Re: [PATCH v4 2/2] drm/i915/vlv: WA to fix Voltage not getting dropped to Vmin when Gfx is power gated.

From: Daniel Vetter <daniel@ffwll.ch>
To: "S, Deepak" <deepak.s@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH v4 2/2] drm/i915/vlv: WA to fix Voltage not getting dropped to Vmin when Gfx is power gated.
Date: Tue, 28 Jan 2014 20:33:45 +0100	[thread overview]
Message-ID: <20140128193345.GE7444@phenom.ffwll.local> (raw)
In-Reply-To: <52E7BF98.90404@intel.com>

On Tue, Jan 28, 2014 at 08:02:56PM +0530, S, Deepak wrote:
> 
> 
> On 1/27/2014 10:22 PM, Daniel Vetter wrote:
> >On Mon, Jan 27, 2014 at 09:35:06PM +0530, deepak.s@intel.com wrote:
> >>From: Deepak S <deepak.s@intel.com>
> >>
> >>When we enter RC6 and GFX Clocks are off, the voltage remains higher
> >>than Vmin. When we try to set the freq to RPn, it might fail since the
> >>Gfx clocks are down. So to fix this in Gfx idle, Bring the GFX clock up
> >>and set the freq to RPn then move GFx down.
> >>
> >>v2: remove vlv_update_rps_cur_delay function. Update commit message (Daniel)
> >>
> >>v3: Fix the timeout during wait for gfx clock (Jesse)
> >>
> >>v4: addressed comments on set freq and punit wait (Ville)
> >>
> >>Signed-off-by: Deepak S <deepak.s@intel.com>
> >>---
> >>  drivers/gpu/drm/i915/i915_reg.h |  4 ++++
> >>  drivers/gpu/drm/i915/intel_pm.c | 53 ++++++++++++++++++++++++++++++++++++++++-
> >>  2 files changed, 56 insertions(+), 1 deletion(-)
> >>
> >>diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> >>index 242f540..feaa83b 100644
> >>--- a/drivers/gpu/drm/i915/i915_reg.h
> >>+++ b/drivers/gpu/drm/i915/i915_reg.h
> >>@@ -4944,6 +4944,10 @@
> >>  						 GEN6_PM_RP_DOWN_THRESHOLD | \
> >>  						 GEN6_PM_RP_DOWN_TIMEOUT)
> >>
> >>+#define VLV_GTLC_SURVIVABILITY_REG              0x130098
> >>+#define VLV_GFX_CLK_STATUS_BIT			(1<<3)
> >>+#define VLV_GFX_CLK_FORCE_ON_BIT		(1<<2)
> >>+
> >>  #define GEN6_GT_GFX_RC6_LOCKED			0x138104
> >>  #define VLV_COUNTER_CONTROL			0x138104
> >>  #define   VLV_COUNT_RANGE_HIGH			(1<<15)
> >>diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> >>index c6a07c9..84e20d0 100644
> >>--- a/drivers/gpu/drm/i915/intel_pm.c
> >>+++ b/drivers/gpu/drm/i915/intel_pm.c
> >>@@ -3035,6 +3035,56 @@ void gen6_set_rps(struct drm_device *dev, u8 val)
> >>  	trace_intel_gpu_freq_change(val * 50);
> >>  }
> >>
> >>+/* vlv_set_rps_idle: Set the frequency to Rpn if Gfx clocks are down
> >>+ *
> >>+ * * If Gfx is Idle, then
> >>+ * 1. Mask Turbo interrupts
> >>+ * 2. Bring up Gfx clock
> >>+ * 3. Change the freq to Rpn and wait till P-Unit updates freq
> >>+ * 4. Clear the Force GFX CLK ON bit so that Gfx can down
> >>+ * 5. Unmask Turbo interrupts
> >>+*/
> >>+static void vlv_set_rps_idle(struct drm_i915_private *dev_priv)
> >>+{
> >>+	/*
> >>+	 * When we are idle.  Drop to min voltage state.
> >>+	 */
> >>+
> >>+	if (dev_priv->rps.cur_delay <= dev_priv->rps.min_delay)
> >>+		return;
> >>+
> >>+	/* Mask turbo interrupt so that they will not come in between */
> >>+	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
> >>+
> >>+	/* Bring up the Gfx clock */
> >>+	I915_WRITE(VLV_GTLC_SURVIVABILITY_REG,
> >>+		I915_READ(VLV_GTLC_SURVIVABILITY_REG) |
> >>+				VLV_GFX_CLK_FORCE_ON_BIT);
> >>+
> >>+	if (wait_for_atomic(((VLV_GFX_CLK_STATUS_BIT &
> >
> >Do we really need an atomic register busy loop here? Afaics this is only
> >called from process context, so the normal wait_for macro should be good
> >enough ...
> >-Daniel
> I agree, the reason why i did the _atomic as we observed delay in
> scheduling the wait_for and we ended up spending lot of time here.
> I will wait for other comments, If i dont get any, I will address
> this comment and submit the patch.

