Skip to contentSkip to conversation

Firm up gt park/unpark

PATCH In review Ashutosh Dixit 4 participants 11 messages Oct 11, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover Firm up gt park/unpark

A couple of patches to firm up gt park/unpark.

Ashutosh Dixit (2):
drm/i915/gt: Don't do display work on platforms without display drm/i915/gt: Warn if not in RC6 when GT is parked

 drivers/gpu/drm/i915/gt/intel_gt_pm.c | 42 +++++++++++++++++----------
 drivers/gpu/drm/i915/gt/intel_rc6.c   | 15 ++++++++++
 drivers/gpu/drm/i915/gt/intel_rc6.h   |  1 +
 3 files changed, 42 insertions(+), 16 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH 1/2 drm/i915/gt: Don't do display work on platforms without display

Do display work only on platforms with display. This avoids holding the runtime PM wakeref for an additional 100+ ms after GT has been parked.

---
 drivers/gpu/drm/i915/gt/intel_gt_pm.c | 36 +++++++++++++++------------
 1 file changed, 20 insertions(+), 16 deletions(-)
drivers/gpu/drm/i915/gt/intel_gt_pm.c +20 -16
index f553e2173bdad..26aa2e979a148 100644
@@ -70,19 +70,21 @@ static int __gt_unpark(struct intel_wakeref *wf)
 
 	GT_TRACE(gt, "\n");
 
-	/*
-	 * It seems that the DMC likes to transition between the DC states a lot
-	 * when there are no connected displays (no active power domains) during
-	 * command submission.
-	 *
-	 * This activity has negative impact on the performance of the chip with
-	 * huge latencies observed in the interrupt handler and elsewhere.
-	 *
-	 * Work around it by grabbing a GT IRQ power domain whilst there is any
-	 * GT activity, preventing any DC state transitions.
-	 */
-	gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
-	GEM_BUG_ON(!gt->awake);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {
+		/*
+		 * It seems that the DMC likes to transition between the DC states a lot
+		 * when there are no connected displays (no active power domains) during
+		 * command submission.
+		 *
+		 * This activity has negative impact on the performance of the chip with
+		 * huge latencies observed in the interrupt handler and elsewhere.
+		 *
+		 * Work around it by grabbing a GT IRQ power domain whilst there is any
+		 * GT activity, preventing any DC state transitions.
+		 */
+		gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
+		GEM_BUG_ON(!gt->awake);
+	}
 
 	intel_rc6_unpark(&gt->rc6);
 	intel_rps_unpark(&gt->rps);
@@ -115,9 +117,11 @@ static int __gt_park(struct intel_wakeref *wf)
 	/* Everything switched off, flush any residual interrupt just in case */
 	intel_synchronize_irq(i915);
 
-	/* Defer dropping the display power well for 100ms, it's slow! */
-	GEM_BUG_ON(!wakeref);
-	intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {
+		/* Defer dropping the display power well for 100ms, it's slow! */
+		GEM_BUG_ON(!wakeref);
+		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	}
 
 	return 0;
 }
Signature · 1 line
2.34.1
#6 lore raw

PATCH 2/2 drm/i915/gt: Warn if not in RC6 when GT is parked

In reply to Ashutosh Dixit

Some i915 modules implicitly assume that there is no user, kernel or firmware activity after GT is parked. For example, PMU calculations are incorrect if GT is not in RC6 when GT is parked (outside of the GT wakeref). Therefore check and warn if GT is not in RC6 at the time of parking the GT.

---
 drivers/gpu/drm/i915/gt/intel_gt_pm.c |  6 ++++++
 drivers/gpu/drm/i915/gt/intel_rc6.c   | 15 +++++++++++++++
 drivers/gpu/drm/i915/gt/intel_rc6.h   |  1 +
 3 files changed, 22 insertions(+)
drivers/gpu/drm/i915/gt/intel_gt_pm.c +6 -0
index 26aa2e979a148..5da81e22d0980 100644
@@ -102,6 +102,8 @@ static int __gt_park(struct intel_wakeref *wf)
 	struct intel_gt *gt = container_of(wf, typeof(*gt), wakeref);
 	intel_wakeref_t wakeref = fetch_and_zero(&gt->awake);
 	struct drm_i915_private *i915 = gt->i915;
+	bool in_rc6;
+	int ret;
 
 	GT_TRACE(gt, "\n");
 
@@ -123,6 +125,10 @@ static int __gt_park(struct intel_wakeref *wf)
 		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
 	}
 
+	ret = intel_rc6_in_rc6(&gt->rc6, &in_rc6);
+	if (!ret && !in_rc6)
+		drm_warn_once(&i915->drm, "Parking, but GT is not in RC6!\n");
+
 	return 0;
 }
 
drivers/gpu/drm/i915/gt/intel_rc6.c +15 -0
index f8d0523f4c18e..4517988e704a4 100644
@@ -666,6 +666,21 @@ void intel_rc6_park(struct intel_rc6 *rc6)
 	set(uncore, GEN6_RC_STATE, target << RC_SW_TARGET_STATE_SHIFT);
 }
 
+int intel_rc6_in_rc6(struct intel_rc6 *rc6, bool *in_rc6)
+{
+	u32 gt_core_status;
+
+	/* GEN6_GT_CORE_STATUS exists only for Gen11+ */
+	if (GRAPHICS_VER(rc6_to_i915(rc6)) < 11)
+		return -ENODEV;
+
+	gt_core_status = intel_uncore_read_fw(rc6_to_uncore(rc6), GEN6_GT_CORE_STATUS)
+				& GEN6_RCn_MASK;
+	*in_rc6 = gt_core_status == GEN6_RC6 || gt_core_status == GEN6_RC7;
+
+	return 0;
+}
+
 void intel_rc6_disable(struct intel_rc6 *rc6)
 {
 	if (!rc6->enabled)
drivers/gpu/drm/i915/gt/intel_rc6.h +1 -0
index b6fea71afc223..30da10399028d 100644
@@ -16,6 +16,7 @@ void intel_rc6_fini(struct intel_rc6 *rc6);
 
 void intel_rc6_unpark(struct intel_rc6 *rc6);
 void intel_rc6_park(struct intel_rc6 *rc6);
+int intel_rc6_in_rc6(struct intel_rc6 *rc6, bool *in_rc6);
 
 void intel_rc6_sanitize(struct intel_rc6 *rc6);
 void intel_rc6_enable(struct intel_rc6 *rc6);
Signature · 1 line
2.34.1
#9 lore raw

[Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for Firm up gt park/unpark

In reply to Ashutosh Dixit
== Series Details ==

== Series Details ==

Series: Firm up gt park/unpark
URL : https://patchwork.freedesktop.org/series/109563/
State : warning

== Summary ==

Error: dim checkpatch failed
965f7dcab4c4 drm/i915/gt: Don't do display work on platforms without display bced71f23cb7 drm/i915/gt: Warn if not in RC6 when GT is parked -:56: CHECK:CAMELCASE: Avoid CamelCase: <GEN6_RCn_MASK>
#56: FILE: drivers/gpu/drm/i915/gt/intel_rc6.c:678:
+ & GEN6_RCn_MASK;

total: 0 errors, 0 warnings, 1 checks, 46 lines checked

#10 lore raw

[Intel-gfx] ✗ Fi.CI.BAT: failure for Firm up gt park/unpark

In reply to Ashutosh Dixit
== Series Details ==

== Series Details ==

Series: Firm up gt park/unpark
URL : https://patchwork.freedesktop.org/series/109563/
State : failure

== Summary ==

CI Bug Log - changes from CI_DRM_12231 -> Patchwork_109563v1
====================================================

Summary
-------

**FAILURE**

Serious unknown changes coming with Patchwork_109563v1 absolutely need to be verified manually.

If you think the reported changes have nothing to do with the changes introduced in Patchwork_109563v1, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.

External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/index.html

Participating hosts (46 -> 47)
------------------------------

Additional (3): fi-kbl-soraka bat-atsm-1 bat-adls-5
Missing (2): fi-ctg-p8600 fi-tgl-mst

Possible new issues
-------------------

Here are the unknown changes that may have been introduced in Patchwork_109563v1:

### IGT changes ###

#### Possible regressions ####

* igt@i915_module_load@reload:
- bat-dg1-5: [PASS][1] -> [DMESG-WARN][2] +16 similar issues
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg1-5/igt@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg1-5/igt@[email protected]

* igt@i915_pm_rpm@module-reload:
- bat-adlp-4: [PASS][3] -> [DMESG-WARN][4] +16 similar issues
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adlp-4/igt@[email protected]
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlp-4/igt@[email protected]

* igt@i915_selftest@live@gt_lrc:
- fi-icl-u2: [PASS][5] -> [DMESG-WARN][6] +16 similar issues
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-icl-u2/igt@i915_selftest@live@gt_lrc.html
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-icl-u2/igt@i915_selftest@live@gt_lrc.html

* igt@i915_selftest@live@gt_mocs:
- fi-rkl-guc: [PASS][7] -> [DMESG-WARN][8] +16 similar issues
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-rkl-guc/igt@i915_selftest@live@gt_mocs.html
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-rkl-guc/igt@i915_selftest@live@gt_mocs.html

