| Message ID | 20240510020346.12528-1-tao1.su@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] KVM: selftests: x86: Prioritize getting max_gfn from GuestPhysBits | expand |
On 5/10/2024 10:03 AM, 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. > > Adjust max_gfn in vm_compute_max_gfn() since x86 selftests already > overrides vm_compute_max_gfn() specifically to deal with goofy edge cases. > > Signed-off-by: Tao Su <tao1.su@linux.intel.com> > Tested-by: Yi Lai <yi1.lai@intel.com> > --- > This patch is based on https://github.com/kvm-x86/linux/commit/b628cb523c65 > > Changelog: > v1 -> v2: > - Only adjust vm->max_gfn in vm_compute_max_gfn() > - Add Yi Lai's Tested-by > > v1: https://lore.kernel.org/all/20240508064205.15301-1-tao1.su@linux.intel.com/ > --- > tools/testing/selftests/kvm/include/x86_64/processor.h | 1 + > tools/testing/selftests/kvm/lib/x86_64/processor.c | 10 ++++++++-- > 2 files changed, 9 insertions(+), 2 deletions(-) > > 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..aa9966ead543 100644 > --- a/tools/testing/selftests/kvm/lib/x86_64/processor.c > +++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c > @@ -1293,10 +1293,16 @@ const struct kvm_cpuid2 *vcpu_get_supported_hv_cpuid(struct kvm_vcpu *vcpu) > unsigned long vm_compute_max_gfn(struct kvm_vm *vm) > { > const unsigned long num_ht_pages = 12 << (30 - vm->page_shift); /* 12 GiB */ > - unsigned long ht_gfn, max_gfn, max_pfn; > + unsigned long ht_gfn, max_gfn, max_pfn, max_bits = 0; > uint8_t maxphyaddr; > > - max_gfn = (1ULL << (vm->pa_bits - vm->page_shift)) - 1; > + if (kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR)) > + max_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR); We can get rid of the kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR) check and call kvm_cpu_property() unconditionally. As a bonus, we don't need to init max_bits as 0. BTW, could we just name it guest_pa_bits? Otherwise, Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com> > + > + if (!max_bits) > + max_bits = vm->pa_bits; > + > + max_gfn = (1ULL << (max_bits - vm->page_shift)) - 1; > > /* Avoid reserved HyperTransport region on AMD processors. */ > if (!host_cpu_is_amd) > > base-commit: 448b3fe5a0eab5b625a7e15c67c7972169e47ff8
On Sat, May 11, 2024 at 03:08:16PM +0800, Xiaoyao Li wrote: [...] > > diff --git a/tools/testing/selftests/kvm/lib/x86_64/processor.c b/tools/testing/selftests/kvm/lib/x86_64/processor.c > > index 74a4c736c9ae..aa9966ead543 100644 > > --- a/tools/testing/selftests/kvm/lib/x86_64/processor.c > > +++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c > > @@ -1293,10 +1293,16 @@ const struct kvm_cpuid2 *vcpu_get_supported_hv_cpuid(struct kvm_vcpu *vcpu) > > unsigned long vm_compute_max_gfn(struct kvm_vm *vm) > > { > > const unsigned long num_ht_pages = 12 << (30 - vm->page_shift); /* 12 GiB */ > > - unsigned long ht_gfn, max_gfn, max_pfn; > > + unsigned long ht_gfn, max_gfn, max_pfn, max_bits = 0; > > uint8_t maxphyaddr; > > - max_gfn = (1ULL << (vm->pa_bits - vm->page_shift)) - 1; > > + if (kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR)) > > + max_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR); > > We can get rid of the kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR) check > and call kvm_cpu_property() unconditionally. As a bonus, we don't need to > init max_bits as 0. Thanks, good suggestion! > > BTW, could we just name it guest_pa_bits? Yes, it will be more accurate.