Now that we're using Chris' idle-clamping rps logic this would be run from
a work queue. So I hope that the potential scheduling delays here won't
affect things badly. If it does you can stick with the _atomic, but then
it needs a comment.

But I really hope that we only call this from worker threads, so shouldn't
matter too badly if there's a bit a scheduler delay.
-Daniel

> 
> >>+		I915_READ(VLV_GTLC_SURVIVABILITY_REG)) != 0), 5)) {
> >>+			DRM_ERROR("GFX_CLK_ON request timed out\n");
> >>+		return;
> >>+	}
> >>+
> >>+	vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, dev_priv->rps.min_delay);
> >>+
> >>+	if (wait_for_atomic(((vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS))
> >>+				& GENFREQSTATUS) == 0, 5))
> >>+		DRM_DEBUG_DRIVER("timed out waiting for Punit\n");
> >>+
> >>+	/* Release the Gfx clock */
> >>+	I915_WRITE(VLV_GTLC_SURVIVABILITY_REG,
> >>+		I915_READ(VLV_GTLC_SURVIVABILITY_REG) &
> >>+				~VLV_GFX_CLK_FORCE_ON_BIT);
> >>+
> >>+	/* Unmask Turbo interrupts */
> >>+	I915_WRITE(GEN6_PMINTRMSK, ~(GEN6_PM_RPS_EVENTS |
> >>+					GEN6_PM_RP_UP_EI_EXPIRED));
> >>+}
> >>+
> >>+
> >>+
> >>  void gen6_rps_idle(struct drm_i915_private *dev_priv)
> >>  {
> >>  	struct drm_device *dev = dev_priv->dev;
> >>@@ -3042,7 +3092,7 @@ void gen6_rps_idle(struct drm_i915_private *dev_priv)
> >>  	mutex_lock(&dev_priv->rps.hw_lock);
> >>  	if (dev_priv->rps.enabled) {
> >>  		if (IS_VALLEYVIEW(dev))
> >>-			valleyview_set_rps(dev_priv->dev, dev_priv->rps.min_delay);
> >>+			vlv_set_rps_idle(dev_priv);
> >>  		else
> >>  			gen6_set_rps(dev_priv->dev, dev_priv->rps.min_delay);
> >>  		dev_priv->rps.last_adj = 0;
> >>@@ -4273,6 +4323,7 @@ void intel_gpu_ips_teardown(void)
> >>  	i915_mch_dev = NULL;
> >>  	spin_unlock_irq(&mchdev_lock);
> >>  }
> >>+
> >>  static void intel_init_emon(struct drm_device *dev)
> >>  {
> >>  	struct drm_i915_private *dev_priv = dev->dev_private;
> >>--
> >>1.8.5.2
> >>
> >>_______________________________________________
> >>Intel-gfx mailing list
> >>Intel-gfx@lists.freedesktop.org
> >>http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> >

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch