[igt-dev] [PATCH i-g-t 1/3] tests/perf_pmu: Tighten busy measurement

[Tvrtko Ursulin]

From: Tvrtko Ursulin <tvrtko.ursulin at intel.com>

In cases where we manually terminate the busy batch, we always want to
sample busyness while the batch is running, just before we will
terminate it, and not the other way around. This way we make the window
for unwated idleness getting sampled smaller.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
---
 tests/perf_pmu.c | 28 +++++++++++++---------------
 1 file changed, 13 insertions(+), 15 deletions(-)

diff --git a/tests/perf_pmu.c b/tests/perf_pmu.c
index 2f7d33414a53..bf16e5e8b1f9 100644
--- a/tests/perf_pmu.c
+++ b/tests/perf_pmu.c
@@ -146,10 +146,9 @@ single(int gem_fd, const struct intel_execution_engine2 *e, bool busy)
 		spin = NULL;
 
 	slept = measured_usleep(batch_duration_ns / 1000);
-	igt_spin_batch_end(spin);
-
 	val = pmu_read_single(fd);
 
+	igt_spin_batch_end(spin);
 	igt_spin_batch_free(gem_fd, spin);
 	close(fd);
 
@@ -256,7 +255,7 @@ busy_double_start(int gem_fd, const struct intel_execution_engine2 *e)
 	gem_quiescent_gpu(gem_fd);
 }
 
-static void log_busy(int fd, unsigned int num_engines, uint64_t *val)
+static void log_busy(unsigned int num_engines, uint64_t *val)
 {
 	char buf[1024];
 	int rem = sizeof(buf);
@@ -303,14 +302,14 @@ busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
 
 	spin = igt_spin_batch_new(gem_fd, 0, e2ring(gem_fd, e), 0);
 	slept = measured_usleep(batch_duration_ns / 1000);
-	igt_spin_batch_end(spin);
-
 	pmu_read_multi(fd[0], num_engines, val);
-	log_busy(fd[0], num_engines, val);
 
+	igt_spin_batch_end(spin);
 	igt_spin_batch_free(gem_fd, spin);
 	close(fd[0]);
 
+	log_busy(num_engines, val);
+
 	assert_within_epsilon(val[busy_idx], slept, tolerance);
 	for (i = 0; i < num_engines; i++) {
 		if (i == busy_idx)
@@ -364,14 +363,14 @@ most_busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
 		fd[i] = open_group(val[i], fd[0]);
 
 	slept = measured_usleep(batch_duration_ns / 1000);
-	igt_spin_batch_end(spin);
-
 	pmu_read_multi(fd[0], num_engines, val);
-	log_busy(fd[0], num_engines, val);
 
+	igt_spin_batch_end(spin);
 	igt_spin_batch_free(gem_fd, spin);
 	close(fd[0]);
 
+	log_busy(num_engines, val);
+
 	for (i = 0; i < num_engines; i++) {
 		if (i == idle_idx)
 			assert_within_epsilon(val[i], 0.0f, tolerance);
@@ -420,14 +419,14 @@ all_busy_check_all(int gem_fd, const unsigned int num_engines)
 		fd[i] = open_group(val[i], fd[0]);
 
 	slept = measured_usleep(batch_duration_ns / 1000);
-	igt_spin_batch_end(spin);
-
 	pmu_read_multi(fd[0], num_engines, val);
-	log_busy(fd[0], num_engines, val);
 
+	igt_spin_batch_end(spin);
 	igt_spin_batch_free(gem_fd, spin);
 	close(fd[0]);
 
+	log_busy(num_engines, val);
+
 	for (i = 0; i < num_engines; i++)
 		assert_within_epsilon(val[i], slept, tolerance);
 	gem_quiescent_gpu(gem_fd);
@@ -903,12 +902,11 @@ static void cpu_hotplug(int gem_fd)
 
 	igt_waitchildren();
 
-	igt_spin_batch_end(spin);
-	gem_sync(gem_fd, spin->handle);
-
 	ref = igt_nsec_elapsed(&start);
 	val = pmu_read_single(fd);
 
+	igt_spin_batch_end(spin);
+	gem_sync(gem_fd, spin->handle);
 	igt_spin_batch_free(gem_fd, spin);
 	close(fd);
 
-- 
2.14.1