[igt-dev] [PATCH i-g-t] tests/xe_vm: make sure test can finish within time

[Niranjana Vishwanathapura]

On Tue, Apr 11, 2023 at 07:45:55PM +0000, Chang, Bruce wrote:
>the BAT takes a lot longer (> 48s) than expected (~5s). This seems
>due to the system is too busy to wake up the main thread which is in
>sleep.
>
>So, change the test to run in a timeout instead of relying on main
>thread.
>
>on another hand, also change the logic to make sure all threads start
>together.
>
>Cc: Niranjana Vishwanathapura <niranjana.vishwanathapura at intel.com>
>Signed-off-by: Bruce Chang <yu.bruce.chang at intel.com>

LGTM.
Interesting to see that main thread (with same thread priority) never
got schduled, given the threads are doing a lot of ioctl calls.
In any case, this change is a better approach.

Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura at intel.com>

>---
> tests/xe/xe_prime_self_import.c | 25 ++++++++++++++-----------
> 1 file changed, 14 insertions(+), 11 deletions(-)
>
>diff --git a/tests/xe/xe_prime_self_import.c b/tests/xe/xe_prime_self_import.c
>index 5710cff92..fb677b50b 100644
>--- a/tests/xe/xe_prime_self_import.c
>+++ b/tests/xe/xe_prime_self_import.c
>@@ -58,7 +58,8 @@ IGT_TEST_DESCRIPTION("Check whether prime import/export works on the same"
> #define BO_SIZE (16*1024)
>
> static char counter;
>-volatile int pls_die = 0;
>+static int g_time_out = 5;
>+static pthread_barrier_t g_barrier;
>
> static void
> check_bo(int fd1, uint32_t handle1, int fd2, uint32_t handle2)
>@@ -272,7 +273,9 @@ static void *thread_fn_reimport_vs_close(void *p)
> 	int dma_buf_fd = fds[1];
> 	uint32_t handle;
>
>-	while (!pls_die) {
>+	pthread_barrier_wait(&g_barrier);
>+
>+	igt_until_timeout(g_time_out) {
> 		handle = prime_fd_to_handle(fd, dma_buf_fd);
>
> 		close_bo.handle = handle;
>@@ -306,6 +309,7 @@ static void *thread_fn_reimport_vs_close(void *p)
> 	obj_count = 0;
>
> 	num_threads = sysconf(_SC_NPROCESSORS_ONLN);
>+	igt_info("create %d threads\n", num_threads);
>
> 	threads = calloc(num_threads, sizeof(pthread_t));
>
>@@ -315,6 +319,7 @@ static void *thread_fn_reimport_vs_close(void *p)
> 	handle = xe_bo_create(fds[0], 0, 0, BO_SIZE);
>
> 	fds[1] = prime_handle_to_fd(fds[0], handle);
>+	pthread_barrier_init(&g_barrier, NULL, num_threads);
>
> 	for (i = 0; i < num_threads; i++) {
> 		r = pthread_create(&threads[i], NULL,
>@@ -323,15 +328,12 @@ static void *thread_fn_reimport_vs_close(void *p)
> 		igt_assert_eq(r, 0);
> 	}
>
>-	sleep(5);
>-
>-	pls_die = 1;
>-
> 	for (i = 0;  i < num_threads; i++) {
> 		pthread_join(threads[i], &status);
> 		igt_assert(status == 0);
> 	}
>
>+	pthread_barrier_destroy(&g_barrier);
> 	xe_device_put(fds[0]);
> 	close(fds[0]);
> 	close(fds[1]);
>@@ -353,7 +355,9 @@ static void *thread_fn_export_vs_close(void *p)
> 	int fd = (uintptr_t)p;
> 	uint32_t handle;
>
>-	while (!pls_die) {
>+	pthread_barrier_wait(&g_barrier);
>+
>+	igt_until_timeout(g_time_out) {
> 		/* We want to race gem close against prime export on handle one.*/
> 		handle = xe_bo_create(fd, 0, 0, 4096);
> 		if (handle != 1)
>@@ -395,6 +399,7 @@ static void test_export_close_race(void)
> 	int fake;
>
> 	num_threads = sysconf(_SC_NPROCESSORS_ONLN);
>+	igt_info("create %d threads\n", num_threads);
>
> 	threads = calloc(num_threads, sizeof(pthread_t));
>
>@@ -408,6 +413,7 @@ static void test_export_close_race(void)
>
> 	fd = drm_open_driver(DRIVER_XE);
> 	xe_device_get(fd);
>+	pthread_barrier_init(&g_barrier, NULL, num_threads);
>
> 	for (i = 0; i < num_threads; i++) {
> 		r = pthread_create(&threads[i], NULL,
>@@ -416,15 +422,12 @@ static void test_export_close_race(void)
> 		igt_assert_eq(r, 0);
> 	}
>
>-	sleep(5);
>-
>-	pls_die = 1;
>-
> 	for (i = 0;  i < num_threads; i++) {
> 		pthread_join(threads[i], &status);
> 		igt_assert(status == 0);
> 	}
>
>+	pthread_barrier_destroy(&g_barrier);
> 	xe_device_put(fd);
> 	close(fd);
>
>-- 
>2.25.1
>