* igt@i915_selftest@live@mman:
- bat-dg1-5: [PASS][9] -> [INCOMPLETE][10]
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg1-5/igt@i915_selftest@[email protected]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg1-5/igt@i915_selftest@[email protected]
- fi-pnv-d510: [PASS][11] -> [INCOMPLETE][12]
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-pnv-d510/igt@i915_selftest@[email protected]
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-pnv-d510/igt@i915_selftest@[email protected]
- fi-glk-j4005: [PASS][13] -> [INCOMPLETE][14]
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-glk-j4005/igt@i915_selftest@[email protected]
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-glk-j4005/igt@i915_selftest@[email protected]
- fi-kbl-7567u: [PASS][15] -> [INCOMPLETE][16]
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-kbl-7567u/igt@i915_selftest@[email protected]
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-7567u/igt@i915_selftest@[email protected]
- fi-rkl-11600: [PASS][17] -> [INCOMPLETE][18]
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-rkl-11600/igt@i915_selftest@[email protected]
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-rkl-11600/igt@i915_selftest@[email protected]
- fi-adl-ddr5: [PASS][19] -> [INCOMPLETE][20]
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-adl-ddr5/igt@i915_selftest@[email protected]
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-adl-ddr5/igt@i915_selftest@[email protected]
- fi-cfl-guc: [PASS][21] -> [INCOMPLETE][22]
[21]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-cfl-guc/igt@i915_selftest@[email protected]
[22]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-cfl-guc/igt@i915_selftest@[email protected]
- fi-skl-6700k2: [PASS][23] -> [INCOMPLETE][24]
[23]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-skl-6700k2/igt@i915_selftest@[email protected]
[24]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-skl-6700k2/igt@i915_selftest@[email protected]
- fi-cfl-8109u: [PASS][25] -> [INCOMPLETE][26]
[25]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-cfl-8109u/igt@i915_selftest@[email protected]
[26]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-cfl-8109u/igt@i915_selftest@[email protected]
- fi-ilk-650: [PASS][27] -> [INCOMPLETE][28]
[27]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-ilk-650/igt@i915_selftest@[email protected]
[28]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-ilk-650/igt@i915_selftest@[email protected]
- fi-blb-e6850: [PASS][29] -> [INCOMPLETE][30]
[29]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-blb-e6850/igt@i915_selftest@[email protected]
[30]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-blb-e6850/igt@i915_selftest@[email protected]
- fi-skl-6600u: [PASS][31] -> [INCOMPLETE][32]
[31]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-skl-6600u/igt@i915_selftest@[email protected]
[32]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-skl-6600u/igt@i915_selftest@[email protected]
- fi-apl-guc: [PASS][33] -> [INCOMPLETE][34]
[33]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-apl-guc/igt@i915_selftest@[email protected]
[34]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-apl-guc/igt@i915_selftest@[email protected]
- fi-bdw-5557u: [PASS][35] -> [INCOMPLETE][36]
[35]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-bdw-5557u/igt@i915_selftest@[email protected]
[36]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bdw-5557u/igt@i915_selftest@[email protected]
- fi-snb-2520m: [PASS][37] -> [INCOMPLETE][38]
[37]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-snb-2520m/igt@i915_selftest@[email protected]
[38]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-snb-2520m/igt@i915_selftest@[email protected]
- fi-rkl-guc: [PASS][39] -> [INCOMPLETE][40]
[39]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-rkl-guc/igt@i915_selftest@[email protected]
[40]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-rkl-guc/igt@i915_selftest@[email protected]
- fi-skl-guc: [PASS][41] -> [INCOMPLETE][42]
[41]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-skl-guc/igt@i915_selftest@[email protected]
[42]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-skl-guc/igt@i915_selftest@[email protected]
- fi-kbl-soraka: NOTRUN -> [INCOMPLETE][43]
[43]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@i915_selftest@[email protected]
- fi-cfl-8700k: [PASS][44] -> [INCOMPLETE][45]
[44]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-cfl-8700k/igt@i915_selftest@[email protected]
[45]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-cfl-8700k/igt@i915_selftest@[email protected]
- fi-bsw-nick: [PASS][46] -> [INCOMPLETE][47]
[46]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-bsw-nick/igt@i915_selftest@[email protected]
[47]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bsw-nick/igt@i915_selftest@[email protected]
- fi-hsw-g3258: [PASS][48] -> [INCOMPLETE][49]
[48]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-hsw-g3258/igt@i915_selftest@[email protected]
[49]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-hsw-g3258/igt@i915_selftest@[email protected]
- fi-bdw-gvtdvm: [PASS][50] -> [INCOMPLETE][51]
[50]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-bdw-gvtdvm/igt@i915_selftest@[email protected]
[51]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bdw-gvtdvm/igt@i915_selftest@[email protected]
- fi-hsw-4770: [PASS][52] -> [INCOMPLETE][53]
[52]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-hsw-4770/igt@i915_selftest@[email protected]
[53]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-hsw-4770/igt@i915_selftest@[email protected]
- bat-adlp-4: [PASS][54] -> [INCOMPLETE][55]
[54]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adlp-4/igt@i915_selftest@[email protected]
[55]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlp-4/igt@i915_selftest@[email protected]
- fi-ivb-3770: [PASS][56] -> [INCOMPLETE][57]
[56]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-ivb-3770/igt@i915_selftest@[email protected]
[57]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-ivb-3770/igt@i915_selftest@[email protected]
- fi-elk-e7500: [PASS][58] -> [INCOMPLETE][59]
[58]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-elk-e7500/igt@i915_selftest@[email protected]
[59]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-elk-e7500/igt@i915_selftest@[email protected]
- fi-snb-2600: [PASS][60] -> [INCOMPLETE][61]
[60]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-snb-2600/igt@i915_selftest@[email protected]
[61]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-snb-2600/igt@i915_selftest@[email protected]
- fi-bxt-dsi: [PASS][62] -> [INCOMPLETE][63]
[62]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-bxt-dsi/igt@i915_selftest@[email protected]
[63]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bxt-dsi/igt@i915_selftest@[email protected]

* igt@i915_selftest@live@sanitycheck:
- fi-rkl-11600: [PASS][64] -> [DMESG-WARN][65] +16 similar issues
[64]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-rkl-11600/igt@i915_selftest@[email protected]
[65]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-rkl-11600/igt@i915_selftest@[email protected]
- fi-adl-ddr5: [PASS][66] -> [DMESG-WARN][67] +16 similar issues
[66]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-adl-ddr5/igt@i915_selftest@[email protected]
[67]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-adl-ddr5/igt@i915_selftest@[email protected]

#### Suppressed ####

The following results come from untrusted machines, tests, or statuses. They do not affect the overall result.

* igt@i915_module_load@load:
- {bat-atsm-1}: NOTRUN -> [DMESG-WARN][68]
[68]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-atsm-1/igt@[email protected]

* igt@i915_pm_rpm@module-reload:
- {bat-adlp-6}: [PASS][69] -> [DMESG-WARN][70] +16 similar issues
[69]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adlp-6/igt@[email protected]
[70]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlp-6/igt@[email protected]
- {bat-jsl-3}: [PASS][71] -> [DMESG-WARN][72] +16 similar issues
[71]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-jsl-3/igt@[email protected]
[72]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-jsl-3/igt@[email protected]
- {bat-dg2-8}: [PASS][73] -> [DMESG-WARN][74] +16 similar issues
[73]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg2-8/igt@[email protected]
[74]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg2-8/igt@[email protected]

* igt@i915_selftest@live@active:
- {bat-adlm-1}: [PASS][75] -> [DMESG-WARN][76] +16 similar issues
[75]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adlm-1/igt@i915_selftest@[email protected]
[76]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlm-1/igt@i915_selftest@[email protected]

* igt@i915_selftest@live@gt_contexts:
- {bat-jsl-1}: [PASS][77] -> [DMESG-WARN][78] +16 similar issues
[77]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-jsl-1/igt@i915_selftest@live@gt_contexts.html
[78]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-jsl-1/igt@i915_selftest@live@gt_contexts.html
- {fi-jsl-1}: [PASS][79] -> [DMESG-WARN][80] +16 similar issues
[79]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-jsl-1/igt@i915_selftest@live@gt_contexts.html
[80]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-jsl-1/igt@i915_selftest@live@gt_contexts.html

* igt@i915_selftest@live@gt_engines:
- {bat-rpls-2}: [PASS][81] -> [DMESG-WARN][82] +14 similar issues
[81]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-rpls-2/igt@i915_selftest@live@gt_engines.html
[82]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-rpls-2/igt@i915_selftest@live@gt_engines.html

* igt@i915_selftest@live@gt_heartbeat:
- {bat-adln-1}: [PASS][83] -> [DMESG-WARN][84] +16 similar issues
[83]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adln-1/igt@i915_selftest@live@gt_heartbeat.html
[84]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adln-1/igt@i915_selftest@live@gt_heartbeat.html

* igt@i915_selftest@live@gt_lrc:
- {fi-ehl-2}: [PASS][85] -> [DMESG-WARN][86] +16 similar issues
[85]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-ehl-2/igt@i915_selftest@live@gt_lrc.html
[86]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-ehl-2/igt@i915_selftest@live@gt_lrc.html

* igt@i915_selftest@live@gt_mocs:
- {bat-dg2-11}: [PASS][87] -> [DMESG-WARN][88] +16 similar issues
[87]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg2-11/igt@i915_selftest@live@gt_mocs.html
[88]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg2-11/igt@i915_selftest@live@gt_mocs.html

* igt@i915_selftest@live@gt_timelines:
- {bat-dg2-9}: [PASS][89] -> [DMESG-WARN][90] +16 similar issues
[89]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg2-9/igt@i915_selftest@live@gt_timelines.html
[90]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg2-9/igt@i915_selftest@live@gt_timelines.html

* igt@i915_selftest@live@mman:
- {fi-tgl-dsi}: [PASS][91] -> [INCOMPLETE][92]
[91]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-tgl-dsi/igt@i915_selftest@[email protected]
[92]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-tgl-dsi/igt@i915_selftest@[email protected]
- {bat-rplp-1}: [PASS][93] -> [INCOMPLETE][94]
[93]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-rplp-1/igt@i915_selftest@[email protected]
[94]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-rplp-1/igt@i915_selftest@[email protected]
- {bat-adls-5}: NOTRUN -> [INCOMPLETE][95]
[95]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adls-5/igt@i915_selftest@[email protected]
- {bat-dg2-9}: [PASS][96] -> [INCOMPLETE][97]
[96]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg2-9/igt@i915_selftest@[email protected]
[97]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg2-9/igt@i915_selftest@[email protected]
- {bat-adln-1}: [PASS][98] -> [INCOMPLETE][99]
[98]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adln-1/igt@i915_selftest@[email protected]
[99]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adln-1/igt@i915_selftest@[email protected]
- {bat-rpls-2}: [PASS][100] -> [INCOMPLETE][101]
[100]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-rpls-2/igt@i915_selftest@[email protected]
[101]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-rpls-2/igt@i915_selftest@[email protected]
- {bat-adlp-6}: [PASS][102] -> [INCOMPLETE][103]
[102]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adlp-6/igt@i915_selftest@[email protected]
[103]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlp-6/igt@i915_selftest@[email protected]
- {fi-ehl-2}: [PASS][104] -> [INCOMPLETE][105]
[104]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-ehl-2/igt@i915_selftest@[email protected]
[105]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-ehl-2/igt@i915_selftest@[email protected]
- {bat-dg2-11}: [PASS][106] -> [INCOMPLETE][107]
[106]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg2-11/igt@i915_selftest@[email protected]
[107]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg2-11/igt@i915_selftest@[email protected]
- {bat-jsl-3}: [PASS][108] -> [INCOMPLETE][109]
[108]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-jsl-3/igt@i915_selftest@[email protected]
[109]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-jsl-3/igt@i915_selftest@[email protected]
- {bat-dg2-8}: [PASS][110] -> [INCOMPLETE][111]
[110]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-dg2-8/igt@i915_selftest@[email protected]
[111]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg2-8/igt@i915_selftest@[email protected]
- {bat-adlm-1}: [PASS][112] -> [INCOMPLETE][113]
[112]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-adlm-1/igt@i915_selftest@[email protected]
[113]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlm-1/igt@i915_selftest@[email protected]
- {bat-jsl-1}: [PASS][114] -> [INCOMPLETE][115]
[114]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-jsl-1/igt@i915_selftest@[email protected]
[115]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-jsl-1/igt@i915_selftest@[email protected]
- {fi-jsl-1}: [PASS][116] -> [INCOMPLETE][117]
[116]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-jsl-1/igt@i915_selftest@[email protected]
[117]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-jsl-1/igt@i915_selftest@[email protected]
- {bat-rpls-1}: [PASS][118] -> [INCOMPLETE][119]
[118]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-rpls-1/igt@i915_selftest@[email protected]
[119]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-rpls-1/igt@i915_selftest@[email protected]

* igt@i915_selftest@live@sanitycheck:
- {fi-tgl-dsi}: [PASS][120] -> [DMESG-WARN][121] +12 similar issues
[120]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-tgl-dsi/igt@i915_selftest@[email protected]
[121]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-tgl-dsi/igt@i915_selftest@[email protected]
- {bat-rplp-1}: [PASS][122] -> [DMESG-WARN][123] +16 similar issues
[122]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-rplp-1/igt@i915_selftest@[email protected]
[123]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-rplp-1/igt@i915_selftest@[email protected]

* igt@i915_selftest@live@workarounds:
- {bat-rpls-1}: [PASS][124] -> [DMESG-WARN][125] +16 similar issues
[124]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/bat-rpls-1/igt@i915_selftest@[email protected]
[125]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-rpls-1/igt@i915_selftest@[email protected]
- {bat-adls-5}: NOTRUN -> [DMESG-WARN][126] +16 similar issues
[126]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adls-5/igt@i915_selftest@[email protected]

Known issues
------------

Here are the changes found in Patchwork_109563v1 that come from known issues:

### IGT changes ###

#### Issues hit ####

* igt@gem_exec_gttfill@basic:
- fi-kbl-soraka: NOTRUN -> [SKIP][127] ([fdo#109271]) +9 similar issues
[127]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@[email protected]

* igt@gem_huc_copy@huc-copy:
- fi-kbl-soraka: NOTRUN -> [SKIP][128] ([fdo#109271] / [i915#2190])
[128]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@[email protected]

* igt@gem_lmem_swapping@basic:
- fi-kbl-soraka: NOTRUN -> [SKIP][129] ([fdo#109271] / [i915#4613]) +3 similar issues
[129]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@[email protected]

* igt@i915_selftest@live@gt_heartbeat:
- fi-bxt-dsi: [PASS][130] -> [DMESG-FAIL][131] ([i915#5334])
[130]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-bxt-dsi/igt@i915_selftest@live@gt_heartbeat.html
[131]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bxt-dsi/igt@i915_selftest@live@gt_heartbeat.html

* igt@i915_selftest@live@gt_pm:
- fi-kbl-soraka: NOTRUN -> [DMESG-FAIL][132] ([i915#1886])
[132]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@i915_selftest@live@gt_pm.html

* igt@i915_selftest@live@mman:
- fi-icl-u2: [PASS][133] -> [INCOMPLETE][134] ([i915#7057])
[133]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12231/fi-icl-u2/igt@i915_selftest@[email protected]
[134]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-icl-u2/igt@i915_selftest@[email protected]

* igt@kms_chamelium@hdmi-hpd-fast:
- fi-kbl-soraka: NOTRUN -> [SKIP][135] ([fdo#109271] / [fdo#111827]) +7 similar issues
[135]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@[email protected]

* igt@runner@aborted:
- bat-adlp-4: NOTRUN -> [FAIL][136] ([i915#4312])
[136]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-adlp-4/igt@[email protected]
- fi-ivb-3770: NOTRUN -> [FAIL][137] ([fdo#109271] / [i915#4312])
[137]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-ivb-3770/igt@[email protected]
- fi-elk-e7500: NOTRUN -> [FAIL][138] ([fdo#109271] / [i915#4312])
[138]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-elk-e7500/igt@[email protected]
- fi-snb-2600: NOTRUN -> [FAIL][139] ([i915#4312])
[139]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-snb-2600/igt@[email protected]
- fi-bxt-dsi: NOTRUN -> [FAIL][140] ([i915#4312])
[140]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bxt-dsi/igt@[email protected]
- bat-dg1-5: NOTRUN -> [FAIL][141] ([i915#4312])
[141]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/bat-dg1-5/igt@[email protected]
- fi-pnv-d510: NOTRUN -> [FAIL][142] ([fdo#109271] / [i915#2403] / [i915#4312])
[142]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-pnv-d510/igt@[email protected]
- fi-glk-j4005: NOTRUN -> [FAIL][143] ([i915#4312])
[143]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-glk-j4005/igt@[email protected]
- fi-kbl-7567u: NOTRUN -> [FAIL][144] ([i915#4312])
[144]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-7567u/igt@[email protected]
- fi-rkl-11600: NOTRUN -> [FAIL][145] ([i915#4312])
[145]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-rkl-11600/igt@[email protected]
- fi-adl-ddr5: NOTRUN -> [FAIL][146] ([i915#4312])
[146]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-adl-ddr5/igt@[email protected]
- fi-cfl-guc: NOTRUN -> [FAIL][147] ([i915#4312])
[147]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-cfl-guc/igt@[email protected]
- fi-skl-6700k2: NOTRUN -> [FAIL][148] ([i915#4312])
[148]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-skl-6700k2/igt@[email protected]
- fi-cfl-8109u: NOTRUN -> [FAIL][149] ([i915#4312])
[149]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-cfl-8109u/igt@[email protected]
- fi-ilk-650: NOTRUN -> [FAIL][150] ([fdo#109271] / [i915#4312])
[150]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-ilk-650/igt@[email protected]
- fi-blb-e6850: NOTRUN -> [FAIL][151] ([fdo#109271] / [i915#2403] / [i915#4312])
[151]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-blb-e6850/igt@[email protected]
- fi-skl-6600u: NOTRUN -> [FAIL][152] ([i915#4312])
[152]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-skl-6600u/igt@[email protected]
- fi-icl-u2: NOTRUN -> [FAIL][153] ([i915#4312] / [i915#4991])
[153]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-icl-u2/igt@[email protected]
- fi-apl-guc: NOTRUN -> [FAIL][154] ([fdo#109271] / [i915#4312] / [i915#6641])
[154]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-apl-guc/igt@[email protected]
- fi-bdw-5557u: NOTRUN -> [FAIL][155] ([i915#4312])
[155]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bdw-5557u/igt@[email protected]
- fi-snb-2520m: NOTRUN -> [FAIL][156] ([i915#4312])
[156]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-snb-2520m/igt@[email protected]
- fi-rkl-guc: NOTRUN -> [FAIL][157] ([i915#4312])
[157]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-rkl-guc/igt@[email protected]
- fi-skl-guc: NOTRUN -> [FAIL][158] ([i915#4312] / [i915#6641])
[158]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-skl-guc/igt@[email protected]
- fi-kbl-soraka: NOTRUN -> [FAIL][159] ([i915#4312] / [i915#6641])
[159]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-kbl-soraka/igt@[email protected]
- fi-cfl-8700k: NOTRUN -> [FAIL][160] ([i915#4312])
[160]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-cfl-8700k/igt@[email protected]
- fi-bsw-nick: NOTRUN -> [FAIL][161] ([fdo#109271] / [i915#4312])
[161]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bsw-nick/igt@[email protected]
- fi-bdw-gvtdvm: NOTRUN -> [FAIL][162] ([i915#4312])
[162]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/fi-bdw-gvtdvm/igt@[email protected]

{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).

[fdo#109271]: https://bugs.freedesktop.org/show_bug.cgi?id=109271
[fdo#109285]: https://bugs.freedesktop.org/show_bug.cgi?id=109285
[fdo#109295]: https://bugs.freedesktop.org/show_bug.cgi?id=109295
[fdo#111827]: https://bugs.freedesktop.org/show_bug.cgi?id=111827
[i915#1886]: https://gitlab.freedesktop.org/drm/intel/issues/1886
[i915#1982]: https://gitlab.freedesktop.org/drm/intel/issues/1982
[i915#2190]: https://gitlab.freedesktop.org/drm/intel/issues/2190
[i915#2403]: https://gitlab.freedesktop.org/drm/intel/issues/2403
[i915#2411]: https://gitlab.freedesktop.org/drm/intel/issues/2411
[i915#2582]: https://gitlab.freedesktop.org/drm/intel/issues/2582
[i915#3282]: https://gitlab.freedesktop.org/drm/intel/issues/3282
[i915#3291]: https://gitlab.freedesktop.org/drm/intel/issues/3291
[i915#3301]: https://gitlab.freedesktop.org/drm/intel/issues/3301
[i915#3555]: https://gitlab.freedesktop.org/drm/intel/issues/3555
[i915#4077]: https://gitlab.freedesktop.org/drm/intel/issues/4077
[i915#4079]: https://gitlab.freedesktop.org/drm/intel/issues/4079
[i915#4083]: https://gitlab.freedesktop.org/drm/intel/issues/4083
[i915#4103]: https://gitlab.freedesktop.org/drm/intel/issues/4103
[i915#4312]: https://gitlab.freedesktop.org/drm/intel/issues/4312
[i915#4613]: https://gitlab.freedesktop.org/drm/intel/issues/4613
[i915#4991]: https://gitlab.freedesktop.org/drm/intel/issues/4991
[i915#5334]: https://gitlab.freedesktop.org/drm/intel/issues/5334
[i915#5537]: https://gitlab.freedesktop.org/drm/intel/issues/5537
[i915#6641]: https://gitlab.freedesktop.org/drm/intel/issues/6641
[i915#7030]: https://gitlab.freedesktop.org/drm/intel/issues/7030
[i915#7057]: https://gitlab.freedesktop.org/drm/intel/issues/7057

Build changes
-------------

* Linux: CI_DRM_12231 -> Patchwork_109563v1

CI-20190529: 20190529
CI_DRM_12231: bb84c1baa34eed834400e9a3cf9642840be002e1 @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_7007: 39a979fb4453c557022f0477c609afe10a049e48 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
Patchwork_109563v1: bb84c1baa34eed834400e9a3cf9642840be002e1 @ git://anongit.freedesktop.org/gfx-ci/linux

### Linux commits

63e047aeba79 drm/i915/gt: Warn if not in RC6 when GT is parked 672b75d1fb78 drm/i915/gt: Don't do display work on platforms without display

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109563v1/index.html

#7 lore raw

Re: [Intel-gfx] [PATCH 2/2] drm/i915/gt: Warn if not in RC6 when GT is parked

In reply to Ashutosh Dixit
Quoted 6 lines · Ashutosh Dixit

Some i915 modules implicitly assume that there is no user, kernel or firmware activity after GT is parked. For example, PMU calculations are incorrect if GT is not in RC6 when GT is parked (outside of the GT wakeref). Therefore check and warn if GT is not in RC6 at the time of parking the GT.

This patch has cause widespread dmesg_warn's in premerge CI so there is no intention of merging this. It just proves that these assumptions in PMU (for quantities such as frequency and RC6 residency) are incorrect and need to be addressed.

Thanks.
--
Ashutosh

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +4 -0
@@ -123,6 +125,10 @@ static int __gt_park(struct intel_wakeref *wf)
 		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
 	}
 
+	ret = intel_rc6_in_rc6(&gt->rc6, &in_rc6);
+	if (!ret && !in_rc6)
+		drm_warn_once(&i915->drm, "Parking, but GT is not in RC6!\n");
+
 	return 0;
 }
maintainer
#2 lore raw

Re: [Intel-gfx] [PATCH 1/2] drm/i915/gt: Don't do display work on platforms without display

In reply to Ashutosh Dixit
Quoted 21 lines · Ashutosh Dixit

Do display work only on platforms with display. This avoids holding the runtime PM wakeref for an additional 100+ ms after GT has been parked.

---
 drivers/gpu/drm/i915/gt/intel_gt_pm.c | 36 +++++++++++++++------------
 1 file changed, 20 insertions(+), 16 deletions(-)
drivers/gpu/drm/i915/gt/intel_gt_pm.c +0 -4
index f553e2173bdad..26aa2e979a148 100644
@@ -70,19 +70,21 @@ static int __gt_unpark(struct intel_wakeref *wf)
 
 	GT_TRACE(gt, "\n");
 
-	/*
-	 * It seems that the DMC likes to transition between the DC states a lot
-	 * when there are no connected displays (no active power domains) during
-	 * command submission.

drivers/gpu/drm/i915/gt/intel_gt_pm.c static int __gt_unpark(struct intel_wakeref *wf) · line 73

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +1 -9
-	 *
-	 * This activity has negative impact on the performance of the chip with
-	 * huge latencies observed in the interrupt handler and elsewhere.
-	 *
-	 * Work around it by grabbing a GT IRQ power domain whilst there is any
-	 * GT activity, preventing any DC state transitions.
-	 */
-	gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
-	GEM_BUG_ON(!gt->awake);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {

Feels like something's wrong if you need both of those.

BR,
Jani.

Quoted 32 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +19 -3
+		/*
+		 * It seems that the DMC likes to transition between the DC states a lot
+		 * when there are no connected displays (no active power domains) during
+		 * command submission.
+		 *
+		 * This activity has negative impact on the performance of the chip with
+		 * huge latencies observed in the interrupt handler and elsewhere.
+		 *
+		 * Work around it by grabbing a GT IRQ power domain whilst there is any
+		 * GT activity, preventing any DC state transitions.
+		 */
+		gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
+		GEM_BUG_ON(!gt->awake);
+	}
 
 	intel_rc6_unpark(&gt->rc6);
 	intel_rps_unpark(&gt->rps);
@@ -115,9 +117,11 @@ static int __gt_park(struct intel_wakeref *wf)
 	/* Everything switched off, flush any residual interrupt just in case */
 	intel_synchronize_irq(i915);
 
-	/* Defer dropping the display power well for 100ms, it's slow! */
-	GEM_BUG_ON(!wakeref);
-	intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {
+		/* Defer dropping the display power well for 100ms, it's slow! */
+		GEM_BUG_ON(!wakeref);
+		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	}
 
 	return 0;
 }
Signature · 1 line
Jani Nikula, Intel Open Source Graphics Center
#3 lore raw

Re: [Intel-gfx] [PATCH 1/2] drm/i915/gt: Don't do display work on platforms without display

Quoted 1 line · Jani Nikula

Hi Jani,

Quoted 18 lines · Ashutosh Dixit
Quoted 17 lines

Do display work only on platforms with display. This avoids holding the runtime PM wakeref for an additional 100+ ms after GT has been parked.

---
 drivers/gpu/drm/i915/gt/intel_gt_pm.c | 36 +++++++++++++++------------
 1 file changed, 20 insertions(+), 16 deletions(-)
drivers/gpu/drm/i915/gt/intel_gt_pm.c +0 -0
index f553e2173bdad..26aa2e979a148 100644
@@ -70,19 +70,21 @@ static int __gt_unpark(struct intel_wakeref *wf)
 
 	GT_TRACE(gt, "\n");
 
Quoted 16 lines
Quoted 4 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +0 -4
-	/*
-	 * It seems that the DMC likes to transition between the DC states a lot
-	 * when there are no connected displays (no active power domains) during
-	 * command submission.

drivers/gpu/drm/i915/gt/intel_gt_pm.c static int __gt_unpark(struct intel_wakeref *wf) · line 73

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +1 -9
-	 *
-	 * This activity has negative impact on the performance of the chip with
-	 * huge latencies observed in the interrupt handler and elsewhere.
-	 *
-	 * Work around it by grabbing a GT IRQ power domain whilst there is any
-	 * GT activity, preventing any DC state transitions.
-	 */
-	gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
-	GEM_BUG_ON(!gt->awake);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {

Feels like something's wrong if you need both of those.

Don't think so:

	/* Only valid when HAS_DISPLAY() is true */
	#define INTEL_DISPLAY_ENABLED(dev_priv) \
		(drm_WARN_ON(&(dev_priv)->drm, !HAS_DISPLAY(dev_priv)),         \
	         !(dev_priv)->params.disable_display &&                         \
	         !intel_opregion_headless_sku(dev_priv))

Maybe inside display code INTEL_DISPLAY_ENABLED is sufficient since code paths have previously invoked HAS_DISPLAY, but not in non-display code.

Thanks.
--
Ashutosh

Quoted 35 lines
Quoted 22 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +14 -1
+		/*
+		 * It seems that the DMC likes to transition between the DC states a lot
+		 * when there are no connected displays (no active power domains) during
+		 * command submission.
+		 *
+		 * This activity has negative impact on the performance of the chip with
+		 * huge latencies observed in the interrupt handler and elsewhere.
+		 *
+		 * Work around it by grabbing a GT IRQ power domain whilst there is any
+		 * GT activity, preventing any DC state transitions.
+		 */
+		gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
+		GEM_BUG_ON(!gt->awake);
+	}
 
 	intel_rc6_unpark(&gt->rc6);
 	intel_rps_unpark(&gt->rps);
@@ -115,9 +117,11 @@ static int __gt_park(struct intel_wakeref *wf)
 	/* Everything switched off, flush any residual interrupt just in case */
 	intel_synchronize_irq(i915);
 
-	/* Defer dropping the display power well for 100ms, it's slow! */

drivers/gpu/drm/i915/gt/intel_gt_pm.c static int __gt_park(struct intel_wakeref *wf) · line 127

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +5 -2
-	GEM_BUG_ON(!wakeref);
-	intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {
+		/* Defer dropping the display power well for 100ms, it's slow! */
+		GEM_BUG_ON(!wakeref);
+		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	}
 
 	return 0;
 }

--
Jani Nikula, Intel Open Source Graphics Center

#4 lore raw

Re: [Intel-gfx] [PATCH 1/2] drm/i915/gt: Don't do display work on platforms without display

Quoted 42 lines · Ashutosh
Quoted 1 line · Jani Nikula

Hi Jani,

Quoted 18 lines · Ashutosh Dixit
Quoted 17 lines

Do display work only on platforms with display. This avoids holding the runtime PM wakeref for an additional 100+ ms after GT has been parked.

---
  drivers/gpu/drm/i915/gt/intel_gt_pm.c | 36 +++++++++++++++------------
  1 file changed, 20 insertions(+), 16 deletions(-)
drivers/gpu/drm/i915/gt/intel_gt_pm.c +0 -0
index f553e2173bdad..26aa2e979a148 100644
@@ -70,19 +70,21 @@ static int __gt_unpark(struct intel_wakeref *wf)
 
 	GT_TRACE(gt, "\n");
 
Quoted 16 lines
Quoted 4 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +0 -4
-	/*
-	 * It seems that the DMC likes to transition between the DC states a lot
-	 * when there are no connected displays (no active power domains) during
-	 * command submission.

drivers/gpu/drm/i915/gt/intel_gt_pm.c static int __gt_unpark(struct intel_wakeref *wf) · line 73

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +1 -9
-	 *
-	 * This activity has negative impact on the performance of the chip with
-	 * huge latencies observed in the interrupt handler and elsewhere.
-	 *
-	 * Work around it by grabbing a GT IRQ power domain whilst there is any
-	 * GT activity, preventing any DC state transitions.
-	 */
-	gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
-	GEM_BUG_ON(!gt->awake);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {

Feels like something's wrong if you need both of those.

Don't think so:

Quoted 8 lines
	/* Only valid when HAS_DISPLAY() is true */
	#define INTEL_DISPLAY_ENABLED(dev_priv) \
		(drm_WARN_ON(&(dev_priv)->drm, !HAS_DISPLAY(dev_priv)),         \
	         !(dev_priv)->params.disable_display &&                         \
	         !intel_opregion_headless_sku(dev_priv))

Maybe inside display code INTEL_DISPLAY_ENABLED is sufficient since code paths have previously invoked HAS_DISPLAY, but not in non-display code.

AFAIR this workaround is only needed when DMC is loaded so I wonder if we could detect that instead?

Although then I am not sure why we haven't done it like that from the start. Maybe there was a good reason and I just can't remember it.

Looking at the history, b68763741aa2 ("drm/i915: Restore GT performance in headless mode with DMC loaded") which added the workaround did not add the 100ms delay. That was added later in 81ff52b70577 ("drm/i915/gt: Ratelimit display power w/a"). That part sounds like it makes sense - if there is cost in these transitions usual approach is too add some hysteresis. (And AFAIR in this particular case it was actually a matter or re-adding the hysteresis which was lost once GEM idle work handler approach was removed.)

So I guess the main question is can we guard this with (ideally something better than) HAS_DMC. Perhaps back then GPUs wo/ display were simply not in our minds? Or obtaining the "DC off" power well was perhaps a no-op in it's own right when there is no display? If that was the case and isn't any more would that be feasible to re-add?

Regards,

Tvrtko

Quoted 40 lines

Thanks.
--
Ashutosh

Quoted 35 lines
Quoted 22 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +14 -1
+		/*
+		 * It seems that the DMC likes to transition between the DC states a lot
+		 * when there are no connected displays (no active power domains) during
+		 * command submission.
+		 *
+		 * This activity has negative impact on the performance of the chip with
+		 * huge latencies observed in the interrupt handler and elsewhere.
+		 *
+		 * Work around it by grabbing a GT IRQ power domain whilst there is any
+		 * GT activity, preventing any DC state transitions.
+		 */
+		gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
+		GEM_BUG_ON(!gt->awake);
+	}
 
 	intel_rc6_unpark(&gt->rc6);
 	intel_rps_unpark(&gt->rps);
@@ -115,9 +117,11 @@ static int __gt_park(struct intel_wakeref *wf)
 	/* Everything switched off, flush any residual interrupt just in case */
 	intel_synchronize_irq(i915);
 
-	/* Defer dropping the display power well for 100ms, it's slow! */

drivers/gpu/drm/i915/gt/intel_gt_pm.c static int __gt_park(struct intel_wakeref *wf) · line 127

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +5 -2
-	GEM_BUG_ON(!wakeref);
-	intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {
+		/* Defer dropping the display power well for 100ms, it's slow! */
+		GEM_BUG_ON(!wakeref);
+		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+	}
 
 	return 0;
  }

--
Jani Nikula, Intel Open Source Graphics Center

#5 lore raw

Re: [Intel-gfx] [PATCH 1/2] drm/i915/gt: Don't do display work on platforms without display

Quoted 58 lines · Tvrtko Ursulin
Quoted 47 lines · Ashutosh
Quoted 1 line · Jani Nikula

Hi Jani,

Quoted 31 lines · Ashutosh Dixit
Quoted 30 lines

Do display work only on platforms with display. This avoids holding the runtime PM wakeref for an additional 100+ ms after GT has been parked.

---
  drivers/gpu/drm/i915/gt/intel_gt_pm.c | 36
+++++++++++++++------------

  1 file changed, 20 insertions(+), 16 deletions(-)
drivers/gpu/drm/i915/gt/intel_gt_pm.c +0 -2
index f553e2173bdad..26aa2e979a148 100644
@@ -70,19 +70,21 @@ static int __gt_unpark(struct intel_wakeref *wf)
 
     GT_TRACE(gt, "\n");
 
-    /*
-     * It seems that the DMC likes to transition between the DC 

states a lot
-     * when there are no connected displays (no active power domains) during
-     * command submission.
-     *
-     * This activity has negative impact on the performance of the chip with
-     * huge latencies observed in the interrupt handler and elsewhere. -     *
-     * Work around it by grabbing a GT IRQ power domain whilst

Quoted 8 lines
Quoted 6 lines

there is any
-     * GT activity, preventing any DC state transitions. -     */
-    gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ); -    GEM_BUG_ON(!gt->awake);
+    if (HAS_DISPLAY(i915) && INTEL_DISPLAY_ENABLED(i915)) {

Feels like something's wrong if you need both of those.

Don't think so:

Quoted 8 lines

    /* Only valid when HAS_DISPLAY() is true */
    #define INTEL_DISPLAY_ENABLED(dev_priv) \
        (drm_WARN_ON(&(dev_priv)->drm, !HAS_DISPLAY(dev_priv)),         \
             !(dev_priv)->params.disable_display &&                         \
             !intel_opregion_headless_sku(dev_priv))

Maybe inside display code INTEL_DISPLAY_ENABLED is sufficient since code paths have previously invoked HAS_DISPLAY, but not in non-display code.

Quoted 7 lines

AFAIR this workaround is only needed when DMC is loaded so I wonder if we could detect that instead?

Although then I am not sure why we haven't done it like that from the start. Maybe there was a good reason and I just can't remember it.

Looking at the history, b68763741aa2 ("drm/i915: Restore GT performance in headless mode with DMC loaded") which added the workaround did not add the 100ms delay. That was added later in 81ff52b70577 ("drm/i915/gt: Ratelimit display power w/a"). That part sounds like it makes sense - if there is cost in these transitions usual approach is too add some hysteresis. (And AFAIR in this particular case it was actually a matter or re-adding the hysteresis which was lost once GEM idle work handler approach was removed.)

So I guess the main question is can we guard this with (ideally something better than) HAS_DMC. Perhaps back then GPUs wo/ display were simply not in our minds? Or obtaining the "DC off" power well was perhaps a no-op in it's own right when there is no display? If that was the case and isn't any more would that be feasible to re-add?

Oops or not - we still need a rpm reference in the current scheme, display or no display! Back in the day that rpm was separate and explicit from this wa... So I guess this code stays as is and only possible improvements can be in the PMU area.

Regards,

Tvrtko

#8 lore raw

Re: [Intel-gfx] [PATCH 2/2] drm/i915/gt: Warn if not in RC6 when GT is parked

In reply to Dixit, Ashutosh
Quoted 8 lines · Ashutosh
Quoted 6 lines · Ashutosh Dixit

Some i915 modules implicitly assume that there is no user, kernel or firmware activity after GT is parked. For example, PMU calculations are incorrect if GT is not in RC6 when GT is parked (outside of the GT wakeref). Therefore check and warn if GT is not in RC6 at the time of parking the GT.

Quoted 4 lines

This patch has cause widespread dmesg_warn's in premerge CI so there is no intention of merging this. It just proves that these assumptions in PMU (for quantities such as frequency and RC6 residency) are incorrect and need to be addressed.

Yeah this probably does not make sense since I am not sure if driver actually controls when hardware goes into RC6.

Regards,

Tvrtko

Quoted 16 lines

Thanks.
--
Ashutosh

Quoted 10 lines
drivers/gpu/drm/i915/gt/intel_gt_pm.c +4 -0
@@ -123,6 +125,10 @@ static int __gt_park(struct intel_wakeref *wf)
 		intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
 	}
 
+	ret = intel_rc6_in_rc6(&gt->rc6, &in_rc6);
+	if (!ret && !in_rc6)
+		drm_warn_once(&i915->drm, "Parking, but GT is not in RC6!\n");
+
 	return 0;
  }