[PATCH v2 15/15] KVM: MTRR: do not map huage page for non-consistent range

[Xiao Guangrong <guangrong.xiao@xxxxxxxxxxxxxxx>]

Based on Intel's SDM, mapping huge page which do not have consistent
memory cache for each 4k page will cause undefined behavior

In order to avoiding this kind of undefined behavior, we force to use
4k pages under this case

Signed-off-by: Xiao Guangrong <guangrong.xiao@xxxxxxxxxxxxxxx>
---
 arch/x86/kvm/mmu.c  | 20 +++++++++++++++++++-
 arch/x86/kvm/mtrr.c | 29 +++++++++++++++++++++++++++++
 arch/x86/kvm/x86.h  |  2 ++
 3 files changed, 50 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 532aad2..f807496 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -3446,6 +3446,16 @@ static bool try_async_pf(struct kvm_vcpu *vcpu, bool prefault, gfn_t gfn,
 	return false;
 }
 
+static bool
+check_hugepage_cache_consistency(struct kvm_vcpu *vcpu, gfn_t gfn, int level)
+{
+	int page_num = KVM_PAGES_PER_HPAGE(level);
+
+	gfn &= ~(page_num - 1);
+
+	return kvm_mtrr_check_gfn_range_consistency(vcpu, gfn, page_num);
+}
+
 static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, u32 error_code,
 			  bool prefault)
 {
@@ -3471,9 +3481,17 @@ static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, u32 error_code,
 	if (r)
 		return r;
 
-	force_pt_level = mapping_level_dirty_bitmap(vcpu, gfn);
+	if (mapping_level_dirty_bitmap(vcpu, gfn) ||
+	    !check_hugepage_cache_consistency(vcpu, gfn, PT_DIRECTORY_LEVEL))
+		force_pt_level = 1;
+	else
+		force_pt_level = 0;
+
 	if (likely(!force_pt_level)) {
 		level = mapping_level(vcpu, gfn);
+		if (level > PT_DIRECTORY_LEVEL &&
+		    !check_hugepage_cache_consistency(vcpu, gfn, level))
+			level = PT_DIRECTORY_LEVEL;
 		gfn &= ~(KVM_PAGES_PER_HPAGE(level) - 1);
 	} else
 		level = PT_PAGE_TABLE_LEVEL;
diff --git a/arch/x86/kvm/mtrr.c b/arch/x86/kvm/mtrr.c
index 097d616..5b26048 100644
--- a/arch/x86/kvm/mtrr.c
+++ b/arch/x86/kvm/mtrr.c
@@ -677,3 +677,32 @@ u8 kvm_mtrr_get_guest_memory_type(struct kvm_vcpu *vcpu, gfn_t gfn)
 	return type;
 }
 EXPORT_SYMBOL_GPL(kvm_mtrr_get_guest_memory_type);
+
+bool kvm_mtrr_check_gfn_range_consistency(struct kvm_vcpu *vcpu, gfn_t gfn,
+					  int page_num)
+{
+	struct kvm_mtrr *mtrr_state = &vcpu->arch.mtrr_state;
+	struct mtrr_iter iter;
+	u64 start, end;
+	int type = -1;
+
+	start = gfn_to_gpa(gfn);
+	end = gfn_to_gpa(gfn + page_num);
+	mtrr_for_each_mem_type(&iter, mtrr_state, start, end) {
+		if (type == -1) {
+			type = iter.mem_type;
+			continue;
+		}
+
+		if (type != iter.mem_type)
+			return false;
+	}
+
+	if (!iter.partial_map)
+		return true;
+
+	if (type == -1)
+		return true;
+
+	return type == mtrr_default_type(mtrr_state);
+}
diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
index 0e4727c..edc8cdc 100644
--- a/arch/x86/kvm/x86.h
+++ b/arch/x86/kvm/x86.h
@@ -167,6 +167,8 @@ u8 kvm_mtrr_get_guest_memory_type(struct kvm_vcpu *vcpu, gfn_t gfn);
 bool kvm_mtrr_valid(struct kvm_vcpu *vcpu, u32 msr, u64 data);
 int kvm_mtrr_set_msr(struct kvm_vcpu *vcpu, u32 msr, u64 data);
 int kvm_mtrr_get_msr(struct kvm_vcpu *vcpu, u32 msr, u64 *pdata);
+bool kvm_mtrr_check_gfn_range_consistency(struct kvm_vcpu *vcpu, gfn_t gfn,
+					  int page_num);
 
 #define KVM_SUPPORTED_XCR0     (XSTATE_FP | XSTATE_SSE | XSTATE_YMM \
 				| XSTATE_BNDREGS | XSTATE_BNDCSR \
-- 
2.1.0

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html