public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
* [PATCH] drm/i915: avoid unclaimed registers when capturing the error state
@ 2013-12-20 17:09 Paulo Zanoni
  2014-01-07 14:07 ` Imre Deak
  2014-06-02 13:35 ` Chris Wilson
  0 siblings, 2 replies; 6+ messages in thread
From: Paulo Zanoni @ 2013-12-20 17:09 UTC (permalink / raw)
  To: intel-gfx; +Cc: Paulo Zanoni

From: Paulo Zanoni <paulo.r.zanoni@intel.com>

We're iterating over the CPU transcoders, so check for the correct
power domain.

This fixes many "unclaimed register" error messages.

This can be reproduced by the IGT test mentioned below, but we still
get a FAIL when we run it.

Testcase: igt/kms_lip/flip-vs-panning-vs-hang
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)


I didn't bisect to check, but it really really looks like a regression from
"drm/i915: add intel_display_power_enabled_sw() for use in atomic ctx".



diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 4d1357a..4d4b4bc 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -11459,7 +11459,8 @@ intel_display_capture_error_state(struct drm_device *dev)
 		enum transcoder cpu_transcoder = transcoders[i];
 
 		error->transcoder[i].power_domain_on =
-			intel_display_power_enabled_sw(dev, POWER_DOMAIN_PIPE(i));
+			intel_display_power_enabled_sw(dev,
+				POWER_DOMAIN_TRANSCODER(cpu_transcoder));
 		if (!error->transcoder[i].power_domain_on)
 			continue;
 
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 6+ messages in thread
* [PATCH] drm/i915: avoid unclaimed registers when capturing the error state
@ 2013-11-01 15:32 Paulo Zanoni
  2013-11-01 18:28 ` Daniel Vetter
  0 siblings, 1 reply; 6+ messages in thread
From: Paulo Zanoni @ 2013-11-01 15:32 UTC (permalink / raw)
  To: intel-gfx; +Cc: Paulo Zanoni

From: Paulo Zanoni <paulo.r.zanoni@intel.com>

Even though we only check for unclaimed registers while we're writing
registers, if we read a bad register we'll still trigger a CPU error
interrupt, and we'll print an "Unclaimed register" DRM_ERROR due to
that. To avoid this error, just avoid touching power domains that are
not enabled.

Use kzalloc so we're sure all the disabled domains will be zeroed on
the error state file. We already print the information that is enough
to discover if the power well is enabled on the error state file, so
this should not be a problem.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=69747
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e841cd7..fba7d9b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -11138,7 +11138,7 @@ intel_display_capture_error_state(struct drm_device *dev)
 	if (INTEL_INFO(dev)->num_pipes == 0)
 		return NULL;
 
-	error = kmalloc(sizeof(*error), GFP_ATOMIC);
+	error = kzalloc(sizeof(*error), GFP_ATOMIC);
 	if (error == NULL)
 		return NULL;
 
@@ -11146,6 +11146,9 @@ intel_display_capture_error_state(struct drm_device *dev)
 		error->power_well_driver = I915_READ(HSW_PWR_WELL_DRIVER);
 
 	for_each_pipe(i) {
+		if (!intel_display_power_enabled(dev, POWER_DOMAIN_PIPE(i)))
+			continue;
+
 		if (INTEL_INFO(dev)->gen <= 6 || IS_VALLEYVIEW(dev)) {
 			error->cursor[i].control = I915_READ(CURCNTR(i));
 			error->cursor[i].position = I915_READ(CURPOS(i));
@@ -11179,6 +11182,10 @@ intel_display_capture_error_state(struct drm_device *dev)
 	for (i = 0; i < error->num_transcoders; i++) {
 		enum transcoder cpu_transcoder = transcoders[i];
 
+		if (!intel_display_power_enabled(dev,
+				POWER_DOMAIN_TRANSCODER(cpu_transcoder)))
+			continue;
+
 		error->transcoder[i].cpu_transcoder = cpu_transcoder;
 
 		error->transcoder[i].conf = I915_READ(PIPECONF(cpu_transcoder));
@@ -11190,12 +11197,6 @@ intel_display_capture_error_state(struct drm_device *dev)
 		error->transcoder[i].vsync = I915_READ(VSYNC(cpu_transcoder));
 	}
 
-	/* In the code above we read the registers without checking if the power
-	 * well was on, so here we have to clear the FPGA_DBG_RM_NOCLAIM bit to
-	 * prevent the next I915_WRITE from detecting it and printing an error
-	 * message. */
-	intel_uncore_clear_errors(dev);
-
 	return error;
 }
 
-- 
1.8.3.1

^ permalink raw reply related	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-06-02 14:03 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-12-20 17:09 [PATCH] drm/i915: avoid unclaimed registers when capturing the error state Paulo Zanoni
2014-01-07 14:07 ` Imre Deak
2014-06-02 13:35 ` Chris Wilson
2014-06-02 14:02   ` Imre Deak
  -- strict thread matches above, loose matches on Subject: below --
2013-11-01 15:32 Paulo Zanoni
2013-11-01 18:28 ` Daniel Vetter

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