drm/i915: Stop using RP_DOWN_EI on Baytrail

commit 8f68d591 upstream.

On Baytrail, we manually calculate busyness over the evaluation interval
to avoid issues with miscaluations with RC6 enabled. However, it turns
out that the DOWN_EI interrupt generator is completely bust - it
operates in two modes, continuous or never. Neither of which are
conducive to good behaviour. Stop unmask the DOWN_EI interrupt and just
compute everything from the UP_EI which does seem to correspond to the
desired interval.

v2: Fixup gen6_rps_pm_mask() as well
v3: Inline vlv_c0_above() to combine the now identical elapsed
calculation for up/down and simplify the threshold testing

Fixes: 43cf3bf0 ("drm/i915: Improved w/a for rps on Baytrail")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170309211232.28878-1-chris@chris-wilson.co.ukReviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170313170617.31564-1-chris@chris-wilson.co.uk
(cherry picked from commit e0e8c7cb)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/gpu/drm/i915/i915_drv.h

@@ -1159,7 +1159,7 @@ struct intel_gen6_power_mgmt {
 	struct intel_rps_client semaphores, mmioflips;
 
 	/* manual wa residency calculations */
-	struct intel_rps_ei up_ei, down_ei;
+	struct intel_rps_ei ei;
 
 	/*
 	 * Protects RPS/RC6 register access and PCU communication.

drivers/gpu/drm/i915/i915_irq.c

@@ -994,68 +994,51 @@ static void vlv_c0_read(struct drm_i915_private *dev_priv,
 	ei->media_c0 = I915_READ(VLV_MEDIA_C0_COUNT);
 }
 
-static bool vlv_c0_above(struct drm_i915_private *dev_priv,
-			 const struct intel_rps_ei *old,
-			 const struct intel_rps_ei *now,
-			 int threshold)
-{
-	u64 time, c0;
-	unsigned int mul = 100;
-
-	if (old->cz_clock == 0)
-		return false;
-
-	if (I915_READ(VLV_COUNTER_CONTROL) & VLV_COUNT_RANGE_HIGH)
-		mul <<= 8;
-
-	time = now->cz_clock - old->cz_clock;
-	time *= threshold * dev_priv->czclk_freq;
-
-	/* Workload can be split between render + media, e.g. SwapBuffers
-	 * being blitted in X after being rendered in mesa. To account for
-	 * this we need to combine both engines into our activity counter.
-	 */
-	c0 = now->render_c0 - old->render_c0;
-	c0 += now->media_c0 - old->media_c0;
-	c0 *= mul * VLV_CZ_CLOCK_TO_MILLI_SEC;
-
-	return c0 >= time;
-}
-
 void gen6_rps_reset_ei(struct drm_i915_private *dev_priv)
 {
-	vlv_c0_read(dev_priv, &dev_priv->rps.down_ei);
-	dev_priv->rps.up_ei = dev_priv->rps.down_ei;
+	memset(&dev_priv->rps.ei, 0, sizeof(dev_priv->rps.ei));
 }
 
 static u32 vlv_wa_c0_ei(struct drm_i915_private *dev_priv, u32 pm_iir)
 {
+	const struct intel_rps_ei *prev = &dev_priv->rps.ei;
 	struct intel_rps_ei now;
 	u32 events = 0;
 
-	if ((pm_iir & (GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED)) == 0)
+	if ((pm_iir & GEN6_PM_RP_UP_EI_EXPIRED) == 0)
 		return 0;
 
 	vlv_c0_read(dev_priv, &now);
 	if (now.cz_clock == 0)
 		return 0;
 
-	if (pm_iir & GEN6_PM_RP_DOWN_EI_EXPIRED) {
-		if (!vlv_c0_above(dev_priv,
-				  &dev_priv->rps.down_ei, &now,
-				  dev_priv->rps.down_threshold))
-			events |= GEN6_PM_RP_DOWN_THRESHOLD;
-		dev_priv->rps.down_ei = now;
-	}
+	if (prev->cz_clock) {
+		u64 time, c0;
+		unsigned int mul;
 
-	if (pm_iir & GEN6_PM_RP_UP_EI_EXPIRED) {
-		if (vlv_c0_above(dev_priv,
-				 &dev_priv->rps.up_ei, &now,
-				 dev_priv->rps.up_threshold))
-			events |= GEN6_PM_RP_UP_THRESHOLD;
-		dev_priv->rps.up_ei = now;
+		mul = VLV_CZ_CLOCK_TO_MILLI_SEC * 100; /* scale to threshold% */
+		if (I915_READ(VLV_COUNTER_CONTROL) & VLV_COUNT_RANGE_HIGH)
+			mul <<= 8;
+
+		time = now.cz_clock - prev->cz_clock;
+		time *= dev_priv->czclk_freq;
+
+		/* Workload can be split between render + media,
+		 * e.g. SwapBuffers being blitted in X after being rendered in
+		 * mesa. To account for this we need to combine both engines
+		 * into our activity counter.
+		 */
+		c0 = now.render_c0 - prev->render_c0;
+		c0 += now.media_c0 - prev->media_c0;
+		c0 *= mul;
+
+		if (c0 > time * dev_priv->rps.up_threshold)
+			events = GEN6_PM_RP_UP_THRESHOLD;
+		else if (c0 < time * dev_priv->rps.down_threshold)
+			events = GEN6_PM_RP_DOWN_THRESHOLD;
 	}
 
+	dev_priv->rps.ei = now;
 	return events;
 }
 
@@ -4390,7 +4373,7 @@ void intel_irq_init(struct drm_i915_private *dev_priv)
 	/* Let's track the enabled rps events */
 	if (IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
 		/* WaGsvRC0ResidencyMethod:vlv */
-		dev_priv->pm_rps_events = GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED;
+		dev_priv->pm_rps_events = GEN6_PM_RP_UP_EI_EXPIRED;
 	else
 		dev_priv->pm_rps_events = GEN6_PM_RPS_EVENTS;
 

drivers/gpu/drm/i915/intel_pm.c

@@ -4411,8 +4411,9 @@ static u32 gen6_rps_pm_mask(struct drm_i915_private *dev_priv, u8 val)
 {
 	u32 mask = 0;
 
+	/* We use UP_EI_EXPIRED interupts for both up/down in manual mode */
 	if (val > dev_priv->rps.min_freq_softlimit)
-		mask |= GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_DOWN_THRESHOLD | GEN6_PM_RP_DOWN_TIMEOUT;
+		mask |= GEN6_PM_RP_UP_EI_EXPIRED | GEN6_PM_RP_DOWN_THRESHOLD | GEN6_PM_RP_DOWN_TIMEOUT;
 	if (val < dev_priv->rps.max_freq_softlimit)
 		mask |= GEN6_PM_RP_UP_EI_EXPIRED | GEN6_PM_RP_UP_THRESHOLD;
 
@@ -4516,7 +4517,7 @@ void gen6_rps_busy(struct drm_i915_private *dev_priv)
 {
 	mutex_lock(&dev_priv->rps.hw_lock);
 	if (dev_priv->rps.enabled) {
-		if (dev_priv->pm_rps_events & (GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED))
+		if (dev_priv->pm_rps_events & GEN6_PM_RP_UP_EI_EXPIRED)
 			gen6_rps_reset_ei(dev_priv);
 		I915_WRITE(GEN6_PMINTRMSK,
 			   gen6_rps_pm_mask(dev_priv, dev_priv->rps.cur_freq));