All of lore.kernel.org
 help / color / mirror / Atom feed
* [Intel-gfx] [PATCH] drm/i915/selftests: Show the full scaling curve on failure
@ 2020-04-21 12:37 Chris Wilson
  2020-04-21 12:46 ` Chris Wilson
                   ` (3 more replies)
  0 siblings, 4 replies; 6+ messages in thread
From: Chris Wilson @ 2020-04-21 12:37 UTC (permalink / raw)
  To: intel-gfx; +Cc: Chris Wilson

If we detect that the RPS end points do not scale perfectly, take the
time to measure all the in between values as well. We are aborting the
test, so we might as well spend the available time gathering critical
debug information instead.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
---
 drivers/gpu/drm/i915/gt/selftest_rps.c | 36 ++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.c b/drivers/gpu/drm/i915/gt/selftest_rps.c
index e0a791eac752..f8c416ab8539 100644
--- a/drivers/gpu/drm/i915/gt/selftest_rps.c
+++ b/drivers/gpu/drm/i915/gt/selftest_rps.c
@@ -484,11 +484,29 @@ int live_rps_frequency_cs(void *arg)
 		if (!scaled_within(max.freq * min.count,
 				   min.freq * max.count,
 				   2, 3)) {
+			int f;
+
 			pr_err("%s: CS did not scale with frequency! scaled min:%llu, max:%llu\n",
 			       engine->name,
 			       max.freq * min.count,
 			       min.freq * max.count);
 			show_pcu_config(rps);
+
+			for (f = min.freq + 1; f <= rps->max_freq; f++) {
+				int act = f;
+				u64 count;
+
+				count = measure_cs_frequency_at(rps, engine, &act);
+				if (act < f)
+					break;
+
+				pr_info("%s: %x:%uMHz: %lluKHz [%d%%]\n",
+					engine->name,
+					act, intel_gpu_freq(rps, act), count,
+					(int)DIV64_U64_ROUND_CLOSEST(100 * min.freq * count,
+								     act * min.count));
+			}
+
 			err = -EINVAL;
 		}
 
@@ -593,11 +611,29 @@ int live_rps_frequency_srm(void *arg)
 		if (!scaled_within(max.freq * min.count,
 				   min.freq * max.count,
 				   1, 2)) {
+			int f;
+
 			pr_err("%s: CS did not scale with frequency! scaled min:%llu, max:%llu\n",
 			       engine->name,
 			       max.freq * min.count,
 			       min.freq * max.count);
 			show_pcu_config(rps);
+
+			for (f = min.freq + 1; f <= rps->max_freq; f++) {
+				int act = f;
+				u64 count;
+
+				count = measure_frequency_at(rps, cntr, &act);
+				if (act < f)
+					break;
+
+				pr_info("%s: %x:%uMHz: %lluKHz [%d%%]\n",
+					engine->name,
+					act, intel_gpu_freq(rps, act), count,
+					(int)DIV64_U64_ROUND_CLOSEST(100 * min.freq * count,
+								     act * min.count));
+			}
+
 			err = -EINVAL;
 		}
 
-- 
2.20.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

^ permalink raw reply related	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2020-04-21 19:26 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-04-21 12:37 [Intel-gfx] [PATCH] drm/i915/selftests: Show the full scaling curve on failure Chris Wilson
2020-04-21 12:46 ` Chris Wilson
2020-04-21 13:54 ` [Intel-gfx] ✓ Fi.CI.BAT: success for drm/i915/selftests: Show the full scaling curve on failure (rev2) Patchwork
2020-04-21 14:00 ` [Intel-gfx] [PATCH] drm/i915/selftests: Show the full scaling curve on failure Mika Kuoppala
2020-04-21 14:09   ` Chris Wilson
2020-04-21 19:26 ` [Intel-gfx] ✓ Fi.CI.IGT: success for drm/i915/selftests: Show the full scaling curve on failure (rev2) Patchwork

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.