On Fri, May 10, 2024 at 10:03:46AM +0800, 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. > >Adjust max_gfn in vm_compute_max_gfn() since x86 selftests already >overrides vm_compute_max_gfn() specifically to deal with goofy edge cases. > >Signed-off-by: Tao Su <tao1.su@linux.intel.com> >Tested-by: Yi Lai <yi1.lai@intel.com> >--- >This patch is based on https://github.com/kvm-x86/linux/commit/b628cb523c65 > >Changelog: >v1 -> v2: > - Only adjust vm->max_gfn in vm_compute_max_gfn() > - Add Yi Lai's Tested-by > >v1: https://lore.kernel.org/all/20240508064205.15301-1-tao1.su@linux.intel.com/ >--- > tools/testing/selftests/kvm/include/x86_64/processor.h | 1 + > tools/testing/selftests/kvm/lib/x86_64/processor.c | 10 ++++++++-- > 2 files changed, 9 insertions(+), 2 deletions(-) > >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..aa9966ead543 100644 >--- a/tools/testing/selftests/kvm/lib/x86_64/processor.c >+++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c >@@ -1293,10 +1293,16 @@ const struct kvm_cpuid2 *vcpu_get_supported_hv_cpuid(struct kvm_vcpu *vcpu) > unsigned long vm_compute_max_gfn(struct kvm_vm *vm) > { > const unsigned long num_ht_pages = 12 << (30 - vm->page_shift); /* 12 GiB */ >- unsigned long ht_gfn, max_gfn, max_pfn; >+ unsigned long ht_gfn, max_gfn, max_pfn, max_bits = 0; nit: max_bits has only 8 bits. so max_bits should be uint8_t. > uint8_t maxphyaddr; > >- max_gfn = (1ULL << (vm->pa_bits - vm->page_shift)) - 1; >+ if (kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR)) >+ max_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR); >+ >+ if (!max_bits) >+ max_bits = vm->pa_bits; >+ >+ max_gfn = (1ULL << (max_bits - vm->page_shift)) - 1; > > /* Avoid reserved HyperTransport region on AMD processors. */ > if (!host_cpu_is_amd) > >base-commit: 448b3fe5a0eab5b625a7e15c67c7972169e47ff8 >-- >2.34.1 >
On Mon, May 13, 2024 at 09:44:32AM +0800, Chao Gao wrote: > On Fri, May 10, 2024 at 10:03:46AM +0800, 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. > > > >Adjust max_gfn in vm_compute_max_gfn() since x86 selftests already > >overrides vm_compute_max_gfn() specifically to deal with goofy edge cases. > > > >Signed-off-by: Tao Su <tao1.su@linux.intel.com> > >Tested-by: Yi Lai <yi1.lai@intel.com> > >--- > >This patch is based on https://github.com/kvm-x86/linux/commit/b628cb523c65 > > > >Changelog: > >v1 -> v2: > > - Only adjust vm->max_gfn in vm_compute_max_gfn() > > - Add Yi Lai's Tested-by > > > >v1: https://lore.kernel.org/all/20240508064205.15301-1-tao1.su@linux.intel.com/ > >--- > > tools/testing/selftests/kvm/include/x86_64/processor.h | 1 + > > tools/testing/selftests/kvm/lib/x86_64/processor.c | 10 ++++++++-- > > 2 files changed, 9 insertions(+), 2 deletions(-) > > > >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..aa9966ead543 100644 > >--- a/tools/testing/selftests/kvm/lib/x86_64/processor.c > >+++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c > >@@ -1293,10 +1293,16 @@ const struct kvm_cpuid2 *vcpu_get_supported_hv_cpuid(struct kvm_vcpu *vcpu) > > unsigned long vm_compute_max_gfn(struct kvm_vm *vm) > > { > > const unsigned long num_ht_pages = 12 << (30 - vm->page_shift); /* 12 GiB */ > >- unsigned long ht_gfn, max_gfn, max_pfn; > >+ unsigned long ht_gfn, max_gfn, max_pfn, max_bits = 0; > > nit: max_bits has only 8 bits. so max_bits should be uint8_t. Because vm->pa_bits is unsigned int, I'm worried that the compiler will complain on stricter compilation, what do you think? > > > uint8_t maxphyaddr; > > > >- max_gfn = (1ULL << (vm->pa_bits - vm->page_shift)) - 1; > >+ if (kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR)) > >+ max_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR); > >+ > >+ if (!max_bits) > >+ max_bits = vm->pa_bits; > >+ > >+ max_gfn = (1ULL << (max_bits - vm->page_shift)) - 1; > > > > /* Avoid reserved HyperTransport region on AMD processors. */ > > if (!host_cpu_is_amd) > > > >base-commit: 448b3fe5a0eab5b625a7e15c67c7972169e47ff8 > >-- > >2.34.1 > >
On Mon, May 13, 2024 at 10:00:35AM +0800, Tao Su wrote: >On Mon, May 13, 2024 at 09:44:32AM +0800, Chao Gao wrote: >> On Fri, May 10, 2024 at 10:03:46AM +0800, 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. >> > >> >Adjust max_gfn in vm_compute_max_gfn() since x86 selftests already >> >overrides vm_compute_max_gfn() specifically to deal with goofy edge cases. >> > >> >Signed-off-by: Tao Su <tao1.su@linux.intel.com> >> >Tested-by: Yi Lai <yi1.lai@intel.com> >> >--- >> >This patch is based on https://github.com/kvm-x86/linux/commit/b628cb523c65 >> > >> >Changelog: >> >v1 -> v2: >> > - Only adjust vm->max_gfn in vm_compute_max_gfn() >> > - Add Yi Lai's Tested-by >> > >> >v1: https://lore.kernel.org/all/20240508064205.15301-1-tao1.su@linux.intel.com/ >> >--- >> > tools/testing/selftests/kvm/include/x86_64/processor.h | 1 + >> > tools/testing/selftests/kvm/lib/x86_64/processor.c | 10 ++++++++-- >> > 2 files changed, 9 insertions(+), 2 deletions(-) >> > >> >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..aa9966ead543 100644 >> >--- a/tools/testing/selftests/kvm/lib/x86_64/processor.c >> >+++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c >> >@@ -1293,10 +1293,16 @@ const struct kvm_cpuid2 *vcpu_get_supported_hv_cpuid(struct kvm_vcpu *vcpu) >> > unsigned long vm_compute_max_gfn(struct kvm_vm *vm) >> > { >> > const unsigned long num_ht_pages = 12 << (30 - vm->page_shift); /* 12 GiB */ >> >- unsigned long ht_gfn, max_gfn, max_pfn; >> >+ unsigned long ht_gfn, max_gfn, max_pfn, max_bits = 0; >> >> nit: max_bits has only 8 bits. so max_bits should be uint8_t. > >Because vm->pa_bits is unsigned int, I'm worried that the compiler will >complain on stricter compilation, what do you think? @maxphyaddr (right below) is in the same situation. And if it was a problem for the compiler, casting vm->page_shift to uint8_t explicitly would be a better solution. > >> >> > uint8_t maxphyaddr; >> > >> >- max_gfn = (1ULL << (vm->pa_bits - vm->page_shift)) - 1; >> >+ if (kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR)) >> >+ max_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR); >> >+ >> >+ if (!max_bits) >> >+ max_bits = vm->pa_bits; >> >+ >> >+ max_gfn = (1ULL << (max_bits - vm->page_shift)) - 1; >> > >> > /* Avoid reserved HyperTransport region on AMD processors. */ >> > if (!host_cpu_is_amd) >> > >> >base-commit: 448b3fe5a0eab5b625a7e15c67c7972169e47ff8 >> >-- >> >2.34.1 >> >
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..aa9966ead543 100644 --- a/tools/testing/selftests/kvm/lib/x86_64/processor.c +++ b/tools/testing/selftests/kvm/lib/x86_64/processor.c @@ -1293,10 +1293,16 @@ const struct kvm_cpuid2 *vcpu_get_supported_hv_cpuid(struct kvm_vcpu *vcpu) unsigned long vm_compute_max_gfn(struct kvm_vm *vm) { const unsigned long num_ht_pages = 12 << (30 - vm->page_shift); /* 12 GiB */ - unsigned long ht_gfn, max_gfn, max_pfn; + unsigned long ht_gfn, max_gfn, max_pfn, max_bits = 0; uint8_t maxphyaddr; - max_gfn = (1ULL << (vm->pa_bits - vm->page_shift)) - 1; + if (kvm_cpu_has_p(X86_PROPERTY_MAX_GUEST_PHY_ADDR)) + max_bits = kvm_cpu_property(X86_PROPERTY_MAX_GUEST_PHY_ADDR); + + if (!max_bits) + max_bits = vm->pa_bits; + + max_gfn = (1ULL << (max_bits - vm->page_shift)) - 1; /* Avoid reserved HyperTransport region on AMD processors. */ if (!host_cpu_is_amd)