[PATCH] KVM: x86: avoid out of bounds indices for fixed performance counters

[PATCH] KVM: x86: avoid out of bounds indices for fixed performance counters

[Paolo Bonzini]

Because IceLake has 4 fixed performance counters but KVM only supports 3,
it is possible for reprogram_fixed_counters to pass to
reprogram_fixed_counter an index that is out of bounds for
the fixed_pmc_events array.

Ultimately intel_find_fixed_event, which is the only place that uses
fixed_pmc_events, handles this correctly because it checks against the
size of fixed_pmc_events anyway.  Every other place
operates on the fixed_counters[] array which is sized
according to INTEL_PMC_MAX_FIXED.  However, it is cleaner if
the unsupported performance counters are culled early on
in reprogram_fixed_counters.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/kvm/vmx/pmu_intel.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
index 1b7456b2177b..d33e9799276e 100644
--- a/arch/x86/kvm/vmx/pmu_intel.c
+++ b/arch/x86/kvm/vmx/pmu_intel.c
@@ -91,7 +91,7 @@ static unsigned intel_find_fixed_event(int idx)
 	u32 event;
 	size_t size = ARRAY_SIZE(fixed_pmc_events);
 
-	if (idx >= size)
+	if (WARN_ON_ONCE(idx >= size))
 		return PERF_COUNT_HW_MAX;
 
 	event = fixed_pmc_events[array_index_nospec(idx, size)];
@@ -500,8 +500,9 @@ static void intel_pmu_refresh(struct kvm_vcpu *vcpu)
 		pmu->nr_arch_fixed_counters = 0;
 	} else {
 		pmu->nr_arch_fixed_counters =
-			min_t(int, edx.split.num_counters_fixed,
-			      x86_pmu.num_counters_fixed);
+			min3(ARRAY_SIZE(fixed_pmc_events),
+			     (size_t) edx.split.num_counters_fixed,
+			     (size_t) x86_pmu.num_counters_fixed);
 		edx.split.bit_width_fixed = min_t(int,
 			edx.split.bit_width_fixed, x86_pmu.bit_width_fixed);
 		pmu->counter_bitmask[KVM_PMC_FIXED] =
-- 
2.31.1

Re: [PATCH] KVM: x86: avoid out of bounds indices for fixed performance counters

[Like Xu]

On 10/12/2021 3:11 am, Paolo Bonzini wrote:
> Because IceLake has 4 fixed performance counters but KVM only supports 3,
> it is possible for reprogram_fixed_counters to pass to

Emm, it's possible for a unwise user space or smart syzkaller.

> reprogram_fixed_counter an index that is out of bounds for
> the fixed_pmc_events array.
> 
> Ultimately intel_find_fixed_event, which is the only place that uses
> fixed_pmc_events, handles this correctly because it checks against the
> size of fixed_pmc_events anyway.  Every other place
> operates on the fixed_counters[] array which is sized
> according to INTEL_PMC_MAX_FIXED.  However, it is cleaner if
> the unsupported performance counters are culled early on
> in reprogram_fixed_counters.

How about introducing a static "struct x86_pmu_capability" variable [1] so that 
we can

(1) setup num_counters_fixed just once in the kvm_init_pmu_capability(), and
(2) avoid repeated calls to perf_get_x86_pmu_capability() ;

?

[1] https://lore.kernel.org/kvm/20210806133802.3528-17-lingshan.zhu@intel.com/
By the way, do you need a re-based version of the guest PBES feature ?

Thanks,
Like Xu

> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>   arch/x86/kvm/vmx/pmu_intel.c | 7 ++++---
>   1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
> index 1b7456b2177b..d33e9799276e 100644
> --- a/arch/x86/kvm/vmx/pmu_intel.c
> +++ b/arch/x86/kvm/vmx/pmu_intel.c
> @@ -91,7 +91,7 @@ static unsigned intel_find_fixed_event(int idx)
>   	u32 event;
>   	size_t size = ARRAY_SIZE(fixed_pmc_events);
>   
> -	if (idx >= size)
> +	if (WARN_ON_ONCE(idx >= size))
>   		return PERF_COUNT_HW_MAX;
>   
>   	event = fixed_pmc_events[array_index_nospec(idx, size)];
> @@ -500,8 +500,9 @@ static void intel_pmu_refresh(struct kvm_vcpu *vcpu)
>   		pmu->nr_arch_fixed_counters = 0;
>   	} else {
>   		pmu->nr_arch_fixed_counters =
> -			min_t(int, edx.split.num_counters_fixed,
> -			      x86_pmu.num_counters_fixed);
> +			min3(ARRAY_SIZE(fixed_pmc_events),
> +			     (size_t) edx.split.num_counters_fixed,
> +			     (size_t) x86_pmu.num_counters_fixed);
>   		edx.split.bit_width_fixed = min_t(int,
>   			edx.split.bit_width_fixed, x86_pmu.bit_width_fixed);
>   		pmu->counter_bitmask[KVM_PMC_FIXED] =

Re: [PATCH] KVM: x86: avoid out of bounds indices for fixed performance counters

[Paolo Bonzini]

On 12/10/21 09:59, Like Xu wrote:
> 
> How about introducing a static "struct x86_pmu_capability" variable [1] 
> so that we can
> 
> (1) setup num_counters_fixed just once in the kvm_init_pmu_capability(), 
> and
> (2) avoid repeated calls to perf_get_x86_pmu_capability() ;

Yes, that works.

> [1] 
> https://lore.kernel.org/kvm/20210806133802.3528-17-lingshan.zhu@intel.com/
> By the way, do you need a re-based version of the guest PBES feature ?

Sure, please send one.

Paolo