Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v5] drm/i915/selftests: Refactor RC6 power measurement and error handling
@ 2025-03-27 19:19 Sk Anirban
  2025-03-27 20:42 ` ✓ i915.CI.BAT: success for drm/i915/selftests: Refactor RC6 power measurement and error handling (rev6) Patchwork
                   ` (3 more replies)
  0 siblings, 4 replies; 6+ messages in thread
From: Sk Anirban @ 2025-03-27 19:19 UTC (permalink / raw)
  To: intel-gfx; +Cc: Sk Anirban, Badal Nilawar

Revise the power measurement logic to save and evaluate energy values.
Previously, the test only checked whether the system had entered the RC6
state, without considering any potential interruptions in that state.
This update introduces a threshold check to ensure that the GPU remains
in the RC6 state properly during the specified sleep duration.

v3:
  - Reorder threshold check (Badal)

v4:
  - Improved commit message (Anshuman)

v5:
  - Rename variables for improved readability (Anshuman)

Signed-off-by: Sk Anirban <sk.anirban@intel.com>
Reviewed-by: Badal Nilawar <badal.nilawar@intel.com>
---
 drivers/gpu/drm/i915/gt/selftest_rc6.c | 51 ++++++++++++++++++--------
 1 file changed, 35 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/selftest_rc6.c b/drivers/gpu/drm/i915/gt/selftest_rc6.c
index 908483ab0bc8..77c1774ddf37 100644
--- a/drivers/gpu/drm/i915/gt/selftest_rc6.c
+++ b/drivers/gpu/drm/i915/gt/selftest_rc6.c
@@ -33,15 +33,22 @@ int live_rc6_manual(void *arg)
 {
 	struct intel_gt *gt = arg;
 	struct intel_rc6 *rc6 = &gt->rc6;
-	u64 rc0_power, rc6_power;
+	struct intel_rps *rps = &gt->rps;
 	intel_wakeref_t wakeref;
+	u64 rc0_sample_energy[2];
+	u64 rc6_sample_energy[2];
+	u64 sleep_time = 1000;
+	u32 rc0_freq = 0;
+	u32 rc6_freq = 0;
+	u64 rc0_power;
+	u64 rc6_power;
 	bool has_power;
+	u64 threshold;
 	ktime_t dt;
 	u64 res[2];
 	int err = 0;
-	u32 rc0_freq = 0;
-	u32 rc6_freq = 0;
-	struct intel_rps *rps = &gt->rps;
+	u64 diff;
+
 
 	/*
 	 * Our claim is that we can "encourage" the GPU to enter rc6 at will.
@@ -65,9 +72,9 @@ int live_rc6_manual(void *arg)
 	res[0] = rc6_residency(rc6);
 
 	dt = ktime_get();
-	rc0_power = librapl_energy_uJ();
-	msleep(1000);
-	rc0_power = librapl_energy_uJ() - rc0_power;
+	rc0_sample_energy[0] = librapl_energy_uJ();
+	msleep(sleep_time);
+	rc0_sample_energy[1] = librapl_energy_uJ() - rc0_sample_energy[0];
 	dt = ktime_sub(ktime_get(), dt);
 	res[1] = rc6_residency(rc6);
 	rc0_freq = intel_rps_read_actual_frequency_fw(rps);
@@ -79,11 +86,12 @@ int live_rc6_manual(void *arg)
 	}
 
 	if (has_power) {
-		rc0_power = div64_u64(NSEC_PER_SEC * rc0_power,
+		rc0_power = div64_u64(NSEC_PER_SEC * rc0_sample_energy[1],
 				      ktime_to_ns(dt));
+
 		if (!rc0_power) {
 			if (rc0_freq)
-				pr_debug("No power measured while in RC0! GPU Freq: %u in RC0\n",
+				pr_debug("No power measured while in RC0! GPU Freq: %uMHz in RC0\n",
 					 rc0_freq);
 			else
 				pr_err("No power and freq measured while in RC0\n");
@@ -98,10 +106,10 @@ int live_rc6_manual(void *arg)
 	res[0] = rc6_residency(rc6);
 	intel_uncore_forcewake_flush(rc6_to_uncore(rc6), FORCEWAKE_ALL);
 	dt = ktime_get();
-	rc6_power = librapl_energy_uJ();
-	msleep(1000);
+	rc6_sample_energy[0] = librapl_energy_uJ();
+	msleep(sleep_time);
 	rc6_freq = intel_rps_read_actual_frequency_fw(rps);
-	rc6_power = librapl_energy_uJ() - rc6_power;
+	rc6_sample_energy[1] = librapl_energy_uJ() - rc6_sample_energy[0];
 	dt = ktime_sub(ktime_get(), dt);
 	res[1] = rc6_residency(rc6);
 	if (res[1] == res[0]) {
@@ -113,13 +121,24 @@ int live_rc6_manual(void *arg)
 	}
 
