[PATCH i-g-t 2/3] tests/intel/xe_pmu: Add PMU test to validate engine activity stats

[Umesh Nerlige Ramappa]

On Fri, Feb 14, 2025 at 10:55:38AM -0800, Umesh Nerlige Ramappa wrote:
>On Wed, Feb 12, 2025 at 03:28:28PM +0530, Riana Tauro wrote:
>>Add a test to validate engine activity by reading PMU counters
>>(engine-active-ticks, engine-total-ticks) when running
>>workload on every engine
>>
>>Signed-off-by: Riana Tauro <riana.tauro at intel.com>
>>---
>>tests/intel/xe_pmu.c | 147 +++++++++++++++++++++++++++++++++++++++++++
>>tests/meson.build    |   1 +
>>2 files changed, 148 insertions(+)
>>create mode 100644 tests/intel/xe_pmu.c
>>
>>diff --git a/tests/intel/xe_pmu.c b/tests/intel/xe_pmu.c
>>new file mode 100644
>>index 000000000..cbb825755
>>--- /dev/null
>>+++ b/tests/intel/xe_pmu.c
>>@@ -0,0 +1,147 @@
>>+// SPDX-License-Identifier: MIT
>>+/*
>>+ * Copyright © 2025 Intel Corporation
>>+ */
>>+
>>+/**
>>+ * TEST: Test Xe PMU(Performance Monitoring Unit) functionality
>>+ * Category: Metrics
>>+ * Functionality: Power/Perf
>>+ * Mega feature: Performance Monitoring Unit
>>+ * Sub-category: Telemetry
>>+ * Test category: Functional tests
>>+ */
>>+
>>+#include "igt.h"
>>+#include "igt_perf.h"
>>+
>>+#include "xe/xe_ioctl.h"
>>+#include "xe/xe_spin.h"
>>+
>>+#define SLEEP_DURATION 2 /* in seconds */
>>+const double tolerance = 0.1;
>>+
>>+#define assert_within_epsilon(x, ref, tolerance) \
>>+	igt_assert_f((double)(x) <= (1.0 + (tolerance)) * (double)(ref) && \
>>+		     (double)(x) >= (1.0 - (tolerance)) * (double)(ref), \
>>+		     "'%s' != '%s' (%f not within +%.1f%%/-%.1f%% tolerance of %f)\n",\
>>+		     #x, #ref, (double)(x), \
>>+		     (tolerance) * 100.0, (tolerance) * 100.0, \
>>+		     (double)(ref))
>>+
>>+static int open_group(int xe, uint64_t config, int group)
>>+{
>>+	int fd;
>>+
>>+	fd = igt_perf_open_group(xe_perf_type_id(xe), config, group);
>>+	igt_skip_on(fd < 0 && errno == ENODEV);
>>+	igt_assert(fd >= 0);
>>+
>>+	return fd;
>>+}
>>+
>>+static uint64_t pmu_read_multi(int fd, unsigned int num, uint64_t *val)
>>+{
>>+	uint64_t buf[2 + num];
>>+	unsigned int i;
>>+
>>+	igt_assert_eq(read(fd, buf, sizeof(buf)), sizeof(buf));
>>+
>>+	for (i = 0; i < num; i++)
>>+		val[i] = buf[2 + i];
>>+
>>+	return buf[1];
>>+}
>>+
>>+static uint64_t get_event_config(int xe, unsigned int gt, struct drm_xe_engine_class_instance *eci,
>>+				 const char *event)
>>+{
>>+	int ret;
>>+	char xe_device[100];
>>+	uint64_t pmu_config;
>>+	u32 start, end;
>
>end is unused, not sure if the c6 patches use them. If unused, then I 
>recommend just dropping it from the helper as well and s/start/shift/.
>
>>+
>>+	xe_perf_device(xe, xe_device, sizeof(xe_device));
>>+	ret = perf_event_config(xe_device, event, &pmu_config);
>>+	igt_assert(ret >= 0);
>
>Well, I commented on Vinay's patch to assert within the helper, so 
>these checks can be removed here. If that happens, then please drop 
>the return value checks here. If not, then checks needed below as well 
>at (1) and (2).
>
>>+	ret = perf_event_format(xe_device, "gt", &start, &end);
>>+	igt_assert(ret >= 0);
>>+	pmu_config |= (uint64_t)gt << start;
>>+
>>+	if (eci) {
>>+		ret = perf_event_format(xe_device, "engine_class", &start, &end);
>(1)
>>+		pmu_config |= (uint64_t)eci->engine_class << start;
>>+		ret = perf_event_format(xe_device, "engine_instance", &start, &end);
>(2)
>>+		pmu_config |= (uint64_t)eci->engine_instance << start;
>>+	}
>>+
>>+	return pmu_config;
>>+}

also s/busy/active/

>
>Thanks,
>Umesh
>>+
>>+/**
>>+ * SUBTEST: engine-activity
>>+ * Description: Test to validate engine activity stats by running a workload and
>>+ *              reading the active ticks and total ticks PMU counters
>>+ */
>>+static void engine_activity(int fd, struct drm_xe_engine_class_instance *eci)
>>+{
>>+	uint64_t config, busy_ticks, total_ticks, before[2], after[2];
>>+	struct xe_cork *cork = NULL;
>>+	uint32_t vm;
>>+	int pmu_fd[2];
>>+
>>+	config = get_event_config(fd, eci->gt_id, eci, "engine-active-ticks");
>>+	pmu_fd[0] = open_group(fd, config, -1);
>>+
>>+	config = get_event_config(fd, eci->gt_id, eci, "engine-total-ticks");
>>+	pmu_fd[1] = open_group(fd, config, pmu_fd[0]);
>>+
>>+	vm = xe_vm_create(fd, 0, 0);
>>+	cork = xe_cork_create_opts(fd, eci, vm, 1, 1);
>>+	xe_cork_sync_start(fd, cork);
>>+
>>+	pmu_read_multi(pmu_fd[0], 2, before);
>>+	usleep(SLEEP_DURATION * USEC_PER_SEC);
>>+	pmu_read_multi(pmu_fd[0], 2, after);
>>+
>>+	xe_cork_sync_end(fd, cork);
>>+
>>+	busy_ticks = after[0] - before[0];
>>+	total_ticks = after[1] - before[1];
>>+
>>+	igt_debug("Engine active ticks:  after %ld, before %ld delta %ld\n", after[0], before[0],
>>+		  busy_ticks);
>>+	igt_debug("Total ticks: after %ld, before %ld delta %ld\n", after[1], before[1],
>>+		  total_ticks);
>>+
>>+	if (cork)
>>+		xe_cork_destroy(fd, cork);
>>+
>>+	xe_vm_destroy(fd, vm);
>>+
>>+	close(pmu_fd[0]);
>>+	close(pmu_fd[1]);
>>+
>>+	assert_within_epsilon(busy_ticks, total_ticks, tolerance);
>>+}
>>+
>>+igt_main
>>+{
>>+	int fd;
>>+	struct drm_xe_engine_class_instance *hwe;
>>+
>>+	igt_fixture {
>>+		fd = drm_open_driver(DRIVER_XE);
>>+	}
>>+
>>+	igt_describe("Validate engine activity with workload running by reading pmu counters");
>>+	igt_subtest_with_dynamic("engine-activity")
>>+		xe_for_each_engine(fd, hwe)
>>+			igt_dynamic_f("engine-%s%d", xe_engine_class_string(hwe->engine_class),
>>+				      hwe->engine_instance)
>>+				engine_activity(fd, hwe);
>>+
>>+	igt_fixture {
>>+		close(fd);
>>+	}
>>+}
>>diff --git a/tests/meson.build b/tests/meson.build
>>index 33dffad31..d20f50766 100644
>>--- a/tests/meson.build
>>+++ b/tests/meson.build
>>@@ -309,6 +309,7 @@ intel_xe_progs = [
>>	'xe_pat',
>>	'xe_peer2peer',
>>	'xe_pm',
>>+	'xe_pmu',
>>	'xe_pm_residency',
>>	'xe_prime_self_import',
>>	'xe_query',
>>-- 
>>2.47.1
>>