KVM: x86/pmu: Add support for AMD Host-Only/Guest-Only bits
Subsystems (5)
v3 cover KVM: x86/pmu: Add support for AMD Host-Only/Guest-Only bits
This series adds support for AMD's Host-Only and Guest-Only performance counter eventsel bits in KVM's mediated PMU passthrough implementation.
These bits allow an nSVM-enabled guest to configure performance counters that count only during L1 execution (Host-Only) or only during L2 execution (Guest-Only).
KVM updates the hardware event selector ENABLE bit at the following state transitions to ensure counters only count in the appropriate mode:
- EFER.SVME changes: Enable/disable Guest-Only counters
- Nested VMRUN: Disable Host-Only, enable Guest-Only counters
- Nested VMEXIT: Enable Host-Only, disable Guest-Only counters
v1: https://lore.kernel.org/kvm/[email protected]/
v2: https://lore.kernel.org/kvm/[email protected]/
v2 -> v3:
* Dropped the svm_enter_guest_mode() and svm_leave_guest_mode() wrappers introduced in v2 [Yosry]
* Introduced a generic nested_transition callback in kvm_x86_ops to avoid confusing SVM-specific wrappers and unnecessary code replication
* Fixed a latent bug with L2 stack alignment, which was triggered by a movdqa instruction in l2_guest_code() that referenced the L2 stack. Note that l2_guest_code() expects the stack to be 16-byte misaligned at function entry. It was not.
Jim Mattson (5):
KVM: x86/pmu: Introduce amd_pmu_set_eventsel_hw()
KVM: x86/pmu: Disable Host-Only/Guest-Only events as appropriate for
vCPU state
KVM: x86/pmu: Refresh Host-Only/Guest-Only eventsel at nested
transitions
KVM: x86/pmu: Allow Host-Only/Guest-Only bits with nSVM and mediated
PMU
KVM: selftests: x86: Add svm_pmu_host_guest_test for
Host-Only/Guest-Only bits
arch/x86/include/asm/kvm-x86-ops.h | 1 + arch/x86/include/asm/kvm_host.h | 2 + arch/x86/include/asm/perf_event.h | 2 + arch/x86/kvm/kvm_cache_regs.h | 2 + arch/x86/kvm/svm/pmu.c | 42 +++- arch/x86/kvm/svm/svm.c | 3 + arch/x86/kvm/svm/svm.h | 5 + arch/x86/kvm/x86.c | 1 + tools/testing/selftests/kvm/Makefile.kvm | 1 + tools/testing/selftests/kvm/include/x86/pmu.h | 6 + .../selftests/kvm/include/x86/processor.h | 2 + .../kvm/x86/svm_pmu_host_guest_test.c | 199 ++++++++++++++++++ 12 files changed, 264 insertions(+), 2 deletions(-) create mode 100644 tools/testing/selftests/kvm/x86/svm_pmu_host_guest_test.c
Signature · 1 line
2.53.0.rc2.204.g2597b5adb4-goog
PATCH v3 1/5 KVM: x86/pmu: Introduce amd_pmu_set_eventsel_hw()
Extract the computation of eventsel_hw from amd_pmu_set_msr() into a separate helper function, amd_pmu_set_eventsel_hw().
No functional change intended.
