Commit e5bc4d46 authored by Paolo Bonzini's avatar Paolo Bonzini
Browse files

Merge branch 'kvm-selftest' into kvm-master 

- Cleanups for the perf test infrastructure and mapping hugepages

- Avoid contention on mmap_sem when the guests start to run

- Add event channel upcall support to xen_shinfo_test
parents 84886c26 e2bd9365

tools/testing/selftests/kvm/access_tracking_perf_test.c

+10 −44
Original line number Diff line number Diff line
@@ -47,7 +47,7 @@ 
#include "guest_modes.h"



/* Global variable used to synchronize all of the vCPU threads. */

static int iteration = -1;

static int iteration;



/* Defines what vCPU threads should do during a given iteration. */

static enum {
@@ -215,12 +215,11 @@ static bool spin_wait_for_next_iteration(int *current_iteration) 
	return true;

}



static void *vcpu_thread_main(void *arg)

static void vcpu_thread_main(struct perf_test_vcpu_args *vcpu_args)

{

	struct perf_test_vcpu_args *vcpu_args = arg;

	struct kvm_vm *vm = perf_test_args.vm;

	int vcpu_id = vcpu_args->vcpu_id;

	int current_iteration = -1;

	int current_iteration = 0;



	while (spin_wait_for_next_iteration(&current_iteration)) {

		switch (READ_ONCE(iteration_work)) {
@@ -235,8 +234,6 @@ static void *vcpu_thread_main(void *arg) 


		vcpu_last_completed_iteration[vcpu_id] = current_iteration;

	}



	return NULL;

}



static void spin_wait_for_vcpu(int vcpu_id, int target_iteration)
@@ -277,8 +274,7 @@ static void run_iteration(struct kvm_vm *vm, int vcpus, const char *description) 
static void access_memory(struct kvm_vm *vm, int vcpus, enum access_type access,

			  const char *description)

{

	perf_test_args.wr_fract = (access == ACCESS_READ) ? INT_MAX : 1;

	sync_global_to_guest(vm, perf_test_args);

	perf_test_set_wr_fract(vm, (access == ACCESS_READ) ? INT_MAX : 1);

	iteration_work = ITERATION_ACCESS_MEMORY;

	run_iteration(vm, vcpus, description);

}
@@ -296,48 +292,16 @@ static void mark_memory_idle(struct kvm_vm *vm, int vcpus) 
	run_iteration(vm, vcpus, "Mark memory idle");

}



static pthread_t *create_vcpu_threads(int vcpus)

{

	pthread_t *vcpu_threads;

	int i;



	vcpu_threads = malloc(vcpus * sizeof(vcpu_threads[0]));

	TEST_ASSERT(vcpu_threads, "Failed to allocate vcpu_threads.");



	for (i = 0; i < vcpus; i++) {

		vcpu_last_completed_iteration[i] = iteration;

		pthread_create(&vcpu_threads[i], NULL, vcpu_thread_main,

			       &perf_test_args.vcpu_args[i]);

	}



	return vcpu_threads;

}



static void terminate_vcpu_threads(pthread_t *vcpu_threads, int vcpus)

{

	int i;



	/* Set done to signal the vCPU threads to exit */

	done = true;



	for (i = 0; i < vcpus; i++)

		pthread_join(vcpu_threads[i], NULL);

}



static void run_test(enum vm_guest_mode mode, void *arg)

{

	struct test_params *params = arg;

	struct kvm_vm *vm;

	pthread_t *vcpu_threads;

	int vcpus = params->vcpus;



	vm = perf_test_create_vm(mode, vcpus, params->vcpu_memory_bytes, 1,

				 params->backing_src);

				 params->backing_src, !overlap_memory_access);



	perf_test_setup_vcpus(vm, vcpus, params->vcpu_memory_bytes,

			      !overlap_memory_access);



	vcpu_threads = create_vcpu_threads(vcpus);

	perf_test_start_vcpu_threads(vcpus, vcpu_thread_main);



	pr_info("\n");

	access_memory(vm, vcpus, ACCESS_WRITE, "Populating memory");
@@ -352,8 +316,10 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
	mark_memory_idle(vm, vcpus);

	access_memory(vm, vcpus, ACCESS_READ, "Reading from idle memory");



	terminate_vcpu_threads(vcpu_threads, vcpus);

	free(vcpu_threads);