 	if (has_power) {
-		rc6_power = div64_u64(NSEC_PER_SEC * rc6_power,
+		rc6_power = div64_u64(NSEC_PER_SEC * rc6_sample_energy[1],
 				      ktime_to_ns(dt));
-		pr_info("GPU consumed %llduW in RC0 and %llduW in RC6\n",
+		pr_info("GPU consumed %lluuW in RC0 and %lluuW in RC6\n",
 			rc0_power, rc6_power);
+
 		if (2 * rc6_power > rc0_power) {
-			pr_err("GPU leaked energy while in RC6! GPU Freq: %u in RC6 and %u in RC0\n",
-			       rc6_freq, rc0_freq);
+			pr_err("GPU leaked energy while in RC6!\n"
+			       "GPU Freq: %uMHz in RC6 and %uMHz in RC0\n"
+			       "RC0 energy before & after sleep respectively: %lluuJ %lluuJ\n"
+			       "RC6 energy before & after sleep respectively: %lluuJ %lluuJ\n",
+			       rc6_freq, rc0_freq, rc0_sample_energy[0], rc0_sample_energy[1],
+			       rc6_sample_energy[0], rc6_sample_energy[1]);
+
+			diff = res[1] - res[0];
+			threshold = (9 * NSEC_PER_MSEC * sleep_time) / 10;
+			if (diff < threshold)
+				pr_err("Did not enter RC6 properly, RC6 start residency=%lluns, RC6 end residency=%lluns\n",
+				       res[0], res[1]);
 			err = -EINVAL;
 			goto out_unlock;
 		}
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread
* [PATCH v5] drm/i915/selftests: Refactor RC6 power measurement and error handling
@ 2025-03-19 17:08 Sk Anirban
  0 siblings, 0 replies; 6+ messages in thread
From: Sk Anirban @ 2025-03-19 17:08 UTC (permalink / raw)
  To: intel-gfx; +Cc: anshuman.gupta, badal.nilawar, sk.anirban

Revise the power measurement logic to save and evaluate energy values.
Previously, the test only checked whether the system had entered the RC6
state, without considering any potential interruptions in that state.
This update introduces a threshold check to ensure that the GPU remains
in the RC6 state properly during the specified sleep duration.

v3:
  - Reorder threshold check (Badal)

v4:
  - Improved commit message (Anshuman)

v5:
  - Rename variables for improved readability (Anshuman)

Signed-off-by: Sk Anirban <sk.anirban@intel.com>
---
 drivers/gpu/drm/i915/gt/selftest_rc6.c | 51 ++++++++++++++++++--------
 1 file changed, 35 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/selftest_rc6.c b/drivers/gpu/drm/i915/gt/selftest_rc6.c
index 908483ab0bc8..77c1774ddf37 100644
--- a/drivers/gpu/drm/i915/gt/selftest_rc6.c
+++ b/drivers/gpu/drm/i915/gt/selftest_rc6.c
@@ -33,15 +33,22 @@ int live_rc6_manual(void *arg)
 {
 	struct intel_gt *gt = arg;
 	struct intel_rc6 *rc6 = &gt->rc6;
-	u64 rc0_power, rc6_power;
+	struct intel_rps *rps = &gt->rps;
 	intel_wakeref_t wakeref;
+	u64 rc0_sample_energy[2];
+	u64 rc6_sample_energy[2];
+	u64 sleep_time = 1000;
+	u32 rc0_freq = 0;
+	u32 rc6_freq = 0;
+	u64 rc0_power;
+	u64 rc6_power;
 	bool has_power;
+	u64 threshold;
 	ktime_t dt;
 	u64 res[2];
 	int err = 0;
-	u32 rc0_freq = 0;
-	u32 rc6_freq = 0;
-	struct intel_rps *rps = &gt->rps;
+	u64 diff;
+
 
