLE that affines
* the VMM to a particular cluster of cores.
*
@@ -893,7 +892,7 @@ static void kvm_arm_set_pmu(struct kvm *kvm, struct arm_pmu *arm_pmu)
* where vCPUs can be scheduled on any core but the guest
* counters could stop working.
*/
-static int kvm_arm_set_default_pmu(struct kvm *kvm)
+int kvm_arm_set_default_pmu(struct kvm *kvm)
{
struct arm_pmu *arm_pmu = kvm_pmu_probe_armpmu();
@@ -946,13 +945,6 @@ int kvm_arm_pmu_v3_set_attr(struct kvm_vcpu *vcpu, struct kvm_device_attr *attr)
if (vcpu->arch.pmu.created)
return -EBUSY;
- if (!kvm->arch.arm_pmu) {
- int ret = kvm_arm_set_default_pmu(kvm);
-
- if (ret)
- return ret;
- }
-
switch (attr->attr) {
case KVM_ARM_VCPU_PMU_V3_IRQ: {
int __user *uaddr = (int __user *)(long)attr->addr;
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 3546ebc469ad..858ed9ce828a 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -101,6 +101,7 @@ void kvm_vcpu_pmu_resync_el0(void);
})
u8 kvm_arm_pmu_get_pmuver_limit(void);
+int kvm_arm_set_default_pmu(struct kvm *kvm);
#else
struct kvm_pmu {
@@ -174,6 +175,11 @@ static inline u8 kvm_arm_pmu_get_pmuver_limit(void)
}
static inline void kvm_vcpu_pmu_resync_el0(void) {}
+static inline int kvm_arm_set_default_pmu(struct kvm *kvm)
+{
+ return -ENODEV;
+}
+
#endif
#endif
--
2.42.0.609.gbb76f46606-goog
Hi Raghu,
On Mon, Oct 09, 2023 at 11:08:48PM +0000, Raghavendra Rao Ananta wrote:
> From: Reiji Watanabe <[email protected]>
>
> The following patches will use the number of counters information
> from the arm_pmu and use this to set the PMCR.N for the guest
> during vCPU reset. However, since the guest is not associated
> with any arm_pmu until userspace configures the vPMU device
> attributes, and a reset can happen before this event, assign a
> default PMU to the guest just before doing the reset.
>
> Signed-off-by: Reiji Watanabe <[email protected]>
> Signed-off-by: Raghavendra Rao Ananta <[email protected]>
> ---
> arch/arm64/kvm/arm.c | 20 ++++++++++++++++++++
> arch/arm64/kvm/pmu-emul.c | 12 ++----------
> include/kvm/arm_pmu.h | 6 ++++++
> 3 files changed, 28 insertions(+), 10 deletions(-)
>
> diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> index 78b0970eb8e6..708a53b70a7b 100644
> --- a/arch/arm64/kvm/arm.c
> +++ b/arch/arm64/kvm/arm.c
> @@ -1313,6 +1313,23 @@ static bool kvm_vcpu_init_changed(struct kvm_vcpu *vcpu,
> KVM_VCPU_MAX_FEATURES);
> }
>
> +static int kvm_vcpu_set_pmu(struct kvm_vcpu *vcpu)
> +{
> + struct kvm *kvm = vcpu->kvm;
> +
> + if (!kvm_arm_support_pmu_v3())
> + return -EINVAL;
This check is pointless; the vCPU feature flags have been sanitised at
this point, and a requirement of having PMUv3 is that this predicate is
true.
> + /*
> + * When the vCPU has a PMU, but no PMU is set for the guest
> + * yet, set the default one.
> + */
> + if (unlikely(!kvm->arch.arm_pmu))
> + return kvm_arm_set_default_pmu(kvm);
> +
> + return 0;
> +}
> +
Apologies, I believe I was unclear last time around as to what I was
wanting here. Let's call this thing kvm_setup_vcpu() such that we can
add other one-time setup activities to it in the future.
Something like:
diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
index 96641e442039..4896a44108e0 100644
--- a/arch/arm64/kvm/arm.c
+++ b/arch/arm64/kvm/arm.c
@@ -1265,19 +1265,17 @@ static bool kvm_vcpu_init_changed(struct kvm_vcpu *vcpu,
KVM_VCPU_MAX_FEATURES);
}
-static int kvm_vcpu_set_pmu(struct kvm_vcpu *vcpu)
+static int kvm_setup_vcpu(struct kvm_vcpu *vcpu)
{
struct kvm *kvm = vcpu->kvm;
- if (!kvm_arm_support_pmu_v3())
- return -EINVAL;
-
/*
* When the vCPU has a PMU, but no PMU is set for the guest
* yet, set the default one.