--- arch/x86/kvm/svm/pmu.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)
arch/x86/kvm/svm/pmu.c +7 -2
@@ -147,6 +147,12 @@ static int amd_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)return 1 ; } +static void amd_pmu_set_eventsel_hw (struct kvm_pmc *pmc) +{ + pmc->eventsel_hw = (pmc->eventsel & ~AMD64_EVENTSEL_HOSTONLY ) | +AMD64_EVENTSEL_GUESTONLY ; +} +static int amd_pmu_set_msr (struct kvm_vcpu *vcpu,struct msr_data *msr_info) {struct kvm_pmu *pmu =vcpu_to_pmu (vcpu);
@@ -166,8 +172,7 @@ static int amd_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) data &= ~pmu->reserved_bits;if (data != pmc->eventsel) { pmc->eventsel = data; - pmc->eventsel_hw = (data & ~AMD64_EVENTSEL_HOSTONLY ) | -AMD64_EVENTSEL_GUESTONLY ; +amd_pmu_set_eventsel_hw (pmc);kvm_pmu_request_counter_reprogram (pmc); }return 0 ;
Signature · 1 line
2.53.0.rc2.204.g2597b5adb4-goog
PATCH v3 2/5 KVM: x86/pmu: Disable Host-Only/Guest-Only events as appropriate for vCPU state
In reply to Jim MattsonUpdate amd_pmu_set_eventsel_hw() to clear the event selector's hardware enable bit when the PMC should not count based on the guest's Host-Only and Guest-Only event selector bits and the current vCPU state.
--- arch/x86/include/asm/perf_event.h | 2 ++ arch/x86/kvm/svm/pmu.c | 18 ++++++++++++++++++ 2 files changed, 20 insertions(+)
arch/x86/include/asm/perf_event.h +2 -0
@@ -58,6 +58,8 @@#define AMD64_EVENTSEL_INT_CORE_ENABLE (1ULL << 36)#define AMD64_EVENTSEL_GUESTONLY (1ULL << 40)#define AMD64_EVENTSEL_HOSTONLY (1ULL << 41) +#define AMD64_EVENTSEL_HOST_GUEST_MASK \ + (AMD64_EVENTSEL_HOSTONLY | AMD64_EVENTSEL_GUESTONLY)#define AMD64_EVENTSEL_INT_CORE_SEL_SHIFT 37#define AMD64_EVENTSEL_INT_CORE_SEL_MASK \
arch/x86/kvm/svm/pmu.c +18 -0
@@ -149,8 +149,26 @@ static int amd_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)static void amd_pmu_set_eventsel_hw (struct kvm_pmc *pmc) { +struct kvm_vcpu *vcpu = pmc->vcpu; +u64 host_guest_bits; + pmc->eventsel_hw = (pmc->eventsel & ~AMD64_EVENTSEL_HOSTONLY ) |AMD64_EVENTSEL_GUESTONLY ; + +if (!(pmc->eventsel &ARCH_PERFMON_EVENTSEL_ENABLE )) +return ; + +if (!(vcpu->arch.efer &EFER_SVME )) +return ; + + host_guest_bits = pmc->eventsel &AMD64_EVENTSEL_HOST_GUEST_MASK ; +if (!host_guest_bits || host_guest_bits ==AMD64_EVENTSEL_HOST_GUEST_MASK ) +return ; + +if (!!(host_guest_bits &AMD64_EVENTSEL_GUESTONLY ) ==is_guest_mode (vcpu)) +return ; + + pmc->eventsel_hw &= ~ARCH_PERFMON_EVENTSEL_ENABLE ; }static int amd_pmu_set_msr (struct kvm_vcpu *vcpu,struct msr_data *msr_info)
Signature · 1 line
2.53.0.rc2.204.g2597b5adb4-goog
Re: [PATCH v3 2/5] KVM: x86/pmu: Disable Host-Only/Guest-Only events as appropriate for vCPU state
Quoted 39 lines · Jim Mattson
Update
amd_pmu_set_eventsel_hw()to clear the event selector's hardware enable bit when the PMC should not count based on the guest's Host-Only and Guest-Only event selector bits and the current vCPU state.Signed-off-by Jim Mattson <[email protected]>--- arch/x86/include/asm/perf_event.h | 2 ++ arch/x86/kvm/svm/pmu.c | 18 ++++++++++++++++++ 2 files changed, 20 insertions(+)arch/x86/include/asm/perf_event.h +2
-0@@ -58,6 +58,8 @@#define AMD64_EVENTSEL_INT_CORE_ENABLE (1ULL << 36)#define AMD64_EVENTSEL_GUESTONLY (1ULL << 40)#define AMD64_EVENTSEL_HOSTONLY (1ULL << 41) +#define AMD64_EVENTSEL_HOST_GUEST_MASK \ + (AMD64_EVENTSEL_HOSTONLY | AMD64_EVENTSEL_GUESTONLY)#define AMD64_EVENTSEL_INT_CORE_SEL_SHIFT 37#define AMD64_EVENTSEL_INT_CORE_SEL_MASK \arch/x86/kvm/svm/pmu.c +6
