Skip to contentSkip to conversation

Atomic watermark updates (v4)

PATCH In review Matt Roper 4 participants 20 messages Sep 15, 2015 – Sep 23, 2015

Subsystems (2)
lore
Share & tools
lore

Matt Roper

author
#0 lore raw

cover Atomic watermark updates (v4)

Previous revision is here:
http://lists.freedesktop.org/archives/intel-gfx/2015-August/074161.html

Key changes:
- Lots of bugfixes from review comments by Ander and Ville (and a couple I found on my own)
- We now keep the active watermarks in intel_crtc, but pre-calculate to future watermarks in intel_crtc_state. The values stored in the state have been renamed to 'optimal' and 'intermediate' for clarity.
- The interface exposed by platforms that can do atomic updates has changed slightly. .initial_watermarks() and .optimize_watermarks() are the two routines called by the device-independent code to program the precomputed watermarks. These functions handle the locking, the copy of values from cstate->wm to intel_crtc->wm.active, and the actual programming (if needed).
- Added the wm_mutex in preparation for async.

A few patches from previous iterations have been left out of this series:
- Async watermark optimization. I had this in an earlier series, but Maarten indicated that it would clash with his ongoing work. I'll bring back the async patch in the future once his work lands.
- SKL/BXT version. I have some gen9 code that should plug into this framework (gen9 only requires one-step programming), but haven't had a chance to test it yet. I should be able to give it a spin on BXT hardware soon, so I'll wait until I've done that before reposting the gen9-specific followups.

Also note that after some mailing list / IRC discussion, we've decided not to perform modesets in a two-step process that was being discussed as a potential prereq for this, so my experimental prereq patches to do that have been dropped.

Matt Roper (13):
drm/i915: Drop redundant watermark programming drm/i915: Eliminate usage of plane_wm_parameters from ILK-style WM

    code (v2)
  drm/i915: Eliminate usage of pipe_wm_parameters from ILK-style WM (v2)
  drm/i915: Determine I915_MAX_PLANES from plane enum
  drm/i915/skl: Simplify wm structures slightly (v2)
  drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM
    (v2)

drm/i915/ivb: Move WaCxSRDisabledForSpriteScaling w/a to atomic check drm/i915: Drop intel_update_sprite_watermarks
drm/i915: Calculate pipe watermarks into CRTC state (v3) drm/i915: Calculate ILK-style watermarks during atomic check (v3) drm/i915: Don't set plane visible during HW readout if CRTC is off drm/i915: Calculate watermark configuration during atomic check (v2) drm/i915: Add two-stage ILK-style watermark programming (v4)

Ville Syrjälä (1):
drm/i915: Refactor ilk_update_wm (v3)

 drivers/gpu/drm/i915/i915_debugfs.c  |   2 +-
 drivers/gpu/drm/i915/i915_drv.h      |  41 +-
 drivers/gpu/drm/i915/intel_atomic.c  |   2 +
 drivers/gpu/drm/i915/intel_display.c | 194 ++++++--
 drivers/gpu/drm/i915/intel_drv.h     |  79 +++-
 drivers/gpu/drm/i915/intel_pm.c      | 849 ++++++++++++++++-------------------
 drivers/gpu/drm/i915/intel_sprite.c  |  15 -
 7 files changed, 637 insertions(+), 545 deletions(-)
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#1 lore raw

PATCH 1/14 drm/i915: Drop redundant watermark programming

In commit

commit e4ca061275ec6a48b66c6edebe08644e666994c0
Author: Patrik Jakobsson <[email protected]>
Date: Wed Jul 8 15:31:52 2015 +0200

drm/i915: Don't forget to mark crtc as inactive after disable

we added extra watermark updates to all of the .crtc_disable() entrypoints to avoid problems problems with system resume on SKL. Those disable entrypoints are currently called in just two places in the
driver: intel_atomic_commit (i.e., during a modeset) and intel_crtc_disable_noatomic (which is called during hardware readout). It seems that this extra watermark recalculation should only be important in the latter case (which happens during a resume operation); the former case should always have appropriate watermark programming happening at other points in the modeset sequence.

Let's move the watermark update out of the .crtc_disable() entrypoints and place it directly in intel_crtc_disable_noatomic() so that it only happens on S3 resume and not during a regular modeset (since the existing watermark handling should properly update watermarks during normal atomic commits).

Cc Patrik Jakobsson <[email protected]>
---
 drivers/gpu/drm/i915/intel_display.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/intel_display.c +2 -9
index fc00867..d144d29 100644
@@ -5031,9 +5031,6 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 
 		ironlake_fdi_pll_disable(intel_crtc);
 	}
-
-	intel_crtc->active = false;
-	intel_update_watermarks(crtc);
 }
 
 static void haswell_crtc_disable(struct drm_crtc *crtc)
@@ -5077,9 +5074,6 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->post_disable)
 			encoder->post_disable(encoder);
-
-	intel_crtc->active = false;
-	intel_update_watermarks(crtc);
 }
 
 static void i9xx_pfit_enable(struct intel_crtc *crtc)
@@ -6178,9 +6172,6 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 
 	if (!IS_GEN2(dev))
 		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
-
-	intel_crtc->active = false;
-	intel_update_watermarks(crtc);
 }
 
 static void intel_crtc_disable_noatomic(struct drm_crtc *crtc)
@@ -6200,6 +6191,8 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc)
 
 	intel_crtc_disable_planes(crtc, crtc->state->plane_mask);
 	dev_priv->display.crtc_disable(crtc);
+	intel_crtc->active = false;
+	intel_update_watermarks(crtc);
 	intel_disable_shared_dpll(intel_crtc);
 
 	domains = intel_crtc->enabled_power_domains;
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#2 lore raw

PATCH 2/14 drm/i915: Eliminate usage of plane_wm_parameters from ILK-style WM code (v2)

In reply to Matt Roper

Just pull the info out of the plane state structure rather than staging it in an additional structure.

v2: Add 'visible' condition to sprites_scaled so that we don't limit the
WM level when the sprite isn't enabled. (Ville)

Reviewed-by(v1): Ander Conselvan de Oliveira <[email protected]>

---
 drivers/gpu/drm/i915/intel_pm.c | 133 +++++++++++++++++++++-------------------
 1 file changed, 70 insertions(+), 63 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +70 -63
index 62de97e..442a1f6 100644
@@ -1799,9 +1799,6 @@ struct ilk_pipe_wm_parameters {
 	bool active;
 	uint32_t pipe_htotal;
 	uint32_t pixel_rate;
-	struct intel_plane_wm_parameters pri;
-	struct intel_plane_wm_parameters spr;
-	struct intel_plane_wm_parameters cur;
 };
 
 struct ilk_wm_maximums {
@@ -1823,25 +1820,25 @@ struct intel_wm_config {
  * mem_value must be in 0.1us units.
  */
 static uint32_t ilk_compute_pri_wm(const struct ilk_pipe_wm_parameters *params,
+				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value,
 				   bool is_lp)
 {
+	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
 	uint32_t method1, method2;
 
-	if (!params->active || !params->pri.enabled)
+	if (!params->active || !pstate->visible)
 		return 0;
 
-	method1 = ilk_wm_method1(params->pixel_rate,
-				 params->pri.bytes_per_pixel,
-				 mem_value);
+	method1 = ilk_wm_method1(params->pixel_rate, bpp, mem_value);
 
 	if (!is_lp)
 		return method1;
 
 	method2 = ilk_wm_method2(params->pixel_rate,
 				 params->pipe_htotal,
-				 params->pri.horiz_pixels,
-				 params->pri.bytes_per_pixel,
+				 drm_rect_width(&pstate->dst),
+				 bpp,
 				 mem_value);
 
 	return min(method1, method2);
@@ -1852,20 +1849,20 @@ static uint32_t ilk_compute_pri_wm(const struct ilk_pipe_wm_parameters *params,
  * mem_value must be in 0.1us units.
  */
 static uint32_t ilk_compute_spr_wm(const struct ilk_pipe_wm_parameters *params,
+				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value)
 {
+	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
 	uint32_t method1, method2;
 
-	if (!params->active || !params->spr.enabled)
+	if (!params->active || !pstate->visible)
 		return 0;
 
-	method1 = ilk_wm_method1(params->pixel_rate,
-				 params->spr.bytes_per_pixel,
-				 mem_value);
+	method1 = ilk_wm_method1(params->pixel_rate, bpp, mem_value);
 	method2 = ilk_wm_method2(params->pixel_rate,
 				 params->pipe_htotal,
-				 params->spr.horiz_pixels,
-				 params->spr.bytes_per_pixel,
+				 drm_rect_width(&pstate->dst),
+				 bpp,
 				 mem_value);
 	return min(method1, method2);
 }
@@ -1875,28 +1872,32 @@ static uint32_t ilk_compute_spr_wm(const struct ilk_pipe_wm_parameters *params,
  * mem_value must be in 0.1us units.
  */
 static uint32_t ilk_compute_cur_wm(const struct ilk_pipe_wm_parameters *params,
+				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value)
 {
-	if (!params->active || !params->cur.enabled)
+	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
+
+	if (!params->active || !pstate->visible)
 		return 0;
 
 	return ilk_wm_method2(params->pixel_rate,
 			      params->pipe_htotal,
-			      params->cur.horiz_pixels,
-			      params->cur.bytes_per_pixel,
+			      drm_rect_width(&pstate->dst),
+			      bpp,
 			      mem_value);
 }
 
 /* Only for WM_LP. */
 static uint32_t ilk_compute_fbc_wm(const struct ilk_pipe_wm_parameters *params,
+				   const struct intel_plane_state *pstate,
 				   uint32_t pri_val)
 {
-	if (!params->active || !params->pri.enabled)
+	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
+
+	if (!params->active || !pstate->visible)
 		return 0;
 
-	return ilk_wm_fbc(pri_val,
-			  params->pri.horiz_pixels,
-			  params->pri.bytes_per_pixel);
+	return ilk_wm_fbc(pri_val, drm_rect_width(&pstate->dst), bpp);
 }
 
 static unsigned int ilk_display_fifo_size(const struct drm_device *dev)
@@ -2061,10 +2062,12 @@ static bool ilk_validate_wm_level(int level,
 }
 
 static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv,
+				 const struct intel_crtc *intel_crtc,
 				 int level,
 				 const struct ilk_pipe_wm_parameters *p,
 				 struct intel_wm_level *result)
 {
+	struct intel_plane *intel_plane;
 	uint16_t pri_latency = dev_priv->wm.pri_latency[level];
 	uint16_t spr_latency = dev_priv->wm.spr_latency[level];
 	uint16_t cur_latency = dev_priv->wm.cur_latency[level];
@@ -2076,10 +2079,29 @@ static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv,
 		cur_latency *= 5;
 	}
 
-	result->pri_val = ilk_compute_pri_wm(p, pri_latency, level);
-	result->spr_val = ilk_compute_spr_wm(p, spr_latency);
-	result->cur_val = ilk_compute_cur_wm(p, cur_latency);
-	result->fbc_val = ilk_compute_fbc_wm(p, result->pri_val);
+	for_each_intel_plane_on_crtc(dev_priv->dev, intel_crtc, intel_plane) {
+		struct intel_plane_state *pstate =
+			to_intel_plane_state(intel_plane->base.state);
+
+		switch (intel_plane->base.type) {
+		case DRM_PLANE_TYPE_PRIMARY:
+			result->pri_val = ilk_compute_pri_wm(p, pstate,
+							     pri_latency,
+							     level);
+			result->fbc_val = ilk_compute_fbc_wm(p, pstate,
+							     result->pri_val);
+			break;
+		case DRM_PLANE_TYPE_OVERLAY:
+			result->spr_val = ilk_compute_spr_wm(p, pstate,
+							     spr_latency);
+			break;
+		case DRM_PLANE_TYPE_CURSOR:
+			result->cur_val = ilk_compute_cur_wm(p, pstate,
+							     cur_latency);
+			break;
+		}
+	}
+
 	result->enable = true;
 }
 
@@ -2341,10 +2363,7 @@ static void skl_setup_wm_latency(struct drm_device *dev)
 static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
 				      struct ilk_pipe_wm_parameters *p)
 {
-	struct drm_device *dev = crtc->dev;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum pipe pipe = intel_crtc->pipe;
-	struct drm_plane *plane;
 
 	if (!intel_crtc->active)
 		return;
@@ -2352,32 +2371,6 @@ static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
 	p->active = true;
 	p->pipe_htotal = intel_crtc->config->base.adjusted_mode.crtc_htotal;
 	p->pixel_rate = ilk_pipe_pixel_rate(intel_crtc->config);
-
-	if (crtc->primary->state->fb)
-		p->pri.bytes_per_pixel =
-			crtc->primary->state->fb->bits_per_pixel / 8;
-	else
-		p->pri.bytes_per_pixel = 4;
-
-	p->cur.bytes_per_pixel = 4;
-	/*
-	 * TODO: for now, assume primary and cursor planes are always enabled.
-	 * Setting them to false makes the screen flicker.
-	 */
-	p->pri.enabled = true;
-	p->cur.enabled = true;
-
-	p->pri.horiz_pixels = intel_crtc->config->pipe_src_w;
-	p->cur.horiz_pixels = intel_crtc->base.cursor->state->crtc_w;
-
-	drm_for_each_legacy_plane(plane, dev) {
-		struct intel_plane *intel_plane = to_intel_plane(plane);
-
-		if (intel_plane->pipe == pipe) {
-			p->spr = intel_plane->wm;
-			break;
-		}
-	}
 }
 
 static void ilk_compute_wm_config(struct drm_device *dev,
@@ -2405,28 +2398,42 @@ static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	const struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_plane *intel_plane;
+	struct intel_plane_state *sprstate = NULL;
 	int level, max_level = ilk_wm_max_level(dev);
 	/* LP0 watermark maximums depend on this pipe alone */
 	struct intel_wm_config config = {
 		.num_pipes_active = 1,
-		.sprites_enabled = params->spr.enabled,
-		.sprites_scaled = params->spr.scaled,
 	};
 	struct ilk_wm_maximums max;
 
+	for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) {
+		if (intel_plane->base.type == DRM_PLANE_TYPE_OVERLAY) {
+			sprstate = to_intel_plane_state(intel_plane->base.state);
+			break;
+		}
+	}
+
+	config.sprites_enabled = sprstate->visible;
+	config.sprites_scaled = sprstate->visible &&
+		(drm_rect_width(&sprstate->dst) != drm_rect_width(&sprstate->src) >> 16 ||
+		drm_rect_height(&sprstate->dst) != drm_rect_height(&sprstate->src) >> 16);
+
+
 	pipe_wm->pipe_enabled = params->active;
-	pipe_wm->sprites_enabled = params->spr.enabled;
-	pipe_wm->sprites_scaled = params->spr.scaled;
+	pipe_wm->sprites_enabled = sprstate->visible;
+	pipe_wm->sprites_scaled = config.sprites_scaled;
 
 	/* ILK/SNB: LP2+ watermarks only w/o sprites */
-	if (INTEL_INFO(dev)->gen <= 6 && params->spr.enabled)
+	if (INTEL_INFO(dev)->gen <= 6 && sprstate->visible)
 		max_level = 1;
 
 	/* ILK/SNB/IVB: LP1+ watermarks only w/o scaling */
-	if (params->spr.scaled)
+	if (config.sprites_scaled)
 		max_level = 0;
 
-	ilk_compute_wm_level(dev_priv, 0, params, &pipe_wm->wm[0]);
+	ilk_compute_wm_level(dev_priv, intel_crtc, 0, params, &pipe_wm->wm[0]);
 
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		pipe_wm->linetime = hsw_compute_linetime_wm(dev, crtc);
@@ -2443,7 +2450,7 @@ static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
 	for (level = 1; level <= max_level; level++) {
 		struct intel_wm_level wm = {};
 
-		ilk_compute_wm_level(dev_priv, level, params, &wm);
+		ilk_compute_wm_level(dev_priv, intel_crtc, level, params, &wm);
 
 		/*
 		 * Disable any watermark level that exceeds the
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#3 lore raw

PATCH 3/14 drm/i915: Eliminate usage of pipe_wm_parameters from ILK-style WM (v2)

In reply to Matt Roper

Just pull the info out of the CRTC state structure rather than staging it in an additional structure.

Note that we use cstate->active rather than intel_crtc->active which may appear to be a change in behavior. However since we're no longer trying to recalculate watermarks during the "pipe off" stage of a modeset, intel_crtc->active and cstate->active should always be identical when watermarks are calculated (at least for ILK-style platforms).

v2: Clarify reasoning for cstate->active and add a WARN_ON to the code

    to assert that it really is always identical to intel_crtc->active
    as expected.
---
 drivers/gpu/drm/i915/intel_pm.c | 84 ++++++++++++++---------------------------
 1 file changed, 29 insertions(+), 55 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +29 -55
index 442a1f6..0caabce 100644
@@ -1795,12 +1795,6 @@ struct skl_pipe_wm_parameters {
 	struct intel_plane_wm_parameters cursor;
 };
 
-struct ilk_pipe_wm_parameters {
-	bool active;
-	uint32_t pipe_htotal;
-	uint32_t pixel_rate;
-};
-
 struct ilk_wm_maximums {
 	uint16_t pri;
 	uint16_t spr;
@@ -1819,7 +1813,7 @@ struct intel_wm_config {
  * For both WM_PIPE and WM_LP.
  * mem_value must be in 0.1us units.
  */
-static uint32_t ilk_compute_pri_wm(const struct ilk_pipe_wm_parameters *params,
+static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate,
 				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value,
 				   bool is_lp)
@@ -1827,16 +1821,16 @@ static uint32_t ilk_compute_pri_wm(const struct ilk_pipe_wm_parameters *params,
 	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
 	uint32_t method1, method2;
 
-	if (!params->active || !pstate->visible)
+	if (!cstate->base.active || !pstate->visible)
 		return 0;
 
-	method1 = ilk_wm_method1(params->pixel_rate, bpp, mem_value);
+	method1 = ilk_wm_method1(ilk_pipe_pixel_rate(cstate), bpp, mem_value);
 
 	if (!is_lp)
 		return method1;
 
-	method2 = ilk_wm_method2(params->pixel_rate,
-				 params->pipe_htotal,
+	method2 = ilk_wm_method2(ilk_pipe_pixel_rate(cstate),
+				 cstate->base.adjusted_mode.crtc_htotal,
 				 drm_rect_width(&pstate->dst),
 				 bpp,
 				 mem_value);
@@ -1848,19 +1842,19 @@ static uint32_t ilk_compute_pri_wm(const struct ilk_pipe_wm_parameters *params,
  * For both WM_PIPE and WM_LP.
  * mem_value must be in 0.1us units.
  */
-static uint32_t ilk_compute_spr_wm(const struct ilk_pipe_wm_parameters *params,
+static uint32_t ilk_compute_spr_wm(const struct intel_crtc_state *cstate,
 				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value)
 {
 	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
 	uint32_t method1, method2;
 
-	if (!params->active || !pstate->visible)
+	if (!cstate->base.active || !pstate->visible)
 		return 0;
 
-	method1 = ilk_wm_method1(params->pixel_rate, bpp, mem_value);
-	method2 = ilk_wm_method2(params->pixel_rate,
-				 params->pipe_htotal,
+	method1 = ilk_wm_method1(ilk_pipe_pixel_rate(cstate), bpp, mem_value);
+	method2 = ilk_wm_method2(ilk_pipe_pixel_rate(cstate),
+				 cstate->base.adjusted_mode.crtc_htotal,
 				 drm_rect_width(&pstate->dst),
 				 bpp,
 				 mem_value);
@@ -1871,30 +1865,30 @@ static uint32_t ilk_compute_spr_wm(const struct ilk_pipe_wm_parameters *params,
  * For both WM_PIPE and WM_LP.
  * mem_value must be in 0.1us units.
  */
-static uint32_t ilk_compute_cur_wm(const struct ilk_pipe_wm_parameters *params,
+static uint32_t ilk_compute_cur_wm(const struct intel_crtc_state *cstate,
 				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value)
 {
 	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
 
-	if (!params->active || !pstate->visible)
+	if (!cstate->base.active || !pstate->visible)
 		return 0;
 
-	return ilk_wm_method2(params->pixel_rate,
-			      params->pipe_htotal,
+	return ilk_wm_method2(ilk_pipe_pixel_rate(cstate),
+			      cstate->base.adjusted_mode.crtc_htotal,
 			      drm_rect_width(&pstate->dst),
 			      bpp,
 			      mem_value);
 }
 
 /* Only for WM_LP. */
-static uint32_t ilk_compute_fbc_wm(const struct ilk_pipe_wm_parameters *params,
+static uint32_t ilk_compute_fbc_wm(const struct intel_crtc_state *cstate,
 				   const struct intel_plane_state *pstate,
 				   uint32_t pri_val)
 {
 	int bpp = pstate->base.fb ? pstate->base.fb->bits_per_pixel / 8 : 0;
 
-	if (!params->active || !pstate->visible)
+	if (!cstate->base.active || !pstate->visible)
 		return 0;
 
 	return ilk_wm_fbc(pri_val, drm_rect_width(&pstate->dst), bpp);
@@ -2064,7 +2058,7 @@ static bool ilk_validate_wm_level(int level,
 static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv,
 				 const struct intel_crtc *intel_crtc,
 				 int level,
-				 const struct ilk_pipe_wm_parameters *p,
+				 struct intel_crtc_state *cstate,
 				 struct intel_wm_level *result)
 {
 	struct intel_plane *intel_plane;
@@ -2085,18 +2079,18 @@ static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv,
 
 		switch (intel_plane->base.type) {
 		case DRM_PLANE_TYPE_PRIMARY:
-			result->pri_val = ilk_compute_pri_wm(p, pstate,
+			result->pri_val = ilk_compute_pri_wm(cstate, pstate,
 							     pri_latency,
 							     level);
-			result->fbc_val = ilk_compute_fbc_wm(p, pstate,
+			result->fbc_val = ilk_compute_fbc_wm(cstate, pstate,
 							     result->pri_val);
 			break;
 		case DRM_PLANE_TYPE_OVERLAY:
-			result->spr_val = ilk_compute_spr_wm(p, pstate,
+			result->spr_val = ilk_compute_spr_wm(cstate, pstate,
 							     spr_latency);
 			break;
 		case DRM_PLANE_TYPE_CURSOR:
-			result->cur_val = ilk_compute_cur_wm(p, pstate,
+			result->cur_val = ilk_compute_cur_wm(cstate, pstate,
 							     cur_latency);
 			break;
 		}
@@ -2360,19 +2354,6 @@ static void skl_setup_wm_latency(struct drm_device *dev)
 	intel_print_wm_latency(dev, "Gen9 Plane", dev_priv->wm.skl_latency);
 }
 
-static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
-				      struct ilk_pipe_wm_parameters *p)
-{
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	if (!intel_crtc->active)
-		return;
-
-	p->active = true;
-	p->pipe_htotal = intel_crtc->config->base.adjusted_mode.crtc_htotal;
-	p->pixel_rate = ilk_pipe_pixel_rate(intel_crtc->config);
-}
-
 static void ilk_compute_wm_config(struct drm_device *dev,
 				  struct intel_wm_config *config)
 {
@@ -2392,10 +2373,10 @@ static void ilk_compute_wm_config(struct drm_device *dev,
 }
 
 /* Compute new watermarks for the pipe */
-static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
-				  const struct ilk_pipe_wm_parameters *params,
+static bool intel_compute_pipe_wm(struct intel_crtc_state *cstate,
 				  struct intel_pipe_wm *pipe_wm)
 {
+	struct drm_crtc *crtc = cstate->base.crtc;
 	struct drm_device *dev = crtc->dev;
 	const struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -2420,8 +2401,7 @@ static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
 		(drm_rect_width(&sprstate->dst) != drm_rect_width(&sprstate->src) >> 16 ||
 		drm_rect_height(&sprstate->dst) != drm_rect_height(&sprstate->src) >> 16);
 
-
-	pipe_wm->pipe_enabled = params->active;
+	pipe_wm->pipe_enabled = cstate->base.active;
 	pipe_wm->sprites_enabled = sprstate->visible;
 	pipe_wm->sprites_scaled = config.sprites_scaled;
 
@@ -2433,7 +2413,7 @@ static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
 	if (config.sprites_scaled)
 		max_level = 0;
 
-	ilk_compute_wm_level(dev_priv, intel_crtc, 0, params, &pipe_wm->wm[0]);
+	ilk_compute_wm_level(dev_priv, intel_crtc, 0, cstate, &pipe_wm->wm[0]);
 
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		pipe_wm->linetime = hsw_compute_linetime_wm(dev, crtc);
@@ -2450,7 +2430,7 @@ static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
 	for (level = 1; level <= max_level; level++) {
 		struct intel_wm_level wm = {};
 
-		ilk_compute_wm_level(dev_priv, intel_crtc, level, params, &wm);
+		ilk_compute_wm_level(dev_priv, intel_crtc, level, cstate, &wm);
 
 		/*
 		 * Disable any watermark level that exceeds the
@@ -3757,19 +3737,19 @@ skl_update_sprite_wm(struct drm_plane *plane, struct drm_crtc *crtc,
 static void ilk_update_wm(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct ilk_wm_maximums max;
-	struct ilk_pipe_wm_parameters params = {};
 	struct ilk_wm_values results = {};
 	enum intel_ddb_partitioning partitioning;
 	struct intel_pipe_wm pipe_wm = {};
 	struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm;
 	struct intel_wm_config config = {};
 
-	ilk_compute_wm_parameters(crtc, &params);
+	WARN_ON(cstate->base.active != intel_crtc->active);
 
-	intel_compute_pipe_wm(crtc, &params, &pipe_wm);
+	intel_compute_pipe_wm(cstate, &pipe_wm);
 
 	if (!memcmp(&intel_crtc->wm.active, &pipe_wm, sizeof(pipe_wm)))
 		return;
@@ -3809,12 +3789,6 @@ ilk_update_sprite_wm(struct drm_plane *plane,
 	struct drm_device *dev = plane->dev;
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 
-	intel_plane->wm.enabled = enabled;
-	intel_plane->wm.scaled = scaled;
-	intel_plane->wm.horiz_pixels = sprite_width;
-	intel_plane->wm.vert_pixels = sprite_width;
-	intel_plane->wm.bytes_per_pixel = pixel_size;
-
 	/*
 	 * IVB workaround: must disable low power watermarks for at least
 	 * one frame before enabling scaling.  LP watermarks can be re-enabled
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#4 lore raw

PATCH 4/14 drm/i915: Determine I915_MAX_PLANES from plane enum

In reply to Matt Roper

Let the compiler figure out what I915_MAX_PLANES is from 'enum plane' so that we don't need a separate #define.

