Merge tag 'kvm-x86-svm-6.11' of https://github.com/kvm-x86/linux into HEAD

KVM SVM changes for 6.11

 - Make per-CPU save_area allocations NUMA-aware.

 - Force sev_es_host_save_area() to be inlined to avoid calling into an
   instrumentable function from noinstr code.

Author: bonzini

arch/x86/kvm/svm/nested.c

@@ -1181,7 +1181,7 @@ int svm_allocate_nested(struct vcpu_svm *svm)
 	if (svm->nested.initialized)
 		return 0;
 
-	vmcb02_page = snp_safe_alloc_page(&svm->vcpu);
+	vmcb02_page = snp_safe_alloc_page();
 	if (!vmcb02_page)
 		return -ENOMEM;
 	svm->nested.vmcb02.ptr = page_address(vmcb02_page);

arch/x86/kvm/svm/sev.c

@@ -4459,13 +4459,13 @@ void sev_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector)
 	}
 }
 
-struct page *snp_safe_alloc_page(struct kvm_vcpu *vcpu)
+struct page *snp_safe_alloc_page_node(int node, gfp_t gfp)
 {
 	unsigned long pfn;
 	struct page *p;
 
 	if (!cc_platform_has(CC_ATTR_HOST_SEV_SNP))
-		return alloc_page(GFP_KERNEL_ACCOUNT | __GFP_ZERO);
+		return alloc_pages_node(node, gfp | __GFP_ZERO, 0);
 
 	/*
 	 * Allocate an SNP-safe page to workaround the SNP erratum where
@@ -4476,7 +4476,7 @@ struct page *snp_safe_alloc_page(struct kvm_vcpu *vcpu)
 	 * Allocate one extra page, choose a page which is not
 	 * 2MB-aligned, and free the other.
 	 */
-	p = alloc_pages(GFP_KERNEL_ACCOUNT | __GFP_ZERO, 1);
+	p = alloc_pages_node(node, gfp | __GFP_ZERO, 1);
 	if (!p)
 		return NULL;
 

arch/x86/kvm/svm/svm.c

@@ -571,6 +571,11 @@ static void __svm_write_tsc_multiplier(u64 multiplier)
 	__this_cpu_write(current_tsc_ratio, multiplier);
 }
 
+static __always_inline struct sev_es_save_area *sev_es_host_save_area(struct svm_cpu_data *sd)
+{
+	return page_address(sd->save_area) + 0x400;
+}
+
 static inline void kvm_cpu_svm_disable(void)
 {
 	uint64_t efer;
@@ -675,12 +680,9 @@ static int svm_hardware_enable(void)
 	 * TSC_AUX field now to avoid a RDMSR on every vCPU run.
 	 */
 	if (boot_cpu_has(X86_FEATURE_V_TSC_AUX)) {
-		struct sev_es_save_area *hostsa;
 		u32 __maybe_unused msr_hi;
 
-		hostsa = (struct sev_es_save_area *)(page_address(sd->save_area) + 0x400);
-
-		rdmsr(MSR_TSC_AUX, hostsa->tsc_aux, msr_hi);
+		rdmsr(MSR_TSC_AUX, sev_es_host_save_area(sd)->tsc_aux, msr_hi);
 	}
 
 	return 0;
@@ -705,7 +707,7 @@ static int svm_cpu_init(int cpu)
 	int ret = -ENOMEM;
 
 	memset(sd, 0, sizeof(struct svm_cpu_data));
-	sd->save_area = snp_safe_alloc_page(NULL);
+	sd->save_area = snp_safe_alloc_page_node(cpu_to_node(cpu), GFP_KERNEL);
 	if (!sd->save_area)
 		return ret;
 
@@ -1431,7 +1433,7 @@ static int svm_vcpu_create(struct kvm_vcpu *vcpu)
 	svm = to_svm(vcpu);
 
 	err = -ENOMEM;
-	vmcb01_page = snp_safe_alloc_page(vcpu);
+	vmcb01_page = snp_safe_alloc_page();
 	if (!vmcb01_page)
 		goto out;
 
@@ -1440,7 +1442,7 @@ static int svm_vcpu_create(struct kvm_vcpu *vcpu)
 		 * SEV-ES guests require a separate VMSA page used to contain
 		 * the encrypted register state of the guest.
 		 */
-		vmsa_page = snp_safe_alloc_page(vcpu);
+		vmsa_page = snp_safe_alloc_page();
 		if (!vmsa_page)
 			goto error_free_vmcb_page;
 	}
@@ -1505,11 +1507,6 @@ static void svm_vcpu_free(struct kvm_vcpu *vcpu)
 	__free_pages(virt_to_page(svm->msrpm), get_order(MSRPM_SIZE));
 }
 
-static struct sev_es_save_area *sev_es_host_save_area(struct svm_cpu_data *sd)
-{
-	return page_address(sd->save_area) + 0x400;
-}
-
 static void svm_prepare_switch_to_guest(struct kvm_vcpu *vcpu)
 {
 	struct vcpu_svm *svm = to_svm(vcpu);
@@ -4968,7 +4965,7 @@ static int svm_vm_init(struct kvm *kvm)
 
 static void *svm_alloc_apic_backing_page(struct kvm_vcpu *vcpu)
 {
-	struct page *page = snp_safe_alloc_page(vcpu);
+	struct page *page = snp_safe_alloc_page();
 
 	if (!page)
 		return NULL;

arch/x86/kvm/svm/svm.h

@@ -726,7 +726,13 @@ void sev_guest_memory_reclaimed(struct kvm *kvm);
 int sev_handle_vmgexit(struct kvm_vcpu *vcpu);
 
 /* These symbols are used in common code and are stubbed below.  */
-struct page *snp_safe_alloc_page(struct kvm_vcpu *vcpu);
+
+struct page *snp_safe_alloc_page_node(int node, gfp_t gfp);
+static inline struct page *snp_safe_alloc_page(void)
+{
+	return snp_safe_alloc_page_node(numa_node_id(), GFP_KERNEL_ACCOUNT);
+}
+
 void sev_free_vcpu(struct kvm_vcpu *vcpu);
 void sev_vm_destroy(struct kvm *kvm);
 void __init sev_set_cpu_caps(void);
@@ -741,8 +747,14 @@ int sev_gmem_prepare(struct kvm *kvm, kvm_pfn_t pfn, gfn_t gfn, int max_order);
 void sev_gmem_invalidate(kvm_pfn_t start, kvm_pfn_t end);
 int sev_private_max_mapping_level(struct kvm *kvm, kvm_pfn_t pfn);
 #else
-static inline struct page *snp_safe_alloc_page(struct kvm_vcpu *vcpu) {
-	return alloc_page(GFP_KERNEL_ACCOUNT | __GFP_ZERO);
+static inline struct page *snp_safe_alloc_page_node(int node, gfp_t gfp)
+{
+	return alloc_pages_node(node, gfp | __GFP_ZERO, 0);
+}
+
+static inline struct page *snp_safe_alloc_page(void)
+{
+	return snp_safe_alloc_page_node(numa_node_id(), GFP_KERNEL_ACCOUNT);
 }
 
 static inline void sev_free_vcpu(struct kvm_vcpu *vcpu) {}