[PATCH 00/18] drm/i915: All sorts of cdclk stuff

[PATCH 00/18] drm/i915: All sorts of cdclk stuff

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

My main motivation here was to get dev_priv->max_cdclk into place on
all platforms so that we can start to use it to validate modes and
whatnot. This series doesn't actually add any new checks like that
apart from the BDW IPS case, and converting over whatever checks we
already had.

The other thing I managed to do is add support for cdclk dynamic
frequency scaling on HSW/BDW, and I also fixed it for CHV.

The HSW stuff we may want to drop actually since I was told it's not
validated, and it doesn't scale the voltage in any case so the benefits
may not be worth it. I did try it out on my HSW here and it seemed to
work just fine. In any case I figured I'll include the patch anyway.

As usual, the CHV case is again a major documentation snafu. Let's
just say I made it work despite the documentation. The 400MHz case
doesn't seem to work actually, or rather the Punit seems to reject
any request above 320MHz. Also the 200MHz case seems as busted as on
VLV. I just get an immediate underrun and a black screen, even though
the pixel clock is definitely below the 90% of cdclk limit.

I pushed the lot (+ a hack to force a higher cdclk via a modparam)
here:
git://gitorious.org/vsyrjala/linux.git cdclk_7

Ville Syrjälä (18):
  drm/i915: Return more precise cdclk for gen2/3
  drm/i915: Fix i855_get_display_clock_speed()
  drm/i915: Fix 852GM/GMV cdclk
  drm/i915: Add cdclk extraction for g33, 965gm and g4x
  drm/i915: ILK cdclk seems to be 450MHz
  drm/i915: Assume 400 MHz cdclk for the rest of gen4-7
  drm/i915: Simplify ilk_get_aux_clock_divider()
  drm/i915: Convert the ddi cdclk code to .get_display_clock_speed()
  drm/i915: Warn when cdclk for the platforms is not known
  drm/i915: Cache the current cdclk frequency in dev_priv
  drm/i915: Use cached cdclk value
  drm/i915: Unify ilk and hsw .get_aux_clock_divider()
  drm/i915: Store max cdclk value in dev_priv
  drm/i915: Don't enable IPS when pixel rate exceeds 95% of cdclk
  drm/i915: Fix chv cdclk support
  drm/i915: HSW cdclk change support
  drm/i915: Add IS_BDW_ULX()
  drm/i915: BDW cdclk change support

 drivers/gpu/drm/i915/i915_drv.h         |   5 +-
 drivers/gpu/drm/i915/i915_reg.h         |  18 +-
 drivers/gpu/drm/i915/intel_ddi.c        | 101 +----
 drivers/gpu/drm/i915/intel_display.c    | 697 +++++++++++++++++++++++++++++---
 drivers/gpu/drm/i915/intel_dp.c         |  26 +-
 drivers/gpu/drm/i915/intel_drv.h        |   2 +-
 drivers/gpu/drm/i915/intel_pm.c         |  18 +-
 drivers/gpu/drm/i915/intel_runtime_pm.c |   2 +-
 8 files changed, 672 insertions(+), 197 deletions(-)

-- 
2.0.4

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

[PATCH 01/18] drm/i915: Return more precise cdclk for gen2/3

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Fill out the lower three digits for gen2 and gen3 cdclk frqeuncy. It's
not clear if these are accurate frquencies or just in the ballpark, but
without docs this is the best we can do.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index dda97b3..e364ca7 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5605,7 +5605,7 @@ static int i945_get_display_clock_speed(struct drm_device *dev)
 
 static int i915_get_display_clock_speed(struct drm_device *dev)
 {
-	return 333000;
+	return 333333;
 }
 
 static int i9xx_misc_get_display_clock_speed(struct drm_device *dev)
@@ -5621,19 +5621,19 @@ static int pnv_get_display_clock_speed(struct drm_device *dev)
 
 	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
 	case GC_DISPLAY_CLOCK_267_MHZ_PNV:
-		return 267000;
+		return 266667;
 	case GC_DISPLAY_CLOCK_333_MHZ_PNV:
-		return 333000;
+		return 333333;
 	case GC_DISPLAY_CLOCK_444_MHZ_PNV:
-		return 444000;
+		return 444444;
 	case GC_DISPLAY_CLOCK_200_MHZ_PNV:
 		return 200000;
 	default:
 		DRM_ERROR("Unknown pnv display core clock 0x%04x\n", gcfgc);
 	case GC_DISPLAY_CLOCK_133_MHZ_PNV:
-		return 133000;
+		return 133333;
 	case GC_DISPLAY_CLOCK_167_MHZ_PNV:
-		return 167000;
+		return 166667;
 	}
 }
 
@@ -5644,11 +5644,11 @@ static int i915gm_get_display_clock_speed(struct drm_device *dev)
 	pci_read_config_word(dev->pdev, GCFGC, &gcfgc);
 
 	if (gcfgc & GC_LOW_FREQUENCY_ENABLE)
-		return 133000;
+		return 133333;
 	else {
 		switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
 		case GC_DISPLAY_CLOCK_333_MHZ:
-			return 333000;
+			return 333333;
 		default:
 		case GC_DISPLAY_CLOCK_190_200_MHZ:
 			return 190000;
@@ -5658,7 +5658,7 @@ static int i915gm_get_display_clock_speed(struct drm_device *dev)
 
 static int i865_get_display_clock_speed(struct drm_device *dev)
 {
-	return 266000;
+	return 266667;
 }
 
 static int i855_get_display_clock_speed(struct drm_device *dev)
@@ -5674,7 +5674,7 @@ static int i855_get_display_clock_speed(struct drm_device *dev)
 	case GC_CLOCK_166_250:
 		return 250000;
 	case GC_CLOCK_100_133:
-		return 133000;
+		return 133333;
 	}
 
 	/* Shouldn't happen */
@@ -5683,7 +5683,7 @@ static int i855_get_display_clock_speed(struct drm_device *dev)
 
 static int i830_get_display_clock_speed(struct drm_device *dev)
 {
-	return 133000;
+	return 133333;
 }
 
 static void
-- 
2.0.4

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

[PATCH 02/18] drm/i915: Fix i855_get_display_clock_speed()

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Actually read the HPLLCC register insted of assuming it's 0. Fix the
HPLLCC bit definitions and all the missing ones from the 852GME spec.

852GME, 854 and 855 all seem to match the same HPLLC encoding even
though only some of the values are valid is some of the platforms.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_reg.h      | 11 ++++++++---
 drivers/gpu/drm/i915/intel_display.c | 15 ++++++++++++---
 2 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index a143127..29af9b3 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -37,12 +37,17 @@
 
 /* PCI config space */
 
-#define HPLLCC	0xc0 /* 855 only */
-#define   GC_CLOCK_CONTROL_MASK		(0xf << 0)
+#define HPLLCC	0xc0 /* 85x only */
+#define   GC_CLOCK_CONTROL_MASK		(0x7 << 0)
 #define   GC_CLOCK_133_200		(0 << 0)
 #define   GC_CLOCK_100_200		(1 << 0)
 #define   GC_CLOCK_100_133		(2 << 0)
-#define   GC_CLOCK_166_250		(3 << 0)
+#define   GC_CLOCK_133_266		(3 << 0)
+#define   GC_CLOCK_133_200_2		(4 << 0)
+#define   GC_CLOCK_133_266_2		(5 << 0)
+#define   GC_CLOCK_166_266		(6 << 0)
+#define   GC_CLOCK_166_250		(7 << 0)
+
 #define GCFGC2	0xda
 #define GCFGC	0xf0 /* 915+ only */
 #define   GC_LOW_FREQUENCY_ENABLE	(1 << 7)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e364ca7..8ed750e 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5661,20 +5661,29 @@ static int i865_get_display_clock_speed(struct drm_device *dev)
 	return 266667;
 }
 
-static int i855_get_display_clock_speed(struct drm_device *dev)
+static int i85x_get_display_clock_speed(struct drm_device *dev)
 {
 	u16 hpllcc = 0;
+
+	pci_bus_read_config_word(dev->pdev->bus,
+				 PCI_DEVFN(0, 3), HPLLCC, &hpllcc);
+
 	/* Assume that the hardware is in the high speed state.  This
 	 * should be the default.
 	 */
 	switch (hpllcc & GC_CLOCK_CONTROL_MASK) {
 	case GC_CLOCK_133_200:
+	case GC_CLOCK_133_200_2:
 	case GC_CLOCK_100_200:
 		return 200000;
 	case GC_CLOCK_166_250:
 		return 250000;
 	case GC_CLOCK_100_133:
 		return 133333;
+	case GC_CLOCK_133_266:
+	case GC_CLOCK_133_266_2:
+	case GC_CLOCK_166_266:
+		return 266667;
 	}
 
 	/* Shouldn't happen */
@@ -12726,8 +12735,8 @@ static void intel_init_display(struct drm_device *dev)
 			i865_get_display_clock_speed;
 	else if (IS_I85X(dev))
 		dev_priv->display.get_display_clock_speed =
-			i855_get_display_clock_speed;
-	else /* 852, 830 */
+			i85x_get_display_clock_speed;
+	else /* 830 */
 		dev_priv->display.get_display_clock_speed =
 			i830_get_display_clock_speed;
 
-- 
2.0.4

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

[PATCH 03/18] drm/i915: Fix 852GM/GMV cdclk

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

It seems 852GM/GMV uses a different HPLLCC encoding than the other
85x platforms. For 852GM/GMV cdclk is always 133MHz. Try to detect that
using the PCI revision (sinc the device ID seems useless for that). I'm
not at all sure this is a good idea, but according to the specs it
should work.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 8ed750e..8070b29 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5665,6 +5665,14 @@ static int i85x_get_display_clock_speed(struct drm_device *dev)
 {
 	u16 hpllcc = 0;
 
+	/*
+	 * 852GM/852GMV only supports 133 MHz and the HPLLCC
+	 * encoding is different :(
+	 * FIXME is this the right way to detect 852GM/852GMV?
+	 */
+	if (dev->pdev->revision == 0x1)
+		return 133333;
+
 	pci_bus_read_config_word(dev->pdev->bus,
 				 PCI_DEVFN(0, 3), HPLLCC, &hpllcc);
 
-- 
2.0.4

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

[PATCH 04/18] drm/i915: Add cdclk extraction for g33, 965gm and g4x

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Implement cdclk extraction for g33, 965gm and g4x platforms. The details
came from configdb. Sadly there isn't anything there for other gen3/gen4
chipsets.