While we're at it, add the cursor plane to the enum. This will cause I915_MAX_PLANES to now include the cursor plane in its count (it didn't
previously). This change is safe since we currently only use this
value in array declarations (never in the actual code logic); we just wind up allocating slightly more memory than we need to. A followup patch will cause various parts of the code to start using the extra array element where appropriate.

(This patch probably should have been squashed with the followup patch, but I couldn't figure out how to get Coccinelle to modify enum declarations...)

---
 drivers/gpu/drm/i915/i915_drv.h | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +6 -6
index 3bf8a9b..c800031 100644
@@ -131,17 +131,17 @@ enum transcoder {
 #define transcoder_name(t) ((t) + 'A')
 
 /*
- * This is the maximum (across all platforms) number of planes (primary +
- * sprites) that can be active at the same time on one pipe.
- *
- * This value doesn't count the cursor plane.
+ * I915_MAX_PLANES in the enum below is the maximum (across all platforms)
+ * number of planes per CRTC.  Not all platforms really have this many planes,
+ * which means some arrays of size I915_MAX_PLANES may have unused entries
+ * between the topmost sprite plane and the cursor plane.
  */
-#define I915_MAX_PLANES	4
-
 enum plane {
 	PLANE_A = 0,
 	PLANE_B,
 	PLANE_C,
+	PLANE_CURSOR,
+	I915_MAX_PLANES,
 };
 #define plane_name(p) ((p) + 'A')
 
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#5 lore raw

PATCH 5/14 drm/i915/skl: Simplify wm structures slightly (v2)

In reply to Matt Roper

A bunch of SKL watermark-related structures have the cursor plane as a separate entry from the rest of the planes. Since a previous patch updated I915_MAX_PLANES such that those plane arrays now have a slot for the cursor, update the code to use the new slot in the existing plane arrays and kill off the cursor-specific structures.

There shouldn't be any functional change here; this is just shuffling around how the data is stored in some of the data structures. The whole patch is generated with Coccinelle via the following semantic patch:

        @@ struct skl_pipe_wm_parameters WMP; @@
        - WMP.cursor
        + WMP.plane[PLANE_CURSOR]

        @@ struct skl_pipe_wm_parameters *WMP; @@
        - WMP->cursor
        + WMP->plane[PLANE_CURSOR]

        @@ @@
        struct skl_pipe_wm_parameters {
        ...
        - struct intel_plane_wm_parameters cursor;
        ...
        };

        @@
        struct skl_ddb_allocation DDB;
        expression E;
        @@
        - DDB.cursor[E]
        + DDB.plane[E][PLANE_CURSOR]

        @@
        struct skl_ddb_allocation *DDB;
        expression E;
        @@
        - DDB->cursor[E]
        + DDB->plane[E][PLANE_CURSOR]

        @@ @@
        struct skl_ddb_allocation {
        ...
        - struct skl_ddb_entry cursor[I915_MAX_PIPES];
        ...
        };

        @@
        struct skl_wm_values WMV;
        expression E1, E2;
        @@
        (
        - WMV.cursor[E1][E2]
        + WMV.plane[E1][PLANE_CURSOR][E2]
        |
        - WMV.cursor_trans[E1]
        + WMV.plane_trans[E1][PLANE_CURSOR]
        )

        @@
        struct skl_wm_values *WMV;
        expression E1, E2;
        @@
        (
        - WMV->cursor[E1][E2]
        + WMV->plane[E1][PLANE_CURSOR][E2]
        |
        - WMV->cursor_trans[E1]
        + WMV->plane_trans[E1][PLANE_CURSOR]
        )

        @@ @@
        struct skl_wm_values {
        ...
        - uint32_t cursor[I915_MAX_PIPES][8];
        ...
        - uint32_t cursor_trans[I915_MAX_PIPES];
        ...
        };

        @@ struct skl_wm_level WML; @@
        (
        - WML.cursor_en
        + WML.plane_en[PLANE_CURSOR]
        |
        - WML.cursor_res_b
        + WML.plane_res_b[PLANE_CURSOR]
        |
        - WML.cursor_res_l
        + WML.plane_res_l[PLANE_CURSOR]
        )

        @@ struct skl_wm_level *WML; @@
        (
        - WML->cursor_en
        + WML->plane_en[PLANE_CURSOR]
        |
        - WML->cursor_res_b
        + WML->plane_res_b[PLANE_CURSOR]
        |
        - WML->cursor_res_l
        + WML->plane_res_l[PLANE_CURSOR]
        )

        @@ @@
        struct skl_wm_level {
        ...
        - bool cursor_en;
        ...
        - uint16_t cursor_res_b;
        - uint8_t cursor_res_l;
        ...
        };

v2: Use a PLANE_CURSOR enum entry rather than making the code reference
I915_MAX_PLANES or I915_MAX_PLANES+1, which was confusing. (Ander)

---
 drivers/gpu/drm/i915/i915_debugfs.c  |  2 +-
 drivers/gpu/drm/i915/i915_drv.h      |  8 +---
 drivers/gpu/drm/i915/intel_display.c |  4 +-
 drivers/gpu/drm/i915/intel_pm.c      | 87 +++++++++++++++++++-----------------
 4 files changed, 49 insertions(+), 52 deletions(-)
drivers/gpu/drm/i915/i915_debugfs.c +1 -1
index 72ae347..52ee2a5 100644
@@ -3144,7 +3144,7 @@ static int i915_ddb_info(struct seq_file *m, void *unused)
 				   skl_ddb_entry_size(entry));
 		}
 
-		entry = &ddb->cursor[pipe];
+		entry = &ddb->plane[pipe][PLANE_CURSOR];
 		seq_printf(m, "  %-13s%8u%8u%8u\n", "Cursor", entry->start,
 			   entry->end, skl_ddb_entry_size(entry));
 	}
drivers/gpu/drm/i915/i915_drv.h +1 -7
index c800031..539bfa2 100644
@@ -1585,8 +1585,7 @@ static inline bool skl_ddb_entry_equal(const struct skl_ddb_entry *e1,
 struct skl_ddb_allocation {
 	struct skl_ddb_entry pipe[I915_MAX_PIPES];
 	struct skl_ddb_entry plane[I915_MAX_PIPES][I915_MAX_PLANES]; /* packed/uv */
-	struct skl_ddb_entry y_plane[I915_MAX_PIPES][I915_MAX_PLANES]; /* y-plane */
-	struct skl_ddb_entry cursor[I915_MAX_PIPES];
+	struct skl_ddb_entry y_plane[I915_MAX_PIPES][I915_MAX_PLANES];
 };
 
 struct skl_wm_values {
@@ -1594,18 +1593,13 @@ struct skl_wm_values {
 	struct skl_ddb_allocation ddb;
 	uint32_t wm_linetime[I915_MAX_PIPES];
 	uint32_t plane[I915_MAX_PIPES][I915_MAX_PLANES][8];
-	uint32_t cursor[I915_MAX_PIPES][8];
 	uint32_t plane_trans[I915_MAX_PIPES][I915_MAX_PLANES];
-	uint32_t cursor_trans[I915_MAX_PIPES];
 };
 
 struct skl_wm_level {
 	bool plane_en[I915_MAX_PLANES];
-	bool cursor_en;
 	uint16_t plane_res_b[I915_MAX_PLANES];
 	uint8_t plane_res_l[I915_MAX_PLANES];
-	uint16_t cursor_res_b;
-	uint8_t cursor_res_l;
 };
 
 /*
drivers/gpu/drm/i915/intel_display.c +2 -2
index d144d29..017641a 100644
@@ -12527,8 +12527,8 @@ static void check_wm_state(struct drm_device *dev)
 		}
 
 		/* cursor */
-		hw_entry = &hw_ddb.cursor[pipe];
-		sw_entry = &sw_ddb->cursor[pipe];
+		hw_entry = &hw_ddb.plane[pipe][PLANE_CURSOR];
+		sw_entry = &sw_ddb->plane[pipe][PLANE_CURSOR];
 
 		if (skl_ddb_entry_equal(hw_entry, sw_entry))
 			continue;
drivers/gpu/drm/i915/intel_pm.c +45 -42

205 lines not on this page · raw message

Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#6 lore raw

PATCH 6/14 drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v2)

In reply to Matt Roper

Just pull the info out of the state structures rather than staging it in an additional set of structures. To make this more straightforward, we change the signature of several internal WM functions to take the crtc state as a parameter.

v2:
- Don't forget to skip cursor planes on a loop in the DDB allocation function to match original behavior. (Ander)
- Change a use of intel_crtc->active to cstate->active. They should be identical, but it's better to be consistent. (Ander)
- Rework more function signatures to pass states rather than crtc for consistency. (Ander)

---
 drivers/gpu/drm/i915/intel_pm.c | 326 +++++++++++++++++++---------------------
 1 file changed, 151 insertions(+), 175 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +151 -175
index 7115368..51dbda7 100644
@@ -1787,13 +1787,6 @@ static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels,
 	return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2;
 }
 
-struct skl_pipe_wm_parameters {
-	bool active;
-	uint32_t pipe_htotal;
-	uint32_t pixel_rate; /* in KHz */
-	struct intel_plane_wm_parameters plane[I915_MAX_PLANES];
-};
-
 struct ilk_wm_maximums {
 	uint16_t pri;
 	uint16_t spr;
@@ -2834,18 +2827,40 @@ static bool ilk_disable_lp_wm(struct drm_device *dev)
 #define SKL_DDB_SIZE		896	/* in blocks */
 #define BXT_DDB_SIZE		512
 
+/*
+ * Return the index of a plane in the SKL DDB and wm result arrays.  Primary
+ * plane is always in slot 0, cursor is always in slot I915_MAX_PLANES-1, and
+ * other universal planes are in indices 1..n.  Note that this may leave unused
+ * indices between the top "sprite" plane and the cursor.
+ */
+static int
+skl_wm_plane_id(const struct intel_plane *plane)
+{
+	switch (plane->base.type) {
+	case DRM_PLANE_TYPE_PRIMARY:
+		return 0;
+	case DRM_PLANE_TYPE_CURSOR:
+		return PLANE_CURSOR;
+	case DRM_PLANE_TYPE_OVERLAY:
+		return plane->plane;
+	default:
+		MISSING_CASE(plane->base.type);
+		return plane->plane;
+	}
+}
+
 static void
 skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
-				   struct drm_crtc *for_crtc,
+				   const struct intel_crtc_state *cstate,
 				   const struct intel_wm_config *config,
-				   const struct skl_pipe_wm_parameters *params,
 				   struct skl_ddb_entry *alloc /* out */)
 {
+	struct drm_crtc *for_crtc = cstate->base.crtc;
 	struct drm_crtc *crtc;
 	unsigned int pipe_size, ddb_size;
 	int nth_active_pipe;
 
-	if (!params->active) {
+	if (!cstate->base.active) {
 		alloc->start = 0;
 		alloc->end = 0;
 		return;
@@ -2911,19 +2926,29 @@ void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
 }
 
 static unsigned int
-skl_plane_relative_data_rate(const struct intel_plane_wm_parameters *p, int y)
+skl_plane_relative_data_rate(const struct intel_crtc_state *cstate,
+			     const struct drm_plane_state *pstate,
+			     int y)
 {
+	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
+	struct drm_framebuffer *fb = pstate->fb;
 
 	/* for planar format */
-	if (p->y_bytes_per_pixel) {
+	if (fb->pixel_format == DRM_FORMAT_NV12) {
 		if (y)  /* y-plane data rate */
-			return p->horiz_pixels * p->vert_pixels * p->y_bytes_per_pixel;
+			return intel_crtc->config->pipe_src_w *
+				intel_crtc->config->pipe_src_h *
+				drm_format_plane_cpp(fb->pixel_format, 0);
 		else    /* uv-plane data rate */
-			return (p->horiz_pixels/2) * (p->vert_pixels/2) * p->bytes_per_pixel;
+			return (intel_crtc->config->pipe_src_w/2) *
+				(intel_crtc->config->pipe_src_h/2) *
+				drm_format_plane_cpp(fb->pixel_format, 1);
 	}
 
 	/* for packed formats */
-	return p->horiz_pixels * p->vert_pixels * p->bytes_per_pixel;
+	return intel_crtc->config->pipe_src_w *
+		intel_crtc->config->pipe_src_h *
+		drm_format_plane_cpp(fb->pixel_format, 1);
 }
 
 /*
@@ -2932,46 +2957,51 @@ skl_plane_relative_data_rate(const struct intel_plane_wm_parameters *p, int y)
  *   3 * 4096 * 8192  * 4 < 2^32
  */
 static unsigned int
-skl_get_total_relative_data_rate(struct intel_crtc *intel_crtc,
-				 const struct skl_pipe_wm_parameters *params)
+skl_get_total_relative_data_rate(const struct intel_crtc_state *cstate)
 {
+	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
+	struct drm_device *dev = intel_crtc->base.dev;
+	const struct intel_plane *intel_plane;
 	unsigned int total_data_rate = 0;
-	int plane;
 
-	for (plane = 0; plane < intel_num_planes(intel_crtc); plane++) {
-		const struct intel_plane_wm_parameters *p;
+	for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) {
+		const struct drm_plane_state *pstate = intel_plane->base.state;
 
-		p = &params->plane[plane];
-		if (!p->enabled)
+		if (WARN_ON(pstate->fb == NULL))
 			continue;
 
-		total_data_rate += skl_plane_relative_data_rate(p, 0); /* packed/uv */
-		if (p->y_bytes_per_pixel) {
-			total_data_rate += skl_plane_relative_data_rate(p, 1); /* y-plane */
-		}
+		/* packed/uv */
+		total_data_rate += skl_plane_relative_data_rate(cstate,
+								pstate,
+								0);
+
+		if (pstate->fb->pixel_format == DRM_FORMAT_NV12)
+			/* y-plane */
+			total_data_rate += skl_plane_relative_data_rate(cstate,
+									pstate,
+									1);
 	}
 
 	return total_data_rate;
 }
 
 static void
-skl_allocate_pipe_ddb(struct drm_crtc *crtc,
+skl_allocate_pipe_ddb(struct intel_crtc_state *cstate,
 		      const struct intel_wm_config *config,
-		      const struct skl_pipe_wm_parameters *params,
 		      struct skl_ddb_allocation *ddb /* out */)
 {
+	struct drm_crtc *crtc = cstate->base.crtc;
 	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_plane *intel_plane;
 	enum pipe pipe = intel_crtc->pipe;
 	struct skl_ddb_entry *alloc = &ddb->pipe[pipe];
 	uint16_t alloc_size, start, cursor_blocks;
 	uint16_t minimum[I915_MAX_PLANES];
 	uint16_t y_minimum[I915_MAX_PLANES];
 	unsigned int total_data_rate;
-	int plane;
 
-	skl_ddb_get_pipe_allocation_limits(dev, crtc, config, params, alloc);
+	skl_ddb_get_pipe_allocation_limits(dev, cstate, config, alloc);
 	alloc_size = skl_ddb_entry_size(alloc);
 	if (alloc_size == 0) {
 		memset(ddb->plane[pipe], 0, sizeof(ddb->plane[pipe]));
@@ -2988,17 +3018,20 @@ skl_allocate_pipe_ddb(struct drm_crtc *crtc,
 	alloc->end -= cursor_blocks;
 
 	/* 1. Allocate the mininum required blocks for each active plane */
-	for_each_plane(dev_priv, pipe, plane) {
-		const struct intel_plane_wm_parameters *p;
+	for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) {
+		struct drm_plane *plane = &intel_plane->base;
+		struct drm_framebuffer *fb = plane->fb;
+		int id = skl_wm_plane_id(intel_plane);
 
-		p = &params->plane[plane];
-		if (!p->enabled)
+		if (fb == NULL)
+			continue;
+		if (plane->type == DRM_PLANE_TYPE_CURSOR)
 			continue;
 
-		minimum[plane] = 8;
-		alloc_size -= minimum[plane];
-		y_minimum[plane] = p->y_bytes_per_pixel ? 8 : 0;
-		alloc_size -= y_minimum[plane];
+		minimum[id] = 8;
+		alloc_size -= minimum[id];
+		y_minimum[id] = (fb->pixel_format == DRM_FORMAT_NV12) ? 8 : 0;
+		alloc_size -= y_minimum[id];
 	}
 
 	/*
@@ -3007,45 +3040,50 @@ skl_allocate_pipe_ddb(struct drm_crtc *crtc,
 	 *
 	 * FIXME: we may not allocate every single block here.
 	 */
-	total_data_rate = skl_get_total_relative_data_rate(intel_crtc, params);
+	total_data_rate = skl_get_total_relative_data_rate(cstate);
 
 	start = alloc->start;
-	for (plane = 0; plane < intel_num_planes(intel_crtc); plane++) {
-		const struct intel_plane_wm_parameters *p;
+	for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) {
+		struct drm_plane *plane = &intel_plane->base;
+		struct drm_plane_state *pstate = intel_plane->base.state;
 		unsigned int data_rate, y_data_rate;
 		uint16_t plane_blocks, y_plane_blocks = 0;
+		int id = skl_wm_plane_id(intel_plane);
 
-		p = &params->plane[plane];
-		if (!p->enabled)
+		if (pstate->fb == NULL)
+			continue;
+		if (plane->type == DRM_PLANE_TYPE_CURSOR)
 			continue;
 
-		data_rate = skl_plane_relative_data_rate(p, 0);
+		data_rate = skl_plane_relative_data_rate(cstate, pstate, 0);
 
 		/*
 		 * allocation for (packed formats) or (uv-plane part of planar format):
 		 * promote the expression to 64 bits to avoid overflowing, the
 		 * result is < available as data_rate / total_data_rate < 1
 		 */
-		plane_blocks = minimum[plane];
+		plane_blocks = minimum[id];
 		plane_blocks += div_u64((uint64_t)alloc_size * data_rate,
 					total_data_rate);
 
-		ddb->plane[pipe][plane].start = start;
-		ddb->plane[pipe][plane].end = start + plane_blocks;
+		ddb->plane[pipe][id].start = start;
+		ddb->plane[pipe][id].end = start + plane_blocks;
 
 		start += plane_blocks;
 
 		/*
 		 * allocation for y_plane part of planar format:
 		 */
-		if (p->y_bytes_per_pixel) {
-			y_data_rate = skl_plane_relative_data_rate(p, 1);
-			y_plane_blocks = y_minimum[plane];
+		if (pstate->fb->pixel_format == DRM_FORMAT_NV12) {
+			y_data_rate = skl_plane_relative_data_rate(cstate,
+								   pstate,
+								   1);
+			y_plane_blocks = y_minimum[id];
 			y_plane_blocks += div_u64((uint64_t)alloc_size * y_data_rate,
 						total_data_rate);
 
-			ddb->y_plane[pipe][plane].start = start;
-			ddb->y_plane[pipe][plane].end = start + y_plane_blocks;
+			ddb->y_plane[pipe][id].start = start;
+			ddb->y_plane[pipe][id].end = start + y_plane_blocks;
 
 			start += y_plane_blocks;
 		}
@@ -3146,73 +3184,16 @@ static void skl_compute_wm_global_parameters(struct drm_device *dev,
 	}
 }
 
-static void skl_compute_wm_pipe_parameters(struct drm_crtc *crtc,
-					   struct skl_pipe_wm_parameters *p)
-{
-	struct drm_device *dev = crtc->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum pipe pipe = intel_crtc->pipe;
-	struct drm_plane *plane;
-	struct drm_framebuffer *fb;
-	int i = 1; /* Index for sprite planes start */
-
-	p->active = intel_crtc->active;
-	if (p->active) {
-		p->pipe_htotal = intel_crtc->config->base.adjusted_mode.crtc_htotal;
-		p->pixel_rate = skl_pipe_pixel_rate(intel_crtc->config);
-
-		fb = crtc->primary->state->fb;
-		/* For planar: Bpp is for uv plane, y_Bpp is for y plane */
-		if (fb) {
-			p->plane[0].enabled = true;
-			p->plane[0].bytes_per_pixel = fb->pixel_format == DRM_FORMAT_NV12 ?
-				drm_format_plane_cpp(fb->pixel_format, 1) :
-				drm_format_plane_cpp(fb->pixel_format, 0);
-			p->plane[0].y_bytes_per_pixel = fb->pixel_format == DRM_FORMAT_NV12 ?
-				drm_format_plane_cpp(fb->pixel_format, 0) : 0;
-			p->plane[0].tiling = fb->modifier[0];
-		} else {
-			p->plane[0].enabled = false;
-			p->plane[0].bytes_per_pixel = 0;
-			p->plane[0].y_bytes_per_pixel = 0;
-			p->plane[0].tiling = DRM_FORMAT_MOD_NONE;
-		}
-		p->plane[0].horiz_pixels = intel_crtc->config->pipe_src_w;
-		p->plane[0].vert_pixels = intel_crtc->config->pipe_src_h;
-		p->plane[0].rotation = crtc->primary->state->rotation;
-
-		fb = crtc->cursor->state->fb;
-		p->plane[PLANE_CURSOR].y_bytes_per_pixel = 0;
-		if (fb) {
-			p->plane[PLANE_CURSOR].enabled = true;
-			p->plane[PLANE_CURSOR].bytes_per_pixel = fb->bits_per_pixel / 8;
-			p->plane[PLANE_CURSOR].horiz_pixels = crtc->cursor->state->crtc_w;
-			p->plane[PLANE_CURSOR].vert_pixels = crtc->cursor->state->crtc_h;
-		} else {
-			p->plane[PLANE_CURSOR].enabled = false;
-			p->plane[PLANE_CURSOR].bytes_per_pixel = 0;
-			p->plane[PLANE_CURSOR].horiz_pixels = 64;
-			p->plane[PLANE_CURSOR].vert_pixels = 64;
-		}
-	}
-
-	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
-		struct intel_plane *intel_plane = to_intel_plane(plane);
-
-		if (intel_plane->pipe == pipe &&
-			plane->type == DRM_PLANE_TYPE_OVERLAY)
-			p->plane[i++] = intel_plane->wm;
-	}
-}
-
 static bool skl_compute_plane_wm(const struct drm_i915_private *dev_priv,
-				 struct skl_pipe_wm_parameters *p,
-				 struct intel_plane_wm_parameters *p_params,
+				 struct intel_crtc_state *cstate,
+				 struct intel_plane *intel_plane,
 				 uint16_t ddb_allocation,
 				 int level,
 				 uint16_t *out_blocks, /* out */
 				 uint8_t *out_lines /* out */)
 {
+	struct drm_plane *plane = &intel_plane->base;
+	struct drm_framebuffer *fb = plane->state->fb;
 	uint32_t latency = dev_priv->wm.skl_latency[level];
 	uint32_t method1, method2;
 	uint32_t plane_bytes_per_line, plane_blocks_per_line;
@@ -3220,31 +3201,35 @@ static bool skl_compute_plane_wm(const struct drm_i915_private *dev_priv,
 	uint32_t selected_result;
 	uint8_t bytes_per_pixel;
 
-	if (latency == 0 || !p->active || !p_params->enabled)
+	if (latency == 0 || !cstate->base.active || !fb)
 		return false;
 
-	bytes_per_pixel = p_params->y_bytes_per_pixel ?
-		p_params->y_bytes_per_pixel :
-		p_params->bytes_per_pixel;
-	method1 = skl_wm_method1(p->pixel_rate,
+	bytes_per_pixel = (fb->pixel_format == DRM_FORMAT_NV12) ?
+		drm_format_plane_cpp(DRM_FORMAT_NV12, 0) :
+		drm_format_plane_cpp(DRM_FORMAT_NV12, 1);
+	method1 = skl_wm_method1(skl_pipe_pixel_rate(cstate),
 				 bytes_per_pixel,
 				 latency);
-	method2 = skl_wm_method2(p->pixel_rate,
-				 p->pipe_htotal,
-				 p_params->horiz_pixels,
+	method2 = skl_wm_method2(skl_pipe_pixel_rate(cstate),
+				 cstate->base.adjusted_mode.crtc_htotal,
+				 cstate->pipe_src_w,
 				 bytes_per_pixel,
-				 p_params->tiling,
+				 fb->modifier[0],
 				 latency);
 
-	plane_bytes_per_line = p_params->horiz_pixels * bytes_per_pixel;
+	plane_bytes_per_line = cstate->pipe_src_w * bytes_per_pixel;
 	plane_blocks_per_line = DIV_ROUND_UP(plane_bytes_per_line, 512);
 
-	if (p_params->tiling == I915_FORMAT_MOD_Y_TILED ||
-	    p_params->tiling == I915_FORMAT_MOD_Yf_TILED) {
+	if (fb->modifier[0] == I915_FORMAT_MOD_Y_TILED ||
+	    fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED) {
 		uint32_t min_scanlines = 4;
 		uint32_t y_tile_minimum;
-		if (intel_rotation_90_or_270(p_params->rotation)) {
-			switch (p_params->bytes_per_pixel) {
+		if (intel_rotation_90_or_270(plane->state->rotation)) {
+			int bpp = (fb->pixel_format == DRM_FORMAT_NV12) ?
+				drm_format_plane_cpp(fb->pixel_format, 1) :
+				drm_format_plane_cpp(fb->pixel_format, 0);
+
+			switch (bpp) {
 			case 1:
 				min_scanlines = 16;
 				break;
@@ -3268,8 +3253,8 @@ static bool skl_compute_plane_wm(const struct drm_i915_private *dev_priv,
 	res_lines = DIV_ROUND_UP(selected_result, plane_blocks_per_line);
 
 	if (level >= 1 && level <= 7) {
-		if (p_params->tiling == I915_FORMAT_MOD_Y_TILED ||
-		    p_params->tiling == I915_FORMAT_MOD_Yf_TILED)
+		if (fb->modifier[0] == I915_FORMAT_MOD_Y_TILED ||
+		    fb->modifier[0] == I915_FORMAT_MOD_Yf_TILED)
 			res_lines += 4;
 		else
 			res_blocks++;
@@ -3286,84 +3271,80 @@ static bool skl_compute_plane_wm(const struct drm_i915_private *dev_priv,
 
 static void skl_compute_wm_level(const struct drm_i915_private *dev_priv,
 				 struct skl_ddb_allocation *ddb,
-				 struct skl_pipe_wm_parameters *p,
-				 enum pipe pipe,
+				 struct intel_crtc_state *cstate,
 				 int level,
-				 int num_planes,
 				 struct skl_wm_level *result)
 {
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
+	struct intel_plane *intel_plane;
 	uint16_t ddb_blocks;
-	int i;
+	enum pipe pipe = intel_crtc->pipe;
+
+	for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) {
+		int i = skl_wm_plane_id(intel_plane);
 
-	for (i = 0; i < num_planes; i++) {
 		ddb_blocks = skl_ddb_entry_size(&ddb->plane[pipe][i]);
 
 		result->plane_en[i] = skl_compute_plane_wm(dev_priv,
-						p, &p->plane[i],
+						cstate,
+						intel_plane,
 						ddb_blocks,
 						level,
 						&result->plane_res_b[i],
 						&result->plane_res_l[i]);
 	}
-
-	ddb_blocks = skl_ddb_entry_size(&ddb->plane[pipe][PLANE_CURSOR]);
-	result->plane_en[PLANE_CURSOR] = skl_compute_plane_wm(dev_priv, p,
-						 &p->plane[PLANE_CURSOR],
-						 ddb_blocks, level,
-						 &result->plane_res_b[PLANE_CURSOR],
-						 &result->plane_res_l[PLANE_CURSOR]);
 }
 
 static uint32_t
-skl_compute_linetime_wm(struct drm_crtc *crtc, struct skl_pipe_wm_parameters *p)
+skl_compute_linetime_wm(struct intel_crtc_state *cstate)
 {
-	if (!to_intel_crtc(crtc)->active)
+	if (!cstate->base.active)
 		return 0;
 
-	if (WARN_ON(p->pixel_rate == 0))
+	if (WARN_ON(skl_pipe_pixel_rate(cstate) == 0))
 		return 0;
 
-	return DIV_ROUND_UP(8 * p->pipe_htotal * 1000, p->pixel_rate);
+	return DIV_ROUND_UP(8 * cstate->base.adjusted_mode.crtc_htotal * 1000,
+			    skl_pipe_pixel_rate(cstate));
 }
 
-static void skl_compute_transition_wm(struct drm_crtc *crtc,
-				      struct skl_pipe_wm_parameters *params,
+static void skl_compute_transition_wm(struct intel_crtc_state *cstate,
 				      struct skl_wm_level *trans_wm /* out */)
 {
+	struct drm_crtc *crtc = cstate->base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int i;
+	struct intel_plane *intel_plane;
 
-	if (!params->active)
+	if (!cstate->base.active)
 		return;
 
 	/* Until we know more, just disable transition WMs */
-	for (i = 0; i < intel_num_planes(intel_crtc); i++)
+	for_each_intel_plane_on_crtc(crtc->dev, intel_crtc, intel_plane) {
+		int i = skl_wm_plane_id(intel_plane);
+
 		trans_wm->plane_en[i] = false;
-	trans_wm->plane_en[PLANE_CURSOR] = false;
+	}
 }
 
-static void skl_compute_pipe_wm(struct drm_crtc *crtc,
+static void skl_compute_pipe_wm(struct intel_crtc_state *cstate,
 				struct skl_ddb_allocation *ddb,
-				struct skl_pipe_wm_parameters *params,
 				struct skl_pipe_wm *pipe_wm)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = cstate->base.crtc->dev;
 	const struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int level, max_level = ilk_wm_max_level(dev);
 
 	for (level = 0; level <= max_level; level++) {
-		skl_compute_wm_level(dev_priv, ddb, params, intel_crtc->pipe,
-				     level, intel_num_planes(intel_crtc),
-				     &pipe_wm->wm[level]);
+		skl_compute_wm_level(dev_priv, ddb, cstate,
+				     level, &pipe_wm->wm[level]);
 	}
-	pipe_wm->linetime = skl_compute_linetime_wm(crtc, params);
+	pipe_wm->linetime = skl_compute_linetime_wm(cstate);
 
-	skl_compute_transition_wm(crtc, params, &pipe_wm->trans_wm);
+	skl_compute_transition_wm(cstate, &pipe_wm->trans_wm);
 }
 
 static void skl_compute_wm_results(struct drm_device *dev,
-				   struct skl_pipe_wm_parameters *p,
 				   struct skl_pipe_wm *p_wm,
 				   struct skl_wm_values *r,
 				   struct intel_crtc *intel_crtc)
@@ -3607,16 +3588,15 @@ static void skl_flush_wm_values(struct drm_i915_private *dev_priv,
 }
 
 static bool skl_update_pipe_wm(struct drm_crtc *crtc,
-			       struct skl_pipe_wm_parameters *params,
 			       struct intel_wm_config *config,
 			       struct skl_ddb_allocation *ddb, /* out */
 			       struct skl_pipe_wm *pipe_wm /* out */)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
 
-	skl_compute_wm_pipe_parameters(crtc, params);
-	skl_allocate_pipe_ddb(crtc, config, params, ddb);
-	skl_compute_pipe_wm(crtc, ddb, params, pipe_wm);
+	skl_allocate_pipe_ddb(cstate, config, ddb);
+	skl_compute_pipe_wm(cstate, ddb, pipe_wm);
 
 	if (!memcmp(&intel_crtc->wm.skl_active, pipe_wm, sizeof(*pipe_wm)))
 		return false;
@@ -3649,7 +3629,6 @@ static void skl_update_other_pipe_wm(struct drm_device *dev,
 	 */
 	list_for_each_entry(intel_crtc, &dev->mode_config.crtc_list,
 				base.head) {
-		struct skl_pipe_wm_parameters params = {};
 		struct skl_pipe_wm pipe_wm = {};
 		bool wm_changed;
 
@@ -3659,8 +3638,7 @@ static void skl_update_other_pipe_wm(struct drm_device *dev,
 		if (!intel_crtc->active)
 			continue;
 
-		wm_changed = skl_update_pipe_wm(&intel_crtc->base,
-						&params, config,
+		wm_changed = skl_update_pipe_wm(&intel_crtc->base, config,
 						&r->ddb, &pipe_wm);
 
 		/*
@@ -3670,7 +3648,7 @@ static void skl_update_other_pipe_wm(struct drm_device *dev,
 		 */
 		WARN_ON(!wm_changed);
 
-		skl_compute_wm_results(dev, &params, &pipe_wm, r, intel_crtc);
+		skl_compute_wm_results(dev, &pipe_wm, r, intel_crtc);
 		r->dirty[intel_crtc->pipe] = true;
 	}
 }
@@ -3680,7 +3658,6 @@ static void skl_update_wm(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct skl_pipe_wm_parameters params = {};
 	struct skl_wm_values *results = &dev_priv->wm.skl_results;
 	struct skl_pipe_wm pipe_wm = {};
 	struct intel_wm_config config = {};
@@ -3689,11 +3666,10 @@ static void skl_update_wm(struct drm_crtc *crtc)
 
 	skl_compute_wm_global_parameters(dev, &config);
 
-	if (!skl_update_pipe_wm(crtc, &params, &config,
-				&results->ddb, &pipe_wm))
+	if (!skl_update_pipe_wm(crtc, &config, &results->ddb, &pipe_wm))
 		return;
 
-	skl_compute_wm_results(dev, &params, &pipe_wm, results, intel_crtc);
+	skl_compute_wm_results(dev, &pipe_wm, results, intel_crtc);
 	results->dirty[intel_crtc->pipe] = true;
 
 	skl_update_other_pipe_wm(dev, crtc, &config, results);
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v1 6/14
#7 lore raw

Re: [PATCH 06/14] drm/i915/skl: Eliminate usage of pipe_wm_parameters from SKL-style WM (v2)

Hey,

Op 15-09-15 om 04:19 schreef Matt Roper:

Quoted 46 lines

Just pull the info out of the state structures rather than staging it in an additional set of structures. To make this more straightforward, we change the signature of several internal WM functions to take the crtc state as a parameter.

v2:
- Don't forget to skip cursor planes on a loop in the DDB allocation function to match original behavior. (Ander)
- Change a use of intel_crtc->active to cstate->active. They should be identical, but it's better to be consistent. (Ander)
- Rework more function signatures to pass states rather than crtc for consistency. (Ander)

---
 drivers/gpu/drm/i915/intel_pm.c | 326 +++++++++++++++++++---------------------
 1 file changed, 151 insertions(+), 175 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +6 -7
index 7115368..51dbda7 100644
@@ -1787,13 +1787,6 @@ static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels,
 	return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2;
 }
 
-struct skl_pipe_wm_parameters {
-	bool active;
-	uint32_t pipe_htotal;
-	uint32_t pixel_rate; /* in KHz */
-	struct intel_plane_wm_parameters plane[I915_MAX_PLANES];
-};
-
 struct ilk_wm_maximums {
 	uint16_t pri;
 	uint16_t spr;
@@ -2834,18 +2827,40 @@ static bool ilk_disable_lp_wm(struct drm_device *dev)
 #define SKL_DDB_SIZE		896	/* in blocks */
 #define BXT_DDB_SIZE		512
 
+/*
+ * Return the index of a plane in the SKL DDB and wm result arrays.  Primary
+ * plane is always in slot 0, cursor is always in slot I915_MAX_PLANES-1, and
+ * other universal planes are in indices 1..n.  Note that this may leave unused
+ * indices between the top "sprite" plane and the cursor.
+ */

drivers/gpu/drm/i915/intel_pm.c static bool ilk_disable_lp_wm(struct drm_device *dev) · line 2845

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+static int
+skl_wm_plane_id(const struct intel_plane *plane)
+{
+	switch (plane->base.type) {
+	case DRM_PLANE_TYPE_PRIMARY:
+		return 0;
+	case DRM_PLANE_TYPE_CURSOR:
+		return PLANE_CURSOR;
+	case DRM_PLANE_TYPE_OVERLAY:
+		return plane->plane;

Shouldn't this be plane + 1 on skylake? Though it's not the only code in i915 that gets it wrong, assert_sprites_disabled has the same issue on SKL.. O:-)

Quoted 46 lines
drivers/gpu/drm/i915/intel_pm.c +18 -7
+	default:
+		MISSING_CASE(plane->base.type);
+		return plane->plane;
+	}
+}
+
 static void
 skl_ddb_get_pipe_allocation_limits(struct drm_device *dev,
-				   struct drm_crtc *for_crtc,
+				   const struct intel_crtc_state *cstate,
 				   const struct intel_wm_config *config,
-				   const struct skl_pipe_wm_parameters *params,
 				   struct skl_ddb_entry *alloc /* out */)
 {
+	struct drm_crtc *for_crtc = cstate->base.crtc;
 	struct drm_crtc *crtc;
 	unsigned int pipe_size, ddb_size;
 	int nth_active_pipe;
 
-	if (!params->active) {
+	if (!cstate->base.active) {
 		alloc->start = 0;
 		alloc->end = 0;
 		return;
@@ -2911,19 +2926,29 @@ void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
 }
 
 static unsigned int
-skl_plane_relative_data_rate(const struct intel_plane_wm_parameters *p, int y)
+skl_plane_relative_data_rate(const struct intel_crtc_state *cstate,
+			     const struct drm_plane_state *pstate,
+			     int y)
 {
+	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
+	struct drm_framebuffer *fb = pstate->fb;
 
 	/* for planar format */
-	if (p->y_bytes_per_pixel) {
+	if (fb->pixel_format == DRM_FORMAT_NV12) {
 		if (y)  /* y-plane data rate */
-			return p->horiz_pixels * p->vert_pixels * p->y_bytes_per_pixel;
+			return intel_crtc->config->pipe_src_w *
+				intel_crtc->config->pipe_src_h *
+				drm_format_plane_cpp(fb->pixel_format, 0);
 		else    /* uv-plane data rate */
-			return (p->horiz_pixels/2) * (p->vert_pixels/2) * p->bytes_per_pixel;

drivers/gpu/drm/i915/intel_pm.c void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv, · line 2951

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +6 -1
+			return (intel_crtc->config->pipe_src_w/2) *
+				(intel_crtc->config->pipe_src_h/2) *
+				drm_format_plane_cpp(fb->pixel_format, 1);
 	}
 
 	/* for packed formats */
-	return p->horiz_pixels * p->vert_pixels * p->bytes_per_pixel;
+	return intel_crtc->config->pipe_src_w *
+		intel_crtc->config->pipe_src_h *
+		drm_format_plane_cpp(fb->pixel_format, 1);

^Gives a nasty divide by zero error on my skl when total_data_rate ends up being 0.. You want to change last argument to 0 here. ;-)

Quoted 16 lines
drivers/gpu/drm/i915/intel_pm.c +4 -3
 }
 
 /*
@@ -2932,46 +2957,51 @@ skl_plane_relative_data_rate(const struct intel_plane_wm_parameters *p, int y)
  *   3 * 4096 * 8192  * 4 < 2^32
  */
 static unsigned int
-skl_get_total_relative_data_rate(struct intel_crtc *intel_crtc,
-				 const struct skl_pipe_wm_parameters *params)
+skl_get_total_relative_data_rate(const struct intel_crtc_state *cstate)
 {
+	struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc);
+	struct drm_device *dev = intel_crtc->base.dev;
+	const struct intel_plane *intel_plane;
 	unsigned int total_data_rate = 0;
-	int plane;

drivers/gpu/drm/i915/intel_pm.c skl_plane_relative_data_rate(const struct intel_plane_wm_parameters *p, int y) · line 2971

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +3 -4
 
-	for (plane = 0; plane < intel_num_planes(intel_crtc); plane++) {
-		const struct intel_plane_wm_parameters *p;
+	for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) {
+		const struct drm_plane_state *pstate = intel_plane->base.state;
 
-		p = &params->plane[plane];
-		if (!p->enabled)
+		if (WARN_ON(pstate->fb == NULL))
 			continue;

^Why is this a WARN_ON?

All in all, the series looks good.

With those fixes feel free to add this to all patches:

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

author
#8 lore raw

PATCH 7/14 drm/i915/ivb: Move WaCxSRDisabledForSpriteScaling w/a to atomic check

In reply to Matt Roper

Determine whether we need to apply this workaround at atomic check time and just set a flag that will be used by the main watermark update routine.

Moving this workaround into the atomic framework reduces ilk_update_sprite_wm() to just a standard watermark update, so drop it completely and just ensure that ilk_update_wm() is called whenever a sprite plane is updated in a way that would affect watermarks.

---
 drivers/gpu/drm/i915/intel_atomic.c  |  1 +
 drivers/gpu/drm/i915/intel_display.c | 39 +++++++++++++++++++++++++++++-------
 drivers/gpu/drm/i915/intel_drv.h     |  3 +++
 drivers/gpu/drm/i915/intel_pm.c      | 35 +++++++++++---------------------
 4 files changed, 48 insertions(+), 30 deletions(-)
drivers/gpu/drm/i915/intel_atomic.c +1 -0
index f1975f2..05b1203 100644
@@ -94,6 +94,7 @@ intel_crtc_duplicate_state(struct drm_crtc *crtc)
 	__drm_atomic_helper_crtc_duplicate_state(crtc, &crtc_state->base);
 
 	crtc_state->update_pipe = false;
+	crtc_state->disable_lp_wm = false;
 
 	return &crtc_state->base;
 }
drivers/gpu/drm/i915/intel_display.c +32 -7
index 017641a..daa0ede 100644
@@ -11514,18 +11514,32 @@ retry:
 static bool intel_wm_need_update(struct drm_plane *plane,
 				 struct drm_plane_state *state)
 {
-	/* Update watermarks on tiling changes. */
+	struct intel_plane_state *new = to_intel_plane_state(state);
+	struct intel_plane_state *cur = to_intel_plane_state(plane->state);
+
+	/* Update watermarks on tiling or size changes. */
 	if (!plane->state->fb || !state->fb ||
 	    plane->state->fb->modifier[0] != state->fb->modifier[0] ||
-	    plane->state->rotation != state->rotation)
-		return true;
-
-	if (plane->state->crtc_w != state->crtc_w)
+	    plane->state->rotation != state->rotation ||
+	    drm_rect_width(&new->src) != drm_rect_width(&cur->src) ||
+	    drm_rect_height(&new->src) != drm_rect_height(&cur->src) ||
+	    drm_rect_width(&new->dst) != drm_rect_width(&cur->dst) ||
+	    drm_rect_height(&new->dst) != drm_rect_height(&cur->dst))
 		return true;
 
 	return false;
 }
 
+static bool needs_scaling(struct intel_plane_state *state)
+{
+	int src_w = drm_rect_width(&state->src) >> 16;
+	int src_h = drm_rect_height(&state->src) >> 16;
+	int dst_w = drm_rect_width(&state->dst);
+	int dst_h = drm_rect_height(&state->dst);
+
+	return (src_w != dst_w || src_h != dst_h);
+}
+
 int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 				    struct drm_plane_state *plane_state)
 {
@@ -11541,7 +11555,6 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	bool mode_changed = needs_modeset(crtc_state);
 	bool was_crtc_enabled = crtc->state->active;
 	bool is_crtc_enabled = crtc_state->active;
-
 	bool turn_off, turn_on, visible, was_visible;
 	struct drm_framebuffer *fb = plane_state->fb;
 
@@ -11659,11 +11672,23 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	case DRM_PLANE_TYPE_CURSOR:
 		break;
 	case DRM_PLANE_TYPE_OVERLAY:
-		if (turn_off && !mode_changed) {
+		/*
+		 * WaCxSRDisabledForSpriteScaling:ivb
+		 *
+		 * cstate->update_wm was already set above, so this flag will
+		 * take effect when we commit and program watermarks.
+		 */
+		if (IS_IVYBRIDGE(dev) &&
+		    needs_scaling(to_intel_plane_state(plane_state)) &&
+		    !needs_scaling(old_plane_state)) {
+			to_intel_crtc_state(crtc_state)->disable_lp_wm = true;
+		} else if (turn_off && !mode_changed) {
 			intel_crtc->atomic.wait_vblank = true;
 			intel_crtc->atomic.update_sprite_watermarks |=
 				1 << i;
 		}
+
+		break;
 	}
 	return 0;
 }
drivers/gpu/drm/i915/intel_drv.h +3 -0
index 02a755a..3fa2e66 100644
@@ -458,6 +458,9 @@ struct intel_crtc_state {
 
 	/* w/a for waiting 2 vblanks during crtc enable */
 	enum pipe hsw_workaround_pipe;
+
+	/* IVB sprite scaling w/a (WaCxSRDisabledForSpriteScaling:ivb) */
+	bool disable_lp_wm;
 };
 
 struct vlv_wm_state {
drivers/gpu/drm/i915/intel_pm.c +12 -23
index 51dbda7..d601630 100644
@@ -3728,6 +3728,18 @@ static void ilk_update_wm(struct drm_crtc *crtc)
 
 	WARN_ON(cstate->base.active != intel_crtc->active);
 
+	/*
+	 * IVB workaround: must disable low power watermarks for at least
+	 * one frame before enabling scaling.  LP watermarks can be re-enabled
+	 * when scaling is disabled.
+	 *
+	 * WaCxSRDisabledForSpriteScaling:ivb
+	 */
+	if (cstate->disable_lp_wm) {
+		ilk_disable_lp_wm(dev);
+		intel_wait_for_vblank(dev, intel_crtc->pipe);
+	}
+
 	intel_compute_pipe_wm(cstate, &pipe_wm);
 
 	if (!memcmp(&intel_crtc->wm.active, &pipe_wm, sizeof(pipe_wm)))
@@ -3759,28 +3771,6 @@ static void ilk_update_wm(struct drm_crtc *crtc)
 	ilk_write_wm_values(dev_priv, &results);
 }
 
-static void
-ilk_update_sprite_wm(struct drm_plane *plane,
-		     struct drm_crtc *crtc,
-		     uint32_t sprite_width, uint32_t sprite_height,
-		     int pixel_size, bool enabled, bool scaled)
-{
-	struct drm_device *dev = plane->dev;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-
-	/*
-	 * IVB workaround: must disable low power watermarks for at least
-	 * one frame before enabling scaling.  LP watermarks can be re-enabled
-	 * when scaling is disabled.
-	 *
-	 * WaCxSRDisabledForSpriteScaling:ivb
-	 */
-	if (IS_IVYBRIDGE(dev) && scaled && ilk_disable_lp_wm(dev))
-		intel_wait_for_vblank(dev, intel_plane->pipe);
-
-	ilk_update_wm(crtc);
-}
-
 static void skl_pipe_wm_active_state(uint32_t val,
 				     struct skl_pipe_wm *active,
 				     bool is_transwm,
@@ -7091,7 +7081,6 @@ void intel_init_pm(struct drm_device *dev)
 		    (!IS_GEN5(dev) && dev_priv->wm.pri_latency[0] &&
 		     dev_priv->wm.spr_latency[0] && dev_priv->wm.cur_latency[0])) {
 			dev_priv->display.update_wm = ilk_update_wm;
-			dev_priv->display.update_sprite_wm = ilk_update_sprite_wm;
 		} else {
 			DRM_DEBUG_KMS("Failed to read display plane latency. "
 				      "Disable CxSR\n");
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#9 lore raw

PATCH 8/14 drm/i915: Drop intel_update_sprite_watermarks

In reply to Matt Roper

The only platform that still has an update_sprite_wm entrypoint is SKL; on SKL, intel_update_sprite_watermarks just updates intel_plane->wm and then performs a regular watermark update. However intel_plane->wm is only used to update a couple fields in intel_wm_config, and those fields are never used by the SKL code, so on SKL an update_sprite_wm is effectively identical to an update_wm call. Since we're already ensuring that the regular intel_update_wm is called any time we'd try to call intel_update_sprite_watermarks, the whole call is redundant and can be dropped.

---
 drivers/gpu/drm/i915/i915_drv.h      |  4 ---
 drivers/gpu/drm/i915/intel_display.c |  5 ----
 drivers/gpu/drm/i915/intel_drv.h     |  6 ----
 drivers/gpu/drm/i915/intel_pm.c      | 58 ------------------------------------
 drivers/gpu/drm/i915/intel_sprite.c  | 15 ----------
 5 files changed, 88 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +0 -4
index 539bfa2..b197d41 100644
@@ -628,10 +628,6 @@ struct drm_i915_display_funcs {
 			  struct dpll *match_clock,
 			  struct dpll *best_clock);
 	void (*update_wm)(struct drm_crtc *crtc);
-	void (*update_sprite_wm)(struct drm_plane *plane,
-				 struct drm_crtc *crtc,
-				 uint32_t sprite_width, uint32_t sprite_height,
-				 int pixel_size, bool enable, bool scaled);
 	int (*modeset_calc_cdclk)(struct drm_atomic_state *state);
 	void (*modeset_commit_cdclk)(struct drm_atomic_state *state);
 	/* Returns the active state of the crtc, and if the crtc is active,
drivers/gpu/drm/i915/intel_display.c +0 -5
index daa0ede..96ceed7 100644
@@ -4721,7 +4721,6 @@ static void intel_post_plane_update(struct intel_crtc *crtc)
 	struct intel_crtc_atomic_commit *atomic = &crtc->atomic;
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_plane *plane;
 
 	if (atomic->wait_vblank)
 		intel_wait_for_vblank(dev, crtc->pipe);
@@ -4740,10 +4739,6 @@ static void intel_post_plane_update(struct intel_crtc *crtc)
 	if (atomic->post_enable_primary)
 		intel_post_enable_primary(&crtc->base);
 
-	drm_for_each_plane_mask(plane, dev, atomic->update_sprite_watermarks)
-		intel_update_sprite_watermarks(plane, &crtc->base,
-					       0, 0, 0, false, false);
-
 	memset(atomic, 0, sizeof(*atomic));
 }
 
drivers/gpu/drm/i915/intel_drv.h +0 -6
index 3fa2e66..9885b25 100644
@@ -1380,12 +1380,6 @@ void intel_init_clock_gating(struct drm_device *dev);
 void intel_suspend_hw(struct drm_device *dev);
 int ilk_wm_max_level(const struct drm_device *dev);
 void intel_update_watermarks(struct drm_crtc *crtc);
-void intel_update_sprite_watermarks(struct drm_plane *plane,
-				    struct drm_crtc *crtc,
-				    uint32_t sprite_width,
-				    uint32_t sprite_height,
-				    int pixel_size,
-				    bool enabled, bool scaled);
 void intel_init_pm(struct drm_device *dev);
 void intel_pm_setup(struct drm_device *dev);
 void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
drivers/gpu/drm/i915/intel_pm.c +0 -58

85 lines not on this page · raw message

drivers/gpu/drm/i915/intel_sprite.c +0 -15
index 4d27243..fb096e5 100644
@@ -182,7 +182,6 @@ skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc,
 	const int pipe = intel_plane->pipe;
 	const int plane = intel_plane->plane + 1;
 	u32 plane_ctl, stride_div, stride;
-	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
 	const struct drm_intel_sprite_colorkey *key =
 		&to_intel_plane_state(drm_plane->state)->ckey;
 	unsigned long surf_addr;
@@ -201,10 +200,6 @@ skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc,
 	rotation = drm_plane->state->rotation;
 	plane_ctl |= skl_plane_ctl_rotation(rotation);
 
-	intel_update_sprite_watermarks(drm_plane, crtc, src_w, src_h,
-				       pixel_size, true,
-				       src_w != crtc_w || src_h != crtc_h);
-
 	stride_div = intel_fb_stride_alignment(dev, fb->modifier[0],
 					       fb->pixel_format);
 
@@ -286,8 +281,6 @@ skl_disable_plane(struct drm_plane *dplane, struct drm_crtc *crtc)
 
 	I915_WRITE(PLANE_SURF(pipe, plane), 0);
 	POSTING_READ(PLANE_SURF(pipe, plane));
-
-	intel_update_sprite_watermarks(dplane, crtc, 0, 0, 0, false, false);
 }
 
 static void
@@ -530,10 +523,6 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (IS_HASWELL(dev) || IS_BROADWELL(dev))
 		sprctl |= SPRITE_PIPE_CSC_ENABLE;
 
-	intel_update_sprite_watermarks(plane, crtc, src_w, src_h, pixel_size,
-				       true,
-				       src_w != crtc_w || src_h != crtc_h);
-
 	/* Sizes are 0 based */
 	src_w--;
 	src_h--;
@@ -667,10 +656,6 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (IS_GEN6(dev))
 		dvscntr |= DVS_TRICKLE_FEED_DISABLE; /* must disable */
 
-	intel_update_sprite_watermarks(plane, crtc, src_w, src_h,
-				       pixel_size, true,
-				       src_w != crtc_w || src_h != crtc_h);
-
 	/* Sizes are 0 based */
 	src_w--;
 	src_h--;
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#10 lore raw

PATCH 9/14 drm/i915: Refactor ilk_update_wm (v3)

In reply to Matt Roper

From: Ville Syrjälä <[email protected]>

Split ilk_update_wm() into two parts; one doing the programming and the other the calculations.

v2: Fix typo in commit message

v3 (by Matt): Heavily rebased for current codebase.

---
 drivers/gpu/drm/i915/intel_pm.c | 64 ++++++++++++++++++++++-------------------
 1 file changed, 35 insertions(+), 29 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +35 -29
index 3f9e0e2..20d91bd 100644
@@ -3671,39 +3671,14 @@ static void skl_update_wm(struct drm_crtc *crtc)
 	dev_priv->wm.skl_hw = *results;
 }
 
-static void ilk_update_wm(struct drm_crtc *crtc)
+static void ilk_program_watermarks(struct drm_i915_private *dev_priv)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm;
 	struct ilk_wm_maximums max;
+	struct intel_wm_config config = {};
 	struct ilk_wm_values results = {};
 	enum intel_ddb_partitioning partitioning;
-	struct intel_pipe_wm pipe_wm = {};
-	struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm;
-	struct intel_wm_config config = {};
-
-	WARN_ON(cstate->base.active != intel_crtc->active);
-
-	/*
-	 * IVB workaround: must disable low power watermarks for at least
-	 * one frame before enabling scaling.  LP watermarks can be re-enabled
-	 * when scaling is disabled.
-	 *
-	 * WaCxSRDisabledForSpriteScaling:ivb
-	 */
-	if (cstate->disable_lp_wm) {
-		ilk_disable_lp_wm(dev);
-		intel_wait_for_vblank(dev, intel_crtc->pipe);
-	}
-
-	intel_compute_pipe_wm(cstate, &pipe_wm);
-
-	if (!memcmp(&intel_crtc->wm.active, &pipe_wm, sizeof(pipe_wm)))
-		return;
-
-	intel_crtc->wm.active = pipe_wm;
 
 	ilk_compute_wm_config(dev, &config);
 
@@ -3729,6 +3704,37 @@ static void ilk_update_wm(struct drm_crtc *crtc)
 	ilk_write_wm_values(dev_priv, &results);
 }
 
+static void ilk_update_wm(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state);
+	struct intel_pipe_wm pipe_wm = {};
+
+	WARN_ON(cstate->base.active != intel_crtc->active);
+
+	/*
+	 * IVB workaround: must disable low power watermarks for at least
+	 * one frame before enabling scaling.  LP watermarks can be re-enabled
+	 * when scaling is disabled.
+	 *
+	 * WaCxSRDisabledForSpriteScaling:ivb
+	 */
+	if (cstate->disable_lp_wm) {
+		ilk_disable_lp_wm(crtc->dev);
+		intel_wait_for_vblank(crtc->dev, intel_crtc->pipe);
+	}
+
+	intel_compute_pipe_wm(cstate, &pipe_wm);
+
+	if (!memcmp(&intel_crtc->wm.active, &pipe_wm, sizeof(pipe_wm)))
+		return;
+
+	intel_crtc->wm.active = pipe_wm;
+
+	ilk_program_watermarks(dev_priv);
+}
+
 static void skl_pipe_wm_active_state(uint32_t val,
 				     struct skl_pipe_wm *active,
 				     bool is_transwm,
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#11 lore raw

PATCH 10/14 drm/i915: Calculate pipe watermarks into CRTC state (v3)

In reply to Matt Roper

A future patch will calculate these during the atomic 'check' phase rather than at WM programming time, so let's store the watermark values we're planning to use in the CRTC state; the values actually active on the hardware remains in intel_crtc.

While we're at it, do some minor restructuring to keep ILK and SKL values in a union.

v2: Don't move cxsr_allowed to state (Maarten)

v3: Only calculate watermarks in state. Still keep active watermarks in
intel_crtc itself. (Ville)

---
 drivers/gpu/drm/i915/intel_drv.h | 48 +++++++++++++++++++++++++---------------
 drivers/gpu/drm/i915/intel_pm.c  | 44 ++++++++++++++++++++++--------------
 2 files changed, 57 insertions(+), 35 deletions(-)
drivers/gpu/drm/i915/intel_drv.h +30 -18
index 9885b25..6d6b275 100644
@@ -324,6 +324,21 @@ struct intel_crtc_scaler_state {
 /* drm_mode->private_flags */
 #define I915_MODE_FLAG_INHERITED 1
 
+struct intel_pipe_wm {
+	struct intel_wm_level wm[5];
+	uint32_t linetime;
+	bool fbc_wm_enabled;
+	bool pipe_enabled;
+	bool sprites_enabled;
+	bool sprites_scaled;
+};
+
+struct skl_pipe_wm {
+	struct skl_wm_level wm[8];
+	struct skl_wm_level trans_wm;
+	uint32_t linetime;
+};
+
 struct intel_crtc_state {
 	struct drm_crtc_state base;
 
@@ -461,6 +476,17 @@ struct intel_crtc_state {
 
 	/* IVB sprite scaling w/a (WaCxSRDisabledForSpriteScaling:ivb) */
 	bool disable_lp_wm;
+
+	struct {
+		/*
+		 * optimal watermarks, programmed post-vblank when this state
+		 * is committed
+		 */
+		union {
+			struct intel_pipe_wm ilk;
+			struct skl_pipe_wm skl;
+		} optimal;
+	} wm;
 };
 
 struct vlv_wm_state {
@@ -472,15 +498,6 @@ struct vlv_wm_state {
 	bool cxsr;
 };
 
-struct intel_pipe_wm {
-	struct intel_wm_level wm[5];
-	uint32_t linetime;
-	bool fbc_wm_enabled;
-	bool pipe_enabled;
-	bool sprites_enabled;
-	bool sprites_scaled;
-};
-
 struct intel_mmio_flip {
 	struct work_struct work;
 	struct drm_i915_private *i915;
@@ -488,12 +505,6 @@ struct intel_mmio_flip {
 	struct intel_crtc *crtc;
 };
 
-struct skl_pipe_wm {
-	struct skl_wm_level wm[8];
-	struct skl_wm_level trans_wm;
-	uint32_t linetime;
-};
-
 /*
  * Tracking of operations that need to be performed at the beginning/end of an
  * atomic commit, outside the atomic section where interrupts are disabled.
@@ -559,9 +570,10 @@ struct intel_crtc {
 	/* per-pipe watermark state */
 	struct {
 		/* watermarks currently being used  */
-		struct intel_pipe_wm active;
-		/* SKL wm values currently in use */
-		struct skl_pipe_wm skl_active;
+		union {
+			struct intel_pipe_wm ilk;
+			struct skl_pipe_wm skl;
+		} active;
 		/* allow CxSR on this pipe */
 		bool cxsr_allowed;
 	} wm;
drivers/gpu/drm/i915/intel_pm.c +27 -17

127 lines not on this page · raw message

Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#12 lore raw

PATCH 11/14 drm/i915: Calculate ILK-style watermarks during atomic check (v3)

In reply to Matt Roper

Calculate pipe watermarks during atomic calculation phase, based on the contents of the atomic transaction's state structure. We still program the watermarks at the same time we did before, but the computation now happens much earlier.

While this patch isn't too exciting by itself, it paves the way for future patches. The eventual goal (which will be realized in future patches in this series) is to calculate multiple sets up watermark values up front, and then program them at different times (pre- vs post-vblank) on the platforms that need a two-step watermark update.

While we're at it, s/intel_compute_pipe_wm/ilk_compute_pipe_wm/ since this function only applies to ILK-style watermarks and we have a completely different function for SKL-style watermarks.

Note that the original code had a memcmp() in ilk_update_wm() to avoid calling ilk_program_watermarks() if the watermarks hadn't changed. This memcmp vanishes here, which means we may do some unnecessary result generation and merging in cases where watermarks didn't change, but the lower-level function ilk_write_wm_values already makes sure that we don't actually try to program the watermark registers again.

v2: Squash a few commits from the original series together; no longer
leave pre-calculated wm's in a separate temporary structure since it's easier to follow the logic if we just cut over to using the pre-calculated values directly.

v3:
- Pass intel_crtc instead of drm_crtc to .compute_pipe_wm() entrypoint and use intel_atomic_get_crtc_state() to avoid need for extra casting. (Ander)
- Drop unused intel_check_crtc() function prototype. (Ander)

---
 drivers/gpu/drm/i915/i915_drv.h      |  2 +
 drivers/gpu/drm/i915/intel_display.c |  6 +++
 drivers/gpu/drm/i915/intel_pm.c      | 88 ++++++++++++++++++------------------
 3 files changed, 51 insertions(+), 45 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +2 -0
index b197d41..2696522 100644
@@ -627,6 +627,8 @@ struct drm_i915_display_funcs {
 			  int target, int refclk,
 			  struct dpll *match_clock,
 			  struct dpll *best_clock);
+	int (*compute_pipe_wm)(struct intel_crtc *crtc,
+			       struct drm_atomic_state *state);
 	void (*update_wm)(struct drm_crtc *crtc);
 	int (*modeset_calc_cdclk)(struct drm_atomic_state *state);
 	void (*modeset_commit_cdclk)(struct drm_atomic_state *state);
drivers/gpu/drm/i915/intel_display.c +6 -0
index 96ceed7..0f513e5 100644
@@ -11768,6 +11768,12 @@ static int intel_crtc_atomic_check(struct drm_crtc *crtc,
 	}
 
 	ret = 0;
+	if (dev_priv->display.compute_pipe_wm) {
+		ret = dev_priv->display.compute_pipe_wm(intel_crtc, state);
+		if (ret)
+			return ret;
+	}
+
 	if (INTEL_INFO(dev)->gen >= 9) {
 		if (mode_changed)
 			ret = skl_update_scaler_crtc(pipe_config);
drivers/gpu/drm/i915/intel_pm.c +43 -45

163 lines not on this page · raw message

Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#13 lore raw

PATCH 12/14 drm/i915: Don't set plane visible during HW readout if CRTC is off

In reply to Matt Roper

We already ensure that pstate->visible = false when crtc->active = false during runtime programming; make sure we follow the same logic when reading out initial hardware state.

---
 drivers/gpu/drm/i915/intel_display.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_display.c +1 -1
index 0f513e5..36eee4a 100644
@@ -15108,7 +15108,7 @@ static void readout_plane_state(struct intel_crtc *crtc,
 
 		plane_state = to_intel_plane_state(p->base.state);
 
-		if (p->base.type == DRM_PLANE_TYPE_PRIMARY)
+		if (p->base.type == DRM_PLANE_TYPE_PRIMARY && active)
 			plane_state->visible = primary_get_hw_state(crtc);
 		else {
 			if (active)
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#14 lore raw

PATCH 13/14 drm/i915: Calculate watermark configuration during atomic check (v2)

In reply to Matt Roper

v2: Don't forget to actually check the cstate->active value when
tallying up the number of active CRTC's. (Ander)

---
 drivers/gpu/drm/i915/i915_drv.h      | 10 ++++++
 drivers/gpu/drm/i915/intel_display.c | 52 +++++++++++++++++++++++++++--
 drivers/gpu/drm/i915/intel_drv.h     |  1 +
 drivers/gpu/drm/i915/intel_pm.c      | 64 +++++++-----------------------------
 4 files changed, 72 insertions(+), 55 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +10 -0
index 2696522..e8ef328 100644
@@ -1699,6 +1699,13 @@ struct i915_execbuffer_params {
 	struct drm_i915_gem_request     *request;
 };
 
+/* used in computing the new watermarks state */
+struct intel_wm_config {
+	unsigned int num_pipes_active;
+	bool sprites_enabled;
+	bool sprites_scaled;
+};
+
 struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *objects;
@@ -1918,6 +1925,9 @@ struct drm_i915_private {
 		 */
 		uint16_t skl_latency[8];
 
+		/* Committed wm config */
+		struct intel_wm_config config;
+
 		/*
 		 * The skl_wm_values structure is a bit too big for stack
 		 * allocation, so we keep the staging struct where we store
drivers/gpu/drm/i915/intel_display.c +49 -3

77 lines not on this page · raw message

drivers/gpu/drm/i915/intel_drv.h +1 -0
index 6d6b275..c40b079 100644
@@ -240,6 +240,7 @@ struct intel_atomic_state {
 	unsigned int cdclk;
 	bool dpll_set;
 	struct intel_shared_dpll_config shared_dpll[I915_NUM_PLLS];
+	struct intel_wm_config wm_config;
 };
 
 struct intel_plane_state {
drivers/gpu/drm/i915/intel_pm.c +12 -52

142 lines not on this page · raw message

Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#15 lore raw

PATCH 14/14 drm/i915: Add two-stage ILK-style watermark programming (v4)

In reply to Matt Roper

In addition to calculating final watermarks, let's also pre-calculate a set of intermediate watermark values at atomic check time. These intermediate watermarks are a combination of the watermarks for the old state and the new state; they should satisfy the requirements of both states which means they can be programmed immediately when we commit the atomic state (without waiting for a vblank). Once the vblank does happen, we can then re-program watermarks to the more optimal final value.

v2: Significant rebasing/rewriting.

v3:
- Move 'need_postvbl_update' flag to CRTC state (Daniel)
- Don't forget to check intermediate watermark values for validity (Maarten)
- Don't due async watermark optimization; just do it at the end of the atomic transaction, after waiting for vblanks. We do want it to be async eventually, but adding that now will cause more trouble for Maarten's in-progress work. (Maarten)
- Don't allocate space in crtc_state for intermediate watermarks on platforms that don't need it (gen9+).
- Move WaCxSRDisabledForSpriteScaling:ivb into intel_begin_crtc_commit now that ilk_update_wm is gone.

v4:
- Add a wm_mutex to cover updates to intel_crtc->active and the need_postvbl_update flag. Since we don't have async yet it isn't terribly important yet, but might as well add it now.
- Change interface to program watermarks. Platforms will now expose .initial_watermarks() and .optimize_watermarks() functions to do watermark programming. These should lock wm_mutex, copy the appropriate state values into intel_crtc->active, and then call the internal program watermarks function.

---
 drivers/gpu/drm/i915/i915_drv.h      |   5 ++
 drivers/gpu/drm/i915/intel_atomic.c  |   1 +
 drivers/gpu/drm/i915/intel_display.c |  77 +++++++++++++++++-
 drivers/gpu/drm/i915/intel_drv.h     |  25 +++++-
 drivers/gpu/drm/i915/intel_pm.c      | 154 +++++++++++++++++++++++++----------
 5 files changed, 213 insertions(+), 49 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +5 -0
index e8ef328..6d08d47 100644
@@ -629,6 +629,11 @@ struct drm_i915_display_funcs {
 			  struct dpll *best_clock);
 	int (*compute_pipe_wm)(struct intel_crtc *crtc,
 			       struct drm_atomic_state *state);
+	int (*compute_intermediate_wm)(struct drm_device *dev,
+				       struct intel_crtc *intel_crtc,
+				       struct intel_crtc_state *newstate);
+	void (*initial_watermarks)(struct intel_crtc_state *cstate);
+	void (*optimize_watermarks)(struct intel_crtc_state *cstate);
 	void (*update_wm)(struct drm_crtc *crtc);
 	int (*modeset_calc_cdclk)(struct drm_atomic_state *state);
 	void (*modeset_commit_cdclk)(struct drm_atomic_state *state);
drivers/gpu/drm/i915/intel_atomic.c +1 -0
index 05b1203..28630f8 100644
@@ -95,6 +95,7 @@ intel_crtc_duplicate_state(struct drm_crtc *crtc)
 
 	crtc_state->update_pipe = false;
 	crtc_state->disable_lp_wm = false;
+	crtc_state->wm.need_postvbl_update = false;
 
 	return &crtc_state->base;
 }
drivers/gpu/drm/i915/intel_display.c +75 -2

120 lines not on this page · raw message

drivers/gpu/drm/i915/intel_drv.h +23 -2
index c40b079..067100d 100644
@@ -480,13 +480,29 @@ struct intel_crtc_state {
 
 	struct {
 		/*
-		 * optimal watermarks, programmed post-vblank when this state
-		 * is committed
+		 * Optimal watermarks, programmed post-vblank when this state
+		 * is committed.
 		 */
 		union {
 			struct intel_pipe_wm ilk;
 			struct skl_pipe_wm skl;
 		} optimal;
+
+		/*
+		 * Intermediate watermarks; these can be programmed immediately
+		 * since they satisfy both the current configuration we're
+		 * switching away from and the new configuration we're switching
+		 * to.
+		 */
+		struct intel_pipe_wm intermediate;
+
+		/*
+		 * Platforms with two-step watermark programming will need to
+		 * update watermark programming post-vblank to switch from the
+		 * safe intermediate watermarks to the optimal final
+		 * watermarks.
+		 */
+		bool need_postvbl_update;
 	} wm;
 };
 
@@ -575,8 +591,12 @@ struct intel_crtc {
 			struct intel_pipe_wm ilk;
 			struct skl_pipe_wm skl;
 		} active;
+
 		/* allow CxSR on this pipe */
 		bool cxsr_allowed;
+
+		/* Protects active and need_postvbl_update */
+		struct mutex wm_mutex;
 	} wm;
 
 	int scanline_offset;
@@ -1418,6 +1438,7 @@ 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_state *pipe_config);
+bool ilk_disable_lp_wm(struct drm_device *dev);
 
 /* intel_sdvo.c */
 bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob);
drivers/gpu/drm/i915/intel_pm.c +109 -45

232 lines not on this page · raw message

Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#16 lore raw

Re: [PATCH 14/14] drm/i915: Add two-stage ILK-style watermark programming (v4)

Hey,

It's worh nothing this series may have a soft dependency on Ville's cleanups, because else pixel_rate == 0 during boot. :(

This patch breaks when the initial sprite watermarks are set on ironlake, but no sprite is enabled.

A naive fix is below, without it my laptop won't power the screen because intermediate wm calculation fails. :(

drivers/gpu/drm/i915/intel_pm.c +1 -1
index 41a4dbaf600b..bf874ff97024 100644
@@ -2457,7 +2457,7 @@ static int ilk_compute_intermediate_wm(struct drm_device *dev,
 
 		a_wm->enable &= b_wm->enable;
 		a_wm->pri_val = max(a_wm->pri_val, b_wm->pri_val);
-		a_wm->spr_val = max(a_wm->spr_val, b_wm->spr_val);
+		a_wm->spr_val = a->sprites_enabled ? max(a_wm->spr_val, b_wm->spr_val) : 0;
 		a_wm->cur_val = max(a_wm->cur_val, b_wm->cur_val);
 		a_wm->fbc_val = max(a_wm->fbc_val, b_wm->fbc_val);
 	}

Op 15-09-15 om 04:19 schreef Matt Roper:

Quoted 8 lines

<snip>

(quoted diff) +3 -0
@@ -13202,6 +13229,20 @@ static int intel_atomic_commit(struct drm_device *dev,
 	/* FIXME: add subpixel order */
 
 	drm_atomic_helper_wait_for_vblanks(dev, state);
+
+	/*
+	 * Now that the vblank has passed, we can go ahead and program the

Review static int intel_atomic_commit(struct drm_device *dev, · line 13244

Quoted 10 lines
(quoted diff) +10 -0
+	 * optimal watermarks on platforms that need two-step watermark
+	 * programming.
+	 *
+	 * TODO: Move this (and other cleanup) to an async worker eventually.
+	 */
+	for_each_crtc_in_state(state, crtc, crtc_state, i) {
+		intel_cstate = to_intel_crtc_state(crtc->state);
+
+		dev_priv->display.optimize_watermarks(intel_cstate);
+	}

^Breaks on my skylake with a null pointer deref; I think you might want to hide it behind a if. :-)

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#17 lore raw

Re: [PATCH 14/14] drm/i915: Add two-stage ILK-style watermark programming (v4)

Quoted 14 lines · Maarten Lankhorst

Hey,

It's worh nothing this series may have a soft dependency on Ville's cleanups, because else pixel_rate == 0 during boot. :(

This patch breaks when the initial sprite watermarks are set on ironlake, but no sprite is enabled.

A naive fix is below, without it my laptop won't power the screen because intermediate wm calculation fails. :(

drivers/gpu/drm/i915/intel_pm.c +0 -0
index 41a4dbaf600b..bf874ff97024 100644

drivers/gpu/drm/i915/intel_pm.c static int ilk_compute_intermediate_wm(struct drm_device *dev, · line 2463

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +1 -1
@@ -2457,7 +2457,7 @@ static int ilk_compute_intermediate_wm(struct drm_device *dev,
 
 		a_wm->enable &= b_wm->enable;
 		a_wm->pri_val = max(a_wm->pri_val, b_wm->pri_val);
-		a_wm->spr_val = max(a_wm->spr_val, b_wm->spr_val);
+		a_wm->spr_val = a->sprites_enabled ? max(a_wm->spr_val, b_wm->spr_val) : 0;
 		a_wm->cur_val = max(a_wm->cur_val, b_wm->cur_val);
 		a_wm->fbc_val = max(a_wm->fbc_val, b_wm->fbc_val);
 	}

Hmm. That could work, but feels a bit wrong. Would be nice if we could always keep the watermarks totally consistent with the plane configuration.

One thing we could do during init is simply calculate the optimal watermarks for the current config and just blast them in. Then when the first plane update/modeset happens the current state would actually be sane wrt. the current plane config.

Quoted 27 lines

Op 15-09-15 om 04:19 schreef Matt Roper:

Quoted 8 lines

<snip>

(quoted diff) +3 -0
@@ -13202,6 +13229,20 @@ static int intel_atomic_commit(struct drm_device *dev,
 	/* FIXME: add subpixel order */
 
 	drm_atomic_helper_wait_for_vblanks(dev, state);
+
+	/*
+	 * Now that the vblank has passed, we can go ahead and program the

Review static int intel_atomic_commit(struct drm_device *dev, · line 13244

Quoted 10 lines
(quoted diff) +10 -0
+	 * optimal watermarks on platforms that need two-step watermark
+	 * programming.
+	 *
+	 * TODO: Move this (and other cleanup) to an async worker eventually.
+	 */
+	for_each_crtc_in_state(state, crtc, crtc_state, i) {
+		intel_cstate = to_intel_crtc_state(crtc->state);
+
+		dev_priv->display.optimize_watermarks(intel_cstate);
+	}

^Breaks on my skylake with a null pointer deref; I think you might want to hide it behind a if. :-)

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#18 lore raw

Re: [PATCH 14/14] drm/i915: Add two-stage ILK-style watermark programming (v4)

Op 17-09-15 om 15:05 schreef Ville Syrjälä:

Quoted 27 lines · Maarten Lankhorst
Quoted 14 lines

Hey,

It's worh nothing this series may have a soft dependency on Ville's cleanups, because else pixel_rate == 0 during boot. :(

This patch breaks when the initial sprite watermarks are set on ironlake, but no sprite is enabled.

A naive fix is below, without it my laptop won't power the screen because intermediate wm calculation fails. :(

drivers/gpu/drm/i915/intel_pm.c +0 -0
index 41a4dbaf600b..bf874ff97024 100644

drivers/gpu/drm/i915/intel_pm.c static int ilk_compute_intermediate_wm(struct drm_device *dev, · line 2463

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +1 -1
@@ -2457,7 +2457,7 @@ static int ilk_compute_intermediate_wm(struct drm_device *dev,
 
 		a_wm->enable &= b_wm->enable;
 		a_wm->pri_val = max(a_wm->pri_val, b_wm->pri_val);
-		a_wm->spr_val = max(a_wm->spr_val, b_wm->spr_val);
+		a_wm->spr_val = a->sprites_enabled ? max(a_wm->spr_val, b_wm->spr_val) : 0;
 		a_wm->cur_val = max(a_wm->cur_val, b_wm->cur_val);
 		a_wm->fbc_val = max(a_wm->fbc_val, b_wm->fbc_val);
 	}

Hmm. That could work, but feels a bit wrong. Would be nice if we could always keep the watermarks totally consistent with the plane configuration.

Quoted 5 lines

One thing we could do during init is simply calculate the optimal watermarks for the current config and just blast them in. Then when the first plane update/modeset happens the current state would actually be sane wrt. the current plane config.

That would work too, as long as we end up with something sane I'm all for it. :-)
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

#19 lore raw

Re: [PATCH 14/14] drm/i915: Add two-stage ILK-style watermark programming (v4)

Quoted 28 lines · Maarten Lankhorst

Op 17-09-15 om 15:05 schreef Ville Syrjälä:

Quoted 27 lines · Maarten Lankhorst
Quoted 14 lines

Hey,

It's worh nothing this series may have a soft dependency on Ville's cleanups, because else pixel_rate == 0 during boot. :(

This patch breaks when the initial sprite watermarks are set on ironlake, but no sprite is enabled.

A naive fix is below, without it my laptop won't power the screen because intermediate wm calculation fails. :(

drivers/gpu/drm/i915/intel_pm.c +0 -0
index 41a4dbaf600b..bf874ff97024 100644

drivers/gpu/drm/i915/intel_pm.c static int ilk_compute_intermediate_wm(struct drm_device *dev, · line 2463

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +1 -1
@@ -2457,7 +2457,7 @@ static int ilk_compute_intermediate_wm(struct drm_device *dev,
 
 		a_wm->enable &= b_wm->enable;
 		a_wm->pri_val = max(a_wm->pri_val, b_wm->pri_val);
-		a_wm->spr_val = max(a_wm->spr_val, b_wm->spr_val);
+		a_wm->spr_val = a->sprites_enabled ? max(a_wm->spr_val, b_wm->spr_val) : 0;
 		a_wm->cur_val = max(a_wm->cur_val, b_wm->cur_val);
 		a_wm->fbc_val = max(a_wm->fbc_val, b_wm->fbc_val);
 	}

Hmm. That could work, but feels a bit wrong. Would be nice if we could always keep the watermarks totally consistent with the plane configuration.

Quoted 6 lines
Quoted 5 lines

One thing we could do during init is simply calculate the optimal watermarks for the current config and just blast them in. Then when the first plane update/modeset happens the current state would actually be sane wrt. the current plane config.

That would work too, as long as we end up with something sane I'm all for it. :-)

Yeah blasting in our own preferred config is probably simpler - adding all kinds of hw state readout for this seems overkill to me. And adding special cases all over is imo bad since we're really not any good at testing all the various corner-cases ...
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx