KVM: do not inject #PF in (read|write)_emulated() callbacks

Return error to x86 emulator instead of injection exception behind its back.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>

arch/x86/include/asm/kvm_emulate.h

@@ -94,6 +94,7 @@ struct x86_emulate_ops {
 	int (*read_emulated)(unsigned long addr,
 			     void *val,
 			     unsigned int bytes,
+			     unsigned int *error,
 			     struct kvm_vcpu *vcpu);
 
 	/*
@@ -106,6 +107,7 @@ struct x86_emulate_ops {
 	int (*write_emulated)(unsigned long addr,
 			      const void *val,
 			      unsigned int bytes,
+			      unsigned int *error,
 			      struct kvm_vcpu *vcpu);
 
 	/*
@@ -120,6 +122,7 @@ struct x86_emulate_ops {
 				const void *old,
 				const void *new,
 				unsigned int bytes,
+				unsigned int *error,
 				struct kvm_vcpu *vcpu);
 
 	int (*pio_in_emulated)(int size, unsigned short port, void *val,

arch/x86/kvm/emulate.c

@@ -1277,6 +1277,7 @@ static int read_emulated(struct x86_emulate_ctxt *ctxt,
 {
 	int rc;
 	struct read_cache *mc = &ctxt->decode.mem_read;
+	u32 err;
 
 	while (size) {
 		int n = min(size, 8u);
@@ -1284,7 +1285,10 @@ static int read_emulated(struct x86_emulate_ctxt *ctxt,
 		if (mc->pos < mc->end)
 			goto read_cached;
 
-		rc = ops->read_emulated(addr, mc->data + mc->end, n, ctxt->vcpu);
+		rc = ops->read_emulated(addr, mc->data + mc->end, n, &err,
+					ctxt->vcpu);
+		if (rc == X86EMUL_PROPAGATE_FAULT)
+			kvm_inject_page_fault(ctxt->vcpu, addr, err);
 		if (rc != X86EMUL_CONTINUE)
 			return rc;
 		mc->end += n;
@@ -1789,6 +1793,7 @@ static inline int writeback(struct x86_emulate_ctxt *ctxt,
 {
 	int rc;
 	struct decode_cache *c = &ctxt->decode;
+	u32 err;
 
 	switch (c->dst.type) {
 	case OP_REG:
@@ -1817,13 +1822,18 @@ static inline int writeback(struct x86_emulate_ctxt *ctxt,
 					&c->dst.orig_val,
 					&c->dst.val,
 					c->dst.bytes,
+					&err,
 					ctxt->vcpu);
 		else
 			rc = ops->write_emulated(
 					(unsigned long)c->dst.ptr,
 					&c->dst.val,
 					c->dst.bytes,
+					&err,
 					ctxt->vcpu);
+		if (rc == X86EMUL_PROPAGATE_FAULT)
+			kvm_inject_page_fault(ctxt->vcpu,
+					      (unsigned long)c->dst.ptr, err);
 		if (rc != X86EMUL_CONTINUE)
 			return rc;
 		break;

arch/x86/kvm/x86.c

@@ -3346,10 +3346,10 @@ static int kvm_write_guest_virt_system(gva_t addr, void *val,
 static int emulator_read_emulated(unsigned long addr,
 				  void *val,
 				  unsigned int bytes,
+				  unsigned int *error_code,
 				  struct kvm_vcpu *vcpu)
 {
 	gpa_t                 gpa;
-	u32 error_code;
 
 	if (vcpu->mmio_read_completed) {
 		memcpy(val, vcpu->mmio_data, bytes);
@@ -3359,12 +3359,10 @@ static int emulator_read_emulated(unsigned long addr,
 		return X86EMUL_CONTINUE;
 	}
 
-	gpa = kvm_mmu_gva_to_gpa_read(vcpu, addr, &error_code);
+	gpa = kvm_mmu_gva_to_gpa_read(vcpu, addr, error_code);
 
-	if (gpa == UNMAPPED_GVA) {
-		kvm_inject_page_fault(vcpu, addr, error_code);
+	if (gpa == UNMAPPED_GVA)
 		return X86EMUL_PROPAGATE_FAULT;
-	}
 
 	/* For APIC access vmexit */
 	if ((gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
@@ -3409,17 +3407,15 @@ int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
 static int emulator_write_emulated_onepage(unsigned long addr,
 					   const void *val,
 					   unsigned int bytes,
+					   unsigned int *error_code,
 					   struct kvm_vcpu *vcpu)
 {
 	gpa_t                 gpa;
-	u32 error_code;
 
-	gpa = kvm_mmu_gva_to_gpa_write(vcpu, addr, &error_code);
+	gpa = kvm_mmu_gva_to_gpa_write(vcpu, addr, error_code);
 
-	if (gpa == UNMAPPED_GVA) {
-		kvm_inject_page_fault(vcpu, addr, error_code);
+	if (gpa == UNMAPPED_GVA)
 		return X86EMUL_PROPAGATE_FAULT;
-	}
 
 	/* For APIC access vmexit */
 	if ((gpa & PAGE_MASK) == APIC_DEFAULT_PHYS_BASE)
@@ -3449,6 +3445,7 @@ static int emulator_write_emulated_onepage(unsigned long addr,
 int emulator_write_emulated(unsigned long addr,
 			    const void *val,
 			    unsigned int bytes,
+			    unsigned int *error_code,
 			    struct kvm_vcpu *vcpu)
 {
 	/* Crossing a page boundary? */
@@ -3456,14 +3453,16 @@ int emulator_write_emulated(unsigned long addr,
 		int rc, now;
 
 		now = -addr & ~PAGE_MASK;
-		rc = emulator_write_emulated_onepage(addr, val, now, vcpu);
+		rc = emulator_write_emulated_onepage(addr, val, now, error_code,
+						     vcpu);
 		if (rc != X86EMUL_CONTINUE)
 			return rc;
 		addr += now;
 		val += now;
 		bytes -= now;
 	}
-	return emulator_write_emulated_onepage(addr, val, bytes, vcpu);
+	return emulator_write_emulated_onepage(addr, val, bytes, error_code,
+					       vcpu);
 }
 
 #define CMPXCHG_TYPE(t, ptr, old, new) \
@@ -3480,6 +3479,7 @@ static int emulator_cmpxchg_emulated(unsigned long addr,
 				     const void *old,
 				     const void *new,
 				     unsigned int bytes,
+				     unsigned int *error_code,
 				     struct kvm_vcpu *vcpu)
 {
 	gpa_t gpa;
@@ -3533,7 +3533,7 @@ static int emulator_cmpxchg_emulated(unsigned long addr,
 emul_write:
 	printk_once(KERN_WARNING "kvm: emulating exchange as write\n");
 
-	return emulator_write_emulated(addr, new, bytes, vcpu);
+	return emulator_write_emulated(addr, new, bytes, error_code, vcpu);
 }
 
 static int kernel_pio(struct kvm_vcpu *vcpu, void *pd)
@@ -4293,7 +4293,7 @@ int kvm_fix_hypercall(struct kvm_vcpu *vcpu)
 
 	kvm_x86_ops->patch_hypercall(vcpu, instruction);
 
-	return emulator_write_emulated(rip, instruction, 3, vcpu);
+	return emulator_write_emulated(rip, instruction, 3, NULL, vcpu);
 }
 
 void realmode_lgdt(struct kvm_vcpu *vcpu, u16 limit, unsigned long base)