From: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
It is unfortunate and inflexible for kvm_mmu_do_page_fault() to call
kvm_mem_is_private(), eventually looking up memory attributes. Later
__kvm_faultin_pfn() looks up memory attributes again. There is a race
condition that other threads can change memory attributes due to not
gaining the mmu lock. SNP-SEV and TDX define theri way to indicate that
the page fault is private.
Add two PFERR codes to designate that the page fault is private and that
it requires looking up memory attributes.
Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
---
arch/x86/include/asm/kvm_host.h | 4 ++++
arch/x86/kvm/mmu/mmu.c | 9 +++++++--
arch/x86/kvm/mmu/mmu_internal.h | 4 ++--
3 files changed, 13 insertions(+), 4 deletions(-)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 8ae131dc645d..2763f9837a0b 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -255,7 +255,9 @@ enum x86_intercept_stage;
#define PFERR_SGX_BIT 15
#define PFERR_GUEST_FINAL_BIT 32
#define PFERR_GUEST_PAGE_BIT 33
+#define PFERR_GUEST_ENC_BIT 34
#define PFERR_IMPLICIT_ACCESS_BIT 48
+#define PFERR_HASATTR_BIT 63
#define PFERR_PRESENT_MASK BIT(PFERR_PRESENT_BIT)
#define PFERR_WRITE_MASK BIT(PFERR_WRITE_BIT)
@@ -266,7 +268,9 @@ enum x86_intercept_stage;
#define PFERR_SGX_MASK BIT(PFERR_SGX_BIT)
#define PFERR_GUEST_FINAL_MASK BIT_ULL(PFERR_GUEST_FINAL_BIT)
#define PFERR_GUEST_PAGE_MASK BIT_ULL(PFERR_GUEST_PAGE_BIT)
+#define PFERR_GUEST_ENC_MASK BIT_ULL(PFERR_GUEST_ENC_BIT)
#define PFERR_IMPLICIT_ACCESS BIT_ULL(PFERR_IMPLICIT_ACCESS_BIT)
+#define PFERR_HASATTR_MASK BIT_ULL(PFERR_HASATTR_BIT)
#define PFERR_NESTED_GUEST_PAGE (PFERR_GUEST_PAGE_MASK | \
PFERR_WRITE_MASK | \
diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index b8ba7f11c3cb..e9c9780bab89 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -4358,6 +4358,7 @@ static int __kvm_faultin_pfn(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
{
struct kvm_memory_slot *slot = fault->slot;
bool async;
+ bool is_private;
/*
* Retry the page fault if the gfn hit a memslot that is being deleted
@@ -4386,8 +4387,12 @@ static int __kvm_faultin_pfn(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault
return RET_PF_EMULATE;
}
- if (fault->is_private != kvm_mem_is_private(vcpu->kvm, fault->gfn))
- return kvm_do_memory_fault_exit(vcpu, fault);
+ is_private = kvm_mem_is_private(vcpu->kvm, fault->gfn);
+ if (fault->error_code & PFERR_HASATTR_MASK) {
+ if (fault->is_private != is_private)
+ return kvm_do_memory_fault_exit(vcpu, fault);
+ } else
+ fault->is_private = is_private;
if (fault->is_private)
return kvm_faultin_pfn_private(vcpu, fault);
diff --git a/arch/x86/kvm/mmu/mmu_internal.h b/arch/x86/kvm/mmu/mmu_internal.h
index 7f9ec1e5b136..22f2cd60cabf 100644
--- a/arch/x86/kvm/mmu/mmu_internal.h
+++ b/arch/x86/kvm/mmu/mmu_internal.h
@@ -203,7 +203,7 @@ struct kvm_page_fault {
/* Derived from mmu and global state. */
const bool is_tdp;
- const bool is_private;
+ bool is_private;
const bool nx_huge_page_workaround_enabled;
/*
@@ -301,7 +301,7 @@ static inline int kvm_mmu_do_page_fault(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa,
.max_level = KVM_MAX_HUGEPAGE_LEVEL,
.req_level = PG_LEVEL_4K,
.goal_level = PG_LEVEL_4K,
- .is_private = kvm_mem_is_private(vcpu->kvm, cr2_or_gpa >> PAGE_SHIFT),
+ .is_private = err & PFERR_GUEST_ENC_MASK,
};
int r;
--
2.25.1
