drm/i915/bdw: BDW Software Turbo

BDW supports GT C0 residency reporting in constant time unit. Driver
calculates GT utilization based on C0 residency and adjusts RP
frequency up/down accordingly. For offscreen workload specificly,
set frequency to RP0.

Offscreen task is not restricted by frame rate, it can be
executed as soon as possible. Transcoding and serilized workload
between CPU and GPU both need high GT performance, RP0 is a good
option in this case. RC6 will kick in to compensate power
consumption when GT is not active.

v2: Rebase on recent drm-intel-nightly
v3: Add flip timerout monitor, when no flip is deteced within
100ms, set frequency to RP0.
Signed-off-by: NDaisy Sun <daisy.sun@intel.com>
[torourke: rebased on latest and resolved conflict]
Signed-off-by: NTom O'Rourke <Tom.O'Rourke@intel.com>
Acked-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

drivers/gpu/drm/i915/i915_drv.h

@@ -936,6 +936,23 @@ struct intel_rps_ei {
 	u32 media_c0;
 };
 
+struct intel_rps_bdw_cal {
+	u32 it_threshold_pct; /* interrupt, in percentage */
+	u32 eval_interval; /* evaluation interval, in us */
+	u32 last_ts;
+	u32 last_c0;
+	bool is_up;
+};
+
+struct intel_rps_bdw_turbo {
+	struct intel_rps_bdw_cal up;
+	struct intel_rps_bdw_cal down;
+	struct timer_list flip_timer;
+	u32 timeout;
+	atomic_t flip_received;
+	struct work_struct work_max_freq;
+};
+
 struct intel_gen6_power_mgmt {
 	/* work and pm_iir are protected by dev_priv->irq_lock */
 	struct work_struct work;
@@ -969,6 +986,9 @@ struct intel_gen6_power_mgmt {
 	bool enabled;
 	struct delayed_work delayed_resume_work;
 
+	bool is_bdw_sw_turbo;	/* Switch of BDW software turbo */
+	struct intel_rps_bdw_turbo sw_turbo; /* Calculate RP interrupt timing */
+
 	/* manual wa residency calculations */
 	struct intel_rps_ei up_ei, down_ei;
 
@@ -2791,6 +2811,8 @@ extern void intel_disable_fbc(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
 extern void intel_init_pch_refclk(struct drm_device *dev);
 extern void gen6_set_rps(struct drm_device *dev, u8 val);
+extern void bdw_software_turbo(struct drm_device *dev);
+extern void gen8_flip_interrupt(struct drm_device *dev);
 extern void valleyview_set_rps(struct drm_device *dev, u8 val);
 extern void intel_set_memory_cxsr(struct drm_i915_private *dev_priv,
 				  bool enable);

drivers/gpu/drm/i915/i915_irq.c

@@ -1979,6 +1979,27 @@ static void i9xx_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe)
 				     res1, res2);
 }
 
+void gen8_flip_interrupt(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!dev_priv->rps.is_bdw_sw_turbo)
+		return;
+
+	if(atomic_read(&dev_priv->rps.sw_turbo.flip_received)) {
+		mod_timer(&dev_priv->rps.sw_turbo.flip_timer,
+				usecs_to_jiffies(dev_priv->rps.sw_turbo.timeout) + jiffies);
+	}
+	else {
+		dev_priv->rps.sw_turbo.flip_timer.expires =
+				usecs_to_jiffies(dev_priv->rps.sw_turbo.timeout) + jiffies;
+		add_timer(&dev_priv->rps.sw_turbo.flip_timer);
+		atomic_set(&dev_priv->rps.sw_turbo.flip_received, true);
+	}
+
+	bdw_software_turbo(dev);
+}
+
 /* The RPS events need forcewake, so we add them to a work queue and mask their
  * IMR bits until the work is done. Other interrupts can be processed without
  * the work queue. */

drivers/gpu/drm/i915/i915_reg.h

@@ -5562,6 +5562,10 @@ enum punit_power_well {
 #define GEN8_UCGCTL6				0x9430
 #define   GEN8_SDEUNIT_CLOCK_GATE_DISABLE	(1<<14)
 
