KVM: selftests: Stop conflating vCPU index and ID in perf tests

Track vCPUs by their 'struct kvm_vcpu' object, and stop assuming that a
vCPU's ID is the same as its index when referencing a vCPU's metadata.

Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

2 files changed, 51 insertions, 45 deletions

diff --git a/tools/testing/selftests/kvm/lib/perf_test_util.c b/tools/testing/selftests/kvm/lib/perf_test_util.c
index 05f65c9292eb..91a89553afdd 100644
--- a/tools/testing/selftests/kvm/lib/perf_test_util.c
+++ b/tools/testing/selftests/kvm/lib/perf_test_util.c
@@ -17,8 +17,8 @@ struct perf_test_args perf_test_args;
 static uint64_t guest_test_virt_mem = DEFAULT_GUEST_TEST_MEM;
 
 struct vcpu_thread {
-	/* The id of the vCPU. */
-	int vcpu_id;
+	/* The index of the vCPU. */
+	int vcpu_idx;
 
 	/* The pthread backing the vCPU. */
 	pthread_t thread;
@@ -36,24 +36,26 @@ static void (*vcpu_thread_fn)(struct perf_test_vcpu_args *);
 /* Set to true once all vCPU threads are up and running. */
 static bool all_vcpu_threads_running;
 
+static struct kvm_vcpu *vcpus[KVM_MAX_VCPUS];
+
 /*
  * Continuously write to the first 8 bytes of each page in the
  * specified region.
  */
-void perf_test_guest_code(uint32_t vcpu_id)
+void perf_test_guest_code(uint32_t vcpu_idx)
 {
 	struct perf_test_args *pta = &perf_test_args;
-	struct perf_test_vcpu_args *vcpu_args = &pta->vcpu_args[vcpu_id];
+	struct perf_test_vcpu_args *vcpu_args = &pta->vcpu_args[vcpu_idx];
 	uint64_t gva;
 	uint64_t pages;
 	int i;
 
-	/* Make sure vCPU args data structure is not corrupt. */
-	GUEST_ASSERT(vcpu_args->vcpu_id == vcpu_id);
-
 	gva = vcpu_args->gva;
 	pages = vcpu_args->pages;
 
+	/* Make sure vCPU args data structure is not corrupt. */
+	GUEST_ASSERT(vcpu_args->vcpu_idx == vcpu_idx);
+
 	while (true) {
 		for (i = 0; i < pages; i++) {
 			uint64_t addr = gva + (i * pta->guest_page_size);
@@ -68,40 +70,43 @@ void perf_test_guest_code(uint32_t vcpu_id)
 	}
 }
 
-void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus,
+void perf_test_setup_vcpus(struct kvm_vm *vm, int nr_vcpus,
+			   struct kvm_vcpu *vcpus[],
 			   uint64_t vcpu_memory_bytes,
 			   bool partition_vcpu_memory_access)
 {
 	struct perf_test_args *pta = &perf_test_args;
 	struct perf_test_vcpu_args *vcpu_args;
-	int vcpu_id;
+	int i;
+
+	for (i = 0; i < nr_vcpus; i++) {
+		vcpu_args = &pta->vcpu_args[i];
 
-	for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) {
-		vcpu_args = &pta->vcpu_args[vcpu_id];
+		vcpu_args->vcpu = vcpus[i];
+		vcpu_args->vcpu_idx = i;
 
-		vcpu_args->vcpu_id = vcpu_id;
 		if (partition_vcpu_memory_access) {
 			vcpu_args->gva = guest_test_virt_mem +
-					 (vcpu_id * vcpu_memory_bytes);
+					 (i * vcpu_memory_bytes);
 			vcpu_args->pages = vcpu_memory_bytes /
 					   pta->guest_page_size;
-			vcpu_args->gpa = pta->gpa + (vcpu_id * vcpu_memory_bytes);
+			vcpu_args->gpa = pta->gpa + (i * vcpu_memory_bytes);
 		} else {
 			vcpu_args->gva = guest_test_virt_mem;
-			vcpu_args->pages = (vcpus * vcpu_memory_bytes) /
+			vcpu_args->pages = (nr_vcpus * vcpu_memory_bytes) /
 					   pta->guest_page_size;
 			vcpu_args->gpa = pta->gpa;
 		}
 