*/
- if (unlikely(!kvm->arch.arm_pmu))
- return kvm_arm_set_default_pmu(kvm);
+ if (kvm_vcpu_has_pmu(vcpu) && !kvm->arch.arm_pmu &&
+ kvm_arm_set_default_pmu(kvm))
+ return -EINVAL;
return 0;
}
@@ -1297,7 +1295,8 @@ static int __kvm_vcpu_set_target(struct kvm_vcpu *vcpu,
bitmap_copy(kvm->arch.vcpu_features, &features, KVM_VCPU_MAX_FEATURES);
- if (kvm_vcpu_has_pmu(vcpu) && kvm_vcpu_set_pmu(vcpu))
+ ret = kvm_setup_vcpu(vcpu);
+ if (ret)
goto out_unlock;
/* Now we know what it is, we can reset it. */
--
Thanks,
Oliver
On Tue, Oct 10, 2023 at 3:25 PM Oliver Upton <[email protected]> wrote:
>
> Hi Raghu,
>
> On Mon, Oct 09, 2023 at 11:08:48PM +0000, Raghavendra Rao Ananta wrote:
> > From: Reiji Watanabe <[email protected]>
> >
> > The following patches will use the number of counters information
> > from the arm_pmu and use this to set the PMCR.N for the guest
> > during vCPU reset. However, since the guest is not associated
> > with any arm_pmu until userspace configures the vPMU device
> > attributes, and a reset can happen before this event, assign a
> > default PMU to the guest just before doing the reset.
> >
> > Signed-off-by: Reiji Watanabe <[email protected]>
> > Signed-off-by: Raghavendra Rao Ananta <[email protected]>
> > ---
> > arch/arm64/kvm/arm.c | 20 ++++++++++++++++++++
> > arch/arm64/kvm/pmu-emul.c | 12 ++----------
> > include/kvm/arm_pmu.h | 6 ++++++
> > 3 files changed, 28 insertions(+), 10 deletions(-)
> >
> > diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> > index 78b0970eb8e6..708a53b70a7b 100644
> > --- a/arch/arm64/kvm/arm.c
> > +++ b/arch/arm64/kvm/arm.c
> > @@ -1313,6 +1313,23 @@ static bool kvm_vcpu_init_changed(struct kvm_vcpu *vcpu,
> > KVM_VCPU_MAX_FEATURES);
> > }
> >
> > +static int kvm_vcpu_set_pmu(struct kvm_vcpu *vcpu)
> > +{
> > + struct kvm *kvm = vcpu->kvm;
> > +
> > + if (!kvm_arm_support_pmu_v3())
> > + return -EINVAL;
>
> This check is pointless; the vCPU feature flags have been sanitised at
> this point, and a requirement of having PMUv3 is that this predicate is
> true.
>
Oh yes. I'll avoid this in v8.
> > + /*
> > + * When the vCPU has a PMU, but no PMU is set for the guest
> > + * yet, set the default one.
> > + */
> > + if (unlikely(!kvm->arch.arm_pmu))
> > + return kvm_arm_set_default_pmu(kvm);
> > +
> > + return 0;
> > +}
> > +
>
> Apologies, I believe I was unclear last time around as to what I was
> wanting here. Let's call this thing kvm_setup_vcpu() such that we can
> add other one-time setup activities to it in the future.
>
> Something like:
>
> diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> index 96641e442039..4896a44108e0 100644
> --- a/arch/arm64/kvm/arm.c
> +++ b/arch/arm64/kvm/arm.c
> @@ -1265,19 +1265,17 @@ static bool kvm_vcpu_init_changed(struct kvm_vcpu *vcpu,
> KVM_VCPU_MAX_FEATURES);
> }
>
> -static int kvm_vcpu_set_pmu(struct kvm_vcpu *vcpu)
> +static int kvm_setup_vcpu(struct kvm_vcpu *vcpu)
> {
> struct kvm *kvm = vcpu->kvm;
>
> - if (!kvm_arm_support_pmu_v3())
> - return -EINVAL;
> -
> /*
> * When the vCPU has a PMU, but no PMU is set for the guest
> * yet, set the default one.
> */
> - if (unlikely(!kvm->arch.arm_pmu))
> - return kvm_arm_set_default_pmu(kvm);
> + if (kvm_vcpu_has_pmu(vcpu) && !kvm->arch.arm_pmu &&
> + kvm_arm_set_default_pmu(kvm))
> + return -EINVAL;
>
> return 0;
> }
> @@ -1297,7 +1295,8 @@ static int __kvm_vcpu_set_target(struct kvm_vcpu *vcpu,
>
> bitmap_copy(kvm->arch.vcpu_features, &features, KVM_VCPU_MAX_FEATURES);
>
> - if (kvm_vcpu_has_pmu(vcpu) && kvm_vcpu_set_pmu(vcpu))
> + ret = kvm_setup_vcpu(vcpu);
> + if (ret)
> goto out_unlock;
>
> /* Now we know what it is, we can reset it. */
>
Introducing kvm_setup_vcpu() seems better than directly calling
kvm_vcpu_set_pmu(), which feels like it's crashing a party.
Thank you.
Raghavendra
> --
> Thanks,
> Oliver