So far I've tested this on one ELK where it gave me a HPLL VCO of 5333
MHz and cdclk of 444 MHz which seems perfectly sane for this machine.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_reg.h      |   3 +
 drivers/gpu/drm/i915/intel_display.c | 186 ++++++++++++++++++++++++++++++++++-
 2 files changed, 185 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 29af9b3..1b4d93a 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -2102,6 +2102,9 @@ enum punit_power_well {
 #define CLKCFG_MEM_800					(3 << 4)
 #define CLKCFG_MEM_MASK					(7 << 4)
 
+#define HPLLVCO				(MCHBAR_MIRROR_BASE + 0xc38)
+#define HPLLVCO_MOBILE			(MCHBAR_MIRROR_BASE + 0xc0f)
+
 #define TSC1			0x11001
 #define   TSE			(1<<0)
 #define TR1			0x11006
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 8070b29..2e7beeb 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5703,6 +5703,175 @@ static int i830_get_display_clock_speed(struct drm_device *dev)
 	return 133333;
 }
 
+static unsigned int intel_hpll_vco(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	static const unsigned int blb_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+		[4] = 6400000,
+	};
+	static const unsigned int pnv_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+		[4] = 2666667,
+	};
+	static const unsigned int cl_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 6400000,
+		[4] = 3333333,
+		[5] = 3566667,
+		[6] = 4266667,
+	};
+	static const unsigned int elk_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+	};
+	static const unsigned int ctg_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 6400000,
+		[4] = 2666667,
+		[5] = 4266667,
+	};
+	const unsigned int *vco_table;
+	unsigned int vco;
+	uint8_t tmp = 0;
+
+	/* FIXME other chipsets? */
+	if (IS_GM45(dev))
+		vco_table = ctg_vco;
+	else if (IS_G4X(dev))
+		vco_table = elk_vco;
+	else if (IS_CRESTLINE(dev))
+		vco_table = cl_vco;
+	else if (IS_PINEVIEW(dev))
+		vco_table = pnv_vco;
+	else if (IS_G33(dev))
+		vco_table = blb_vco;
+	else
+		return 0;
+
+	tmp = I915_READ(IS_MOBILE(dev) ? HPLLVCO_MOBILE : HPLLVCO);
+
+	vco = vco_table[tmp & 0x7];
+	if (vco == 0)
+		DRM_ERROR("Bad HPLL VCO (HPLLVCO=0x%02x)\n", tmp);
+	else
+		DRM_DEBUG_KMS("HPLL VCO %u kHz\n", vco);
+
+	return vco;
+}
+
+static int gm45_get_display_clock_speed(struct drm_device *dev)
+{
+	unsigned int cdclk_sel, vco = intel_hpll_vco(dev);
+	uint16_t tmp = 0;
+
+	pci_read_config_word(dev->pdev, GCFGC, &tmp);
+
+	cdclk_sel = (tmp >> 12) & 0x1;
+
+	switch (vco) {
+	case 2666667:
+	case 4000000:
+	case 5333333:
+		return cdclk_sel ? 333333 : 222222;
+	case 3200000:
+		return cdclk_sel ? 320000 : 228571;
+	default:
+		DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u, CFGC=0x%04x\n", vco, tmp);
+		return 222222;
+	}
+}
+
+static int i965gm_get_display_clock_speed(struct drm_device *dev)
+{
+	static const uint8_t div_3200[] = { 16, 10,  8 };
+	static const uint8_t div_4000[] = { 20, 12, 10 };
+	static const uint8_t div_5333[] = { 24, 16, 14 };
+	const uint8_t *div_table;
+	unsigned int cdclk_sel, vco = intel_hpll_vco(dev);
+	uint16_t tmp = 0;
+
+	pci_read_config_word(dev->pdev, GCFGC, &tmp);
+
+	cdclk_sel = ((tmp >> 8) & 0x1f) - 1;
+
+	if (cdclk_sel >= ARRAY_SIZE(div_3200))
+		goto fail;
+
+	switch (vco) {
+	case 3200000:
+		div_table = div_3200;
+		break;
+	case 4000000:
+		div_table = div_4000;
+		break;
+	case 5333333:
+		div_table = div_5333;
+		break;
+	default:
+		goto fail;
+	}
+
+	return DIV_ROUND_CLOSEST(vco, div_table[cdclk_sel]);
+
+ fail:
+	DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%04x\n", vco, tmp);
+	return 200000;
+}
+
+static int g33_get_display_clock_speed(struct drm_device *dev)
+{
+	static const uint8_t div_3200[] = { 12, 10,  8,  7, 5, 16 };
+	static const uint8_t div_4000[] = { 14, 12, 10,  8, 6, 20 };
+	static const uint8_t div_4800[] = { 20, 14, 12, 10, 8, 24 };
+	static const uint8_t div_5333[] = { 20, 16, 12, 12, 8, 28 };
+	const uint8_t *div_table;
+	unsigned int cdclk_sel, vco = intel_hpll_vco(dev);
+	uint16_t tmp = 0;
+
+	pci_read_config_word(dev->pdev, GCFGC, &tmp);
+
+	cdclk_sel = (tmp >> 4) & 0x7;
+
+	if (cdclk_sel >= ARRAY_SIZE(div_3200))
+		goto fail;
+
+	switch (vco) {
+	case 3200000:
+		div_table = div_3200;
+		break;
+	case 4000000:
+		div_table = div_4000;
+		break;
+	case 4800000:
+		div_table = div_4800;
+		break;
+	case 5333333:
+		div_table = div_5333;
+		break;
+	default:
+		goto fail;
+	}
+
+	return DIV_ROUND_CLOSEST(vco, div_table[cdclk_sel]);
+
+ fail:
+	DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%08x\n", vco, tmp);
+	return 190476;
+}
+
 static void
 intel_reduce_m_n_ratio(uint32_t *num, uint32_t *den)
 {
@@ -12723,7 +12892,19 @@ static void intel_init_display(struct drm_device *dev)
 	if (IS_VALLEYVIEW(dev))
 		dev_priv->display.get_display_clock_speed =
 			valleyview_get_display_clock_speed;
-	else if (IS_I945G(dev) || (IS_G33(dev) && !IS_PINEVIEW_M(dev)))
+	else if (IS_GM45(dev))
+		dev_priv->display.get_display_clock_speed =
+			gm45_get_display_clock_speed;
+	else if (IS_CRESTLINE(dev))
+		dev_priv->display.get_display_clock_speed =
+			i965gm_get_display_clock_speed;
+	else if (IS_PINEVIEW(dev))
+		dev_priv->display.get_display_clock_speed =
+			pnv_get_display_clock_speed;
+	else if (IS_G33(dev) || IS_G4X(dev))
+		dev_priv->display.get_display_clock_speed =
+			g33_get_display_clock_speed;
+	else if (IS_I945G(dev))
 		dev_priv->display.get_display_clock_speed =
 			i945_get_display_clock_speed;
 	else if (IS_I915G(dev))
@@ -12732,9 +12913,6 @@ static void intel_init_display(struct drm_device *dev)
 	else if (IS_I945GM(dev) || IS_845G(dev))
 		dev_priv->display.get_display_clock_speed =
 			i9xx_misc_get_display_clock_speed;
-	else if (IS_PINEVIEW(dev))
-		dev_priv->display.get_display_clock_speed =
-			pnv_get_display_clock_speed;
 	else if (IS_I915GM(dev))
 		dev_priv->display.get_display_clock_speed =
 			i915gm_get_display_clock_speed;
-- 
2.0.4

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

[PATCH 05/18] drm/i915: ILK cdclk seems to be 450MHz

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Based on the BIOS DP A AUX 2x clock divider the cdclk frequency
on ILK is 450Mhz. At least that holds on my ILK and it matches
how we program the divider.

Supposedly cdclk is 400MHz on SNB and IVB, again based on the AUX 2x
clock divider. Note that I don't have a SNB or IVB machine with
eDP so I couldn't verify what the BIOS used, so this notion is
purely based on our current code,

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 2e7beeb..f14868d 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5598,6 +5598,11 @@ static int valleyview_get_display_clock_speed(struct drm_device *dev)
 	return DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, divider + 1);
 }
 
+static int ilk_get_display_clock_speed(struct drm_device *dev)
+{
+	return 450000;
+}
+
 static int i945_get_display_clock_speed(struct drm_device *dev)
 {
 	return 400000;
@@ -12892,6 +12897,9 @@ static void intel_init_display(struct drm_device *dev)
 	if (IS_VALLEYVIEW(dev))
 		dev_priv->display.get_display_clock_speed =
 			valleyview_get_display_clock_speed;
+	else if (IS_GEN5(dev))
+		dev_priv->display.get_display_clock_speed =
+			ilk_get_display_clock_speed;
 	else if (IS_GM45(dev))
 		dev_priv->display.get_display_clock_speed =
 			gm45_get_display_clock_speed;
-- 
2.0.4

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

[PATCH 06/18] drm/i915: Assume 400 MHz cdclk for the rest of gen4-7

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

We don't currently have cdclk extraction code for 965g,snb,ivb.
Let's assumee 400 MHz until we know better. That seems to match hints
in various vague documents. Whether that's good enough is not
entirely clear.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index f14868d..b86b989 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -12912,7 +12912,8 @@ static void intel_init_display(struct drm_device *dev)
 	else if (IS_G33(dev) || IS_G4X(dev))
 		dev_priv->display.get_display_clock_speed =
 			g33_get_display_clock_speed;
-	else if (IS_I945G(dev))
+	else if (IS_I945G(dev) || IS_BROADWATER(dev) ||
+		 IS_GEN6(dev) || IS_IVYBRIDGE(dev))
 		dev_priv->display.get_display_clock_speed =
 			i945_get_display_clock_speed;
 	else if (IS_I915G(dev))
-- 
2.0.4

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

[PATCH 07/18] drm/i915: Simplify ilk_get_aux_clock_divider()

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Now that we are "extracting" the cdclk frequency on ILK-IVB we
can also simplify ilk_get_aux_clock_divider() to calculate the
divider based on cdclk instead of hardcoding the values.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_dp.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 45b53ff..f76e04f 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -698,15 +698,13 @@ static uint32_t ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	if (index)
 		return 0;
 
 	if (intel_dig_port->port == PORT_A) {
-		if (IS_GEN6(dev) || IS_GEN7(dev))
-			return 200; /* SNB & IVB eDP input clock at 400Mhz */
-		else
-			return 225; /* eDP input clock at 450Mhz */
+		return DIV_ROUND_UP(dev_priv->display.get_display_clock_speed(dev), 2000);
 	} else {
 		return DIV_ROUND_UP(intel_pch_rawclk(dev), 2);
 	}
-- 
2.0.4

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

[PATCH 08/18] drm/i915: Convert the ddi cdclk code to .get_display_clock_speed()

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Unify the HSW/BDW/SKL cdclk extraction code to conform to the same
.get_display_clock_speed() mold that all the other platforms
use.

v2: Update due to SKL code getting added

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_ddi.c        | 101 +-------------------------------
 drivers/gpu/drm/i915/intel_display.c    |  98 ++++++++++++++++++++++++++++++-
 drivers/gpu/drm/i915/intel_dp.c         |   2 +-
 drivers/gpu/drm/i915/intel_drv.h        |   1 -
 drivers/gpu/drm/i915/intel_pm.c         |   2 +-
 drivers/gpu/drm/i915/intel_runtime_pm.c |   2 +-
 6 files changed, 101 insertions(+), 105 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index ca33ee9..3b7e81d 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -1614,105 +1614,6 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	}
 }
 