 	/*
 	 * Our claim is that we can "encourage" the GPU to enter rc6 at will.
@@ -65,9 +72,9 @@ int live_rc6_manual(void *arg)
 	res[0] = rc6_residency(rc6);
 
 	dt = ktime_get();
-	rc0_power = librapl_energy_uJ();
-	msleep(1000);
-	rc0_power = librapl_energy_uJ() - rc0_power;
+	rc0_sample_energy[0] = librapl_energy_uJ();
+	msleep(sleep_time);
+	rc0_sample_energy[1] = librapl_energy_uJ() - rc0_sample_energy[0];
 	dt = ktime_sub(ktime_get(), dt);
 	res[1] = rc6_residency(rc6);
 	rc0_freq = intel_rps_read_actual_frequency_fw(rps);
@@ -79,11 +86,12 @@ int live_rc6_manual(void *arg)
 	}
 
 	if (has_power) {
-		rc0_power = div64_u64(NSEC_PER_SEC * rc0_power,
+		rc0_power = div64_u64(NSEC_PER_SEC * rc0_sample_energy[1],
 				      ktime_to_ns(dt));
+
 		if (!rc0_power) {
 			if (rc0_freq)
-				pr_debug("No power measured while in RC0! GPU Freq: %u in RC0\n",
+				pr_debug("No power measured while in RC0! GPU Freq: %uMHz in RC0\n",
 					 rc0_freq);
 			else
 				pr_err("No power and freq measured while in RC0\n");
@@ -98,10 +106,10 @@ int live_rc6_manual(void *arg)
 	res[0] = rc6_residency(rc6);
 	intel_uncore_forcewake_flush(rc6_to_uncore(rc6), FORCEWAKE_ALL);
 	dt = ktime_get();
-	rc6_power = librapl_energy_uJ();
-	msleep(1000);
+	rc6_sample_energy[0] = librapl_energy_uJ();
+	msleep(sleep_time);
 	rc6_freq = intel_rps_read_actual_frequency_fw(rps);
-	rc6_power = librapl_energy_uJ() - rc6_power;
+	rc6_sample_energy[1] = librapl_energy_uJ() - rc6_sample_energy[0];
 	dt = ktime_sub(ktime_get(), dt);
 	res[1] = rc6_residency(rc6);
 	if (res[1] == res[0]) {
@@ -113,13 +121,24 @@ int live_rc6_manual(void *arg)
 	}
 
 	if (has_power) {
-		rc6_power = div64_u64(NSEC_PER_SEC * rc6_power,
+		rc6_power = div64_u64(NSEC_PER_SEC * rc6_sample_energy[1],
 				      ktime_to_ns(dt));
-		pr_info("GPU consumed %llduW in RC0 and %llduW in RC6\n",
+		pr_info("GPU consumed %lluuW in RC0 and %lluuW in RC6\n",
 			rc0_power, rc6_power);
+
 		if (2 * rc6_power > rc0_power) {
-			pr_err("GPU leaked energy while in RC6! GPU Freq: %u in RC6 and %u in RC0\n",
-			       rc6_freq, rc0_freq);
+			pr_err("GPU leaked energy while in RC6!\n"
+			       "GPU Freq: %uMHz in RC6 and %uMHz in RC0\n"
+			       "RC0 energy before & after sleep respectively: %lluuJ %lluuJ\n"
+			       "RC6 energy before & after sleep respectively: %lluuJ %lluuJ\n",
+			       rc6_freq, rc0_freq, rc0_sample_energy[0], rc0_sample_energy[1],
+			       rc6_sample_energy[0], rc6_sample_energy[1]);
+
+			diff = res[1] - res[0];
+			threshold = (9 * NSEC_PER_MSEC * sleep_time) / 10;
+			if (diff < threshold)
+				pr_err("Did not enter RC6 properly, RC6 start residency=%lluns, RC6 end residency=%lluns\n",
+				       res[0], res[1]);
 			err = -EINVAL;
 			goto out_unlock;
 		}
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-04-01  7:24 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-03-27 19:19 [PATCH v5] drm/i915/selftests: Refactor RC6 power measurement and error handling Sk Anirban
2025-03-27 20:42 ` ✓ i915.CI.BAT: success for drm/i915/selftests: Refactor RC6 power measurement and error handling (rev6) Patchwork
2025-03-28  0:21 ` ✗ i915.CI.Full: failure " Patchwork
2025-03-28  6:08 ` ✓ i915.CI.Full: success " Patchwork
2025-04-01  7:24 ` [PATCH v5] drm/i915/selftests: Refactor RC6 power measurement and error handling Andi Shyti
  -- strict thread matches above, loose matches on Subject: below --
2025-03-19 17:08 Sk Anirban

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox