Skip to contentSkip to conversation

SAGV support for Gen12+

PATCH Superseded Stanislav Lisovskiy 3 participants 24 messages May 5, 2020 – May 6, 2020

Subsystems (3)

v27 of 30 Superseded Latest is v30 →

lore
Share & tools
lore
#0 lore raw

v27 cover SAGV support for Gen12+

For Gen11+ platforms BSpec suggests disabling specific QGV points separately, depending on bandwidth limitations and current display configuration. Thus it required adding a new PCode request for disabling QGV points and some refactoring of already existing SAGV code.
Also had to refactor intel_can_enable_sagv function, as current seems to be outdated and using skl specific workarounds, also not following BSpec for Gen11+.

v25: Rebased patch series as part was merged already
v26: Had to resend the whole series as one more mid patch was added
v27: Patches 2,3,7 were pushed, have to resend the series to prevent
build failure.

Stanislav Lisovskiy (6):
drm/i915: Introduce skl_plane_wm_level accessor. drm/i915: Separate icl and skl SAGV checking drm/i915: Add TGL+ SAGV support
drm/i915: Added required new PCode commands
drm/i915: Restrict qgv points which don't have enough bandwidth. drm/i915: Enable SAGV support for Gen12

 drivers/gpu/drm/i915/display/intel_bw.c       | 139 ++++++--
 drivers/gpu/drm/i915/display/intel_bw.h       |   9 +
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   6 +
 drivers/gpu/drm/i915/i915_reg.h               |   4 +
 drivers/gpu/drm/i915/intel_pm.c               | 303 ++++++++++++++++--
 drivers/gpu/drm/i915/intel_pm.h               |   2 +
 drivers/gpu/drm/i915/intel_sideband.c         |   2 +
 8 files changed, 407 insertions(+), 66 deletions(-)
Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

PATCH v27 1/6 drm/i915: Introduce skl_plane_wm_level accessor.

For future Gen12 SAGV implementation we need to seemlessly alter wm levels calculated, depending on whether we are allowed to enable SAGV or not.

So this accessor will give additional flexibility to do that.

Currently this accessor is still simply working as "pass-through" function. This will be changed in next coming patches from this series.

v2: - plane_id -> plane->id(Ville Syrjälä)
- Moved wm_level var to have more local scope (Ville Syrjälä)
- Renamed yuv to color_plane(Ville Syrjälä) in skl_plane_wm_level

v3: - plane->id -> plane_id(this time for real, Ville Syrjälä)
- Changed colorplane id type from boolean to int as index (Ville Syrjälä)
- Moved crtc_state param so that it is first now (Ville Syrjälä)
- Moved wm_level declaration to tigher scope in skl_write_plane_wm(Ville Syrjälä)

v4: - Started to use enum values for color plane
- Do sizeof for a type what we are memset'ing
- Zero out wm_uv as well(Ville Syrjälä)

v5: - Fixed rebase conflict caused by COLOR_PLANE_*
enum removal

---
 drivers/gpu/drm/i915/intel_pm.c | 85 ++++++++++++++++++++++++++-------
 1 file changed, 67 insertions(+), 18 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +67 -18
index 416cb1a1e7cb..da567fac7c93 100644
@@ -4632,6 +4632,18 @@ icl_get_total_relative_data_rate(struct intel_crtc_state *crtc_state,
 	return total_data_rate;
 }
 
+static const struct skl_wm_level *
+skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
+		   enum plane_id plane_id,
+		   int level,
+		   int color_plane)
+{
+	const struct skl_plane_wm *wm =
+		&crtc_state->wm.skl.optimal.planes[plane_id];
+
+	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -4691,22 +4703,28 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 	 */
 	for (level = ilk_wm_max_level(dev_priv); level >= 0; level--) {
 		blocks = 0;
+
 		for_each_plane_id_on_crtc(crtc, plane_id) {
-			const struct skl_plane_wm *wm =
-				&crtc_state->wm.skl.optimal.planes[plane_id];
+			const struct skl_wm_level *wm_level;
+			const struct skl_wm_level *wm_uv_level;
+
+			wm_level = skl_plane_wm_level(crtc_state, plane_id,
+						      level, 0);
+			wm_uv_level = skl_plane_wm_level(crtc_state, plane_id,
+							 level, 1);
 
 			if (plane_id == PLANE_CURSOR) {
-				if (wm->wm[level].min_ddb_alloc > total[PLANE_CURSOR]) {
+				if (wm_level->min_ddb_alloc > total[PLANE_CURSOR]) {
 					drm_WARN_ON(&dev_priv->drm,
-						    wm->wm[level].min_ddb_alloc != U16_MAX);
+						    wm_level->min_ddb_alloc != U16_MAX);
 					blocks = U32_MAX;
 					break;
 				}
 				continue;
 			}
 
-			blocks += wm->wm[level].min_ddb_alloc;
-			blocks += wm->uv_wm[level].min_ddb_alloc;
+			blocks += wm_level->min_ddb_alloc;
+			blocks += wm_uv_level->min_ddb_alloc;
 		}
 
 		if (blocks <= alloc_size) {
@@ -4729,11 +4747,16 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 	 * proportional to its relative data rate.
 	 */
 	for_each_plane_id_on_crtc(crtc, plane_id) {
-		const struct skl_plane_wm *wm =
-			&crtc_state->wm.skl.optimal.planes[plane_id];
+		const struct skl_wm_level *wm_level;
+		const struct skl_wm_level *wm_uv_level;
 		u64 rate;
 		u16 extra;
 
+		wm_level = skl_plane_wm_level(crtc_state, plane_id,
+					      level, 0);
+		wm_uv_level = skl_plane_wm_level(crtc_state, plane_id,
+						 level, 1);
+
 		if (plane_id == PLANE_CURSOR)
 			continue;
 
@@ -4748,7 +4771,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 		extra = min_t(u16, alloc_size,
 			      DIV64_U64_ROUND_UP(alloc_size * rate,
 						 total_data_rate));
-		total[plane_id] = wm->wm[level].min_ddb_alloc + extra;
+		total[plane_id] = wm_level->min_ddb_alloc + extra;
 		alloc_size -= extra;
 		total_data_rate -= rate;
 
@@ -4759,7 +4782,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 		extra = min_t(u16, alloc_size,
 			      DIV64_U64_ROUND_UP(alloc_size * rate,
 						 total_data_rate));
-		uv_total[plane_id] = wm->uv_wm[level].min_ddb_alloc + extra;
+		uv_total[plane_id] = wm_uv_level->min_ddb_alloc + extra;
 		alloc_size -= extra;
 		total_data_rate -= rate;
 	}
@@ -4802,9 +4825,16 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 	 */
 	for (level++; level <= ilk_wm_max_level(dev_priv); level++) {
 		for_each_plane_id_on_crtc(crtc, plane_id) {
+			const struct skl_wm_level *wm_level;
+			const struct skl_wm_level *wm_uv_level;
 			struct skl_plane_wm *wm =
 				&crtc_state->wm.skl.optimal.planes[plane_id];
 
+			wm_level = skl_plane_wm_level(crtc_state, plane_id,
+						      level, 0);
+			wm_uv_level = skl_plane_wm_level(crtc_state, plane_id,
+							 level, 1);
+
 			/*
 			 * We only disable the watermarks for each plane if
 			 * they exceed the ddb allocation of said plane. This
@@ -4817,9 +4847,13 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 			 *  planes must be enabled before the level will be used."
 			 * So this is actually safe to do.
 			 */
-			if (wm->wm[level].min_ddb_alloc > total[plane_id] ||
-			    wm->uv_wm[level].min_ddb_alloc > uv_total[plane_id])
-				memset(&wm->wm[level], 0, sizeof(wm->wm[level]));
+			if (wm_level->min_ddb_alloc > total[plane_id] ||
+			    wm_uv_level->min_ddb_alloc > uv_total[plane_id]) {
+				memset(&wm->wm[level], 0,
+				       sizeof(wm->wm[level]));
+				memset(&wm->uv_wm[level], 0,
+				       sizeof(wm->uv_wm[level]));
+			}
 
 			/*
 			 * Wa_1408961008:icl, ehl
@@ -4827,9 +4861,14 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 			 */
 			if (IS_GEN(dev_priv, 11) &&
 			    level == 1 && wm->wm[0].plane_en) {
-				wm->wm[level].plane_res_b = wm->wm[0].plane_res_b;
-				wm->wm[level].plane_res_l = wm->wm[0].plane_res_l;
-				wm->wm[level].ignore_lines = wm->wm[0].ignore_lines;
+				wm_level = skl_plane_wm_level(crtc_state, plane_id,
+							      0, 0);
+				wm->wm[level].plane_res_b =
+					wm_level->plane_res_b;
+				wm->wm[level].plane_res_l =
+					wm_level->plane_res_l;
+				wm->wm[level].ignore_lines =
+					wm_level->ignore_lines;
 			}
 		}
 	}
@@ -5439,8 +5478,13 @@ void skl_write_plane_wm(struct intel_plane *plane,
 		&crtc_state->wm.skl.plane_ddb_uv[plane_id];
 
 	for (level = 0; level <= max_level; level++) {
+		const struct skl_wm_level *wm_level;
+		int color_plane = 0;
+
+		wm_level = skl_plane_wm_level(crtc_state, plane_id, level, color_plane);
+
 		skl_write_wm_level(dev_priv, PLANE_WM(pipe, plane_id, level),
-				   &wm->wm[level]);
+				   wm_level);
 	}
 	skl_write_wm_level(dev_priv, PLANE_WM_TRANS(pipe, plane_id),
 			   &wm->trans_wm);
@@ -5473,8 +5517,13 @@ void skl_write_cursor_wm(struct intel_plane *plane,
 		&crtc_state->wm.skl.plane_ddb_y[plane_id];
 
 	for (level = 0; level <= max_level; level++) {
+		const struct skl_wm_level *wm_level;
+		int color_plane = 0;
+
+		wm_level = skl_plane_wm_level(crtc_state, plane_id, level, color_plane);
+
 		skl_write_wm_level(dev_priv, CUR_WM(pipe, level),
-				   &wm->wm[level]);
+				   wm_level);
 	}
 	skl_write_wm_level(dev_priv, CUR_WM_TRANS(pipe), &wm->trans_wm);
 
Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

PATCH v27 2/6 drm/i915: Separate icl and skl SAGV checking

In reply to Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +28 -2
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3863,22 +3881,30 @@ bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 
 static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 {
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
 
 	for_each_new_intel_crtc_in_state(state, crtc,
 					 new_crtc_state, i) {
+		bool can_sagv;
+
 		new_bw_state = intel_atomic_get_bw_state(state);
 		if (IS_ERR(new_bw_state))
 			return PTR_ERR(new_bw_state);
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (intel_crtc_can_enable_sagv(new_crtc_state))
+		if (INTEL_GEN(dev_priv) >= 11)
+			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
+		else
+			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
+
+		if (can_sagv)
 			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
 		else
 			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

Quoted 39 lines
drivers/gpu/drm/i915/intel_pm.c +11 -2
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3863,22 +3881,30 @@ bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 
 static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 {
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
 
 	for_each_new_intel_crtc_in_state(state, crtc,
 					 new_crtc_state, i) {
+		bool can_sagv;
+
 		new_bw_state = intel_atomic_get_bw_state(state);
 		if (IS_ERR(new_bw_state))
 			return PTR_ERR(new_bw_state);
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (intel_crtc_can_enable_sagv(new_crtc_state))
+		if (INTEL_GEN(dev_priv) >= 11)
+			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
+		else
+			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
+
+		if (can_sagv)
 			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
 		else
 			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 32 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

Quoted 13 lines

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

Do we even need an icl variant actually? Does it use the skl or tgl way of checking for sagv yes vs. no?

Quoted 48 lines
Quoted 40 lines
drivers/gpu/drm/i915/intel_pm.c +11 -2
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3863,22 +3881,30 @@ bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 
 static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 {
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
 
 	for_each_new_intel_crtc_in_state(state, crtc,
 					 new_crtc_state, i) {
+		bool can_sagv;
+
 		new_bw_state = intel_atomic_get_bw_state(state);
 		if (IS_ERR(new_bw_state))
 			return PTR_ERR(new_bw_state);
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (intel_crtc_can_enable_sagv(new_crtc_state))
+		if (INTEL_GEN(dev_priv) >= 11)
+			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
+		else
+			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
+
+		if (can_sagv)
 			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
 		else
 			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 47 lines · Ville Syrjälä
Quoted 32 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

Quoted 13 lines

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

Quoted 2 lines

Do we even need an icl variant actually? Does it use the skl or tgl way of checking for sagv yes vs. no?

As I undestand icl implementation should be pretty much the same as skl, except that icl doesn't have this one active pipe limitation thing.

Stan

Quoted 41 lines
Quoted 41 lines
Quoted 41 lines
drivers/gpu/drm/i915/intel_pm.c +11 -2
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3863,22 +3881,30 @@ bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 
 static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 {
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
 
 	for_each_new_intel_crtc_in_state(state, crtc,
 					 new_crtc_state, i) {
+		bool can_sagv;
+
 		new_bw_state = intel_atomic_get_bw_state(state);
 		if (IS_ERR(new_bw_state))
 			return PTR_ERR(new_bw_state);
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (intel_crtc_can_enable_sagv(new_crtc_state))
+		if (INTEL_GEN(dev_priv) >= 11)
+			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
+		else
+			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
+
+		if (can_sagv)
 			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
 		else
 			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
Signature · 1 line
2.24.1.485.gad05a3d8e5
Quoted 12 lines
Quoted 8 lines
Signature · 7 lines
-- 
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 2 lines
Ville Syrjälä
Intel

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

#6 lore raw

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 48 lines · Stanislav
Quoted 47 lines · Ville Syrjälä
Quoted 45 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

Quoted 6 lines
Quoted 2 lines

Do we even need an icl variant actually? Does it use the skl or tgl way of checking for sagv yes vs. no?

As I undestand icl implementation should be pretty much the same as skl, except that icl doesn't have this one active pipe limitation thing.

Right, and since that's now moved into intel_can_enable_sagv() we can use skl_crtc_can_enable_sagv() for all pre-tgl.

Quoted 56 lines

Stan

Quoted 53 lines
Quoted 49 lines
Quoted 41 lines
drivers/gpu/drm/i915/intel_pm.c +11 -2
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3863,22 +3881,30 @@ bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 
 static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 {
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
 
 	for_each_new_intel_crtc_in_state(state, crtc,
 					 new_crtc_state, i) {
+		bool can_sagv;
+
 		new_bw_state = intel_atomic_get_bw_state(state);
 		if (IS_ERR(new_bw_state))
 			return PTR_ERR(new_bw_state);
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (intel_crtc_can_enable_sagv(new_crtc_state))
+		if (INTEL_GEN(dev_priv) >= 11)
+			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
+		else
+			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
+
+		if (can_sagv)
 			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
 		else
 			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#7 lore raw

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

In reply to Ville Syrjälä
Quoted 32 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

Quoted 13 lines

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

This is done so, because icl and skl checking share the same code to check if SAGV can be enabled, except active_pipes > 1 thing.

So that icl and skl can share the same code avoiding duplicating,
i.e if I put code from intel_crtc_can_enable_sagv to
skl_crtc_can_enable_sagv, I will have to
1) either duplicate this code to icl_crtc_can_enable_sagv(if I add remaining active_pipes check to skl)
2) use skl_crtc_can_enable_sagv from icl_crtc_can_enable_sagv,
but this active_pipes check will be still outside of this skl function,
which I don't find nice - to me the best way is to keep all skl
specific checks in a correspondent function.

So that is why I preferred to extract some common code to some separate universal function which can be then used from both icl and skl functions:
from icl it is used "as is" and from skl it is intel_crtc_can_enable_sagv
+ this active_pipes check.

Currently anyway we of course have that active_pipes check in intel_can_enable_sagv i.e already outside of skl_crtc_can_enable_sagv(where it should be imo), so was your intention to leave it outside anyway?

Stan

Quoted 28 lines
Quoted 28 lines
drivers/gpu/drm/i915/intel_pm.c +5 -1
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3863,22 +3881,30 @@ bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 
 static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 {
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
 
 	for_each_new_intel_crtc_in_state(state, crtc,
 					 new_crtc_state, i) {
+		bool can_sagv;
+
 		new_bw_state = intel_atomic_get_bw_state(state);
 		if (IS_ERR(new_bw_state))
 			return PTR_ERR(new_bw_state);
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
Quoted 16 lines
Quoted 12 lines
drivers/gpu/drm/i915/intel_pm.c +6 -1
-		if (intel_crtc_can_enable_sagv(new_crtc_state))
+		if (INTEL_GEN(dev_priv) >= 11)
+			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
+		else
+			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
+
+		if (can_sagv)
 			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
 		else
 			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 2 lines
Ville Syrjälä
Intel

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

#8 lore raw

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 60 lines · Stanislav
Quoted 32 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

Quoted 13 lines

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

This is done so, because icl and skl checking share the same code to check if SAGV can be enabled, except active_pipes > 1 thing.

So that icl and skl can share the same code avoiding duplicating,
i.e if I put code from intel_crtc_can_enable_sagv to
skl_crtc_can_enable_sagv, I will have to
1) either duplicate this code to icl_crtc_can_enable_sagv(if I add remaining active_pipes check to skl)
2) use skl_crtc_can_enable_sagv from icl_crtc_can_enable_sagv,
but this active_pipes check will be still outside of this skl function,
which I don't find nice - to me the best way is to keep all skl
specific checks in a correspondent function.

Quoted 8 lines

So that is why I preferred to extract some common code to some separate universal function which can be then used from both icl and skl functions:
from icl it is used "as is" and from skl it is intel_crtc_can_enable_sagv
+ this active_pipes check.

Currently anyway we of course have that active_pipes check in intel_can_enable_sagv i.e already outside of skl_crtc_can_enable_sagv(where it should be imo), so was your intention to leave it outside anyway?

My intention is that we eventually remove it altogeher. In the meantime intel_can_enable_sagv() looks like the right place for it.

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

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 66 lines · Ville Syrjälä
Quoted 65 lines · Stanislav
Quoted 32 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

Quoted 13 lines

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

This is done so, because icl and skl checking share the same code to check if SAGV can be enabled, except active_pipes > 1 thing.

So that icl and skl can share the same code avoiding duplicating,
i.e if I put code from intel_crtc_can_enable_sagv to
skl_crtc_can_enable_sagv, I will have to
1) either duplicate this code to icl_crtc_can_enable_sagv(if I add remaining active_pipes check to skl)
2) use skl_crtc_can_enable_sagv from icl_crtc_can_enable_sagv,
but this active_pipes check will be still outside of this skl function,
which I don't find nice - to me the best way is to keep all skl
specific checks in a correspondent function.

So that is why I preferred to extract some common code to some separate universal function which can be then used from both icl and skl functions:
from icl it is used "as is" and from skl it is intel_crtc_can_enable_sagv
+ this active_pipes check.

Quoted 7 lines
Quoted 3 lines

Currently anyway we of course have that active_pipes check in intel_can_enable_sagv i.e already outside of skl_crtc_can_enable_sagv(where it should be imo), so was your intention to leave it outside anyway?

My intention is that we eventually remove it altogeher. In the meantime intel_can_enable_sagv() looks like the right place for it.

Wow, you were really fast replying :) Ok, then I just leave skl_crtc_can_enable_sagv, use it for icl and active_pipes check stays in intel_can_enable_sagv, however probably it will now need somekind of INTEL_GEN < 11 check, I guess.

Stan

Quoted 4 lines
Signature · 3 lines
-- 
Ville Syrjälä
Intel

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

#10 lore raw

Re: [Intel-gfx] [PATCH v27 2/6] drm/i915: Separate icl and skl SAGV checking

Quoted 71 lines · Stanislav
Quoted 70 lines · Ville Syrjälä
Quoted 68 lines · Stanislav
Quoted 32 lines · Ville Syrjälä
Quoted 17 lines · Stanislav Lisovskiy

Introduce platform dependent SAGV checking in
combination with bandwidth state pipe SAGV mask.

v2, v3, v4, v5, v6: Fix rebase conflict

---
 drivers/gpu/drm/i915/intel_pm.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -0
index da567fac7c93..c7d726a656b2 100644
@@ -3853,6 +3853,24 @@ static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state
 	return true;
 }

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3864

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +9 -0
 
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	/*
+	 * SKL+ workaround: bspec recommends we disable SAGV when we have
+	 * more then one pipe enabled
+	 */
+	if (hweight8(state->active_pipes) > 1)
+		return false;

That stuff should no longer be here since we now have it done properly in intel_can_eanble_sagv().

Quoted 13 lines

drivers/gpu/drm/i915/intel_pm.c static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state · line 3872

Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +8 -0
+
+	return intel_crtc_can_enable_sagv(crtc_state);
+}
+
+static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	return intel_crtc_can_enable_sagv(crtc_state);
+}

This looks the wrong way around. IMO intel_crtc_can_enable_sagv() should rather call the skl vs. icl variants as needed. Although we don't yet have the icl variant so the oerdering of the patches is a bit weird.

This is done so, because icl and skl checking share the same code to check if SAGV can be enabled, except active_pipes > 1 thing.

So that icl and skl can share the same code avoiding duplicating,
i.e if I put code from intel_crtc_can_enable_sagv to
skl_crtc_can_enable_sagv, I will have to
1) either duplicate this code to icl_crtc_can_enable_sagv(if I add remaining active_pipes check to skl)
2) use skl_crtc_can_enable_sagv from icl_crtc_can_enable_sagv,
but this active_pipes check will be still outside of this skl function,
which I don't find nice - to me the best way is to keep all skl
specific checks in a correspondent function.

So that is why I preferred to extract some common code to some separate universal function which can be then used from both icl and skl functions:
from icl it is used "as is" and from skl it is intel_crtc_can_enable_sagv
+ this active_pipes check.

Currently anyway we of course have that active_pipes check in intel_can_enable_sagv i.e already outside of skl_crtc_can_enable_sagv(where it should be imo), so was your intention to leave it outside anyway?

Quoted 7 lines
Quoted 3 lines

My intention is that we eventually remove it altogeher. In the meantime intel_can_enable_sagv() looks like the right place for it.

Wow, you were really fast replying :) Ok, then I just leave skl_crtc_can_enable_sagv, use it for icl and active_pipes check stays in intel_can_enable_sagv, however probably it will now need somekind of INTEL_GEN < 11 check, I guess.

If we want to enable sagv for multi-pipe on just icl+ first, yes. I guess we can do that. I suggest making that a standalone patch so we can easily revert it if things blow up. We can then let that bake for a while and follow up later with nuking the single pipe restriction for all platforms.

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

PATCH v27 3/6 drm/i915: Add TGL+ SAGV support

In reply to Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +3 -0
index 9488449e4b94..32cbbf7dddc6 100644
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +121 -7

213 lines not on this page · raw message

Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

Re: [Intel-gfx] [PATCH v27 3/6] drm/i915: Add TGL+ SAGV support

Quoted 46 lines · Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
index 9488449e4b94..32cbbf7dddc6 100644

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 692

Quoted 5 lines
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;

As mentioned before uv_wm is not a thing on icl+, so nuke this.

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 698

Quoted 6 lines
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;

I would call it use_sagv_wm or somesuch to make it actually clear what it does.

Quoted 45 lines
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +11 -2
index c7d726a656b2..1b9925b6672c 100644
@@ -3871,6 +3871,9 @@ static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 	return intel_crtc_can_enable_sagv(crtc_state);
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state);
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3884,7 +3887,7 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
@@ -3899,7 +3902,9 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (INTEL_GEN(dev_priv) >= 11)
+		if (INTEL_GEN(dev_priv) >= 12)
+			can_sagv = tgl_crtc_can_enable_sagv(new_crtc_state);
+		else if (INTEL_GEN(dev_priv) >= 11)
 			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
 		else
 			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
@@ -3921,6 +3926,24 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 			return ret;
 	}
 
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+

drivers/gpu/drm/i915/intel_pm.c static int intel_compute_sagv_mask(struct intel_atomic_state *state) · line 3942

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+		/*
+		 * Due to drm limitation at commit state, when
+		 * changes are written the whole atomic state is
+		 * zeroed away => which prevents from using it,
+		 * so just sticking it into pipe wm state for
+		 * keeping it simple - anyway this is related to wm.
+		 * Proper way in ideal universe would be of course not
+		 * to lose parent atomic state object from child crtc_state,
+		 * and stick to OOP programming principles, which had been
+		 * scientifically proven to work.

Seems like random ramblings. I'd just nuke the comment.

Quoted 180 lines
drivers/gpu/drm/i915/intel_pm.c +100 -5
+		 */
+		pipe_wm->can_sagv = intel_can_enable_sagv(new_bw_state);
+	}
+
 	if (intel_can_enable_sagv(new_bw_state) != intel_can_enable_sagv(old_bw_state)) {
 		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
 		if (ret)
@@ -4664,12 +4687,39 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 		   int level,
 		   int color_plane)
 {
-	const struct skl_plane_wm *wm =
-		&crtc_state->wm.skl.optimal.planes[plane_id];
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (!level) {
+		if (pipe_wm->can_sagv)
+			return color_plane == 0 ? &wm->sagv_wm0 : &wm->uv_sagv_wm0;
+	}
 
 	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *plane_alloc =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (skl_ddb_entry_size(plane_alloc) < wm->sagv_wm0.min_ddb_alloc)
+			return false;
+	}
+
+	return true;
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -5251,10 +5301,17 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
 static void
 skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      const struct skl_wm_params *wm_params,
-		      struct skl_wm_level *levels)
+		      struct skl_plane_wm *plane_wm,
+		      int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int level, max_level = ilk_wm_max_level(dev_priv);
+	/*
+	 * Check which kind of plane is it and based on that calculate
+	 * correspondent WM levels.
+	 */
+	struct skl_wm_level *levels = color_plane == 1 ?
+				      plane_wm->uv_wm : plane_wm->wm;
 	struct skl_wm_level *result_prev = &levels[0];
 
 	for (level = 0; level <= max_level; level++) {
@@ -5268,6 +5325,40 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 	}
 }
 
+static void skl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm,
+				int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = color_plane == 1 ?
+				&plane_wm->uv_sagv_wm0 : &plane_wm->sagv_wm0;
+	struct skl_wm_level *levels = color_plane == 1 ?
+				plane_wm->uv_wm : plane_wm->wm;
+
+	/*
+	 * For Gen12 if it is an L0 we need to also
+	 * consider sagv_block_time when calculating
+	 * L0 watermark - we will need that when making
+	 * a decision whether enable SAGV or not.
+	 * For older gens we agreed to copy L0 value for
+	 * compatibility.
+	 */
+	if ((INTEL_GEN(dev_priv) >= 12)) {
+		u32 latency = dev_priv->wm.skl_latency[0];
+
+		latency += dev_priv->sagv_block_time_us;
+		skl_compute_plane_wm(crtc_state, 0, latency,
+				     wm_params, &levels[0],
+				     sagv_wm);
+		DRM_DEBUG_KMS("%d L0 blocks required for SAGV vs %d for non-SAGV\n",
+			      sagv_wm->min_ddb_alloc, levels[0].min_ddb_alloc);
+	} else {
+		/* Since all members are POD */
+		*sagv_wm = levels[0];
+	}
+}
+
 static void skl_compute_transition_wm(const struct intel_crtc_state *crtc_state,
 				      const struct skl_wm_params *wp,
 				      struct skl_plane_wm *wm)
@@ -5344,7 +5435,8 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 0);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 0);
 	skl_compute_transition_wm(crtc_state, &wm_params, wm);
 
 	return 0;
@@ -5366,7 +5458,8 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->uv_wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 1);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 1);
 
 	return 0;
 }
@@ -5751,6 +5844,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].plane_res_l,
 				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.plane_res_l);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 lines %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_l,
+				    new_wm->sagv_wm0.plane_res_l);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5766,6 +5865,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[6].plane_res_b, new_wm->wm[7].plane_res_b,
 				    new_wm->trans_wm.plane_res_b);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 blocks %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_b,
+				    new_wm->sagv_wm0.plane_res_b);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5780,6 +5885,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
 				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
 				    new_wm->trans_wm.min_ddb_alloc);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 min ddb %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.min_ddb_alloc,
+				    new_wm->sagv_wm0.min_ddb_alloc);
 		}
 	}
 }
@@ -6072,6 +6183,9 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			skl_wm_level_from_reg_val(val, &wm->wm[level]);
 		}
 
+		memcpy(&wm->sagv_wm0, &wm->wm[0],
+		       sizeof(struct skl_wm_level));
+
 		if (plane_id != PLANE_CURSOR)
 			val = I915_READ(PLANE_WM_TRANS(pipe, plane_id));
 		else
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#13 lore raw

Re: [Intel-gfx] [PATCH v27 3/6] drm/i915: Add TGL+ SAGV support

Quoted 41 lines · Ville Syrjälä
Quoted 40 lines · Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
Quoted 13 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 692

Quoted 11 lines
drivers/gpu/drm/i915/display/intel_display.c +0 -0
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
index 9488449e4b94..32cbbf7dddc6 100644
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;

As mentioned before uv_wm is not a thing on icl+, so nuke this.

This is used in skl_plane_wm_level accessor, which is used for all platforms, not just icl+. I remember we had agreed that for all platforms before tgl I simply copy sagv_wm0 values from regular wm0, so that this behaviour is unified(remember that your comment about memcpy which I changed to assignment, see skl_compute_sagv_wm).

So if I remove it and this is called for pre-icl platforms, guess this would still need wm_uv data?

Stan

Quoted 241 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 698

Quoted 7 lines
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;

I would call it use_sagv_wm or somesuch to make it actually clear what it does.

Quoted 45 lines
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +11 -2
index c7d726a656b2..1b9925b6672c 100644
@@ -3871,6 +3871,9 @@ static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 	return intel_crtc_can_enable_sagv(crtc_state);
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state);
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3884,7 +3887,7 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
@@ -3899,7 +3902,9 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (INTEL_GEN(dev_priv) >= 11)
+		if (INTEL_GEN(dev_priv) >= 12)
+			can_sagv = tgl_crtc_can_enable_sagv(new_crtc_state);
+		else if (INTEL_GEN(dev_priv) >= 11)
 			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
 		else
 			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
@@ -3921,6 +3926,24 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 			return ret;
 	}
 
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+

drivers/gpu/drm/i915/intel_pm.c static int intel_compute_sagv_mask(struct intel_atomic_state *state) · line 3942

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+		/*
+		 * Due to drm limitation at commit state, when
+		 * changes are written the whole atomic state is
+		 * zeroed away => which prevents from using it,
+		 * so just sticking it into pipe wm state for
+		 * keeping it simple - anyway this is related to wm.
+		 * Proper way in ideal universe would be of course not
+		 * to lose parent atomic state object from child crtc_state,
+		 * and stick to OOP programming principles, which had been
+		 * scientifically proven to work.

Seems like random ramblings. I'd just nuke the comment.

Quoted 172 lines
drivers/gpu/drm/i915/intel_pm.c +97 -5
+		 */
+		pipe_wm->can_sagv = intel_can_enable_sagv(new_bw_state);
+	}
+
 	if (intel_can_enable_sagv(new_bw_state) != intel_can_enable_sagv(old_bw_state)) {
 		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
 		if (ret)
@@ -4664,12 +4687,39 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 		   int level,
 		   int color_plane)
 {
-	const struct skl_plane_wm *wm =
-		&crtc_state->wm.skl.optimal.planes[plane_id];
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (!level) {
+		if (pipe_wm->can_sagv)
+			return color_plane == 0 ? &wm->sagv_wm0 : &wm->uv_sagv_wm0;
+	}
 
 	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *plane_alloc =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (skl_ddb_entry_size(plane_alloc) < wm->sagv_wm0.min_ddb_alloc)
+			return false;
+	}
+
+	return true;
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -5251,10 +5301,17 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
 static void
 skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      const struct skl_wm_params *wm_params,
-		      struct skl_wm_level *levels)
+		      struct skl_plane_wm *plane_wm,
+		      int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int level, max_level = ilk_wm_max_level(dev_priv);
+	/*
+	 * Check which kind of plane is it and based on that calculate
+	 * correspondent WM levels.
+	 */
+	struct skl_wm_level *levels = color_plane == 1 ?
+				      plane_wm->uv_wm : plane_wm->wm;
 	struct skl_wm_level *result_prev = &levels[0];
 
 	for (level = 0; level <= max_level; level++) {
@@ -5268,6 +5325,40 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 	}
 }
 