+#define TIMESTAMP_CTR		0x44070
+#define FREQ_1_28_US(us)	(((us) * 100) >> 7)
+#define MCHBAR_PCU_C0		(MCHBAR_MIRROR_BASE_SNB + 0x5960)
+
 #define GEN6_GFXPAUSE				0xA000
 #define GEN6_RPNSWREQ				0xA008
 #define   GEN6_TURBO_DISABLE			(1<<31)

drivers/gpu/drm/i915/intel_display.c

@@ -9757,6 +9757,9 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	unsigned long flags;
 	int ret;
 
+	//trigger software GT busyness calculation
+	gen8_flip_interrupt(dev);
+
 	/*
 	 * drm_mode_page_flip_ioctl() should already catch this, but double
 	 * check to be safe.  In the future we may enable pageflipping from

drivers/gpu/drm/i915/intel_pm.c

@@ -2258,7 +2258,6 @@ int ilk_wm_max_level(const struct drm_device *dev)
 	else
 		return 2;
 }
-
 static void intel_print_wm_latency(struct drm_device *dev,
 				   const char *name,
 				   const uint16_t wm[5])
@@ -3227,6 +3226,9 @@ static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val)
 {
 	int new_power;
 
+	if (dev_priv->rps.is_bdw_sw_turbo)
+		return;
+
 	new_power = dev_priv->rps.power;
 	switch (dev_priv->rps.power) {
 	case LOW_POWER:
@@ -3434,8 +3436,11 @@ void gen6_rps_idle(struct drm_i915_private *dev_priv)
 			valleyview_set_rps(dev_priv->dev, dev_priv->rps.min_freq_softlimit);
 		else if (IS_VALLEYVIEW(dev))
 			vlv_set_rps_idle(dev_priv);
-		else
+		else if (!dev_priv->rps.is_bdw_sw_turbo
+					|| atomic_read(&dev_priv->rps.sw_turbo.flip_received)){
 			gen6_set_rps(dev_priv->dev, dev_priv->rps.min_freq_softlimit);
+		}
+
 		dev_priv->rps.last_adj = 0;
 	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
@@ -3449,8 +3454,11 @@ void gen6_rps_boost(struct drm_i915_private *dev_priv)
 	if (dev_priv->rps.enabled) {
 		if (IS_VALLEYVIEW(dev))
 			valleyview_set_rps(dev_priv->dev, dev_priv->rps.max_freq_softlimit);
-		else
+		else if (!dev_priv->rps.is_bdw_sw_turbo
+					|| atomic_read(&dev_priv->rps.sw_turbo.flip_received)){
 			gen6_set_rps(dev_priv->dev, dev_priv->rps.max_freq_softlimit);
+		}
+
 		dev_priv->rps.last_adj = 0;
 	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
@@ -3481,21 +3489,26 @@ void valleyview_set_rps(struct drm_device *dev, u8 val)
 static void gen8_disable_rps_interrupts(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	if (IS_BROADWELL(dev) && dev_priv->rps.is_bdw_sw_turbo){
+		if (atomic_read(&dev_priv->rps.sw_turbo.flip_received))
+			del_timer(&dev_priv->rps.sw_turbo.flip_timer);
+		dev_priv-> rps.is_bdw_sw_turbo = false;
+	} else {
+		I915_WRITE(GEN6_PMINTRMSK, ~GEN8_PMINTR_REDIRECT_TO_NON_DISP);
+		I915_WRITE(GEN8_GT_IER(2), I915_READ(GEN8_GT_IER(2)) &
+					   ~dev_priv->pm_rps_events);
+		/* Complete PM interrupt masking here doesn't race with the rps work
+		 * item again unmasking PM interrupts because that is using a different
+		 * register (GEN8_GT_IMR(2)) to mask PM interrupts. The only risk is in
+		 * leaving stale bits in GEN8_GT_IIR(2) and GEN8_GT_IMR(2) which
+		 * gen8_enable_rps will clean up. */
 
