[Intel-gfx] [PATCH igt] igt/perf_pmu: Keep batch_duration_ns as the minimum measurement duration

[Chris Wilson]

We have chosen batch_duration_ns to be the minimum duration we need to
meet our accuracy requirements for legacy ringbuffer PMU sampling. As
such, we need to be careful to use multiples of it during tests, and not
split it into different phases with a test, like multi_client does.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
---
 tests/perf_pmu.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/tests/perf_pmu.c b/tests/perf_pmu.c
index 8d58ecea..e527ba58 100644
--- a/tests/perf_pmu.c
+++ b/tests/perf_pmu.c
@@ -638,9 +638,9 @@ multi_client(int gem_fd, const struct intel_execution_engine2 *e)
 	fd[1] = open_pmu(config);
 
 	spin = igt_spin_batch_new(gem_fd, 0, e2ring(gem_fd, e), 0);
-	igt_spin_batch_set_timeout(spin, batch_duration_ns);
+	igt_spin_batch_set_timeout(spin, 2 * batch_duration_ns);
 
-	slept = measured_usleep(batch_duration_ns / 3000);
+	slept = measured_usleep(batch_duration_ns / 1000);
 	val[1] = pmu_read_single(fd[1]);
 	close(fd[1]);
 
@@ -651,7 +651,7 @@ multi_client(int gem_fd, const struct intel_execution_engine2 *e)
 	igt_spin_batch_free(gem_fd, spin);
 	close(fd[0]);
 
-	assert_within_epsilon(val[0], batch_duration_ns, tolerance);
+	assert_within_epsilon(val[0], 2 * batch_duration_ns, tolerance);
 	assert_within_epsilon(val[1], slept, tolerance);
 }
 
-- 
2.15.0