-		vcpu_args_set(vm, vcpu_id, 1, vcpu_id);
+		vcpu_args_set(vm, vcpus[i]->id, 1, i);
 
 		pr_debug("Added VCPU %d with test mem gpa [%lx, %lx)\n",
-			 vcpu_id, vcpu_args->gpa, vcpu_args->gpa +
+			 i, vcpu_args->gpa, vcpu_args->gpa +
 			 (vcpu_args->pages * pta->guest_page_size));
 	}
 }
 
-struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus,
+struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int nr_vcpus,
 				   uint64_t vcpu_memory_bytes, int slots,
 				   enum vm_mem_backing_src_type backing_src,
 				   bool partition_vcpu_memory_access)
@@ -125,7 +130,7 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus,
 	pta->guest_page_size = vm_guest_mode_params[mode].page_size;
 
 	guest_num_pages = vm_adjust_num_guest_pages(mode,
-				(vcpus * vcpu_memory_bytes) / pta->guest_page_size);
+				(nr_vcpus * vcpu_memory_bytes) / pta->guest_page_size);
 
 	TEST_ASSERT(vcpu_memory_bytes % getpagesize() == 0,
 		    "Guest memory size is not host page size aligned.");
@@ -140,16 +145,16 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus,
 	 * in-memory data structures.
 	 */
 	if (pta->nested)
-		slot0_pages += perf_test_nested_pages(vcpus);
+		slot0_pages += perf_test_nested_pages(nr_vcpus);
 
 	/*
 	 * Pass guest_num_pages to populate the page tables for test memory.
 	 * The memory is also added to memslot 0, but that's a benign side
 	 * effect as KVM allows aliasing HVAs in meslots.
 	 */
-	vm = __vm_create_with_vcpus(mode, vcpus, DEFAULT_GUEST_PHY_PAGES,
+	vm = __vm_create_with_vcpus(mode, nr_vcpus, DEFAULT_GUEST_PHY_PAGES,
 				    slot0_pages + guest_num_pages, 0,
-				    perf_test_guest_code, NULL);
+				    perf_test_guest_code, vcpus);
 
 	pta->vm = vm;
 
@@ -171,11 +176,10 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus,
 	TEST_ASSERT(guest_num_pages < region_end_gfn,
 		    "Requested more guest memory than address space allows.\n"
 		    "    guest pages: %" PRIx64 " max gfn: %" PRIx64
-		    " vcpus: %d wss: %" PRIx64 "]\n",
-		    guest_num_pages, region_end_gfn - 1, vcpus,
-		    vcpu_memory_bytes);
+		    " nr_vcpus: %d wss: %" PRIx64 "]\n",
+		    guest_num_pages, region_end_gfn - 1, nr_vcpus, vcpu_memory_bytes);
 
-	pta->gpa = (region_end_gfn - guest_num_pages) * pta->guest_page_size;
+	pta->gpa = (region_end_gfn - guest_num_pages - 1) * pta->guest_page_size;
 	pta->gpa = align_down(pta->gpa, backing_src_pagesz);
 #ifdef __s390x__
 	/* Align to 1M (segment size) */
@@ -198,11 +202,12 @@ struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus,
 	/* Do mapping for the demand paging memory slot */
 	virt_map(vm, guest_test_virt_mem, pta->gpa, guest_num_pages);
 
-	perf_test_setup_vcpus(vm, vcpus, vcpu_memory_bytes, partition_vcpu_memory_access);
+	perf_test_setup_vcpus(vm, nr_vcpus, vcpus, vcpu_memory_bytes,
+			      partition_vcpu_memory_access);
 
 	if (pta->nested) {
 		pr_info("Configuring vCPUs to run in L2 (nested).\n");
-		perf_test_setup_nested(vm, vcpus);
+		perf_test_setup_nested(vm, nr_vcpus, vcpus);
 	}
 
 	ucall_init(vm, NULL);
@@ -230,7 +235,7 @@ uint64_t __weak perf_test_nested_pages(int nr_vcpus)
 	return 0;
 }
 