-	I915_WRITE(GEN6_PMINTRMSK, ~GEN8_PMINTR_REDIRECT_TO_NON_DISP);
-	I915_WRITE(GEN8_GT_IER(2), I915_READ(GEN8_GT_IER(2)) &
-				   ~dev_priv->pm_rps_events);
-	/* Complete PM interrupt masking here doesn't race with the rps work
-	 * item again unmasking PM interrupts because that is using a different
-	 * register (GEN8_GT_IMR(2)) to mask PM interrupts. The only risk is in
-	 * leaving stale bits in GEN8_GT_IIR(2) and GEN8_GT_IMR(2) which
-	 * gen8_enable_rps will clean up. */
-
-	spin_lock_irq(&dev_priv->irq_lock);
-	dev_priv->rps.pm_iir = 0;
-	spin_unlock_irq(&dev_priv->irq_lock);
+		spin_lock_irq(&dev_priv->irq_lock);
+		dev_priv->rps.pm_iir = 0;
+		spin_unlock_irq(&dev_priv->irq_lock);
 
-	I915_WRITE(GEN8_GT_IIR(2), dev_priv->pm_rps_events);
+		I915_WRITE(GEN8_GT_IIR(2), dev_priv->pm_rps_events);
+	}
 }
 
 static void gen6_disable_rps_interrupts(struct drm_device *dev)
@@ -3653,13 +3666,111 @@ static void parse_rp_state_cap(struct drm_i915_private *dev_priv, u32 rp_state_c
 		dev_priv->rps.min_freq_softlimit = dev_priv->rps.min_freq;
 }
 
+static void bdw_sw_calculate_freq(struct drm_device *dev,
+		struct intel_rps_bdw_cal *c, u32 *cur_time, u32 *c0)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u64 busy = 0;
+	u32 busyness_pct = 0;
+	u32 elapsed_time = 0;
+	u16 new_freq = 0;
+
+	if (!c || !cur_time || !c0)
+		return;
+
+	if (0 == c->last_c0)
+		goto out;
+
+	/* Check Evaluation interval */
+	elapsed_time = *cur_time - c->last_ts;
+	if (elapsed_time < c->eval_interval)
+		return;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+
+	/*
+	 * c0 unit in 32*1.28 usec, elapsed_time unit in 1 usec.
+	 * Whole busyness_pct calculation should be
+	 *     busy = ((u64)(*c0 - c->last_c0) << 5 << 7) / 100;
+	 *     busyness_pct = (u32)(busy * 100 / elapsed_time);
+	 * The final formula is to simplify CPU calculation
+	 */
+	busy = (u64)(*c0 - c->last_c0) << 12;
+	do_div(busy, elapsed_time);
+	busyness_pct = (u32)busy;
+
+	if (c->is_up && busyness_pct >= c->it_threshold_pct)
+		new_freq = (u16)dev_priv->rps.cur_freq + 3;
+	if (!c->is_up && busyness_pct <= c->it_threshold_pct)
+		new_freq = (u16)dev_priv->rps.cur_freq - 1;
+
+	/* Adjust to new frequency busyness and compare with threshold */
+	if (0 != new_freq) {
+		if (new_freq > dev_priv->rps.max_freq_softlimit)
+			new_freq = dev_priv->rps.max_freq_softlimit;
+		else if (new_freq < dev_priv->rps.min_freq_softlimit)
+			new_freq = dev_priv->rps.min_freq_softlimit;
+
+		gen6_set_rps(dev, new_freq);
+	}
+
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+out:
+	c->last_c0 = *c0;
+	c->last_ts = *cur_time;
+}
+
+static void gen8_set_frequency_RP0(struct work_struct *work)
+{
+	struct intel_rps_bdw_turbo *p_bdw_turbo =
+			container_of(work, struct intel_rps_bdw_turbo, work_max_freq);
+	struct intel_gen6_power_mgmt *p_power_mgmt =
+			container_of(p_bdw_turbo, struct intel_gen6_power_mgmt, sw_turbo);
+	struct drm_i915_private *dev_priv =
+			container_of(p_power_mgmt, struct drm_i915_private, rps);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	gen6_set_rps(dev_priv->dev, dev_priv->rps.rp0_freq);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
+static void flip_active_timeout_handler(unsigned long var)
+{
+	struct drm_i915_private *dev_priv = (struct drm_i915_private *) var;
+
+	del_timer(&dev_priv->rps.sw_turbo.flip_timer);
+	atomic_set(&dev_priv->rps.sw_turbo.flip_received, false);
+
+	queue_work(dev_priv->wq, &dev_priv->rps.sw_turbo.work_max_freq);
+}
+
+void bdw_software_turbo(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	u32 current_time = I915_READ(TIMESTAMP_CTR); /* unit in usec */
+	u32 current_c0 = I915_READ(MCHBAR_PCU_C0); /* unit in 32*1.28 usec */
+
+	bdw_sw_calculate_freq(dev, &dev_priv->rps.sw_turbo.up,
+			&current_time, &current_c0);
+	bdw_sw_calculate_freq(dev, &dev_priv->rps.sw_turbo.down,
+			&current_time, &current_c0);
+}
+
 static void gen8_enable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_engine_cs *ring;
 	uint32_t rc6_mask = 0, rp_state_cap;
+	uint32_t threshold_up_pct, threshold_down_pct;
+	uint32_t ei_up, ei_down; /* up and down evaluation interval */
+	u32 rp_ctl_flag;
 	int unused;
 
+	/* Use software Turbo for BDW */
+	dev_priv->rps.is_bdw_sw_turbo = IS_BROADWELL(dev);
+
 	/* 1a: Software RC state - RC0 */
 	I915_WRITE(GEN6_RC_STATE, 0);
 
@@ -3703,35 +3814,74 @@ static void gen8_enable_rps(struct drm_device *dev)
 		   HSW_FREQUENCY(dev_priv->rps.rp1_freq));
 	I915_WRITE(GEN6_RC_VIDEO_FREQ,
 		   HSW_FREQUENCY(dev_priv->rps.rp1_freq));
-	/* NB: Docs say 1s, and 1000000 - which aren't equivalent */
-	I915_WRITE(GEN6_RP_DOWN_TIMEOUT, 100000000 / 128); /* 1 second timeout */
+	ei_up = 84480; /* 84.48ms */
+	ei_down = 448000;
+	threshold_up_pct = 90; /* x percent busy */
+	threshold_down_pct = 70;
+
+	if (dev_priv->rps.is_bdw_sw_turbo) {
+		dev_priv->rps.sw_turbo.up.it_threshold_pct = threshold_up_pct;
+		dev_priv->rps.sw_turbo.up.eval_interval = ei_up;
+		dev_priv->rps.sw_turbo.up.is_up = true;
+		dev_priv->rps.sw_turbo.up.last_ts = 0;
+		dev_priv->rps.sw_turbo.up.last_c0 = 0;
+
+		dev_priv->rps.sw_turbo.down.it_threshold_pct = threshold_down_pct;
+		dev_priv->rps.sw_turbo.down.eval_interval = ei_down;
+		dev_priv->rps.sw_turbo.down.is_up = false;
+		dev_priv->rps.sw_turbo.down.last_ts = 0;
+		dev_priv->rps.sw_turbo.down.last_c0 = 0;
+
+		/* Start the timer to track if flip comes*/
+		dev_priv->rps.sw_turbo.timeout = 200*1000; /* in us */
+
+		init_timer(&dev_priv->rps.sw_turbo.flip_timer);
+		dev_priv->rps.sw_turbo.flip_timer.function = flip_active_timeout_handler;
+		dev_priv->rps.sw_turbo.flip_timer.data  = (unsigned long) dev_priv;
+		dev_priv->rps.sw_turbo.flip_timer.expires =
+			usecs_to_jiffies(dev_priv->rps.sw_turbo.timeout) + jiffies;
+		add_timer(&dev_priv->rps.sw_turbo.flip_timer);
+		INIT_WORK(&dev_priv->rps.sw_turbo.work_max_freq, gen8_set_frequency_RP0);
+
+		atomic_set(&dev_priv->rps.sw_turbo.flip_received, true);
+	} else {
+		/* NB: Docs say 1s, and 1000000 - which aren't equivalent
+		 * 1 second timeout*/
+		I915_WRITE(GEN6_RP_DOWN_TIMEOUT, FREQ_1_28_US(1000000));
 
-	/* Docs recommend 900MHz, and 300 MHz respectively */
-	I915_WRITE(GEN6_RP_INTERRUPT_LIMITS,
-		   dev_priv->rps.max_freq_softlimit << 24 |
-		   dev_priv->rps.min_freq_softlimit << 16);
+		/* Docs recommend 900MHz, and 300 MHz respectively */
+		I915_WRITE(GEN6_RP_INTERRUPT_LIMITS,
+			   dev_priv->rps.max_freq_softlimit << 24 |
+			   dev_priv->rps.min_freq_softlimit << 16);
 
-	I915_WRITE(GEN6_RP_UP_THRESHOLD, 7600000 / 128); /* 76ms busyness per EI, 90% */
-	I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 31300000 / 128); /* 313ms busyness per EI, 70%*/
-	I915_WRITE(GEN6_RP_UP_EI, 66000); /* 84.48ms, XXX: random? */
-	I915_WRITE(GEN6_RP_DOWN_EI, 350000); /* 448ms, XXX: random? */
+		I915_WRITE(GEN6_RP_UP_THRESHOLD,
+			FREQ_1_28_US(ei_up * threshold_up_pct / 100));
+		I915_WRITE(GEN6_RP_DOWN_THRESHOLD,
+			FREQ_1_28_US(ei_down * threshold_down_pct / 100));
+		I915_WRITE(GEN6_RP_UP_EI,
+			FREQ_1_28_US(ei_up));
+		I915_WRITE(GEN6_RP_DOWN_EI,
+			FREQ_1_28_US(ei_down));
 
-	I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
+		I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
+	}
 
 	/* 5: Enable RPS */
-	I915_WRITE(GEN6_RP_CONTROL,
-		   GEN6_RP_MEDIA_TURBO |
-		   GEN6_RP_MEDIA_HW_NORMAL_MODE |
-		   GEN6_RP_MEDIA_IS_GFX |
-		   GEN6_RP_ENABLE |
-		   GEN6_RP_UP_BUSY_AVG |
-		   GEN6_RP_DOWN_IDLE_AVG);
-
-	/* 6: Ring frequency + overclocking (our driver does this later */
-
+	rp_ctl_flag = GEN6_RP_MEDIA_TURBO |
+					GEN6_RP_MEDIA_HW_NORMAL_MODE |
+					GEN6_RP_MEDIA_IS_GFX |
+					GEN6_RP_UP_BUSY_AVG |
+					GEN6_RP_DOWN_IDLE_AVG;
+	if (!dev_priv->rps.is_bdw_sw_turbo)
+		rp_ctl_flag |= GEN6_RP_ENABLE;
+
+	I915_WRITE(GEN6_RP_CONTROL, rp_ctl_flag);
+
+	/* 6: Ring frequency + overclocking
+	 * (our driver does this later */
 	gen6_set_rps(dev, (I915_READ(GEN6_GT_PERF_STATUS) & 0xff00) >> 8);
-
-	gen8_enable_rps_interrupts(dev);
+	if (!dev_priv->rps.is_bdw_sw_turbo)
+		gen8_enable_rps_interrupts(dev);
 
 	gen6_gt_force_wake_put(dev_priv, FORCEWAKE_ALL);
 }
@@ -5199,6 +5349,8 @@ static void intel_gen6_powersave_work(struct work_struct *work)
 			     rps.delayed_resume_work.work);
 	struct drm_device *dev = dev_priv->dev;
 
+	dev_priv->rps.is_bdw_sw_turbo = false;
+
 	mutex_lock(&dev_priv->rps.hw_lock);
 
 	if (IS_CHERRYVIEW(dev)) {