-0@@ -149,8 +149,26 @@ static int amd_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)static void amd_pmu_set_eventsel_hw (struct kvm_pmc *pmc) { +struct kvm_vcpu *vcpu = pmc->vcpu; +u64 host_guest_bits; + pmc->eventsel_hw = (pmc->eventsel & ~AMD64_EVENTSEL_HOSTONLY ) |AMD64_EVENTSEL_GUESTONLY ; + +if (!(pmc->eventsel &ARCH_PERFMON_EVENTSEL_ENABLE )) +return ;
arch/x86/kvm/svm/pmu.c static int amd_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) · line 169
Quoted 10 lines
arch/x86/kvm/svm/pmu.c +10
-0+ +if (!(vcpu->arch.efer &EFER_SVME )) +return ; + + host_guest_bits = pmc->eventsel &AMD64_EVENTSEL_HOST_GUEST_MASK ; +if (!host_guest_bits || host_guest_bits ==AMD64_EVENTSEL_HOST_GUEST_MASK ) +return ; + +if (!!(host_guest_bits &AMD64_EVENTSEL_GUESTONLY ) ==is_guest_mode (vcpu)) +return ;
This seems to disable the PMCs after exits from an L2 guest to the L0 hypervisor. For such transitions, the corresponding L1 vCPU's PMC has GuestOnly set but is_guest_mode() is false as this function is called at the very end of leave_guest_mode() after vcpu->stat.guest_mode is set to 0.
Is this a correct interpretation of the condition above?
Quoted 5 lines
arch/x86/kvm/svm/pmu.c +2
-0+ + pmc->eventsel_hw &= ~ARCH_PERFMON_EVENTSEL_ENABLE ; }static int amd_pmu_set_msr (struct kvm_vcpu *vcpu,struct msr_data *msr_info)
Re: [PATCH v3 2/5] KVM: x86/pmu: Disable Host-Only/Guest-Only events as appropriate for vCPU state
Quoted 57 lines · Sandipan Das
Quoted 39 lines · Jim Mattson
Update
amd_pmu_set_eventsel_hw()to clear the event selector's hardware enable bit when the PMC should not count based on the guest's Host-Only and Guest-Only event selector bits and the current vCPU state.Signed-off-by Jim Mattson <[email protected]>--- arch/x86/include/asm/perf_event.h | 2 ++ arch/x86/kvm/svm/pmu.c | 18 ++++++++++++++++++ 2 files changed, 20 insertions(+)arch/x86/include/asm/perf_event.h +2
-0@@ -58,6 +58,8 @@#define AMD64_EVENTSEL_INT_CORE_ENABLE (1ULL << 36)#define AMD64_EVENTSEL_GUESTONLY (1ULL << 40)#define AMD64_EVENTSEL_HOSTONLY (1ULL << 41) +#define AMD64_EVENTSEL_HOST_GUEST_MASK \ + (AMD64_EVENTSEL_HOSTONLY | AMD64_EVENTSEL_GUESTONLY)#define AMD64_EVENTSEL_INT_CORE_SEL_SHIFT 37#define AMD64_EVENTSEL_INT_CORE_SEL_MASK \arch/x86/kvm/svm/pmu.c +6
-0@@ -149,8 +149,26 @@ static int amd_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)static void amd_pmu_set_eventsel_hw (struct kvm_pmc *pmc) { +struct kvm_vcpu *vcpu = pmc->vcpu; +u64 host_guest_bits; + pmc->eventsel_hw = (pmc->eventsel & ~AMD64_EVENTSEL_HOSTONLY ) |AMD64_EVENTSEL_GUESTONLY ; + +if (!(pmc->eventsel &ARCH_PERFMON_EVENTSEL_ENABLE )) +return ;arch/x86/kvm/svm/pmu.c static int amd_pmu_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) · line 169
Quoted 10 lines
arch/x86/kvm/svm/pmu.c +10
-0+ +if (!(vcpu->arch.efer &EFER_SVME )) +return ; + + host_guest_bits = pmc->eventsel &AMD64_EVENTSEL_HOST_GUEST_MASK ; +if (!host_guest_bits || host_guest_bits ==AMD64_EVENTSEL_HOST_GUEST_MASK ) +return ; + +if (!!(host_guest_bits &AMD64_EVENTSEL_GUESTONLY ) ==is_guest_mode (vcpu)) +return ;This seems to disable the PMCs after exits from an L2 guest to the L0 hypervisor. For such transitions, the corresponding L1 vCPU's PMC has GuestOnly set but
is_guest_mode()is false as this function is called at the very end ofleave_guest_mode()aftervcpu->stat.guest_modeis set to 0.
Quoted 1 line
Is this a correct interpretation of the condition above?
I think you are confusing the VMCB02 bits (that always have Guest-Only set) and the VMCB12 bits (which are under the control of L1.
If L1 sets only Guest-Only, that indicates a desire to monitor only L2 execution. So, yes, after emulating an exit from L2 to L1, the PMC is disabled.
Quoted 6 lines
Quoted 5 lines
arch/x86/kvm/svm/pmu.c +2
-0+ + pmc->eventsel_hw &= ~ARCH_PERFMON_EVENTSEL_ENABLE ; }static int amd_pmu_set_msr (struct kvm_vcpu *vcpu,struct msr_data *msr_info)
PATCH v3 3/5 KVM: x86/pmu: Refresh Host-Only/Guest-Only eventsel at nested transitions
In reply to Jim MattsonAdd amd_pmu_refresh_host_guest_eventsel_hw() to recalculate eventsel_hw for all PMCs based on the current vCPU state. This is needed because Host-Only and Guest-Only counters must be enabled/disabled at:
- SVME changes: When EFER.SVME is modified, counters with Guest-Only bits need their hardware enable state updated.
- Nested transitions: When entering or leaving guest mode, Host-Only counters should be disabled/enabled and Guest-Only counters should be enabled/disabled accordingly.
Add a nested_transition() callback to kvm_x86_ops and call it from enter_guest_mode() and leave_guest_mode() to ensure the PMU state stays synchronized with guest mode transitions.
--- arch/x86/include/asm/kvm-x86-ops.h | 1 + arch/x86/include/asm/kvm_host.h | 2 ++ arch/x86/kvm/kvm_cache_regs.h | 2 ++ arch/x86/kvm/svm/pmu.c | 12 ++++++++++++ arch/x86/kvm/svm/svm.c | 3 +++ arch/x86/kvm/svm/svm.h | 5 +++++ arch/x86/kvm/x86.c | 1 + 7 files changed, 26 insertions(+)
arch/x86/include/asm/kvm-x86-ops.h +1 -0
@@ -108,6 +108,7 @@ KVM_X86_OP(get_entry_info) KVM_X86_OP(check_intercept) KVM_X86_OP(handle_exit_irqoff) KVM_X86_OP_OPTIONAL(update_cpu_dirty_logging) +KVM_X86_OP_OPTIONAL(nested_transition) KVM_X86_OP_OPTIONAL(vcpu_blocking) KVM_X86_OP_OPTIONAL(vcpu_unblocking) KVM_X86_OP_OPTIONAL(pi_update_irte)
arch/x86/include/asm/kvm_host.h +2 -0
@@ -1901,6 +1901,8 @@ struct kvm_x86_ops {void (*update_cpu_dirty_logging)(struct kvm_vcpu *vcpu); +void (*nested_transition)(struct kvm_vcpu *vcpu); +const struct kvm_x86_nested_ops *nested_ops;void (*vcpu_blocking)(struct kvm_vcpu *vcpu);
arch/x86/kvm/kvm_cache_regs.h +2 -0
@@ -227,6 +227,7 @@ static inline void enter_guest_mode(struct kvm_vcpu *vcpu) { vcpu->arch.hflags |=HF_GUEST_MASK ; vcpu->stat.guest_mode =1 ; +kvm_x86_call (nested_transition)(vcpu); }static inline void leave_guest_mode (struct kvm_vcpu *vcpu)
@@ -239,6 +240,7 @@ static inline void leave_guest_mode(struct kvm_vcpu *vcpu) } vcpu->stat.guest_mode =0 ; +kvm_x86_call (nested_transition)(vcpu); }static inline bool is_guest_mode (struct kvm_vcpu *vcpu)
arch/x86/kvm/svm/pmu.c +12 -0
@@ -171,6 +171,18 @@ static void amd_pmu_set_eventsel_hw(struct kvm_pmc *pmc) pmc->eventsel_hw &= ~ARCH_PERFMON_EVENTSEL_ENABLE ; } +void amd_pmu_refresh_host_guest_eventsel_hw (struct kvm_vcpu *vcpu) +{ +struct kvm_pmu *pmu =vcpu_to_pmu (vcpu); +int i; + +if (pmu->reserved_bits &AMD64_EVENTSEL_HOST_GUEST_MASK ) +return ; + +for (i =0 ; i < pmu->nr_arch_gp_counters; i++) +amd_pmu_set_eventsel_hw (&pmu->gp_counters[i]); +} +static int amd_pmu_set_msr (struct kvm_vcpu *vcpu,struct msr_data *msr_info) {struct kvm_pmu *pmu =vcpu_to_pmu (vcpu);
arch/x86/kvm/svm/svm.c +3 -0
@@ -244,6 +244,8 @@ int svm_set_efer(struct kvm_vcpu *vcpu, u64 efer)if (svm_gp_erratum_intercept && !sev_guest (vcpu->kvm))set_exception_intercept (svm,GP_VECTOR ); } + +amd_pmu_refresh_host_guest_eventsel_hw (vcpu); } svm->vmcb->save.efer = efer |EFER_SVME ;
@@ -5222,6 +5224,7 @@ struct kvm_x86_ops svm_x86_ops __initdata = { .check_intercept = svm_check_intercept, .handle_exit_irqoff = svm_handle_exit_irqoff, + .nested_transition = amd_pmu_refresh_host_guest_eventsel_hw, .nested_ops = &svm_nested_ops,
arch/x86/kvm/svm/svm.h +5 -0
@@ -864,6 +864,11 @@ void sev_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector);void sev_es_prepare_switch_to_guest (struct vcpu_svm *svm,struct sev_es_save_area *hostsa);void sev_es_unmap_ghcb (struct vcpu_svm *svm); + +/* pmu.c */ +void amd_pmu_refresh_host_guest_eventsel_hw (struct kvm_vcpu *vcpu); + +#ifdef CONFIG_KVM_AMD_SEV int sev_mem_enc_ioctl (struct kvm *kvm,void __user *argp); int sev_mem_enc_register_region(struct kvm *kvm,
arch/x86/kvm/x86.c +1 -0
@@ -150,6 +150,7 @@ struct kvm_x86_ops kvm_x86_ops __read_mostly;#include <asm/kvm-x86-ops.h> EXPORT_STATIC_CALL_GPL (kvm_x86_get_cs_db_l_bits);EXPORT_STATIC_CALL_GPL (kvm_x86_cache_reg); +EXPORT_STATIC_CALL_GPL (kvm_x86_nested_transition);static bool __read_mostly ignore_msrs =0 ;module_param (ignore_msrs, bool,0644 );
Signature · 1 line
2.53.0.rc2.204.g2597b5adb4-goog
Sean Christopherson
maintainerRe: [PATCH v3 3/5] KVM: x86/pmu: Refresh Host-Only/Guest-Only eventsel at nested transitions
Quoted 7 lines · Jim Mattson
Add
amd_pmu_refresh_host_guest_eventsel_hw()to recalculateeventsel_hwfor all PMCs based on the current vCPU state. This is needed because Host-Only and Guest-Only counters must be enabled/disabled at:- SVME changes: When EFER.SVME is modified, counters with Guest-Only bits need their hardware enable state updated.
Quoted 7 lines
- Nested transitions: When entering or leaving guest mode, Host-Only counters should be disabled/enabled and Guest-Only counters should be enabled/disabled accordingly.
Add a
nested_transition()callback tokvm_x86_opsand call it fromenter_guest_mode()andleave_guest_mode()to ensure the PMU state stays synchronized with guest mode transitions.
Blech, I'm not a fan of this kvm_x86_ops hook. I especially don't like calling out to vendor code from {enter,leave}_guest_mode(). The subtle dependency on vcpu-arch.efer being up-to-date in svm_set_efer() is a little nasty too.
More importantly, I think this series is actively buggy, as I don't see anything in amd_pmu_refresh_host_guest_eventsel_hw() that restricts it to the mediated PMU. And I'm pretty sure that path will bypass the PMU event filter. And I believe kvm_pmu_recalc_pmc_emulation() also needs to be invoked so that emulated instructions are counted correctly.
To avoid ordering issues and bugs where event filtering and guest/host handling clobber each other, I think we should funnel all processing through KVM_REQ_PMU, and then do something like this:
arch/x86/kvm/kvm_cache_regs.h +4 -2
@@ -227,7 +227,8 @@ static inline void enter_guest_mode(struct kvm_vcpu *vcpu) { vcpu->arch.hflags |=HF_GUEST_MASK ; vcpu->stat.guest_mode =1 ; -kvm_x86_call (nested_transition)(vcpu); + +kvm_pmu_handle_nested_transition (); }static inline void leave_guest_mode (struct kvm_vcpu *vcpu)
@@ -240,7 +241,8 @@ static inline void leave_guest_mode(struct kvm_vcpu *vcpu) } vcpu->stat.guest_mode =0 ; -kvm_x86_call (nested_transition)(vcpu); + +kvm_pmu_handle_nested_transition (); }static inline bool is_guest_mode (struct kvm_vcpu *vcpu)
arch/x86/kvm/pmu.h +12 -0
@@ -244,6 +244,18 @@ static inline bool kvm_pmu_is_fastpath_emulation_allowed(struct kvm_vcpu *vcpu) X86_PMC_IDX_MAX); } +static inline void kvm_pmu_handle_nested_transition (struct kvm_vcpu *vcpu) +{ +if (!kvm_vcpu_has_mediated_pmu (vcpu)) +return ; + +if (vcpu_to_pmu (vcpu)->reserved_bits &AMD64_EVENTSEL_HOST_GUEST_MASK ) +return ; + +atomic64_set (&vcpu_to_pmu (vcpu)->__reprogram_pmi,-1ull ); +kvm_make_request (KVM_REQ_PMU , vcpu); +} +void kvm_pmu_deliver_pmi (struct kvm_vcpu *vcpu);void kvm_pmu_handle_event (struct kvm_vcpu *vcpu);int kvm_pmu_rdpmc (struct kvm_vcpu *vcpu,unsigned pmc,u64 *data);
PATCH v3 4/5 KVM: x86/pmu: Allow Host-Only/Guest-Only bits with nSVM and mediated PMU
In reply to Jim MattsonIf the vCPU advertises SVM and uses the mediated PMU, allow the guest to set the Host-Only and Guest-Only bits in the event selector MSRs.
--- arch/x86/kvm/svm/pmu.c | 3 +++ 1 file changed, 3 insertions(+)
arch/x86/kvm/svm/pmu.c +3 -0
@@ -243,6 +243,9 @@ static void amd_pmu_refresh(struct kvm_vcpu *vcpu) pmu->counter_bitmask[KVM_PMC_GP ] =BIT_ULL (48 ) -1 ; pmu->reserved_bits =0xfffffff000280000ull ; +if (guest_cpu_cap_has (vcpu,X86_FEATURE_SVM ) && +kvm_vcpu_has_mediated_pmu (vcpu)) + pmu->reserved_bits &= ~AMD64_EVENTSEL_HOST_GUEST_MASK ; pmu->raw_event_mask =AMD64_RAW_EVENT_MASK ;/* not applicable to AMD; but clean them to prevent any fall out */ pmu->counter_bitmask[KVM_PMC_FIXED ] =0 ;
Signature · 1 line
2.53.0.rc2.204.g2597b5adb4-goog
PATCH v3 5/5 KVM: selftests: x86: Add svm_pmu_host_guest_test for Host-Only/Guest-Only bits
In reply to Jim MattsonAdd a selftest to verify KVM correctly virtualizes the AMD PMU Host-Only (bit 41) and Guest-Only (bit 40) event selector bits across all relevant SVM state transitions.
The test programs 4 PMCs simultaneously with all combinations of the Host-Only and Guest-Only bits, then verifies correct counting behavior:
1. SVME=0: all counters count (Host-Only/Guest-Only bits ignored)
2. Set SVME=1: Host-Only and neither/both count; Guest-Only stops
3. VMRUN to L2: Guest-Only and neither/both count; Host-Only stops
4. VMEXIT to L1: Host-Only and neither/both count; Guest-Only stops
5. Clear SVME=0: all counters count (bits ignored again)
--- tools/testing/selftests/kvm/Makefile.kvm | 1 + tools/testing/selftests/kvm/include/x86/pmu.h | 6 + .../selftests/kvm/include/x86/processor.h | 2 + .../kvm/x86/svm_pmu_host_guest_test.c | 199 ++++++++++++++++++ 4 files changed, 208 insertions(+) create mode 100644 tools/testing/selftests/kvm/x86/svm_pmu_host_guest_test.c
tools/testing/selftests/kvm/Makefile.kvm +1 -0
@@ -112,6 +112,7 @@ TEST_GEN_PROGS_x86 += x86/svm_vmcall_test TEST_GEN_PROGS_x86 += x86/svm_int_ctl_test TEST_GEN_PROGS_x86 += x86/svm_nested_shutdown_test TEST_GEN_PROGS_x86 += x86/svm_nested_soft_inject_test +TEST_GEN_PROGS_x86 += x86/svm_pmu_host_guest_test TEST_GEN_PROGS_x86 += x86/tsc_scaling_sync TEST_GEN_PROGS_x86 += x86/sync_regs_test TEST_GEN_PROGS_x86 += x86/ucna_injection_test
tools/testing/selftests/kvm/include/x86/pmu.h +6 -0
@@ -38,6 +38,12 @@#define ARCH_PERFMON_EVENTSEL_INV BIT_ULL(23)#define ARCH_PERFMON_EVENTSEL_CMASK GENMASK_ULL(31, 24) +/* +* These are AMD-specific bits. +*/ +#define AMD64_EVENTSEL_GUESTONLY BIT_ULL(40) +#define AMD64_EVENTSEL_HOSTONLY BIT_ULL(41) +/* RDPMC control flags, Intel only. */ #define INTEL_RDPMC_METRICS BIT_ULL(29)#define INTEL_RDPMC_FIXED BIT_ULL(30)
tools/testing/selftests/kvm/include/x86/processor.h +2 -0
@@ -19,6 +19,8 @@#include "kvm_util.h" #include "ucall_common.h" +#define __stack_aligned__ __aligned(16) +extern bool host_cpu_is_intel;extern bool host_cpu_is_amd;extern uint64_t guest_tsc_khz;
tools/testing/selftests/kvm/x86/svm_pmu_host_guest_test.c new file +199 -0
199 lines not on this page · raw message
Signature · 1 line
2.53.0.rc2.204.g2597b5adb4-goog