	/* Set done to signal the vCPU threads to exit */

	done = true;



	perf_test_join_vcpu_threads(vcpus);

	perf_test_destroy_vm(vm);

}

tools/testing/selftests/kvm/demand_paging_test.c

+9 −47
Original line number Diff line number Diff line
@@ -42,10 +42,9 @@ static uint64_t guest_percpu_mem_size = DEFAULT_PER_VCPU_MEM_SIZE; 
static size_t demand_paging_size;

static char *guest_data_prototype;



static void *vcpu_worker(void *data)

static void vcpu_worker(struct perf_test_vcpu_args *vcpu_args)

{

	int ret;

	struct perf_test_vcpu_args *vcpu_args = (struct perf_test_vcpu_args *)data;

	int vcpu_id = vcpu_args->vcpu_id;

	struct kvm_vm *vm = perf_test_args.vm;

	struct kvm_run *run;
@@ -68,8 +67,6 @@ static void *vcpu_worker(void *data) 
	ts_diff = timespec_elapsed(start);

	PER_VCPU_DEBUG("vCPU %d execution time: %ld.%.9lds\n", vcpu_id,

		       ts_diff.tv_sec, ts_diff.tv_nsec);



	return NULL;

}



static int handle_uffd_page_request(int uffd_mode, int uffd, uint64_t addr)
@@ -282,7 +279,6 @@ struct test_params { 
static void run_test(enum vm_guest_mode mode, void *arg)

{

	struct test_params *p = arg;

	pthread_t *vcpu_threads;

	pthread_t *uffd_handler_threads = NULL;

	struct uffd_handler_args *uffd_args = NULL;

	struct timespec start;
@@ -293,9 +289,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
	int r;



	vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size, 1,

				 p->src_type);



	perf_test_args.wr_fract = 1;

				 p->src_type, p->partition_vcpu_memory_access);



	demand_paging_size = get_backing_src_pagesz(p->src_type);
@@ -304,12 +298,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
		    "Failed to allocate buffer for guest data pattern");

	memset(guest_data_prototype, 0xAB, demand_paging_size);



	vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads));

	TEST_ASSERT(vcpu_threads, "Memory allocation failed");



	perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size,

			      p->partition_vcpu_memory_access);



	if (p->uffd_mode) {

		uffd_handler_threads =

			malloc(nr_vcpus * sizeof(*uffd_handler_threads));
@@ -322,26 +310,15 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
		TEST_ASSERT(pipefds, "Unable to allocate memory for pipefd");



		for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) {

			vm_paddr_t vcpu_gpa;

			struct perf_test_vcpu_args *vcpu_args;

			void *vcpu_hva;

			void *vcpu_alias;

			uint64_t vcpu_mem_size;





			if (p->partition_vcpu_memory_access) {

				vcpu_gpa = guest_test_phys_mem +

					   (vcpu_id * guest_percpu_mem_size);

				vcpu_mem_size = guest_percpu_mem_size;

			} else {

				vcpu_gpa = guest_test_phys_mem;

				vcpu_mem_size = guest_percpu_mem_size * nr_vcpus;

			}

			PER_VCPU_DEBUG("Added VCPU %d with test mem gpa [%lx, %lx)\n",

				       vcpu_id, vcpu_gpa, vcpu_gpa + vcpu_mem_size);

			vcpu_args = &perf_test_args.vcpu_args[vcpu_id];



			/* Cache the host addresses of the region */

			vcpu_hva = addr_gpa2hva(vm, vcpu_gpa);

			vcpu_alias = addr_gpa2alias(vm, vcpu_gpa);

			vcpu_hva = addr_gpa2hva(vm, vcpu_args->gpa);

			vcpu_alias = addr_gpa2alias(vm, vcpu_args->gpa);



			/*

			 * Set up user fault fd to handle demand paging
@@ -355,32 +332,18 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
					    pipefds[vcpu_id * 2], p->uffd_mode,

					    p->uffd_delay, &uffd_args[vcpu_id],

					    vcpu_hva, vcpu_alias,

					    vcpu_mem_size);

					    vcpu_args->pages * perf_test_args.guest_page_size);

		}

	}



	/* Export the shared variables to the guest */

	sync_global_to_guest(vm, perf_test_args);



	pr_info("Finished creating vCPUs and starting uffd threads\n");



	clock_gettime(CLOCK_MONOTONIC, &start);



	for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) {

		pthread_create(&vcpu_threads[vcpu_id], NULL, vcpu_worker,

			       &perf_test_args.vcpu_args[vcpu_id]);

	}



	perf_test_start_vcpu_threads(nr_vcpus, vcpu_worker);

	pr_info("Started all vCPUs\n");



	/* Wait for the vcpu threads to quit */

	for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) {

		pthread_join(vcpu_threads[vcpu_id], NULL);

		PER_VCPU_DEBUG("Joined thread for vCPU %d\n", vcpu_id);

	}



	perf_test_join_vcpu_threads(nr_vcpus);

	ts_diff = timespec_elapsed(start);



	pr_info("All vCPU threads joined\n");



	if (p->uffd_mode) {
@@ -404,7 +367,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
	perf_test_destroy_vm(vm);



	free(guest_data_prototype);

	free(vcpu_threads);

	if (p->uffd_mode) {

		free(uffd_handler_threads);

		free(uffd_args);

tools/testing/selftests/kvm/dirty_log_perf_test.c

+7 −22
Original line number Diff line number Diff line
@@ -31,7 +31,7 @@ static bool host_quit; 
static int iteration;

static int vcpu_last_completed_iteration[KVM_MAX_VCPUS];



static void *vcpu_worker(void *data)

static void vcpu_worker(struct perf_test_vcpu_args *vcpu_args)

{

	int ret;

	struct kvm_vm *vm = perf_test_args.vm;
@@ -41,7 +41,6 @@ static void *vcpu_worker(void *data) 
	struct timespec ts_diff;

	struct timespec total = (struct timespec){0};

	struct timespec avg;

	struct perf_test_vcpu_args *vcpu_args = (struct perf_test_vcpu_args *)data;

	int vcpu_id = vcpu_args->vcpu_id;



	run = vcpu_state(vm, vcpu_id);
@@ -83,8 +82,6 @@ static void *vcpu_worker(void *data) 
	pr_debug("\nvCPU %d dirtied 0x%lx pages over %d iterations in %ld.%.9lds. (Avg %ld.%.9lds/iteration)\n",

		vcpu_id, pages_count, vcpu_last_completed_iteration[vcpu_id],

		total.tv_sec, total.tv_nsec, avg.tv_sec, avg.tv_nsec);



	return NULL;

}



struct test_params {
@@ -170,7 +167,6 @@ static void free_bitmaps(unsigned long *bitmaps[], int slots) 
static void run_test(enum vm_guest_mode mode, void *arg)

{

	struct test_params *p = arg;

	pthread_t *vcpu_threads;

	struct kvm_vm *vm;

	unsigned long **bitmaps;

	uint64_t guest_num_pages;
@@ -186,9 +182,10 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
	struct timespec clear_dirty_log_total = (struct timespec){0};



	vm = perf_test_create_vm(mode, nr_vcpus, guest_percpu_mem_size,

				 p->slots, p->backing_src);

				 p->slots, p->backing_src,

				 p->partition_vcpu_memory_access);



	perf_test_args.wr_fract = p->wr_fract;

	perf_test_set_wr_fract(vm, p->wr_fract);



	guest_num_pages = (nr_vcpus * guest_percpu_mem_size) >> vm_get_page_shift(vm);

	guest_num_pages = vm_adjust_num_guest_pages(mode, guest_num_pages);
@@ -203,25 +200,15 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
		vm_enable_cap(vm, &cap);

	}



	vcpu_threads = malloc(nr_vcpus * sizeof(*vcpu_threads));

	TEST_ASSERT(vcpu_threads, "Memory allocation failed");



	perf_test_setup_vcpus(vm, nr_vcpus, guest_percpu_mem_size,

			      p->partition_vcpu_memory_access);



	sync_global_to_guest(vm, perf_test_args);



	/* Start the iterations */

	iteration = 0;

	host_quit = false;



	clock_gettime(CLOCK_MONOTONIC, &start);

	for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++) {

	for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++)

		vcpu_last_completed_iteration[vcpu_id] = -1;



		pthread_create(&vcpu_threads[vcpu_id], NULL, vcpu_worker,

			       &perf_test_args.vcpu_args[vcpu_id]);

	}

	perf_test_start_vcpu_threads(nr_vcpus, vcpu_worker);



	/* Allow the vCPUs to populate memory */

	pr_debug("Starting iteration %d - Populating\n", iteration);
@@ -290,8 +277,7 @@ static void run_test(enum vm_guest_mode mode, void *arg) 


	/* Tell the vcpu thread to quit */

	host_quit = true;

	for (vcpu_id = 0; vcpu_id < nr_vcpus; vcpu_id++)

		pthread_join(vcpu_threads[vcpu_id], NULL);

	perf_test_join_vcpu_threads(nr_vcpus);



	avg = timespec_div(get_dirty_log_total, p->iterations);

	pr_info("Get dirty log over %lu iterations took %ld.%.9lds. (Avg %ld.%.9lds/iteration)\n",
@@ -306,7 +292,6 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
	}



	free_bitmaps(bitmaps, p->slots);

	free(vcpu_threads);

	perf_test_destroy_vm(vm);

}

tools/testing/selftests/kvm/dirty_log_test.c

+3 −3
Original line number Diff line number Diff line
@@ -115,7 +115,7 @@ static void guest_code(void) 
			addr = guest_test_virt_mem;

			addr += (READ_ONCE(random_array[i]) % guest_num_pages)

				* guest_page_size;

			addr &= ~(host_page_size - 1);

			addr = align_down(addr, host_page_size);

			*(uint64_t *)addr = READ_ONCE(iteration);

		}
@@ -737,14 +737,14 @@ static void run_test(enum vm_guest_mode mode, void *arg) 
	if (!p->phys_offset) {

		guest_test_phys_mem = (vm_get_max_gfn(vm) -

				       guest_num_pages) * guest_page_size;

		guest_test_phys_mem &= ~(host_page_size - 1);

		guest_test_phys_mem = align_down(guest_test_phys_mem, host_page_size);

	} else {

		guest_test_phys_mem = p->phys_offset;

	}



#ifdef __s390x__

	/* Align to 1M (segment size) */

	guest_test_phys_mem &= ~((1 << 20) - 1);

	guest_test_phys_mem = align_down(guest_test_phys_mem, 1 << 20);

#endif



	pr_info("guest physical test memory offset: 0x%lx\n", guest_test_phys_mem);

tools/testing/selftests/kvm/include/perf_test_util.h

+11 −12
Original line number Diff line number Diff line
@@ -8,6 +8,8 @@ 
#ifndef SELFTEST_KVM_PERF_TEST_UTIL_H

#define SELFTEST_KVM_PERF_TEST_UTIL_H



#include <pthread.h>



#include "kvm_util.h"



/* Default guest test virtual memory offset */
@@ -18,6 +20,7 @@ 
#define PERF_TEST_MEM_SLOT_INDEX	1



struct perf_test_vcpu_args {

	uint64_t gpa;

	uint64_t gva;

	uint64_t pages;
@@ -27,7 +30,7 @@ struct perf_test_vcpu_args { 


struct perf_test_args {

	struct kvm_vm *vm;

	uint64_t host_page_size;

	uint64_t gpa;

	uint64_t guest_page_size;

	int wr_fract;
@@ -36,19 +39,15 @@ struct perf_test_args { 


extern struct perf_test_args perf_test_args;



/*

 * Guest physical memory offset of the testing memory slot.

 * This will be set to the topmost valid physical address minus

 * the test memory size.

 */

extern uint64_t guest_test_phys_mem;



struct kvm_vm *perf_test_create_vm(enum vm_guest_mode mode, int vcpus,

				   uint64_t vcpu_memory_bytes, int slots,

				   enum vm_mem_backing_src_type backing_src);

void perf_test_destroy_vm(struct kvm_vm *vm);

void perf_test_setup_vcpus(struct kvm_vm *vm, int vcpus,

			   uint64_t vcpu_memory_bytes,

				   enum vm_mem_backing_src_type backing_src,

				   bool partition_vcpu_memory_access);

void perf_test_destroy_vm(struct kvm_vm *vm);



void perf_test_set_wr_fract(struct kvm_vm *vm, int wr_fract);



void perf_test_start_vcpu_threads(int vcpus, void (*vcpu_fn)(struct perf_test_vcpu_args *));

void perf_test_join_vcpu_threads(int vcpus);



#endif /* SELFTEST_KVM_PERF_TEST_UTIL_H */