+static void skl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm,
+				int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = color_plane == 1 ?
+				&plane_wm->uv_sagv_wm0 : &plane_wm->sagv_wm0;
+	struct skl_wm_level *levels = color_plane == 1 ?
+				plane_wm->uv_wm : plane_wm->wm;
+
+	/*
+	 * For Gen12 if it is an L0 we need to also
+	 * consider sagv_block_time when calculating
+	 * L0 watermark - we will need that when making
+	 * a decision whether enable SAGV or not.
+	 * For older gens we agreed to copy L0 value for
+	 * compatibility.
+	 */
+	if ((INTEL_GEN(dev_priv) >= 12)) {
+		u32 latency = dev_priv->wm.skl_latency[0];
+
+		latency += dev_priv->sagv_block_time_us;
+		skl_compute_plane_wm(crtc_state, 0, latency,
+				     wm_params, &levels[0],
+				     sagv_wm);
+		DRM_DEBUG_KMS("%d L0 blocks required for SAGV vs %d for non-SAGV\n",
+			      sagv_wm->min_ddb_alloc, levels[0].min_ddb_alloc);
+	} else {
+		/* Since all members are POD */
+		*sagv_wm = levels[0];
+	}
+}
+
 static void skl_compute_transition_wm(const struct intel_crtc_state *crtc_state,
 				      const struct skl_wm_params *wp,
 				      struct skl_plane_wm *wm)
@@ -5344,7 +5435,8 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 0);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 0);
 	skl_compute_transition_wm(crtc_state, &wm_params, wm);
 
 	return 0;
@@ -5366,7 +5458,8 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->uv_wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 1);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 1);
 
 	return 0;
 }
@@ -5751,6 +5844,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].plane_res_l,
 				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.plane_res_l);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 lines %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_l,
+				    new_wm->sagv_wm0.plane_res_l);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5766,6 +5865,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[6].plane_res_b, new_wm->wm[7].plane_res_b,
 				    new_wm->trans_wm.plane_res_b);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 blocks %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_b,
+				    new_wm->sagv_wm0.plane_res_b);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5780,6 +5885,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
 				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
 				    new_wm->trans_wm.min_ddb_alloc);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 min ddb %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.min_ddb_alloc,
+				    new_wm->sagv_wm0.min_ddb_alloc);
 		}
 	}
 }
@@ -6072,6 +6183,9 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			skl_wm_level_from_reg_val(val, &wm->wm[level]);
 		}
 
Quoted 12 lines
Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +3 -0
+		memcpy(&wm->sagv_wm0, &wm->wm[0],
+		       sizeof(struct skl_wm_level));
+
 		if (plane_id != PLANE_CURSOR)
 			val = I915_READ(PLANE_WM_TRANS(pipe, plane_id));
 		else
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 2 lines
Ville Syrjälä
Intel

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

#14 lore raw

Re: [Intel-gfx] [PATCH v27 3/6] drm/i915: Add TGL+ SAGV support

Quoted 56 lines · Stanislav
Quoted 41 lines · Ville Syrjälä
Quoted 40 lines · Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
Quoted 13 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 692

Quoted 11 lines
drivers/gpu/drm/i915/display/intel_display.c +0 -0
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
index 9488449e4b94..32cbbf7dddc6 100644
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;

As mentioned before uv_wm is not a thing on icl+, so nuke this.

Quoted 5 lines

This is used in skl_plane_wm_level accessor, which is used for all platforms, not just icl+. I remember we had agreed that for all platforms before tgl I simply copy sagv_wm0 values from regular wm0, so that this behaviour is unified(remember that your comment about memcpy which I changed to assignment, see skl_compute_sagv_wm).

I think having that duplicated is just making things more confusing. Also uv_wm is never used by the hardware even on pre-icl, so having the accessor thing use it for the hw programming just doesn't make any sense.

For the compute side I think all we should really need is something like:

tgl_compute_sagv_wm()

{
	skl_compute_wm_level(sagv_wm0, latency + whatever);
}

skl_build_plane_wm_single()

{
	...
	skl_compute_transition_wm();

+ if (gen >= 12)
+ tgl_compute_sagv_wm();
}

And for the progamming side we should just pick the right wm0 based on the crtc_state->use_sagv_wm or whatever flag.

Quoted 260 lines

So if I remove it and this is called for pre-icl platforms, guess this would still need wm_uv data?

Stan

Quoted 253 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 698

Quoted 7 lines
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;

I would call it use_sagv_wm or somesuch to make it actually clear what it does.

Quoted 45 lines
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +11 -2
index c7d726a656b2..1b9925b6672c 100644
@@ -3871,6 +3871,9 @@ static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 	return intel_crtc_can_enable_sagv(crtc_state);
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state);
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3884,7 +3887,7 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
@@ -3899,7 +3902,9 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (INTEL_GEN(dev_priv) >= 11)
+		if (INTEL_GEN(dev_priv) >= 12)
+			can_sagv = tgl_crtc_can_enable_sagv(new_crtc_state);
+		else if (INTEL_GEN(dev_priv) >= 11)
 			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
 		else
 			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
@@ -3921,6 +3926,24 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 			return ret;
 	}
 
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+

drivers/gpu/drm/i915/intel_pm.c static int intel_compute_sagv_mask(struct intel_atomic_state *state) · line 3942

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+		/*
+		 * Due to drm limitation at commit state, when
+		 * changes are written the whole atomic state is
+		 * zeroed away => which prevents from using it,
+		 * so just sticking it into pipe wm state for
+		 * keeping it simple - anyway this is related to wm.
+		 * Proper way in ideal universe would be of course not
+		 * to lose parent atomic state object from child crtc_state,
+		 * and stick to OOP programming principles, which had been
+		 * scientifically proven to work.

Seems like random ramblings. I'd just nuke the comment.

Quoted 180 lines
drivers/gpu/drm/i915/intel_pm.c +100 -5
+		 */
+		pipe_wm->can_sagv = intel_can_enable_sagv(new_bw_state);
+	}
+
 	if (intel_can_enable_sagv(new_bw_state) != intel_can_enable_sagv(old_bw_state)) {
 		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
 		if (ret)
@@ -4664,12 +4687,39 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 		   int level,
 		   int color_plane)
 {
-	const struct skl_plane_wm *wm =
-		&crtc_state->wm.skl.optimal.planes[plane_id];
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (!level) {
+		if (pipe_wm->can_sagv)
+			return color_plane == 0 ? &wm->sagv_wm0 : &wm->uv_sagv_wm0;
+	}
 
 	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *plane_alloc =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (skl_ddb_entry_size(plane_alloc) < wm->sagv_wm0.min_ddb_alloc)
+			return false;
+	}
+
+	return true;
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -5251,10 +5301,17 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
 static void
 skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      const struct skl_wm_params *wm_params,
-		      struct skl_wm_level *levels)
+		      struct skl_plane_wm *plane_wm,
+		      int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int level, max_level = ilk_wm_max_level(dev_priv);
+	/*
+	 * Check which kind of plane is it and based on that calculate
+	 * correspondent WM levels.
+	 */
+	struct skl_wm_level *levels = color_plane == 1 ?
+				      plane_wm->uv_wm : plane_wm->wm;
 	struct skl_wm_level *result_prev = &levels[0];
 
 	for (level = 0; level <= max_level; level++) {
@@ -5268,6 +5325,40 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 	}
 }
 
+static void skl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm,
+				int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = color_plane == 1 ?
+				&plane_wm->uv_sagv_wm0 : &plane_wm->sagv_wm0;
+	struct skl_wm_level *levels = color_plane == 1 ?
+				plane_wm->uv_wm : plane_wm->wm;
+
+	/*
+	 * For Gen12 if it is an L0 we need to also
+	 * consider sagv_block_time when calculating
+	 * L0 watermark - we will need that when making
+	 * a decision whether enable SAGV or not.
+	 * For older gens we agreed to copy L0 value for
+	 * compatibility.
+	 */
+	if ((INTEL_GEN(dev_priv) >= 12)) {
+		u32 latency = dev_priv->wm.skl_latency[0];
+
+		latency += dev_priv->sagv_block_time_us;
+		skl_compute_plane_wm(crtc_state, 0, latency,
+				     wm_params, &levels[0],
+				     sagv_wm);
+		DRM_DEBUG_KMS("%d L0 blocks required for SAGV vs %d for non-SAGV\n",
+			      sagv_wm->min_ddb_alloc, levels[0].min_ddb_alloc);
+	} else {
+		/* Since all members are POD */
+		*sagv_wm = levels[0];
+	}
+}
+
 static void skl_compute_transition_wm(const struct intel_crtc_state *crtc_state,
 				      const struct skl_wm_params *wp,
 				      struct skl_plane_wm *wm)
@@ -5344,7 +5435,8 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 0);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 0);
 	skl_compute_transition_wm(crtc_state, &wm_params, wm);
 
 	return 0;
@@ -5366,7 +5458,8 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->uv_wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 1);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 1);
 
 	return 0;
 }
@@ -5751,6 +5844,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].plane_res_l,
 				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.plane_res_l);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 lines %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_l,
+				    new_wm->sagv_wm0.plane_res_l);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5766,6 +5865,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[6].plane_res_b, new_wm->wm[7].plane_res_b,
 				    new_wm->trans_wm.plane_res_b);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 blocks %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_b,
+				    new_wm->sagv_wm0.plane_res_b);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5780,6 +5885,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
 				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
 				    new_wm->trans_wm.min_ddb_alloc);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 min ddb %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.min_ddb_alloc,
+				    new_wm->sagv_wm0.min_ddb_alloc);
 		}
 	}
 }
@@ -6072,6 +6183,9 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			skl_wm_level_from_reg_val(val, &wm->wm[level]);
 		}
 
+		memcpy(&wm->sagv_wm0, &wm->wm[0],
+		       sizeof(struct skl_wm_level));
+
 		if (plane_id != PLANE_CURSOR)
 			val = I915_READ(PLANE_WM_TRANS(pipe, plane_id));
 		else
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#15 lore raw

Re: [Intel-gfx] [PATCH v27 3/6] drm/i915: Add TGL+ SAGV support

Quoted 63 lines · Ville Syrjälä
Quoted 56 lines · Stanislav
Quoted 41 lines · Ville Syrjälä
Quoted 40 lines · Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
Quoted 13 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 692

Quoted 11 lines
drivers/gpu/drm/i915/display/intel_display.c +0 -0
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
index 9488449e4b94..32cbbf7dddc6 100644
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;

As mentioned before uv_wm is not a thing on icl+, so nuke this.

Quoted 5 lines

This is used in skl_plane_wm_level accessor, which is used for all platforms, not just icl+. I remember we had agreed that for all platforms before tgl I simply copy sagv_wm0 values from regular wm0, so that this behaviour is unified(remember that your comment about memcpy which I changed to assignment, see skl_compute_sagv_wm).

Quoted 4 lines

I think having that duplicated is just making things more confusing. Also uv_wm is never used by the hardware even on pre-icl, so having the accessor thing use it for the hw programming just doesn't make any sense.

Problem is that we use it in skl_allocate_pipe_ddb in a few places:

blocks += wm_level->min_ddb_alloc;
blocks += wm_uv_level->min_ddb_alloc;

then those plane data rates:

rate = plane_data_rate[plane_id];
extra = min_t(u16, alloc_size,

      DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate));

total[plane_id] = wm_level->min_ddb_alloc + extra; alloc_size -= extra;
total_data_rate -= rate;

if (total_data_rate == 0)

	break;

rate = uv_plane_data_rate[plane_id];
extra = min_t(u16, alloc_size,

      DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate));

uv_total[plane_id] = wm_uv_level->min_ddb_alloc + extra;

So if I remove this wm_uv from skl_plane_wm_level accessor, the logic in skl_allocate_pipe_ddb will obvisouly change?

Stan

Quoted 268 lines

For the compute side I think all we should really need is something like:

tgl_compute_sagv_wm()

{
	skl_compute_wm_level(sagv_wm0, latency + whatever);
}

skl_build_plane_wm_single()

{
	...
	skl_compute_transition_wm();
(quoted diff) +2 -0
+	if (gen >= 12)
+		tgl_compute_sagv_wm();

}

And for the progamming side we should just pick the right wm0 based on the crtc_state->use_sagv_wm or whatever flag.

Quoted 248 lines

So if I remove it and this is called for pre-icl platforms, guess this would still need wm_uv data?

Stan

Quoted 241 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 698

Quoted 7 lines
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;

I would call it use_sagv_wm or somesuch to make it actually clear what it does.

Quoted 45 lines
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +11 -2
index c7d726a656b2..1b9925b6672c 100644
@@ -3871,6 +3871,9 @@ static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 	return intel_crtc_can_enable_sagv(crtc_state);
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state);
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3884,7 +3887,7 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
@@ -3899,7 +3902,9 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (INTEL_GEN(dev_priv) >= 11)
+		if (INTEL_GEN(dev_priv) >= 12)
+			can_sagv = tgl_crtc_can_enable_sagv(new_crtc_state);
+		else if (INTEL_GEN(dev_priv) >= 11)
 			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
 		else
 			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
@@ -3921,6 +3926,24 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 			return ret;
 	}
 
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+

drivers/gpu/drm/i915/intel_pm.c static int intel_compute_sagv_mask(struct intel_atomic_state *state) · line 3942

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+		/*
+		 * Due to drm limitation at commit state, when
+		 * changes are written the whole atomic state is
+		 * zeroed away => which prevents from using it,
+		 * so just sticking it into pipe wm state for
+		 * keeping it simple - anyway this is related to wm.
+		 * Proper way in ideal universe would be of course not
+		 * to lose parent atomic state object from child crtc_state,
+		 * and stick to OOP programming principles, which had been
+		 * scientifically proven to work.

Seems like random ramblings. I'd just nuke the comment.

Quoted 172 lines
drivers/gpu/drm/i915/intel_pm.c +97 -5
+		 */
+		pipe_wm->can_sagv = intel_can_enable_sagv(new_bw_state);
+	}
+
 	if (intel_can_enable_sagv(new_bw_state) != intel_can_enable_sagv(old_bw_state)) {
 		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
 		if (ret)
@@ -4664,12 +4687,39 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 		   int level,
 		   int color_plane)
 {
-	const struct skl_plane_wm *wm =
-		&crtc_state->wm.skl.optimal.planes[plane_id];
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (!level) {
+		if (pipe_wm->can_sagv)
+			return color_plane == 0 ? &wm->sagv_wm0 : &wm->uv_sagv_wm0;
+	}
 
 	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *plane_alloc =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (skl_ddb_entry_size(plane_alloc) < wm->sagv_wm0.min_ddb_alloc)
+			return false;
+	}
+
+	return true;
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -5251,10 +5301,17 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
 static void
 skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      const struct skl_wm_params *wm_params,
-		      struct skl_wm_level *levels)
+		      struct skl_plane_wm *plane_wm,
+		      int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int level, max_level = ilk_wm_max_level(dev_priv);
+	/*
+	 * Check which kind of plane is it and based on that calculate
+	 * correspondent WM levels.
+	 */
+	struct skl_wm_level *levels = color_plane == 1 ?
+				      plane_wm->uv_wm : plane_wm->wm;
 	struct skl_wm_level *result_prev = &levels[0];
 
 	for (level = 0; level <= max_level; level++) {
@@ -5268,6 +5325,40 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 	}
 }
 
+static void skl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm,
+				int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = color_plane == 1 ?
+				&plane_wm->uv_sagv_wm0 : &plane_wm->sagv_wm0;
+	struct skl_wm_level *levels = color_plane == 1 ?
+				plane_wm->uv_wm : plane_wm->wm;
+
+	/*
+	 * For Gen12 if it is an L0 we need to also
+	 * consider sagv_block_time when calculating
+	 * L0 watermark - we will need that when making
+	 * a decision whether enable SAGV or not.
+	 * For older gens we agreed to copy L0 value for
+	 * compatibility.
+	 */
+	if ((INTEL_GEN(dev_priv) >= 12)) {
+		u32 latency = dev_priv->wm.skl_latency[0];
+
+		latency += dev_priv->sagv_block_time_us;
+		skl_compute_plane_wm(crtc_state, 0, latency,
+				     wm_params, &levels[0],
+				     sagv_wm);
+		DRM_DEBUG_KMS("%d L0 blocks required for SAGV vs %d for non-SAGV\n",
+			      sagv_wm->min_ddb_alloc, levels[0].min_ddb_alloc);
+	} else {
+		/* Since all members are POD */
+		*sagv_wm = levels[0];
+	}
+}
+
 static void skl_compute_transition_wm(const struct intel_crtc_state *crtc_state,
 				      const struct skl_wm_params *wp,
 				      struct skl_plane_wm *wm)
@@ -5344,7 +5435,8 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 0);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 0);
 	skl_compute_transition_wm(crtc_state, &wm_params, wm);
 
 	return 0;
@@ -5366,7 +5458,8 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->uv_wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 1);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 1);
 
 	return 0;
 }
@@ -5751,6 +5844,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].plane_res_l,
 				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.plane_res_l);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 lines %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_l,
+				    new_wm->sagv_wm0.plane_res_l);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5766,6 +5865,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[6].plane_res_b, new_wm->wm[7].plane_res_b,
 				    new_wm->trans_wm.plane_res_b);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 blocks %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_b,
+				    new_wm->sagv_wm0.plane_res_b);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5780,6 +5885,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
 				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
 				    new_wm->trans_wm.min_ddb_alloc);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 min ddb %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.min_ddb_alloc,
+				    new_wm->sagv_wm0.min_ddb_alloc);
 		}
 	}
 }
@@ -6072,6 +6183,9 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			skl_wm_level_from_reg_val(val, &wm->wm[level]);
 		}
 
Quoted 16 lines
Quoted 12 lines
Quoted 12 lines
Quoted 8 lines
drivers/gpu/drm/i915/intel_pm.c +3 -0
+		memcpy(&wm->sagv_wm0, &wm->wm[0],
+		       sizeof(struct skl_wm_level));
+
 		if (plane_id != PLANE_CURSOR)
 			val = I915_READ(PLANE_WM_TRANS(pipe, plane_id));
 		else
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 2 lines
Ville Syrjälä
Intel

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

#16 lore raw

Re: [Intel-gfx] [PATCH v27 3/6] drm/i915: Add TGL+ SAGV support

Quoted 69 lines · Stanislav
Quoted 63 lines · Ville Syrjälä
Quoted 56 lines · Stanislav
Quoted 41 lines · Ville Syrjälä
Quoted 40 lines · Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
Quoted 13 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 692

Quoted 11 lines
drivers/gpu/drm/i915/display/intel_display.c +0 -0
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
index 9488449e4b94..32cbbf7dddc6 100644
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;

As mentioned before uv_wm is not a thing on icl+, so nuke this.