-void __weak perf_test_setup_nested(struct kvm_vm *vm, int nr_vcpus)
+void __weak perf_test_setup_nested(struct kvm_vm *vm, int nr_vcpus, struct kvm_vcpu **vcpus)
 {
 	pr_info("%s() not support on this architecture, skipping.\n", __func__);
 	exit(KSFT_SKIP);
@@ -251,39 +256,40 @@ static void *vcpu_thread_main(void *data)
 	while (!READ_ONCE(all_vcpu_threads_running))
 		;
 
-	vcpu_thread_fn(&perf_test_args.vcpu_args[vcpu->vcpu_id]);
+	vcpu_thread_fn(&perf_test_args.vcpu_args[vcpu->vcpu_idx]);
 
 	return NULL;
 }
 
-void perf_test_start_vcpu_threads(int vcpus, void (*vcpu_fn)(struct perf_test_vcpu_args *))
+void perf_test_start_vcpu_threads(int nr_vcpus,
+				  void (*vcpu_fn)(struct perf_test_vcpu_args *))
 {
-	int vcpu_id;
+	int i;
 
 	vcpu_thread_fn = vcpu_fn;
 	WRITE_ONCE(all_vcpu_threads_running, false);
 
-	for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) {
-		struct vcpu_thread *vcpu = &vcpu_threads[vcpu_id];
+	for (i = 0; i < nr_vcpus; i++) {
+		struct vcpu_thread *vcpu = &vcpu_threads[i];
 
-		vcpu->vcpu_id = vcpu_id;
+		vcpu->vcpu_idx = i;
 		WRITE_ONCE(vcpu->running, false);
 
 		pthread_create(&vcpu->thread, NULL, vcpu_thread_main, vcpu);
 	}
 
-	for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++) {
-		while (!READ_ONCE(vcpu_threads[vcpu_id].running))
+	for (i = 0; i < nr_vcpus; i++) {
+		while (!READ_ONCE(vcpu_threads[i].running))
 			;
 	}
 
 	WRITE_ONCE(all_vcpu_threads_running, true);
 }
 
-void perf_test_join_vcpu_threads(int vcpus)
+void perf_test_join_vcpu_threads(int nr_vcpus)
 {
-	int vcpu_id;
+	int i;
 
-	for (vcpu_id = 0; vcpu_id < vcpus; vcpu_id++)
-		pthread_join(vcpu_threads[vcpu_id].thread, NULL);
+	for (i = 0; i < nr_vcpus; i++)
+		pthread_join(vcpu_threads[i].thread, NULL);
 }
diff --git a/tools/testing/selftests/kvm/lib/x86_64/perf_test_util.c b/tools/testing/selftests/kvm/lib/x86_64/perf_test_util.c
index f525427a37c4..446820a549ba 100644
--- a/tools/testing/selftests/kvm/lib/x86_64/perf_test_util.c
+++ b/tools/testing/selftests/kvm/lib/x86_64/perf_test_util.c
@@ -77,7 +77,7 @@ void perf_test_setup_ept(struct vmx_pages *vmx, struct kvm_vm *vm)
 	nested_identity_map_1g(vmx, vm, start, end - start);
 }
 
-void perf_test_setup_nested(struct kvm_vm *vm, int nr_vcpus)
+void perf_test_setup_nested(struct kvm_vm *vm, int nr_vcpus, struct kvm_vcpu *vcpus[])
 {
 	struct vmx_pages *vmx, *vmx0 = NULL;
 	struct kvm_regs regs;
@@ -103,9 +103,9 @@ void perf_test_setup_nested(struct kvm_vm *vm, int nr_vcpus)
 		 * Override the vCPU to run perf_test_l1_guest_code() which will
 		 * bounce it into L2 before calling perf_test_guest_code().
 		 */
-		vcpu_regs_get(vm, vcpu_id, &regs);
+		vcpu_regs_get(vm, vcpus[vcpu_id]->id, &regs);
 		regs.rip = (unsigned long) perf_test_l1_guest_code;
-		vcpu_regs_set(vm, vcpu_id, &regs);
-		vcpu_args_set(vm, vcpu_id, 2, vmx_gva, vcpu_id);
+		vcpu_regs_set(vm, vcpus[vcpu_id]->id, &regs);
+		vcpu_args_set(vm, vcpus[vcpu_id]->id, 2, vmx_gva, vcpu_id);
 	}
 }