KVM: x86: allocate vcpu->arch.cpuid_entries dynamically

The current limit for guest CPUID leaves (KVM_MAX_CPUID_ENTRIES, 80)
is reported to be insufficient but before we bump it let's switch to
allocating vcpu->arch.cpuid_entries[] array dynamically. Currently,
'struct kvm_cpuid_entry2' is 40 bytes so vcpu->arch.cpuid_entries is
3200 bytes which accounts for 1/4 of the whole 'struct kvm_vcpu_arch'
but having it pre-allocated (for all vCPUs which we also pre-allocate)
gives us no real benefits.

Another plus of the dynamic allocation is that we now do kvm_check_cpuid()
check before we assign anything to vcpu->arch.cpuid_nent/cpuid_entries so
no changes are made in case the check fails.

Opportunistically remove unneeded 'out' labels from
kvm_vcpu_ioctl_set_cpuid()/kvm_vcpu_ioctl_set_cpuid2() and return
directly whenever possible.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20201001130541.1398392-3-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>

arch/x86/include/asm/kvm_host.h

@@ -637,7 +637,7 @@ struct kvm_vcpu_arch {
 	int halt_request; /* real mode on Intel only */
 
 	int cpuid_nent;
-	struct kvm_cpuid_entry2 cpuid_entries[KVM_MAX_CPUID_ENTRIES];
+	struct kvm_cpuid_entry2 *cpuid_entries;
 
 	int maxphyaddr;
 	int max_tdp_level;

arch/x86/kvm/cpuid.c

@@ -217,46 +217,53 @@ int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
 			     struct kvm_cpuid_entry __user *entries)
 {
 	int r, i;
-	struct kvm_cpuid_entry *cpuid_entries = NULL;
+	struct kvm_cpuid_entry *e = NULL;
+	struct kvm_cpuid_entry2 *e2 = NULL;
 
-	r = -E2BIG;
 	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
-		goto out;
+		return -E2BIG;
+
 	if (cpuid->nent) {
-		cpuid_entries = vmemdup_user(entries,
+		e = vmemdup_user(entries, array_size(sizeof(*e), cpuid->nent));
-					     array_size(sizeof(struct kvm_cpuid_entry),
+		if (IS_ERR(e))
-							cpuid->nent));
+			return PTR_ERR(e);
-		if (IS_ERR(cpuid_entries)) {
+
-			r = PTR_ERR(cpuid_entries);
+		e2 = kvmalloc_array(cpuid->nent, sizeof(*e2), GFP_KERNEL_ACCOUNT);
-			goto out;
+		if (!e2) {
+			r = -ENOMEM;
+			goto out_free_cpuid;
 		}
 	}
 	for (i = 0; i < cpuid->nent; i++) {
-		vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
+		e2[i].function = e[i].function;
-		vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
+		e2[i].eax = e[i].eax;
-		vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx;
+		e2[i].ebx = e[i].ebx;
-		vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx;
+		e2[i].ecx = e[i].ecx;
-		vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx;
+		e2[i].edx = e[i].edx;
-		vcpu->arch.cpuid_entries[i].index = 0;
+		e2[i].index = 0;
-		vcpu->arch.cpuid_entries[i].flags = 0;
+		e2[i].flags = 0;
-		vcpu->arch.cpuid_entries[i].padding[0] = 0;
+		e2[i].padding[0] = 0;
-		vcpu->arch.cpuid_entries[i].padding[1] = 0;
+		e2[i].padding[1] = 0;
-		vcpu->arch.cpuid_entries[i].padding[2] = 0;
+		e2[i].padding[2] = 0;
 	}
-	vcpu->arch.cpuid_nent = cpuid->nent;
+
-	r = kvm_check_cpuid(vcpu->arch.cpuid_entries, cpuid->nent);
+	r = kvm_check_cpuid(e2, cpuid->nent);
 	if (r) {
-		vcpu->arch.cpuid_nent = 0;
+		kvfree(e2);
-		kvfree(cpuid_entries);
+		goto out_free_cpuid;
-		goto out;
 	}
 
+	kvfree(vcpu->arch.cpuid_entries);
+	vcpu->arch.cpuid_entries = e2;
+	vcpu->arch.cpuid_nent = cpuid->nent;
+
 	cpuid_fix_nx_cap(vcpu);
 	kvm_update_cpuid_runtime(vcpu);
 	kvm_vcpu_after_set_cpuid(vcpu);
 
-	kvfree(cpuid_entries);
+out_free_cpuid:
-out:
+	kvfree(e);
+
 	return r;
 }
 
@@ -264,26 +271,32 @@ int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
 			      struct kvm_cpuid2 *cpuid,
 			      struct kvm_cpuid_entry2 __user *entries)
 {
+	struct kvm_cpuid_entry2 *e2 = NULL;
 	int r;
 
-	r = -E2BIG;
 	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
-		goto out;
+		return -E2BIG;
-	r = -EFAULT;
+
-	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
+	if (cpuid->nent) {
-			   cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
+		e2 = vmemdup_user(entries, array_size(sizeof(*e2), cpuid->nent));
-		goto out;
+		if (IS_ERR(e2))
-	vcpu->arch.cpuid_nent = cpuid->nent;
+			return PTR_ERR(e2);
-	r = kvm_check_cpuid(vcpu->arch.cpuid_entries, cpuid->nent);
-	if (r) {
-		vcpu->arch.cpuid_nent = 0;
-		goto out;
 	}
 
+	r = kvm_check_cpuid(e2, cpuid->nent);
+	if (r) {
+		kvfree(e2);
+		return r;
+	}
+
+	kvfree(vcpu->arch.cpuid_entries);
+	vcpu->arch.cpuid_entries = e2;
+	vcpu->arch.cpuid_nent = cpuid->nent;
+
 	kvm_update_cpuid_runtime(vcpu);
 	kvm_vcpu_after_set_cpuid(vcpu);
-out:
+
-	return r;
+	return 0;
 }
 
 int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,

arch/x86/kvm/x86.c

@@ -9948,6 +9948,7 @@ void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
 	kvm_mmu_destroy(vcpu);
 	srcu_read_unlock(&vcpu->kvm->srcu, idx);
 	free_page((unsigned long)vcpu->arch.pio_data);
+	kvfree(vcpu->arch.cpuid_entries);
 	if (!lapic_in_kernel(vcpu))
 		static_key_slow_dec(&kvm_no_apic_vcpu);
 }