Quoted 5 lines

This is used in skl_plane_wm_level accessor, which is used for all platforms, not just icl+. I remember we had agreed that for all platforms before tgl I simply copy sagv_wm0 values from regular wm0, so that this behaviour is unified(remember that your comment about memcpy which I changed to assignment, see skl_compute_sagv_wm).

Quoted 4 lines

I think having that duplicated is just making things more confusing. Also uv_wm is never used by the hardware even on pre-icl, so having the accessor thing use it for the hw programming just doesn't make any sense.

Quoted 4 lines

Problem is that we use it in skl_allocate_pipe_ddb in a few places:

blocks += wm_level->min_ddb_alloc;
blocks += wm_uv_level->min_ddb_alloc;

I don't think we should need any accessor in there. If we do what I suggested earlier, which I think was something like (can't find the mail anymore unfortunately):

for reverse wm levels {
blocks = sum(wm[level] blocks for all planes) +
sum(uv_wm[level] blocks for all planes)
...
if (block <= alloc_size)
break;
}

if (level < 0)

	fail;

blocks_sagv = sum(sagv_wm blocks for all planes)

if (blocks_sagv > blocks && blocks_sagv <= alloc_size) {
    blocks = blocks_sagv;
    use_sagv_wm = true;
} else {
    use_sagv_wm = false;
}

alloc_size -= blocks;

for_each_plane_id() {
	total = use_sagv_wm ? sagv_wm.min_ddb_alloc :
		wm[level].min_ddb_alloc;
	...
}
Quoted 310 lines

then those plane data rates:

rate = plane_data_rate[plane_id];
extra = min_t(u16, alloc_size,

      DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate));

total[plane_id] = wm_level->min_ddb_alloc + extra; alloc_size -= extra;
total_data_rate -= rate;

if (total_data_rate == 0)

	break;

rate = uv_plane_data_rate[plane_id];
extra = min_t(u16, alloc_size,

      DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate));

uv_total[plane_id] = wm_uv_level->min_ddb_alloc + extra;

So if I remove this wm_uv from skl_plane_wm_level accessor, the logic in skl_allocate_pipe_ddb will obvisouly change?

Stan

Quoted 284 lines

For the compute side I think all we should really need is something like:

tgl_compute_sagv_wm()

{
	skl_compute_wm_level(sagv_wm0, latency + whatever);
}

skl_build_plane_wm_single()

{
	...
	skl_compute_transition_wm();
(quoted diff) +2 -0
+	if (gen >= 12)
+		tgl_compute_sagv_wm();

}

And for the progamming side we should just pick the right wm0 based on the crtc_state->use_sagv_wm or whatever flag.

Quoted 260 lines

So if I remove it and this is called for pre-icl platforms, guess this would still need wm_uv data?

Stan

Quoted 253 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 698

Quoted 7 lines
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;

I would call it use_sagv_wm or somesuch to make it actually clear what it does.

Quoted 45 lines
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +11 -2
index c7d726a656b2..1b9925b6672c 100644
@@ -3871,6 +3871,9 @@ static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 	return intel_crtc_can_enable_sagv(crtc_state);
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state);
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3884,7 +3887,7 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
@@ -3899,7 +3902,9 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (INTEL_GEN(dev_priv) >= 11)
+		if (INTEL_GEN(dev_priv) >= 12)
+			can_sagv = tgl_crtc_can_enable_sagv(new_crtc_state);
+		else if (INTEL_GEN(dev_priv) >= 11)
 			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
 		else
 			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
@@ -3921,6 +3926,24 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 			return ret;
 	}
 
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+

drivers/gpu/drm/i915/intel_pm.c static int intel_compute_sagv_mask(struct intel_atomic_state *state) · line 3942

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+		/*
+		 * Due to drm limitation at commit state, when
+		 * changes are written the whole atomic state is
+		 * zeroed away => which prevents from using it,
+		 * so just sticking it into pipe wm state for
+		 * keeping it simple - anyway this is related to wm.
+		 * Proper way in ideal universe would be of course not
+		 * to lose parent atomic state object from child crtc_state,
+		 * and stick to OOP programming principles, which had been
+		 * scientifically proven to work.

Seems like random ramblings. I'd just nuke the comment.

Quoted 180 lines
drivers/gpu/drm/i915/intel_pm.c +100 -5
+		 */
+		pipe_wm->can_sagv = intel_can_enable_sagv(new_bw_state);
+	}
+
 	if (intel_can_enable_sagv(new_bw_state) != intel_can_enable_sagv(old_bw_state)) {
 		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
 		if (ret)
@@ -4664,12 +4687,39 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 		   int level,
 		   int color_plane)
 {
-	const struct skl_plane_wm *wm =
-		&crtc_state->wm.skl.optimal.planes[plane_id];
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (!level) {
+		if (pipe_wm->can_sagv)
+			return color_plane == 0 ? &wm->sagv_wm0 : &wm->uv_sagv_wm0;
+	}
 
 	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *plane_alloc =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (skl_ddb_entry_size(plane_alloc) < wm->sagv_wm0.min_ddb_alloc)
+			return false;
+	}
+
+	return true;
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -5251,10 +5301,17 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
 static void
 skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      const struct skl_wm_params *wm_params,
-		      struct skl_wm_level *levels)
+		      struct skl_plane_wm *plane_wm,
+		      int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int level, max_level = ilk_wm_max_level(dev_priv);
+	/*
+	 * Check which kind of plane is it and based on that calculate
+	 * correspondent WM levels.
+	 */
+	struct skl_wm_level *levels = color_plane == 1 ?
+				      plane_wm->uv_wm : plane_wm->wm;
 	struct skl_wm_level *result_prev = &levels[0];
 
 	for (level = 0; level <= max_level; level++) {
@@ -5268,6 +5325,40 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 	}
 }
 
+static void skl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm,
+				int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = color_plane == 1 ?
+				&plane_wm->uv_sagv_wm0 : &plane_wm->sagv_wm0;
+	struct skl_wm_level *levels = color_plane == 1 ?
+				plane_wm->uv_wm : plane_wm->wm;
+
+	/*
+	 * For Gen12 if it is an L0 we need to also
+	 * consider sagv_block_time when calculating
+	 * L0 watermark - we will need that when making
+	 * a decision whether enable SAGV or not.
+	 * For older gens we agreed to copy L0 value for
+	 * compatibility.
+	 */
+	if ((INTEL_GEN(dev_priv) >= 12)) {
+		u32 latency = dev_priv->wm.skl_latency[0];
+
+		latency += dev_priv->sagv_block_time_us;
+		skl_compute_plane_wm(crtc_state, 0, latency,
+				     wm_params, &levels[0],
+				     sagv_wm);
+		DRM_DEBUG_KMS("%d L0 blocks required for SAGV vs %d for non-SAGV\n",
+			      sagv_wm->min_ddb_alloc, levels[0].min_ddb_alloc);
+	} else {
+		/* Since all members are POD */
+		*sagv_wm = levels[0];
+	}
+}
+
 static void skl_compute_transition_wm(const struct intel_crtc_state *crtc_state,
 				      const struct skl_wm_params *wp,
 				      struct skl_plane_wm *wm)
@@ -5344,7 +5435,8 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 0);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 0);
 	skl_compute_transition_wm(crtc_state, &wm_params, wm);
 
 	return 0;
@@ -5366,7 +5458,8 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->uv_wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 1);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 1);
 
 	return 0;
 }
@@ -5751,6 +5844,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].plane_res_l,
 				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.plane_res_l);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 lines %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_l,
+				    new_wm->sagv_wm0.plane_res_l);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5766,6 +5865,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[6].plane_res_b, new_wm->wm[7].plane_res_b,
 				    new_wm->trans_wm.plane_res_b);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 blocks %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_b,
+				    new_wm->sagv_wm0.plane_res_b);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5780,6 +5885,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
 				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
 				    new_wm->trans_wm.min_ddb_alloc);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 min ddb %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.min_ddb_alloc,
+				    new_wm->sagv_wm0.min_ddb_alloc);
 		}
 	}
 }
@@ -6072,6 +6183,9 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			skl_wm_level_from_reg_val(val, &wm->wm[level]);
 		}
 