-static int skl_get_cdclk_freq(struct drm_i915_private *dev_priv)
-{
-	uint32_t lcpll1 = I915_READ(LCPLL1_CTL);
-	uint32_t cdctl = I915_READ(CDCLK_CTL);
-	uint32_t linkrate;
-
-	if (!(lcpll1 & LCPLL_PLL_ENABLE)) {
-		WARN(1, "LCPLL1 not enabled\n");
-		return 24000; /* 24MHz is the cd freq with NSSC ref */
-	}
-
-	if ((cdctl & CDCLK_FREQ_SEL_MASK) == CDCLK_FREQ_540)
-		return 540000;
-
-	linkrate = (I915_READ(DPLL_CTRL1) &
-		    DPLL_CRTL1_LINK_RATE_MASK(SKL_DPLL0)) >> 1;
-
-	if (linkrate == DPLL_CRTL1_LINK_RATE_2160 ||
-	    linkrate == DPLL_CRTL1_LINK_RATE_1080) {
-		/* vco 8640 */
-		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
-		case CDCLK_FREQ_450_432:
-			return 432000;
-		case CDCLK_FREQ_337_308:
-			return 308570;
-		case CDCLK_FREQ_675_617:
-			return 617140;
-		default:
-			WARN(1, "Unknown cd freq selection\n");
-		}
-	} else {
-		/* vco 8100 */
-		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
-		case CDCLK_FREQ_450_432:
-			return 450000;
-		case CDCLK_FREQ_337_308:
-			return 337500;
-		case CDCLK_FREQ_675_617:
-			return 675000;
-		default:
-			WARN(1, "Unknown cd freq selection\n");
-		}
-	}
-
-	/* error case, do as if DPLL0 isn't enabled */
-	return 24000;
-}
-
-static int bdw_get_cdclk_freq(struct drm_i915_private *dev_priv)
-{
-	uint32_t lcpll = I915_READ(LCPLL_CTL);
-	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
-
-	if (lcpll & LCPLL_CD_SOURCE_FCLK)
-		return 800000;
-	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
-		return 450000;
-	else if (freq == LCPLL_CLK_FREQ_450)
-		return 450000;
-	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
-		return 540000;
-	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
-		return 337500;
-	else
-		return 675000;
-}
-
-static int hsw_get_cdclk_freq(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	uint32_t lcpll = I915_READ(LCPLL_CTL);
-	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
-
-	if (lcpll & LCPLL_CD_SOURCE_FCLK)
-		return 800000;
-	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
-		return 450000;
-	else if (freq == LCPLL_CLK_FREQ_450)
-		return 450000;
-	else if (IS_HSW_ULT(dev))
-		return 337500;
-	else
-		return 540000;
-}
-
-int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-
-	if (IS_SKYLAKE(dev))
-		return skl_get_cdclk_freq(dev_priv);
-
-	if (IS_BROADWELL(dev))
-		return bdw_get_cdclk_freq(dev_priv);
-
-	/* Haswell */
-	return hsw_get_cdclk_freq(dev_priv);
-}
-
 static void hsw_ddi_pll_enable(struct drm_i915_private *dev_priv,
 			       struct intel_shared_dpll *pll)
 {
@@ -1899,7 +1800,7 @@ void intel_ddi_pll_init(struct drm_device *dev)
 		hsw_shared_dplls_init(dev_priv);
 
 	DRM_DEBUG_KMS("CDCLK running at %dKHz\n",
-		      intel_ddi_get_cdclk_freq(dev_priv));
+		      dev_priv->display.get_display_clock_speed(dev));
 
 	if (IS_SKYLAKE(dev)) {
 		if (!(I915_READ(LCPLL1_CTL) & LCPLL_PLL_ENABLE))
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index b86b989..ea07418 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5572,6 +5572,93 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 	return 0;
 }
 
+static int skylake_get_display_clock_speed(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	uint32_t lcpll1 = I915_READ(LCPLL1_CTL);
+	uint32_t cdctl = I915_READ(CDCLK_CTL);
+	uint32_t linkrate;
+
+	if (!(lcpll1 & LCPLL_PLL_ENABLE)) {
+		WARN(1, "LCPLL1 not enabled\n");
+		return 24000; /* 24MHz is the cd freq with NSSC ref */
+	}
+
+	if ((cdctl & CDCLK_FREQ_SEL_MASK) == CDCLK_FREQ_540)
+		return 540000;
+
+	linkrate = (I915_READ(DPLL_CTRL1) &
+		    DPLL_CRTL1_LINK_RATE_MASK(SKL_DPLL0)) >> 1;
+
+	if (linkrate == DPLL_CRTL1_LINK_RATE_2160 ||
+	    linkrate == DPLL_CRTL1_LINK_RATE_1080) {
+		/* vco 8640 */
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			return 432000;
+		case CDCLK_FREQ_337_308:
+			return 308570;
+		case CDCLK_FREQ_675_617:
+			return 617140;
+		default:
+			WARN(1, "Unknown cd freq selection\n");
+		}
+	} else {
+		/* vco 8100 */
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			return 450000;
+		case CDCLK_FREQ_337_308:
+			return 337500;
+		case CDCLK_FREQ_675_617:
+			return 675000;
+		default:
+			WARN(1, "Unknown cd freq selection\n");
+		}
+	}
+
+	/* error case, do as if DPLL0 isn't enabled */
+	return 24000;
+}
+
+static int broadwell_get_display_clock_speed(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t lcpll = I915_READ(LCPLL_CTL);
+	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		return 800000;
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
+		return 540000;
+	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
+		return 337500;
+	else
+		return 675000;
+}
+
+static int haswell_get_display_clock_speed(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t lcpll = I915_READ(LCPLL_CTL);
+	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		return 800000;
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		return 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		return 450000;
+	else if (IS_HSW_ULT(dev))
+		return 337500;
+	else
+		return 540000;
+}
+
 static int valleyview_get_display_clock_speed(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -12894,7 +12981,16 @@ static void intel_init_display(struct drm_device *dev)
 	}
 
 	/* Returns the core display clock speed */
-	if (IS_VALLEYVIEW(dev))
+	if (IS_SKYLAKE(dev))
+		dev_priv->display.get_display_clock_speed =
+			skylake_get_display_clock_speed;
+	else if (IS_BROADWELL(dev))
+		dev_priv->display.get_display_clock_speed =
+			broadwell_get_display_clock_speed;
+	else if (IS_HASWELL(dev))
+		dev_priv->display.get_display_clock_speed =
+			haswell_get_display_clock_speed;
+	else if (IS_VALLEYVIEW(dev))
 		dev_priv->display.get_display_clock_speed =
 			valleyview_get_display_clock_speed;
 	else if (IS_GEN5(dev))
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index f76e04f..8479204 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -719,7 +719,7 @@ static uint32_t hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 	if (intel_dig_port->port == PORT_A) {
 		if (index)
 			return 0;
-		return DIV_ROUND_CLOSEST(intel_ddi_get_cdclk_freq(dev_priv), 2000);
+		return DIV_ROUND_CLOSEST(dev_priv->display.get_display_clock_speed(dev), 2000);
 	} else if (dev_priv->pch_id == INTEL_PCH_LPT_DEVICE_ID_TYPE) {
 		/* Workaround for non-ULT HSW */
 		switch (index) {
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index d67c59b..3ff59a9 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -817,7 +817,6 @@ void hsw_fdi_link_train(struct drm_crtc *crtc);
 void intel_ddi_init(struct drm_device *dev, enum port port);
 enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder);
 bool intel_ddi_get_hw_state(struct intel_encoder *encoder, enum pipe *pipe);
-int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv);
 void intel_ddi_pll_init(struct drm_device *dev);
 void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc);
 void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 7558ba2..56ca60c 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -2273,7 +2273,7 @@ hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc)
 	linetime = DIV_ROUND_CLOSEST(mode->crtc_htotal * 1000 * 8,
 				     mode->crtc_clock);
 	ips_linetime = DIV_ROUND_CLOSEST(mode->crtc_htotal * 1000 * 8,
-					 intel_ddi_get_cdclk_freq(dev_priv));
+					 dev_priv->display.get_display_clock_speed(dev));
 
 	return PIPE_WM_LINETIME_IPS_LINETIME(ips_linetime) |
 	       PIPE_WM_LINETIME_TIME(linetime);
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
index f5a78d5..d8841c7 100644
--- a/drivers/gpu/drm/i915/intel_runtime_pm.c
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -1401,6 +1401,6 @@ int i915_get_cdclk_freq(void)
 	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
 				power_domains);
 
-	return intel_ddi_get_cdclk_freq(dev_priv);
+	return dev_priv->display.get_display_clock_speed(dev_priv->dev);
 }
 EXPORT_SYMBOL_GPL(i915_get_cdclk_freq);
-- 
2.0.4

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

[PATCH 09/18] drm/i915: Warn when cdclk for the platforms is not known

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Print a warning if we fall through the .get_display_clock_speed() function
pointer setup. We end up assuming a 133MHz cdclk which should mean that
at least we avoid any 0 deivisions and whatnot. But this could at least
help remind people that they have to provide this function for new platforms.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index ea07418..e76dc56 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -13027,9 +13027,11 @@ static void intel_init_display(struct drm_device *dev)
 	else if (IS_I85X(dev))
 		dev_priv->display.get_display_clock_speed =
 			i85x_get_display_clock_speed;
-	else /* 830 */
+	else { /* 830 */
+		WARN(!IS_I830(dev), "Unknown platform. Assuming 133 MHz CDCLK\n");
 		dev_priv->display.get_display_clock_speed =
 			i830_get_display_clock_speed;
+	}
 
 	if (IS_GEN5(dev)) {
 		dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
-- 
2.0.4

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

[PATCH 10/18] drm/i915: Cache the current cdclk frequency in dev_priv

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Rather that extracting the current cdclk freuqncy every time someone
wants to know it, cache the current value and use that. VLV/CHV already
stored a cached value there so just expand that to cover all platforms.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h      |  2 +-
 drivers/gpu/drm/i915/intel_display.c | 44 ++++++++++++++++++++----------------
 2 files changed, 26 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index c4f2cb6..8b6061d 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1632,7 +1632,7 @@ struct drm_i915_private {
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
 
 	unsigned int fsb_freq, mem_freq, is_ddr3;
-	unsigned int vlv_cdclk_freq;
+	unsigned int cdclk_freq;
 	unsigned int hpll_freq;
 
 	/**
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index e76dc56..0a6c063 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4745,20 +4745,22 @@ static int valleyview_get_vco(struct drm_i915_private *dev_priv)
 	return vco_freq[hpll_freq] * 1000;
 }
 
-static void vlv_update_cdclk(struct drm_device *dev)
+static void intel_update_cdclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	dev_priv->vlv_cdclk_freq = dev_priv->display.get_display_clock_speed(dev);
+	dev_priv->cdclk_freq = dev_priv->display.get_display_clock_speed(dev);
 	DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz\n",
-			 dev_priv->vlv_cdclk_freq);
+			 dev_priv->cdclk_freq);
 
-	/*
-	 * Program the gmbus_freq based on the cdclk frequency.
-	 * BSpec erroneously claims we should aim for 4MHz, but
-	 * in fact 1MHz is the correct frequency.
-	 */
-	I915_WRITE(GMBUSFREQ_VLV, DIV_ROUND_UP(dev_priv->vlv_cdclk_freq, 1000));
+	if (IS_VALLEYVIEW(dev)) {
+		/*
+		 * Program the gmbus_freq based on the cdclk frequency.
+		 * BSpec erroneously claims we should aim for 4MHz, but
+		 * in fact 1MHz is the correct frequency.
+		 */
+		I915_WRITE(GMBUSFREQ_VLV, DIV_ROUND_UP(dev_priv->cdclk_freq, 1000));
+	}
 }
 
 /* Adjust CDclk dividers to allow high res or save power if possible */
@@ -4767,7 +4769,7 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val, cmd;
 
-	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->vlv_cdclk_freq);
+	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->cdclk_freq);
 
 	if (cdclk >= 320000) /* jump to highest voltage for 400MHz too */
 		cmd = 2;
@@ -4823,7 +4825,7 @@ static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
 	vlv_bunit_write(dev_priv, BUNIT_REG_BISOC, val);
 	mutex_unlock(&dev_priv->dpio_lock);
 
-	vlv_update_cdclk(dev);
+	intel_update_cdclk(dev);
 }
 
 static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
@@ -4831,7 +4833,7 @@ static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val, cmd;
 
-	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->vlv_cdclk_freq);
+	WARN_ON(dev_priv->display.get_display_clock_speed(dev) != dev_priv->cdclk_freq);
 
 	switch (cdclk) {
 	case 400000:
@@ -4864,7 +4866,7 @@ static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
 	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
-	vlv_update_cdclk(dev);
+	intel_update_cdclk(dev);
 }
 
 static int valleyview_calc_cdclk(struct drm_i915_private *dev_priv,
@@ -4922,8 +4924,7 @@ static void valleyview_modeset_global_pipes(struct drm_device *dev,
 	struct intel_crtc *intel_crtc;
 	int max_pixclk = intel_mode_max_pixclk(dev_priv);
 
-	if (valleyview_calc_cdclk(dev_priv, max_pixclk) ==
-	    dev_priv->vlv_cdclk_freq)
+	if (valleyview_calc_cdclk(dev_priv, max_pixclk) == dev_priv->cdclk_freq)
 		return;
 
 	/* disable/enable all currently active pipes while we change cdclk */
@@ -4938,7 +4939,7 @@ static void valleyview_modeset_global_resources(struct drm_device *dev)
 	int max_pixclk = intel_mode_max_pixclk(dev_priv);
 	int req_cdclk = valleyview_calc_cdclk(dev_priv, max_pixclk);
 
-	if (req_cdclk != dev_priv->vlv_cdclk_freq) {
+	if (req_cdclk != dev_priv->cdclk_freq) {
 		if (IS_CHERRYVIEW(dev))
 			cherryview_set_cdclk(dev, req_cdclk);
 		else
@@ -8151,6 +8152,8 @@ static void hsw_disable_lcpll(struct drm_i915_private *dev_priv,
 		I915_WRITE(LCPLL_CTL, val);
 		POSTING_READ(LCPLL_CTL);
 	}
+
+	intel_update_cdclk(dev_priv->dev);
 }
 
 /*
@@ -8217,6 +8220,8 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
 	if (--dev_priv->uncore.forcewake_count == 0)
 		dev_priv->uncore.funcs.force_wake_put(dev_priv, FORCEWAKE_ALL);
 	spin_unlock_irq(&dev_priv->uncore.lock);
+
+	intel_update_cdclk(dev_priv->dev);
 }
 
 /*
@@ -11931,6 +11936,8 @@ static void intel_shared_dpll_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	intel_update_cdclk(dev);
+
 	if (HAS_DDI(dev))
 		intel_ddi_pll_init(dev);
 	else if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
@@ -13270,10 +13277,9 @@ static void i915_disable_vga(struct drm_device *dev)
 
 void intel_modeset_init_hw(struct drm_device *dev)
 {
-	intel_prepare_ddi(dev);
+	intel_update_cdclk(dev);
 
-	if (IS_VALLEYVIEW(dev))
-		vlv_update_cdclk(dev);
+	intel_prepare_ddi(dev);
 
 	intel_init_clock_gating(dev);
 
-- 
2.0.4

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

[PATCH 11/18] drm/i915: Use cached cdclk value

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Rather than reading out the current cdclk value use the cached value we
have tucked away in dev_priv.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c    | 3 +--
 drivers/gpu/drm/i915/intel_dp.c         | 4 ++--
 drivers/gpu/drm/i915/intel_pm.c         | 2 +-
 drivers/gpu/drm/i915/intel_runtime_pm.c | 2 +-
 4 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0a6c063..9c6bc82 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5519,8 +5519,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 
 	/* FIXME should check pixel clock limits on all platforms */
 	if (INTEL_INFO(dev)->gen < 4) {
-		int clock_limit =
-			dev_priv->display.get_display_clock_speed(dev);
+		int clock_limit = dev_priv->cdclk_freq;
 
 		/*
 		 * Enable pixel doubling when the dot clock
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 8479204..e376633 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -704,7 +704,7 @@ static uint32_t ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 		return 0;
 
 	if (intel_dig_port->port == PORT_A) {
-		return DIV_ROUND_UP(dev_priv->display.get_display_clock_speed(dev), 2000);
+		return DIV_ROUND_UP(dev_priv->cdclk_freq, 2000);
 	} else {
 		return DIV_ROUND_UP(intel_pch_rawclk(dev), 2);
 	}
@@ -719,7 +719,7 @@ static uint32_t hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 	if (intel_dig_port->port == PORT_A) {
 		if (index)
 			return 0;
-		return DIV_ROUND_CLOSEST(dev_priv->display.get_display_clock_speed(dev), 2000);
+		return DIV_ROUND_CLOSEST(dev_priv->cdclk_freq, 2000);
 	} else if (dev_priv->pch_id == INTEL_PCH_LPT_DEVICE_ID_TYPE) {
 		/* Workaround for non-ULT HSW */
 		switch (index) {
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 56ca60c..58fdfb0 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -2273,7 +2273,7 @@ hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc)
 	linetime = DIV_ROUND_CLOSEST(mode->crtc_htotal * 1000 * 8,
 				     mode->crtc_clock);
 	ips_linetime = DIV_ROUND_CLOSEST(mode->crtc_htotal * 1000 * 8,
-					 dev_priv->display.get_display_clock_speed(dev));
+					 dev_priv->cdclk_freq);
 
 	return PIPE_WM_LINETIME_IPS_LINETIME(ips_linetime) |
 	       PIPE_WM_LINETIME_TIME(linetime);
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
index d8841c7..d23aa05 100644
--- a/drivers/gpu/drm/i915/intel_runtime_pm.c
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -1401,6 +1401,6 @@ int i915_get_cdclk_freq(void)
 	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
 				power_domains);
 
-	return dev_priv->display.get_display_clock_speed(dev_priv->dev);
+	return dev_priv->cdclk_freq;
 }
 EXPORT_SYMBOL_GPL(i915_get_cdclk_freq);
-- 
2.0.4

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

[PATCH 12/18] drm/i915: Unify ilk and hsw .get_aux_clock_divider()

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

ilk_get_aux_clock_divider() is now a subset of
hsw_get_aux_clock_divider() so unify them.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_dp.c | 22 +++-------------------
 1 file changed, 3 insertions(+), 19 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index e376633..95ec6a2 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -700,22 +700,6 @@ static uint32_t ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (index)
-		return 0;
-
-	if (intel_dig_port->port == PORT_A) {
-		return DIV_ROUND_UP(dev_priv->cdclk_freq, 2000);
-	} else {
-		return DIV_ROUND_UP(intel_pch_rawclk(dev), 2);
-	}
-}
-
-static uint32_t hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
-{
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct drm_device *dev = intel_dig_port->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
 	if (intel_dig_port->port == PORT_A) {
 		if (index)
 			return 0;
@@ -728,7 +712,9 @@ static uint32_t hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 		default: return 0;
 		}
 	} else  {
-		return index ? 0 : DIV_ROUND_UP(intel_pch_rawclk(dev), 2);
+		if (index)
+			return 0;
+		return DIV_ROUND_UP(intel_pch_rawclk(dev), 2);
 	}
 }
 
@@ -5346,8 +5332,6 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider;
 	else if (IS_VALLEYVIEW(dev))
 		intel_dp->get_aux_clock_divider = vlv_get_aux_clock_divider;
-	else if (IS_HASWELL(dev) || IS_BROADWELL(dev))
-		intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider;
 	else if (HAS_PCH_SPLIT(dev))
 		intel_dp->get_aux_clock_divider = ilk_get_aux_clock_divider;
 	else
-- 
2.0.4

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

[PATCH 13/18] drm/i915: Store max cdclk value in dev_priv

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Keep the cdclk maximum supported frequency around in dev_priv so that we
can verify certain things against it before actually changing the cdclk
frequency.

For now only VLV/CHV have support changing cdclk frequency, so other
plarforms get to assume cdclk is fixed.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h      |  2 +-
 drivers/gpu/drm/i915/intel_display.c | 20 +++++++++++++++++++-
 2 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 8b6061d..4aecabb 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1632,7 +1632,7 @@ struct drm_i915_private {
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
 
 	unsigned int fsb_freq, mem_freq, is_ddr3;
-	unsigned int cdclk_freq;
+	unsigned int cdclk_freq, max_cdclk_freq;
 	unsigned int hpll_freq;
 
 	/**
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 9c6bc82..5eeb456 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4745,6 +4745,21 @@ static int valleyview_get_vco(struct drm_i915_private *dev_priv)
 	return vco_freq[hpll_freq] * 1000;
 }
 
+static void intel_update_max_cdclk(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (IS_VALLEYVIEW(dev)) {
+		dev_priv->max_cdclk_freq = 400000;
+	} else {
+		/* otherwise assume cdclk is fixed */
+		dev_priv->max_cdclk_freq = dev_priv->cdclk_freq;
+	}
+
+	DRM_DEBUG_DRIVER("Max CD clock rate: %d kHz\n",
+			 dev_priv->max_cdclk_freq);
+}
+
 static void intel_update_cdclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -4761,6 +4776,9 @@ static void intel_update_cdclk(struct drm_device *dev)
 		 */
 		I915_WRITE(GMBUSFREQ_VLV, DIV_ROUND_UP(dev_priv->cdclk_freq, 1000));
 	}
+
+	if (dev_priv->max_cdclk_freq == 0)
+		intel_update_max_cdclk(dev);
 }
 
 /* Adjust CDclk dividers to allow high res or save power if possible */
@@ -5519,7 +5537,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 
 	/* FIXME should check pixel clock limits on all platforms */
 	if (INTEL_INFO(dev)->gen < 4) {
-		int clock_limit = dev_priv->cdclk_freq;
+		int clock_limit = dev_priv->max_cdclk_freq;
 
 		/*
 		 * Enable pixel doubling when the dot clock
-- 
2.0.4

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

[PATCH 14/18] drm/i915: Don't enable IPS when pixel rate exceeds 95% of cdclk

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Bspec says we shouldn't enable IPS on BDW when the pipe pixel rate
exceeds 95% of the core display clock. Apparently this can cause
underruns.

There's no similar restriction listed for HSW, so leave that one alone
for now.

v2: Add pipe_config_supports_ips() (Chris)
v3: Compare against the max cdclk insted of the current cdclk

Tested-by: Timo Aaltonen <tjaalton@ubuntu.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=83497
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 30 ++++++++++++++++++++++++++++--
 drivers/gpu/drm/i915/intel_drv.h     |  1 +
 drivers/gpu/drm/i915/intel_pm.c      | 16 +++++++---------
 3 files changed, 36 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 5eeb456..4e0ffc9 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -5520,12 +5520,38 @@ retry:
 	return setup_ok ? 0 : -EINVAL;
 }
 
+static bool pipe_config_supports_ips(struct drm_i915_private *dev_priv,
+				     struct intel_crtc_config *pipe_config)
+{
+	if (pipe_config->pipe_bpp > 24)
+		return false;
+
+	/* HSW can handle pixel rate up to cdclk? */
+	if (IS_HASWELL(dev_priv->dev))
+		return true;
+
+	/*
+	 * FIXME if we compare against max we should then
+	 * increase the cdclk frequency when the current
+	 * value is too low. The other option is to compare
+	 * against the cdclk frqeuncy we're going have post
+	 * modeset (ie. one we computed using other constraints).
+	 * Need to measure whether using a lower cdclk w/o IPS
+	 * is better or worse than a higher cdclk w/ IPS.
+	 */
+	return ilk_pipe_pixel_rate(pipe_config) <=
+		dev_priv->max_cdclk_freq * 95 / 100;
+}
+
 static void hsw_compute_ips_config(struct intel_crtc *crtc,
 				   struct intel_crtc_config *pipe_config)
 {
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
 	pipe_config->ips_enabled = i915.enable_ips &&
-				   hsw_crtc_supports_ips(crtc) &&
-				   pipe_config->pipe_bpp <= 24;
+		hsw_crtc_supports_ips(crtc) &&
+		pipe_config_supports_ips(dev_priv, pipe_config);
 }
 
 static int intel_crtc_compute_config(struct intel_crtc *crtc,
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 3ff59a9..1a7808f 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -1169,6 +1169,7 @@ void ilk_wm_get_hw_state(struct drm_device *dev);
 void skl_wm_get_hw_state(struct drm_device *dev);
 void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
 			  struct skl_ddb_allocation *ddb /* out */);
+uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_config *pipe_config);
 
 
 /* intel_sdvo.c */
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 58fdfb0..d9bfd5e 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -1892,23 +1892,21 @@ static void i845_update_wm(struct drm_crtc *unused_crtc)
 	I915_WRITE(FW_BLC, fwater_lo);
 }
 
-static uint32_t ilk_pipe_pixel_rate(struct drm_device *dev,
-				    struct drm_crtc *crtc)
+uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_config *pipe_config)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t pixel_rate;
 
-	pixel_rate = intel_crtc->config.adjusted_mode.crtc_clock;
+	pixel_rate = pipe_config->adjusted_mode.crtc_clock;
 
 	/* We only use IF-ID interlacing. If we ever use PF-ID we'll need to
 	 * adjust the pixel_rate here. */
 
-	if (intel_crtc->config.pch_pfit.enabled) {
+	if (pipe_config->pch_pfit.enabled) {
 		uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
-		uint32_t pfit_size = intel_crtc->config.pch_pfit.size;
+		uint32_t pfit_size = pipe_config->pch_pfit.size;
 
-		pipe_w = intel_crtc->config.pipe_src_w;
-		pipe_h = intel_crtc->config.pipe_src_h;
+		pipe_w = pipe_config->pipe_src_w;
+		pipe_h = pipe_config->pipe_src_h;
 		pfit_w = (pfit_size >> 16) & 0xFFFF;
 		pfit_h = pfit_size & 0xFFFF;
 		if (pipe_w < pfit_w)
@@ -2522,7 +2520,7 @@ static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
 
 	p->active = true;
 	p->pipe_htotal = intel_crtc->config.adjusted_mode.crtc_htotal;
-	p->pixel_rate = ilk_pipe_pixel_rate(dev, crtc);
+	p->pixel_rate = ilk_pipe_pixel_rate(&intel_crtc->config);
 	p->pri.bytes_per_pixel = crtc->primary->fb->bits_per_pixel / 8;
 	p->cur.bytes_per_pixel = 4;
 	p->pri.horiz_pixels = intel_crtc->config.pipe_src_w;
-- 
2.0.4

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

[PATCH 15/18] drm/i915: Fix chv cdclk support

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

The specs seem to be full of misinformation wrt. the Punit register
0x36. Some versions still show the old VLV bit layout, some the new
layout, and all of them seem to tell us nonsense about the cdclk
value encoding.

Testing on actual hardware has shown that we simply need to program
the desired CCK divider into the Punit register using the new layout of
the bits. Doing that, the status bit change to indicate the same value,
and the CCK 0x6b register also changes accordingly to indicate that CCK
is now using the new divider.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 22 +++++++---------------
 1 file changed, 7 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 4e0ffc9..709dc34 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4855,23 +4855,23 @@ static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
 
 	switch (cdclk) {
 	case 400000:
-		cmd = 3;
-		break;
 	case 333333:
 	case 320000:
-		cmd = 2;
-		break;
 	case 266667:
-		cmd = 1;
-		break;
 	case 200000:
-		cmd = 0;
 		break;
 	default:
 		WARN_ON(1);
 		return;
 	}
 
+	/*
+	 * Specs are full of misinformation, but testing on actual
+	 * hardware has shown that we just need to write the desired
+	 * CCK divider into the Punit register.
+	 */
+	cmd = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
+
 	mutex_lock(&dev_priv->rps.hw_lock);
 	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
 	val &= ~DSPFREQGUAR_MASK_CHV;
@@ -4892,10 +4892,6 @@ static int valleyview_calc_cdclk(struct drm_i915_private *dev_priv,
 {
 	int freq_320 = (dev_priv->hpll_freq <<  1) % 320000 != 0 ? 333333 : 320000;
 
-	/* FIXME: Punit isn't quite ready yet */
-	if (IS_CHERRYVIEW(dev_priv->dev))
-		return 400000;
-
 	/*
 	 * Really only a few cases to deal with, as only 4 CDclks are supported:
 	 *   200MHz
@@ -5709,10 +5705,6 @@ static int valleyview_get_display_clock_speed(struct drm_device *dev)
 	u32 val;
 	int divider;
 
-	/* FIXME: Punit isn't quite ready yet */
-	if (IS_CHERRYVIEW(dev))
-		return 400000;
-
 	if (dev_priv->hpll_freq == 0)
 		dev_priv->hpll_freq = valleyview_get_vco(dev_priv);
 
-- 
2.0.4

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

[PATCH 16/18] drm/i915: HSW cdclk change support

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Implement support for changing the cdclk frequency during runtime on
HSW. VLV/CHV already have support for this, so we can follow their
example for the most part. Only the actual hardware programming differs,
the rest is pretty much the same.

The pipe pixel rate stuff is handled a bit differently for now due to
the difference in pch vs. gmch pfit handling. Eventually we should unify
that part to eliminate what is essentially duplicated code.

v2: Grab rps.hw_lock around sandybridge_pcode_write()
v3: Rebase due to power well vs. .global_resources() reordering

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_reg.h      |   3 +
 drivers/gpu/drm/i915/intel_display.c | 138 ++++++++++++++++++++++++++++++++++-
 2 files changed, 138 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 1b4d93a..4349543 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -6045,6 +6045,7 @@ enum punit_power_well {
 #define   GEN6_PCODE_WRITE_D_COMP		0x11
 #define   GEN6_ENCODE_RC6_VID(mv)		(((mv) - 245) / 5)
 #define   GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) + 245)
+#define   HSW_PCODE_DE_WRITE_FREQ_REQ		0x17
 #define   DISPLAY_IPS_CONTROL			0x19
 #define GEN6_PCODE_DATA				0x138128
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
@@ -6454,10 +6455,12 @@ enum punit_power_well {
 #define  LCPLL_PLL_LOCK			(1<<30)
 #define  LCPLL_CLK_FREQ_MASK		(3<<26)
 #define  LCPLL_CLK_FREQ_450		(0<<26)
+#define  LCPLL_CLK_FREQ_ALT_HSW		(1<<26) /* 337.5 (ULX) or 540 */
 #define  LCPLL_CLK_FREQ_54O_BDW		(1<<26)
 #define  LCPLL_CLK_FREQ_337_5_BDW	(2<<26)
 #define  LCPLL_CLK_FREQ_675_BDW		(3<<26)
 #define  LCPLL_CD_CLOCK_DISABLE		(1<<25)
+#define  LCPLL_ROOT_CD_CLOCK_DISABLE	(1<<24)
 #define  LCPLL_CD2X_CLOCK_DISABLE	(1<<23)
 #define  LCPLL_POWER_DOWN_ALLOW		(1<<22)
 #define  LCPLL_CD_SOURCE_FCLK		(1<<21)
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 709dc34..b0bda85 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4749,7 +4749,16 @@ static void intel_update_max_cdclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_VALLEYVIEW(dev)) {
+	if (IS_HASWELL(dev)) {
+		if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+			dev_priv->max_cdclk_freq = 450000;
+		else if (IS_HSW_ULX(dev))
+			dev_priv->max_cdclk_freq = 337500;
+		else if (IS_HSW_ULT(dev))
+			dev_priv->max_cdclk_freq = 450000;
+		else
+			dev_priv->max_cdclk_freq = 540000;
+	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->max_cdclk_freq = 400000;
 	} else {
 		/* otherwise assume cdclk is fixed */
@@ -8318,6 +8327,123 @@ void hsw_disable_pc8(struct drm_i915_private *dev_priv)
 	intel_prepare_ddi(dev);
 }
 
+/* compute the max rate for new configuration */
+static int ilk_max_pixel_rate(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_crtc *crtc;
+	int max_pixel_rate = 0;
+
+	for_each_intel_crtc(dev, crtc) {
+		if (crtc->new_enabled)
+			max_pixel_rate = max_t(int, max_pixel_rate,
+					       ilk_pipe_pixel_rate(crtc->new_config));
+	}
+
+	return max_pixel_rate;
+}
+
+static int haswell_calc_cdclk(struct drm_i915_private *dev_priv,
+			      int max_pixel_rate)
+{
+	int cdclk;
+
+	/*
+	 * FIXME should also account for plane ratio
+	 * once 64bpp pixel formats are supported.
+	 */
+	if (max_pixel_rate > 450000)
+		cdclk = 540000;
+	else if (max_pixel_rate > 337500 || !IS_HSW_ULX(dev_priv))
+		cdclk = 450000;
+	else
+		cdclk = 337500;
+
+	/*
+	 * FIXME move the cdclk caclulation to
+	 * compute_config() so we can fail gracegully.
+	 */
+	if (cdclk > dev_priv->max_cdclk_freq) {
+		DRM_ERROR("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
+			  cdclk, dev_priv->max_cdclk_freq);
+		cdclk = dev_priv->max_cdclk_freq;
+	}
+
+	return cdclk;
+}
+
+static void haswell_set_cdclk(struct drm_device *dev, int cdclk)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t val;
+
+	if (WARN((I915_READ(LCPLL_CTL) &
+		  (LCPLL_PLL_DISABLE | LCPLL_PLL_LOCK |
+		   LCPLL_CD_CLOCK_DISABLE | LCPLL_ROOT_CD_CLOCK_DISABLE |
+		   LCPLL_CD2X_CLOCK_DISABLE | LCPLL_POWER_DOWN_ALLOW |
+		   LCPLL_CD_SOURCE_FCLK)) != LCPLL_PLL_LOCK,
+		 "trying to change cdclk frequency with cdclk not enabled\n"))
+		return;
+
+	val = I915_READ(LCPLL_CTL);
+	val &= ~LCPLL_CLK_FREQ_MASK;
+
+	switch (cdclk) {
+	case 450000:
+		val |= LCPLL_CLK_FREQ_450;
+		break;
+	case 337500:
+	case 540000:
+		val |= LCPLL_CLK_FREQ_ALT_HSW;
+		break;
+	default:
+		WARN(1, "invalid cdclk frequency\n");
+		return;
+	}
+
+	I915_WRITE(LCPLL_CTL, val);
+
+	if (IS_HSW_ULX(dev)) {
+		mutex_lock(&dev_priv->rps.hw_lock);
+		sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ,
+					cdclk == 337500);
+		mutex_unlock(&dev_priv->rps.hw_lock);
+	}
+
+	intel_update_cdclk(dev);
+
+	WARN(cdclk != dev_priv->cdclk_freq,
+	     "cdclk requested %d kHz but got %d kHz\n",
+	     cdclk, dev_priv->cdclk_freq);
+}
+
+static void haswell_modeset_global_pipes(struct drm_device *dev,
+					 unsigned *prepare_pipes)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	int max_pixel_rate = ilk_max_pixel_rate(dev_priv);
+
+	if (haswell_calc_cdclk(dev_priv, max_pixel_rate) == dev_priv->cdclk_freq)
+		return;
+
+	/* disable/enable all currently active pipes while we change cdclk */
+	for_each_intel_crtc(dev, crtc)
+		if (crtc->base.enabled)
+			*prepare_pipes |= 1 << crtc->pipe;
+}
+
+static void haswell_modeset_global_resources(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int max_pixel_rate = ilk_max_pixel_rate(dev_priv);
+	int req_cdclk = haswell_calc_cdclk(dev_priv, max_pixel_rate);
+
+	if (req_cdclk != dev_priv->cdclk_freq) {
+		haswell_set_cdclk(dev, req_cdclk);
+	}
+}
+
 static int haswell_crtc_compute_clock(struct intel_crtc *crtc)
 {
 	if (!intel_ddi_pll_select(crtc))
@@ -11272,8 +11398,11 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 	 * mode set on this crtc.  For other crtcs we need to use the
 	 * adjusted_mode bits in the crtc directly.
 	 */
-	if (IS_VALLEYVIEW(dev)) {
-		valleyview_modeset_global_pipes(dev, &prepare_pipes);
+	if (IS_VALLEYVIEW(dev) || IS_HASWELL(dev)) {
+		if (IS_VALLEYVIEW(dev))
+			valleyview_modeset_global_pipes(dev, &prepare_pipes);
+		else
+			haswell_modeset_global_pipes(dev, &prepare_pipes);
 
 		/* may have added more to prepare_pipes than we should */
 		prepare_pipes &= ~disable_pipes;
@@ -13086,6 +13215,9 @@ static void intel_init_display(struct drm_device *dev)
 			ivb_modeset_global_resources;
 	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		dev_priv->display.fdi_link_train = hsw_fdi_link_train;
+		if (IS_HASWELL(dev))
+			dev_priv->display.modeset_global_resources =
+				haswell_modeset_global_resources;
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.modeset_global_resources =
 			valleyview_modeset_global_resources;
-- 
2.0.4

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

[PATCH 17/18] drm/i915: Add IS_BDW_ULX()

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

We need to tell BDW ULT and ULX apart.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 4aecabb..bf3f33d 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -2194,6 +2194,9 @@ struct drm_i915_cmd_table {
 				 ((INTEL_DEVID(dev) & 0xf) == 0x2  || \
 				 (INTEL_DEVID(dev) & 0xf) == 0x6 || \
 				 (INTEL_DEVID(dev) & 0xf) == 0xe))
+/* ULX machines are also considered ULT. */
+#define IS_BDW_ULX(dev)		(IS_BROADWELL(dev) && \
+				 (INTEL_DEVID(dev) & 0xf) == 0xe)
 #define IS_BDW_GT3(dev)		(IS_BROADWELL(dev) && \
 				 (INTEL_DEVID(dev) & 0x00F0) == 0x0020)
 #define IS_HSW_ULT(dev)		(IS_HASWELL(dev) && \
-- 
2.0.4

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

[PATCH 18/18] drm/i915: BDW cdclk change support

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Add support for changing cdclk frequency during runtime on BDW. The
procedure is quite a bit different on BDW from the one on HSW, so
add a separate function for it.

Also with IPS enabled the actual pixel rate mustn't exceed 95% of cdclk,
so take that into account when computing the max pixel rate.

v2: Grab rps.hw_lock around sandybridge_pcode_write()
v3: Rebase due to power well vs. .global_resources() reordering

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_reg.h      |   1 +
 drivers/gpu/drm/i915/intel_display.c | 136 ++++++++++++++++++++++++++++++-----
 2 files changed, 121 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 4349543..6a394ba 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -6046,6 +6046,7 @@ enum punit_power_well {
 #define   GEN6_ENCODE_RC6_VID(mv)		(((mv) - 245) / 5)
 #define   GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) + 245)
 #define   HSW_PCODE_DE_WRITE_FREQ_REQ		0x17
+#define   BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ	0x18
 #define   DISPLAY_IPS_CONTROL			0x19
 #define GEN6_PCODE_DATA				0x138128
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index b0bda85..088e29a 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -4749,7 +4749,22 @@ static void intel_update_max_cdclk(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev)) {
+	if (IS_BROADWELL(dev))  {
+		/*
+		 * FIXME with extra cooling we can allow
+		 * 540 MHz for ULX and 675 Mhz for ULT.
+		 * How can we know if extra cooling is
+		 * available? PCI ID, VTB, something else?
+		 */
+		if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+			dev_priv->max_cdclk_freq = 450000;
+		else if (IS_BDW_ULX(dev))
+			dev_priv->max_cdclk_freq = 450000;
+		else if (IS_BDW_ULT(dev))
+			dev_priv->max_cdclk_freq = 540000;
+		else
+			dev_priv->max_cdclk_freq = 675000;
+	} else if (IS_HASWELL(dev)) {
 		if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
 			dev_priv->max_cdclk_freq = 450000;
 		else if (IS_HSW_ULX(dev))
@@ -5536,12 +5551,11 @@ static bool pipe_config_supports_ips(struct drm_i915_private *dev_priv,
 		return true;
 
 	/*
-	 * FIXME if we compare against max we should then
-	 * increase the cdclk frequency when the current
-	 * value is too low. The other option is to compare
-	 * against the cdclk frqeuncy we're going have post
-	 * modeset (ie. one we computed using other constraints).
-	 * Need to measure whether using a lower cdclk w/o IPS
+	 * We compare against max which means we must take
+	 * the increased cdclk requirement into account when
+	 * claculating the new cdclk.
+	 *
+	 * Should measure whether using a lower cdclk w/o IPS
 	 * is better or worse than a higher cdclk w/ IPS.
 	 */
 	return ilk_pipe_pixel_rate(pipe_config) <=
@@ -8335,9 +8349,18 @@ static int ilk_max_pixel_rate(struct drm_i915_private *dev_priv)
 	int max_pixel_rate = 0;
 
 	for_each_intel_crtc(dev, crtc) {
-		if (crtc->new_enabled)
-			max_pixel_rate = max_t(int, max_pixel_rate,
-					       ilk_pipe_pixel_rate(crtc->new_config));
+		int pixel_rate;
+
+		if (!crtc->new_enabled)
+			continue;
+
+		pixel_rate = ilk_pipe_pixel_rate(crtc->new_config);
+
+		/* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */
+		if (IS_BROADWELL(dev) && crtc->new_config->ips_enabled)
+			pixel_rate = DIV_ROUND_UP(pixel_rate * 100, 95);
+
+		max_pixel_rate = max(max_pixel_rate, pixel_rate);
 	}
 
 	return max_pixel_rate;
@@ -8352,7 +8375,9 @@ static int haswell_calc_cdclk(struct drm_i915_private *dev_priv,
 	 * FIXME should also account for plane ratio
 	 * once 64bpp pixel formats are supported.
 	 */
-	if (max_pixel_rate > 450000)
+	if (max_pixel_rate > 540000)
+		cdclk = 675000;
+	else if (max_pixel_rate > 450000)
 		cdclk = 540000;
 	else if (max_pixel_rate > 337500 || !IS_HSW_ULX(dev_priv))
 		cdclk = 450000;
@@ -8417,6 +8442,83 @@ static void haswell_set_cdclk(struct drm_device *dev, int cdclk)
 	     cdclk, dev_priv->cdclk_freq);
 }
 
+static void broadwell_set_cdclk(struct drm_device *dev, int cdclk)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t val, data;
+	int ret;
+
+	if (WARN((I915_READ(LCPLL_CTL) &
+		  (LCPLL_PLL_DISABLE | LCPLL_PLL_LOCK |
+		   LCPLL_CD_CLOCK_DISABLE | LCPLL_ROOT_CD_CLOCK_DISABLE |
+		   LCPLL_CD2X_CLOCK_DISABLE | LCPLL_POWER_DOWN_ALLOW |
+		   LCPLL_CD_SOURCE_FCLK)) != LCPLL_PLL_LOCK,
+		 "trying to change cdclk frequency with cdclk not enabled\n"))
+		return;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	ret = sandybridge_pcode_write(dev_priv,
+				      BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ, 0x0);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+	if (ret) {
+		DRM_ERROR("failed to inform pcode about cdclk change\n");
+		return;
+	}
+
+	val = I915_READ(LCPLL_CTL);
+	val |= LCPLL_CD_SOURCE_FCLK;
+	I915_WRITE(LCPLL_CTL, val);
+
+	if (wait_for_atomic_us(I915_READ(LCPLL_CTL) &
+			       LCPLL_CD_SOURCE_FCLK_DONE, 1))
+		DRM_ERROR("Switching to FCLK failed\n");
+
+	val = I915_READ(LCPLL_CTL);
+	val &= ~LCPLL_CLK_FREQ_MASK;
+
+	switch (cdclk) {
+	case 450000:
+		val |= LCPLL_CLK_FREQ_450;
+		data = 0;
+		break;
+	case 540000:
+		val |= LCPLL_CLK_FREQ_54O_BDW;
+		data = 1;
+		break;
+	case 337500:
+		val |= LCPLL_CLK_FREQ_337_5_BDW;
+		data = 2;
+		break;
+	case 675000:
+		val |= LCPLL_CLK_FREQ_675_BDW;
+		data = 3;
+		break;
+	default:
+		WARN(1, "invalid cdclk frequency\n");
+		return;
+	}
+
+	I915_WRITE(LCPLL_CTL, val);
+
+	val = I915_READ(LCPLL_CTL);
+	val &= ~LCPLL_CD_SOURCE_FCLK;
+	I915_WRITE(LCPLL_CTL, val);
+
+	if (wait_for_atomic_us((I915_READ(LCPLL_CTL) &
+				LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
+		DRM_ERROR("Switching back to LCPLL failed\n");
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ, data);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	intel_update_cdclk(dev);
+
+	WARN(cdclk != dev_priv->cdclk_freq,
+	     "cdclk requested %d kHz but got %d kHz\n",
+	     cdclk, dev_priv->cdclk_freq);
+}
+
 static void haswell_modeset_global_pipes(struct drm_device *dev,
 					 unsigned *prepare_pipes)
 {
@@ -8440,7 +8542,10 @@ static void haswell_modeset_global_resources(struct drm_device *dev)
 	int req_cdclk = haswell_calc_cdclk(dev_priv, max_pixel_rate);
 
 	if (req_cdclk != dev_priv->cdclk_freq) {
-		haswell_set_cdclk(dev, req_cdclk);
+		if (IS_BROADWELL(dev))
+			broadwell_set_cdclk(dev, req_cdclk);
+		else
+			haswell_set_cdclk(dev, req_cdclk);
 	}
 }
 
@@ -11398,7 +11503,7 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 	 * mode set on this crtc.  For other crtcs we need to use the
 	 * adjusted_mode bits in the crtc directly.
 	 */
-	if (IS_VALLEYVIEW(dev) || IS_HASWELL(dev)) {
+	if (IS_VALLEYVIEW(dev) || IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		if (IS_VALLEYVIEW(dev))
 			valleyview_modeset_global_pipes(dev, &prepare_pipes);
 		else
@@ -13215,9 +13320,8 @@ static void intel_init_display(struct drm_device *dev)
 			ivb_modeset_global_resources;
 	} else if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
 		dev_priv->display.fdi_link_train = hsw_fdi_link_train;
-		if (IS_HASWELL(dev))
-			dev_priv->display.modeset_global_resources =
-				haswell_modeset_global_resources;
+		dev_priv->display.modeset_global_resources =
+			haswell_modeset_global_resources;
 	} else if (IS_VALLEYVIEW(dev)) {
 		dev_priv->display.modeset_global_resources =
 			valleyview_modeset_global_resources;
-- 
2.0.4

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

Re: [PATCH 11/18] drm/i915: Use cached cdclk value

[Daniel Vetter]

On Mon, Nov 17, 2014 at 04:43:45PM +0200, ville.syrjala@linux.intel.com wrote:
> diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> index d8841c7..d23aa05 100644
> --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> @@ -1401,6 +1401,6 @@ int i915_get_cdclk_freq(void)
>  	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
>  				power_domains);
>  
> -	return dev_priv->display.get_display_clock_speed(dev_priv->dev);
> +	return dev_priv->cdclk_freq;

Maybe this here is the reason why hsw shouldn't change cdclck - the audio
side just falls over?
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 13/18] drm/i915: Store max cdclk value in dev_priv

[Daniel Vetter]

On Mon, Nov 17, 2014 at 04:43:47PM +0200, ville.syrjala@linux.intel.com wrote:
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index 9c6bc82..5eeb456 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -4745,6 +4745,21 @@ static int valleyview_get_vco(struct drm_i915_private *dev_priv)
>  	return vco_freq[hpll_freq] * 1000;
>  }
>  
> +static void intel_update_max_cdclk(struct drm_device *dev)
> +{
> +	struct drm_i915_private *dev_priv = dev->dev_private;
> +
> +	if (IS_VALLEYVIEW(dev)) {
> +		dev_priv->max_cdclk_freq = 400000;

I've thought the 400MHz mode is busted? Or is that just Punit bonghits on
SDVs and pre-prod boards?
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 01/18] drm/i915: Return more precise cdclk for gen2/3

[Daniel Vetter]

On Mon, Nov 17, 2014 at 04:43:35PM +0200, ville.syrjala@linux.intel.com wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> Fill out the lower three digits for gen2 and gen3 cdclk frqeuncy. It's
> not clear if these are accurate frquencies or just in the ballpark, but
> without docs this is the best we can do.
> 
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Given that no one complained yet I'm not sure this is worth the trouble.
Otoh it's all below the 10% margin we have already anyway, so one big
wash ;-)
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 06/18] drm/i915: Assume 400 MHz cdclk for the rest of gen4-7

[Daniel Vetter]

On Mon, Nov 17, 2014 at 04:43:40PM +0200, ville.syrjala@linux.intel.com wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> We don't currently have cdclk extraction code for 965g,snb,ivb.
> Let's assumee 400 MHz until we know better. That seems to match hints
> in various vague documents. Whether that's good enough is not
> entirely clear.
> 
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Hm, not sure whether these chips even had outputs which could drive this
high really. And if we now start rejecting modes that previously worked
(really unlikely imo) we'll get the regression reports and can fudge the
numbers some more. So even without more spec hints I'm totally fine with
going forward with this.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 14/18] drm/i915: Don't enable IPS when pixel rate exceeds 95% of cdclk

[Daniel Vetter]

On Mon, Nov 17, 2014 at 3:43 PM,  <ville.syrjala@linux.intel.com> wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
>
> Bspec says we shouldn't enable IPS on BDW when the pipe pixel rate
> exceeds 95% of the core display clock. Apparently this can cause
> underruns.
>
> There's no similar restriction listed for HSW, so leave that one alone
> for now.
>
> v2: Add pipe_config_supports_ips() (Chris)
> v3: Compare against the max cdclk insted of the current cdclk
>
> Tested-by: Timo Aaltonen <tjaalton@ubuntu.com>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=83497
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> ---
>  drivers/gpu/drm/i915/intel_display.c | 30 ++++++++++++++++++++++++++++--
>  drivers/gpu/drm/i915/intel_drv.h     |  1 +
>  drivers/gpu/drm/i915/intel_pm.c      | 16 +++++++---------
>  3 files changed, 36 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index 5eeb456..4e0ffc9 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -5520,12 +5520,38 @@ retry:
>         return setup_ok ? 0 : -EINVAL;
>  }
>
> +static bool pipe_config_supports_ips(struct drm_i915_private *dev_priv,
> +                                    struct intel_crtc_config *pipe_config)
> +{
> +       if (pipe_config->pipe_bpp > 24)
> +               return false;
> +
> +       /* HSW can handle pixel rate up to cdclk? */
> +       if (IS_HASWELL(dev_priv->dev))
> +               return true;
> +
> +       /*
> +        * FIXME if we compare against max we should then
> +        * increase the cdclk frequency when the current
> +        * value is too low. The other option is to compare
> +        * against the cdclk frqeuncy we're going have post
> +        * modeset (ie. one we computed using other constraints).
> +        * Need to measure whether using a lower cdclk w/o IPS
> +        * is better or worse than a higher cdclk w/ IPS.
> +        */
> +       return ilk_pipe_pixel_rate(pipe_config) <=
> +               dev_priv->max_cdclk_freq * 95 / 100;
> +}

Adding Art for insight into this question (lower cdclock or ips) or
whether this just isn't worth the complexity.
-Daniel

> +
>  static void hsw_compute_ips_config(struct intel_crtc *crtc,
>                                    struct intel_crtc_config *pipe_config)
>  {
> +       struct drm_device *dev = crtc->base.dev;
> +       struct drm_i915_private *dev_priv = dev->dev_private;
> +
>         pipe_config->ips_enabled = i915.enable_ips &&
> -                                  hsw_crtc_supports_ips(crtc) &&
> -                                  pipe_config->pipe_bpp <= 24;
> +               hsw_crtc_supports_ips(crtc) &&
> +               pipe_config_supports_ips(dev_priv, pipe_config);
>  }
>
>  static int intel_crtc_compute_config(struct intel_crtc *crtc,
> diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
> index 3ff59a9..1a7808f 100644
> --- a/drivers/gpu/drm/i915/intel_drv.h
> +++ b/drivers/gpu/drm/i915/intel_drv.h
> @@ -1169,6 +1169,7 @@ void ilk_wm_get_hw_state(struct drm_device *dev);
>  void skl_wm_get_hw_state(struct drm_device *dev);
>  void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
>                           struct skl_ddb_allocation *ddb /* out */);
> +uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_config *pipe_config);
>
>
>  /* intel_sdvo.c */
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 58fdfb0..d9bfd5e 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -1892,23 +1892,21 @@ static void i845_update_wm(struct drm_crtc *unused_crtc)
>         I915_WRITE(FW_BLC, fwater_lo);
>  }
>
> -static uint32_t ilk_pipe_pixel_rate(struct drm_device *dev,
> -                                   struct drm_crtc *crtc)
> +uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_config *pipe_config)
>  {
> -       struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
>         uint32_t pixel_rate;
>
> -       pixel_rate = intel_crtc->config.adjusted_mode.crtc_clock;
> +       pixel_rate = pipe_config->adjusted_mode.crtc_clock;
>
>         /* We only use IF-ID interlacing. If we ever use PF-ID we'll need to
>          * adjust the pixel_rate here. */
>
> -       if (intel_crtc->config.pch_pfit.enabled) {
> +       if (pipe_config->pch_pfit.enabled) {
>                 uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
> -               uint32_t pfit_size = intel_crtc->config.pch_pfit.size;
> +               uint32_t pfit_size = pipe_config->pch_pfit.size;
>
> -               pipe_w = intel_crtc->config.pipe_src_w;
> -               pipe_h = intel_crtc->config.pipe_src_h;
> +               pipe_w = pipe_config->pipe_src_w;
> +               pipe_h = pipe_config->pipe_src_h;
>                 pfit_w = (pfit_size >> 16) & 0xFFFF;
>                 pfit_h = pfit_size & 0xFFFF;
>                 if (pipe_w < pfit_w)
> @@ -2522,7 +2520,7 @@ static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
>
>         p->active = true;
>         p->pipe_htotal = intel_crtc->config.adjusted_mode.crtc_htotal;
> -       p->pixel_rate = ilk_pipe_pixel_rate(dev, crtc);
> +       p->pixel_rate = ilk_pipe_pixel_rate(&intel_crtc->config);
>         p->pri.bytes_per_pixel = crtc->primary->fb->bits_per_pixel / 8;
>         p->cur.bytes_per_pixel = 4;
>         p->pri.horiz_pixels = intel_crtc->config.pipe_src_w;
> --
> 2.0.4
>
> _______________________________________________
> 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
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 01/18] drm/i915: Return more precise cdclk for gen2/3

[Ville Syrjälä]

On Mon, Nov 17, 2014 at 07:44:30PM +0100, Daniel Vetter wrote:
> On Mon, Nov 17, 2014 at 04:43:35PM +0200, ville.syrjala@linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > 
> > Fill out the lower three digits for gen2 and gen3 cdclk frqeuncy. It's
> > not clear if these are accurate frquencies or just in the ballpark, but
> > without docs this is the best we can do.
> > 
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> Given that no one complained yet I'm not sure this is worth the trouble.
> Otoh it's all below the 10% margin we have already anyway, so one big
> wash ;-)

Yeah <1Mhz is a fairly small error here. But I still prefer to make the
change, if only for consistency. Otherwise it'll keep bugging me and
I'll have to keep fighting the urge to change it every time I see those
numbers.

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 11/18] drm/i915: Use cached cdclk value

[Ville Syrjälä]

On Mon, Nov 17, 2014 at 07:41:44PM +0100, Daniel Vetter wrote:
> On Mon, Nov 17, 2014 at 04:43:45PM +0200, ville.syrjala@linux.intel.com wrote:
> > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > index d8841c7..d23aa05 100644
> > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > @@ -1401,6 +1401,6 @@ int i915_get_cdclk_freq(void)
> >  	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
> >  				power_domains);
> >  
> > -	return dev_priv->display.get_display_clock_speed(dev_priv->dev);
> > +	return dev_priv->cdclk_freq;
> 
> Maybe this here is the reason why hsw shouldn't change cdclck - the audio
> side just falls over?

I must admit that I didn't check whether the audio side will query the
cdclk again after every mode change. It really should, but often reality
doesn't match my expectations :)

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 11/18] drm/i915: Use cached cdclk value

[Daniel Vetter]

On Mon, Nov 17, 2014 at 09:06:53PM +0200, Ville Syrjälä wrote:
> On Mon, Nov 17, 2014 at 07:41:44PM +0100, Daniel Vetter wrote:
> > On Mon, Nov 17, 2014 at 04:43:45PM +0200, ville.syrjala@linux.intel.com wrote:
> > > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > > index d8841c7..d23aa05 100644
> > > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > > @@ -1401,6 +1401,6 @@ int i915_get_cdclk_freq(void)
> > >  	dev_priv = container_of(hsw_pwr, struct drm_i915_private,
> > >  				power_domains);
> > >  
> > > -	return dev_priv->display.get_display_clock_speed(dev_priv->dev);
> > > +	return dev_priv->cdclk_freq;
> > 
> > Maybe this here is the reason why hsw shouldn't change cdclck - the audio
> > side just falls over?
> 
> I must admit that I didn't check whether the audio side will query the
> cdclk again after every mode change. It really should, but often reality
> doesn't match my expectations :)

Hm right, we only change cdclk on modeset changes, and those will always
generate an unsolicited event irq. So if we break stuff here that's indeed
just a bug in hda-intel and not something we can't fix.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 01/18] drm/i915: Return more precise cdclk for gen2/3

[Daniel Vetter]

On Mon, Nov 17, 2014 at 09:02:11PM +0200, Ville Syrjälä wrote:
> On Mon, Nov 17, 2014 at 07:44:30PM +0100, Daniel Vetter wrote:
> > On Mon, Nov 17, 2014 at 04:43:35PM +0200, ville.syrjala@linux.intel.com wrote:
> > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > 
> > > Fill out the lower three digits for gen2 and gen3 cdclk frqeuncy. It's
> > > not clear if these are accurate frquencies or just in the ballpark, but
> > > without docs this is the best we can do.
> > > 
> > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > 
> > Given that no one complained yet I'm not sure this is worth the trouble.
> > Otoh it's all below the 10% margin we have already anyway, so one big
> > wash ;-)
> 
> Yeah <1Mhz is a fairly small error here. But I still prefer to make the
> change, if only for consistency. Otherwise it'll keep bugging me and
> I'll have to keep fighting the urge to change it every time I see those
> numbers.

I fully approve of OCD urges ;-) Series overall looks really nifty, I
think it would be best to sign up a senior person from the vpg display
team for to review the details in this - maybe they remember some of the
old lore ...

And one aside for atomic: I guess we need a clocks_lock ww mutex to
protect any dynamic cdclock state. Same for shared dpll state too. This
will be fun to integrate, but since all the code that checks cdclock can
fail with -EINVAL already wiring up -EDEADLK shouldn't cause any fuzz.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 06/18] drm/i915: Assume 400 MHz cdclk for the rest of gen4-7

[Ville Syrjälä]

On Mon, Nov 17, 2014 at 07:46:08PM +0100, Daniel Vetter wrote:
> On Mon, Nov 17, 2014 at 04:43:40PM +0200, ville.syrjala@linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > 
> > We don't currently have cdclk extraction code for 965g,snb,ivb.
> > Let's assumee 400 MHz until we know better. That seems to match hints
> > in various vague documents. Whether that's good enough is not
> > entirely clear.
> > 
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> Hm, not sure whether these chips even had outputs which could drive this
> high really. And if we now start rejecting modes that previously worked
> (really unlikely imo) we'll get the regression reports and can fudge the
> numbers some more. So even without more spec hints I'm totally fine with
> going forward with this.

At least one IVB out there seems to agree with this 400MHz limit:
https://bugs.freedesktop.org/show_bug.cgi?id=85621

But it does look like we're also failing to respect the port limits
which according to BSpec are 315Mhz-388Mhz (depending on the port
and/or bpc) for IVB. So I guess you're right that the cdclk wouldn't
matter all that much if we actually checked the port limits.

Although there are also various sprite scaling/format related limits
we fail to check where the limit ends up being of the form cdclk*N
where N is something < 1.0. So having a decent approximation of the
cdclk around should be useful eventually.

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 13/18] drm/i915: Store max cdclk value in dev_priv

[Ville Syrjälä]

On Mon, Nov 17, 2014 at 07:43:39PM +0100, Daniel Vetter wrote:
> On Mon, Nov 17, 2014 at 04:43:47PM +0200, ville.syrjala@linux.intel.com wrote:
> > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > index 9c6bc82..5eeb456 100644
> > --- a/drivers/gpu/drm/i915/intel_display.c
> > +++ b/drivers/gpu/drm/i915/intel_display.c
> > @@ -4745,6 +4745,21 @@ static int valleyview_get_vco(struct drm_i915_private *dev_priv)
> >  	return vco_freq[hpll_freq] * 1000;
> >  }
> >  
> > +static void intel_update_max_cdclk(struct drm_device *dev)
> > +{
> > +	struct drm_i915_private *dev_priv = dev->dev_private;
> > +
> > +	if (IS_VALLEYVIEW(dev)) {
> > +		dev_priv->max_cdclk_freq = 400000;
> 
> I've thought the 400MHz mode is busted? Or is that just Punit bonghits on
> SDVs and pre-prod boards?

AFAIK it should work. It's just that the Punit doesn't actually support
it, so we have to bypass the Punit and poke at CCK directly to get
there.

-- 
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 14/18] drm/i915: Don't enable IPS when pixel rate exceeds 95% of cdclk

[Runyan, Arthur J]

>> +       /*
>> +        * FIXME if we compare against max we should then
>> +        * increase the cdclk frequency when the current
>> +        * value is too low. The other option is to compare
>> +        * against the cdclk frqeuncy we're going have post
>> +        * modeset (ie. one we computed using other constraints).
>> +        * Need to measure whether using a lower cdclk w/o IPS
>> +        * is better or worse than a higher cdclk w/ IPS.
>> +        */
>> +       return ilk_pipe_pixel_rate(pipe_config) <=
>> +               dev_priv->max_cdclk_freq * 95 / 100;
>> +}
>
>Adding Art for insight into this question (lower cdclock or ips) or
>whether this just isn't worth the complexity.
>-Daniel

IPS enabled probably saves more power than lower cdclk, but it will be dependent on resolution, content, and other features, like FBC and PSR.  I don't have any real measurements.
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 18/18] drm/i915: BDW cdclk change support

[shuang.he]

Tested-By: PRC QA PRTS (Patch Regression Test System Contact: shuang.he@intel.com)
-------------------------------------Summary-------------------------------------
Platform: baseline_drm_intel_nightly_pass_rate->patch_applied_pass_rate
BYT: pass/total=290/291->290/291
PNV: pass/total=356/356->355/356
ILK: pass/total=371/372->367/372
IVB: pass/total=545/546->545/546
SNB: pass/total=370/370->370/370
HSW: pass/total=581/581->579/581
BDW: pass/total=432/435->432/435
-------------------------------------Detailed-------------------------------------
test_platform: test_suite, test_case, result_with_drm_intel_nightly(count, machine_id...)...->result_with_patch_applied(count, machine_id)...
PNV: Intel_gpu_tools, igt_kms_sysfs_edid_timing, PASS(4, M7M23) -> FAIL(1, M23)PASS(3, M23)
ILK: Intel_gpu_tools, igt_kms_flip_flip-vs-modeset-vs-hang, PASS(4, M37M26) -> DMESG_WARN(1, M26)PASS(3, M26)
ILK: Intel_gpu_tools, igt_kms_flip_flip-vs-modeset-vs-hang-interruptible, DMESG_WARN(1, M26)PASS(3, M37M26) -> DMESG_WARN(1, M26)PASS(3, M26)
ILK: Intel_gpu_tools, igt_kms_flip_plain-flip, PASS(4, M37M26) -> DMESG_WARN(1, M26)PASS(3, M26)
ILK: Intel_gpu_tools, igt_kms_flip_wf_vblank-vs-modeset-interruptible, DMESG_WARN(1, M26)PASS(3, M37M26) -> DMESG_WARN(1, M26)PASS(3, M26)
HSW: Intel_gpu_tools, igt_pm_rpm_cursor-dpms, DMESG_WARN(3, M40)PASS(1, M40) -> DMESG_WARN(1, M40)PASS(3, M40)
HSW: Intel_gpu_tools, igt_pm_rpm_dpms-mode-unset-non-lpsp, DMESG_WARN(3, M40)PASS(1, M40) -> DMESG_WARN(1, M40)PASS(3, M40)
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx