[Intel-gfx] [PATCH v2] drm/i915/gt: Pull checking rps->pm_events under the irq_lock

[Intel-gfx] [PATCH v2] drm/i915/gt: Pull checking rps->pm_events under the irq_lock

[Chris Wilson]

Avoid angering kcsan by serialising the read of the pm_events with the
write in rps_diable_interrupts.

[ 6268.713419] BUG: KCSAN: data-race in intel_rps_park [i915] / rps_work [i915]
[ 6268.713437]
[ 6268.713449] write to 0xffff8881eda8efac of 4 bytes by task 1127 on cpu 3:
[ 6268.713680]  intel_rps_park+0x136/0x260 [i915]
[ 6268.713905]  __gt_park+0x61/0xa0 [i915]
[ 6268.714128]  ____intel_wakeref_put_last+0x42/0x90 [i915]
[ 6268.714352]  __intel_wakeref_put_work+0xd3/0xf0 [i915]
[ 6268.714369]  process_one_work+0x3b1/0x690
[ 6268.714384]  worker_thread+0x80/0x670
[ 6268.714398]  kthread+0x19a/0x1e0
[ 6268.714412]  ret_from_fork+0x1f/0x30
[ 6268.714423]
[ 6268.714435] read to 0xffff8881eda8efac of 4 bytes by task 950 on cpu 2:
[ 6268.714664]  rps_work+0xc2/0x680 [i915]
[ 6268.714679]  process_one_work+0x3b1/0x690
[ 6268.714693]  worker_thread+0x80/0x670
[ 6268.714707]  kthread+0x19a/0x1e0
[ 6268.714720]  ret_from_fork+0x1f/0x30

v2: Mark all reads and writes of rpm->pm_events.

The flow of enabling/disabling rps is stronly ordered, so the writes and
interrrupt generation are also strongly ordered -- just this may not be
visible to the compiler, so provide annotations.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/gt/intel_rps.c | 29 +++++++++++++++++------------
 1 file changed, 17 insertions(+), 12 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_rps.c b/drivers/gpu/drm/i915/gt/intel_rps.c
index 506738dede16..c715c02a54c1 100644
--- a/drivers/gpu/drm/i915/gt/intel_rps.c
+++ b/drivers/gpu/drm/i915/gt/intel_rps.c
@@ -57,7 +57,7 @@ static u32 rps_pm_mask(struct intel_rps *rps, u8 val)
 	if (val < rps->max_freq_softlimit)
 		mask |= GEN6_PM_RP_UP_EI_EXPIRED | GEN6_PM_RP_UP_THRESHOLD;
 
-	mask &= rps->pm_events;
+	mask &= READ_ONCE(rps->pm_events);
 
 	return rps_pm_sanitize_mask(rps, ~mask);
 }
@@ -70,17 +70,19 @@ static void rps_reset_ei(struct intel_rps *rps)
 static void rps_enable_interrupts(struct intel_rps *rps)
 {
 	struct intel_gt *gt = rps_to_gt(rps);
+	u32 events;
 
 	rps_reset_ei(rps);
 
 	if (IS_VALLEYVIEW(gt->i915))
 		/* WaGsvRC0ResidencyMethod:vlv */
-		rps->pm_events = GEN6_PM_RP_UP_EI_EXPIRED;
+		events = GEN6_PM_RP_UP_EI_EXPIRED;
 	else
-		rps->pm_events = (GEN6_PM_RP_UP_THRESHOLD |
-				  GEN6_PM_RP_DOWN_THRESHOLD |
-				  GEN6_PM_RP_DOWN_TIMEOUT);
+		events = (GEN6_PM_RP_UP_THRESHOLD |
+			  GEN6_PM_RP_DOWN_THRESHOLD |
+			  GEN6_PM_RP_DOWN_TIMEOUT);
 
+	WRITE_ONCE(rps->pm_events, events);
 	spin_lock_irq(&gt->irq_lock);
 	gen6_gt_pm_enable_irq(gt, rps->pm_events);
 	spin_unlock_irq(&gt->irq_lock);
@@ -117,8 +119,7 @@ static void rps_disable_interrupts(struct intel_rps *rps)
 {
 	struct intel_gt *gt = rps_to_gt(rps);
 
-	rps->pm_events = 0;
-
+	WRITE_ONCE(rps->pm_events, 0);
 	set(gt->uncore, GEN6_PMINTRMSK, rps_pm_sanitize_mask(rps, ~0u));
 
 	spin_lock_irq(&gt->irq_lock);
@@ -1459,12 +1460,12 @@ static void rps_work(struct work_struct *work)
 	u32 pm_iir = 0;
 
 	spin_lock_irq(&gt->irq_lock);
-	pm_iir = fetch_and_zero(&rps->pm_iir);
+	pm_iir = fetch_and_zero(&rps->pm_iir) & READ_ONCE(rps->pm_events);
 	client_boost = atomic_read(&rps->num_waiters);
 	spin_unlock_irq(&gt->irq_lock);
 
 	/* Make sure we didn't queue anything we're not going to process. */
-	if ((pm_iir & rps->pm_events) == 0 && !client_boost)
+	if (!pm_iir && !client_boost)
 		goto out;
 
 	mutex_lock(&rps->lock);
@@ -1560,11 +1561,15 @@ void gen11_rps_irq_handler(struct intel_rps *rps, u32 pm_iir)
 void gen6_rps_irq_handler(struct intel_rps *rps, u32 pm_iir)
 {
 	struct intel_gt *gt = rps_to_gt(rps);
+	u32 events;
 
-	if (pm_iir & rps->pm_events) {
+	events = pm_iir & READ_ONCE(rps->pm_events);
+	if (events) {
 		spin_lock(&gt->irq_lock);
-		gen6_gt_pm_mask_irq(gt, pm_iir & rps->pm_events);
-		rps->pm_iir |= pm_iir & rps->pm_events;
+
+		gen6_gt_pm_mask_irq(gt, events);
+		rps->pm_iir |= events;
+
 		schedule_work(&rps->work);
 		spin_unlock(&gt->irq_lock);
 	}
-- 
2.20.1

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

[Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915/gt: Pull checking rps->pm_events under the irq_lock (rev2)

[Patchwork]

== Series Details ==

Series: drm/i915/gt: Pull checking rps->pm_events under the irq_lock (rev2)
URL   : https://patchwork.freedesktop.org/series/74510/
State : warning

== Summary ==

$ dim checkpatch origin/drm-tip
76ddbcfdb7a0 drm/i915/gt: Pull checking rps->pm_events under the irq_lock
-:7: WARNING:TYPO_SPELLING: 'diable' may be misspelled - perhaps 'disable'?
#7: 
write in rps_diable_interrupts.

-:31: WARNING:TYPO_SPELLING: 'interrrupt' may be misspelled - perhaps 'interrupt'?
#31: 
interrrupt generation are also strongly ordered -- just this may not be

total: 0 errors, 2 warnings, 0 checks, 72 lines checked

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

[Intel-gfx] ✗ Fi.CI.BAT: failure for drm/i915/gt: Pull checking rps->pm_events under the irq_lock (rev2)

[Patchwork]

== Series Details ==

Series: drm/i915/gt: Pull checking rps->pm_events under the irq_lock (rev2)
URL   : https://patchwork.freedesktop.org/series/74510/
State : failure

== Summary ==

CI Bug Log - changes from CI_DRM_8112 -> Patchwork_16908
====================================================

Summary
-------

  **FAILURE**

  Serious unknown changes coming with Patchwork_16908 absolutely need to be
  verified manually.
  
  If you think the reported changes have nothing to do with the changes
  introduced in Patchwork_16908, 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_16908/index.html

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

  Here are the unknown changes that may have been introduced in Patchwork_16908:

### IGT changes ###

#### Possible regressions ####

  * igt@i915_selftest@live@late_gt_pm:
    - fi-kbl-8809g:       [PASS][1] -> [DMESG-WARN][2]
   [1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8112/fi-kbl-8809g/igt@i915_selftest@live@late_gt_pm.html
   [2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_16908/fi-kbl-8809g/igt@i915_selftest@live@late_gt_pm.html

  
#### Suppressed ####

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

  * igt@runner@aborted:
    - {fi-kbl-7560u}:     NOTRUN -> [FAIL][3]
   [3]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_16908/fi-kbl-7560u/igt@runner@aborted.html

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

  Here are the changes found in Patchwork_16908 that come from known issues:

### IGT changes ###

#### Possible fixes ####

  * igt@kms_chamelium@hdmi-hpd-fast:
    - fi-kbl-7500u:       [FAIL][4] ([fdo#111407]) -> [PASS][5]
   [4]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8112/fi-kbl-7500u/igt@kms_chamelium@hdmi-hpd-fast.html
   [5]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_16908/fi-kbl-7500u/igt@kms_chamelium@hdmi-hpd-fast.html

  
#### Warnings ####

  * igt@gem_exec_suspend@basic-s4-devices:
    - fi-tgl-y:           [FAIL][6] ([CI#94]) -> [INCOMPLETE][7] ([CI#94] / [i915#460])
   [6]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8112/fi-tgl-y/igt@gem_exec_suspend@basic-s4-devices.html
   [7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_16908/fi-tgl-y/igt@gem_exec_suspend@basic-s4-devices.html

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

  [CI#94]: https://gitlab.freedesktop.org/gfx-ci/i915-infra/issues/94
  [fdo#111407]: https://bugs.freedesktop.org/show_bug.cgi?id=111407
  [fdo#112259]: https://bugs.freedesktop.org/show_bug.cgi?id=112259
  [i915#460]: https://gitlab.freedesktop.org/drm/intel/issues/460


Participating hosts (44 -> 38)
------------------------------

  Additional (2): fi-skl-6770hq fi-skl-6600u 
  Missing    (8): fi-hsw-4200u fi-hsw-peppy fi-byt-squawks fi-snb-2520m fi-byt-clapper fi-snb-2600 fi-bdw-samus fi-kbl-r 


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

  * CI: CI-20190529 -> None
  * Linux: CI_DRM_8112 -> Patchwork_16908

  CI-20190529: 20190529
  CI_DRM_8112: 032f2fe5c92eb1db6d417738431153c001a41bcc @ git://anongit.freedesktop.org/gfx-ci/linux
  IGT_5504: d6788bf0404f76b66170e18eb26c85004b5ccb25 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
  Patchwork_16908: 76ddbcfdb7a0ba78d7296c88f31ea49f56258ce4 @ git://anongit.freedesktop.org/gfx-ci/linux


== Linux commits ==

76ddbcfdb7a0 drm/i915/gt: Pull checking rps->pm_events under the irq_lock

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_16908/index.html
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx