Re: [PATCH] KVM: selftests: x86: Prioritize getting max_gfn from GuestPhysBits

[Sean Christopherson <seanjc@google.com>]

On Wed, May 08, 2024, Tao Su wrote:
> Use the max mappable GPA via GuestPhysBits advertised by KVM to calculate
> max_gfn. Currently some selftests (e.g. access_tracking_perf_test,
> dirty_log_test...) add RAM regions close to max_gfn, so guest may access
> GPA beyond its mappable range and cause infinite loop.
> 
> Prioritize obtaining pa_bits from the GuestPhysBits field advertised by
> KVM, so max_gfn can be limited to the mappable range.
> 
> Signed-off-by: Tao Su <tao1.su@linux.intel.com>
> ---
> This patch is based on https://github.com/kvm-x86/linux/commit/b628cb523c65
> ---
>  tools/testing/selftests/kvm/include/x86_64/processor.h | 1 +
>  tools/testing/selftests/kvm/lib/x86_64/processor.c     | 4 +++-
>  2 files changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/kvm/include/x86_64/processor.h b/tools/testing/selftests/kvm/include/x86_64/processor.h
> index 81ce37ec407d..ff99f66d81a0 100644
> --- a/tools/testing/selftests/kvm/include/x86_64/processor.h
> +++ b/tools/testing/selftests/kvm/include/x86_64/processor.h
> @@ -282,6 +282,7 @@ struct kvm_x86_cpu_property {
>  #define X86_PROPERTY_MAX_EXT_LEAF		KVM_X86_CPU_PROPERTY(0x80000000, 0, EAX, 0, 31)
>  #define X86_PROPERTY_MAX_PHY_ADDR		KVM_X86_CPU_PROPERTY(0x80000008, 0, EAX, 0, 7)
>  #define X86_PROPERTY_MAX_VIRT_ADDR		KVM_X86_CPU_PROPERTY(0x80000008, 0, EAX, 8, 15)
> +#define X86_PROPERTY_MAX_GUEST_PHY_ADDR		KVM_X86_CPU_PROPERTY(0x80000008, 0, EAX, 16, 23)
>  #define X86_PROPERTY_SEV_C_BIT			KVM_X86_CPU_PROPERTY(0x8000001F, 0, EBX, 0, 5)
>  #define X86_PROPERTY_PHYS_ADDR_REDUCTION	KVM_X86_CPU_PROPERTY(0x8000001F, 0, EBX, 6, 11)
>  
> diff --git a/tools/testing/selftests/kvm/lib/x86_64/processor.c b/tools/testing/selftests/kvm/lib/x86_64/processor.c
> index 74a4c736c9ae..6c69f1dfeed2 100644
> --- a/tools/testing/selftests/kvm/lib/x86_64/processor.c
> +++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c
> @@ -1074,7 +1074,9 @@ void kvm_get_cpu_address_width(unsigned int *pa_bits, unsigned int *va_bits)
>  		*pa_bits = kvm_cpu_has(X86_FEATURE_PAE) ? 36 : 32;
>  		*va_bits = 32;
>  	} else {
> -		*pa_bits = kvm_cpu_property(X86_PROPERTY_MAX_PHY_ADDR);
> +		*pa_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR);
> +		if (*pa_bits == 0)
> +			*pa_bits = kvm_cpu_property(X86_PROPERTY_MAX_PHY_ADDR);

This is the wrong place to incorporaate the max mappable GPA.  The pa_bits field
should reflect the "real" MAXPHYADDR, it's vm->max_gfn that needs to be adjusted,
and x86 selftests already overrides vm_compute_max_gfn() specifically to deal with
goofy edge cases like this.

>  		*va_bits = kvm_cpu_property(X86_PROPERTY_MAX_VIRT_ADDR);
>  	}
>  }
> 
> base-commit: dccb07f2914cdab2ac3a5b6c98406f765acab803
> -- 
> 2.34.1
>