+		memcpy(&wm->sagv_wm0, &wm->wm[0],
+		       sizeof(struct skl_wm_level));
+
 		if (plane_id != PLANE_CURSOR)
 			val = I915_READ(PLANE_WM_TRANS(pipe, plane_id));
 		else
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#17 lore raw

Re: [Intel-gfx] [PATCH v27 3/6] drm/i915: Add TGL+ SAGV support

Quoted 97 lines · Ville Syrjälä
Quoted 69 lines · Stanislav
Quoted 63 lines · Ville Syrjälä
Quoted 56 lines · Stanislav
Quoted 41 lines · Ville Syrjälä
Quoted 40 lines · Stanislav Lisovskiy

Starting from TGL we need to have a separate wm0 values for SAGV and non-SAGV which affects how calculations are done.

v2: Remove long lines
v3: Removed COLOR_PLANE enum references v4, v5, v6: Fixed rebase conflict

---
 drivers/gpu/drm/i915/display/intel_display.c  |   8 +-
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               | 128 +++++++++++++++++-
 3 files changed, 130 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +6 -2
index fd6d63b03489..be5741cb7595 100644
@@ -13961,7 +13961,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
 			drm_err(&dev_priv->drm,
@@ -14016,7 +14018,9 @@ static void verify_wm_state(struct intel_crtc *crtc,
 		/* Watermarks */
 		for (level = 0; level <= max_level; level++) {
 			if (skl_wm_level_equals(&hw_plane_wm->wm[level],
-						&sw_plane_wm->wm[level]))
+						&sw_plane_wm->wm[level]) ||
+			    (level == 0 && skl_wm_level_equals(&hw_plane_wm->wm[level],
+							       &sw_plane_wm->sagv_wm0)))
 				continue;
 
Quoted 13 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 692

Quoted 11 lines
drivers/gpu/drm/i915/display/intel_display.c +0 -0
 			drm_err(&dev_priv->drm,
drivers/gpu/drm/i915/display/intel_display_types.h +2 -0
index 9488449e4b94..32cbbf7dddc6 100644
@@ -688,11 +688,14 @@ struct skl_plane_wm {
 	struct skl_wm_level wm[8];
 	struct skl_wm_level uv_wm[8];
 	struct skl_wm_level trans_wm;
+	struct skl_wm_level sagv_wm0;
+	struct skl_wm_level uv_sagv_wm0;

As mentioned before uv_wm is not a thing on icl+, so nuke this.

Quoted 5 lines

This is used in skl_plane_wm_level accessor, which is used for all platforms, not just icl+. I remember we had agreed that for all platforms before tgl I simply copy sagv_wm0 values from regular wm0, so that this behaviour is unified(remember that your comment about memcpy which I changed to assignment, see skl_compute_sagv_wm).

Quoted 4 lines

I think having that duplicated is just making things more confusing. Also uv_wm is never used by the hardware even on pre-icl, so having the accessor thing use it for the hw programming just doesn't make any sense.

Quoted 4 lines

Problem is that we use it in skl_allocate_pipe_ddb in a few places:

blocks += wm_level->min_ddb_alloc;
blocks += wm_uv_level->min_ddb_alloc;

I don't think we should need any accessor in there. If we do what I suggested earlier, which I think was something like (can't find the mail anymore unfortunately):

for reverse wm levels {
blocks = sum(wm[level] blocks for all planes) +
sum(uv_wm[level] blocks for all planes)
...
if (block <= alloc_size)
break;
}

if (level < 0)

	fail;

blocks_sagv = sum(sagv_wm blocks for all planes)

if (blocks_sagv > blocks && blocks_sagv <= alloc_size) {
    blocks = blocks_sagv;
    use_sagv_wm = true;
} else {
    use_sagv_wm = false;
}
Quoted 9 lines

alloc_size -= blocks;

for_each_plane_id() {
	total = use_sagv_wm ? sagv_wm.min_ddb_alloc :
		wm[level].min_ddb_alloc;
	...
}

It is actually almost same behaviour now:
skl_compute_ddb->skl_sllocate_pipe_ddb is now invoked before we call intel_compute_sagv_mask, which means that use_sagv_wm is set to false, so it will be just transparent leading to same wm->wm[level], wm->uv_wm[level] accesses as done now.
Then in intel_compute_sagv_mask we calculate if each plane ddb entry is big enough to fit SAGV min_ddb_alloc and set use_sagv_wm bit in a pipe_wm struct.
So the only difference here is that use_sagv_wm bit is set in intel_compute_sagv_mask and not skl_allocate_pipe_ddb, which seems logical imo and anyway logic would be the same.

This accessor itself is then conveniently used in skl_write_plane and skl_write_cursor_wm and others, providing proper wm level(otherwise there would be duplicated condition everywhere).

So should I throw away skl_plane_wm_accessor completely? Because we anyway use wm->uv_wm in skl_allocate_pipe_ddb and the logic is otherwise same.

Stan

Quoted 308 lines
Quoted 308 lines

then those plane data rates:

rate = plane_data_rate[plane_id];
extra = min_t(u16, alloc_size,

      DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate));

total[plane_id] = wm_level->min_ddb_alloc + extra; alloc_size -= extra;
total_data_rate -= rate;

if (total_data_rate == 0)

	break;

rate = uv_plane_data_rate[plane_id];
extra = min_t(u16, alloc_size,

      DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate));

uv_total[plane_id] = wm_uv_level->min_ddb_alloc + extra;

So if I remove this wm_uv from skl_plane_wm_level accessor, the logic in skl_allocate_pipe_ddb will obvisouly change?

Stan

Quoted 281 lines

For the compute side I think all we should really need is something like:

tgl_compute_sagv_wm()

{
	skl_compute_wm_level(sagv_wm0, latency + whatever);
}

skl_build_plane_wm_single()

{
	...
	skl_compute_transition_wm();
(quoted diff) +2 -0
+	if (gen >= 12)
+		tgl_compute_sagv_wm();

}

And for the progamming side we should just pick the right wm0 based on the crtc_state->use_sagv_wm or whatever flag.

Quoted 260 lines

So if I remove it and this is called for pre-icl platforms, guess this would still need wm_uv data?

Stan

Quoted 253 lines

drivers/gpu/drm/i915/display/intel_display_types.h struct skl_plane_wm { · line 698

Quoted 7 lines
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
 	bool is_planar;
 };
 
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool can_sagv;

I would call it use_sagv_wm or somesuch to make it actually clear what it does.

Quoted 45 lines
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
 };
 
 enum vlv_wm_level {
drivers/gpu/drm/i915/intel_pm.c +11 -2
index c7d726a656b2..1b9925b6672c 100644
@@ -3871,6 +3871,9 @@ static bool icl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 	return intel_crtc_can_enable_sagv(crtc_state);
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state);
+
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state)
 {
 	if (bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
@@ -3884,7 +3887,7 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	int ret;
 	struct intel_crtc *crtc;
-	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
 	const struct intel_bw_state *old_bw_state = NULL;
 	int i;
@@ -3899,7 +3902,9 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 
 		old_bw_state = intel_atomic_get_old_bw_state(state);
 
-		if (INTEL_GEN(dev_priv) >= 11)
+		if (INTEL_GEN(dev_priv) >= 12)
+			can_sagv = tgl_crtc_can_enable_sagv(new_crtc_state);
+		else if (INTEL_GEN(dev_priv) >= 11)
 			can_sagv = icl_crtc_can_enable_sagv(new_crtc_state);
 		else
 			can_sagv = skl_crtc_can_enable_sagv(new_crtc_state);
@@ -3921,6 +3926,24 @@ static int intel_compute_sagv_mask(struct intel_atomic_state *state)
 			return ret;
 	}
 
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+

drivers/gpu/drm/i915/intel_pm.c static int intel_compute_sagv_mask(struct intel_atomic_state *state) · line 3942

Quoted 10 lines
drivers/gpu/drm/i915/intel_pm.c +10 -0
+		/*
+		 * Due to drm limitation at commit state, when
+		 * changes are written the whole atomic state is
+		 * zeroed away => which prevents from using it,
+		 * so just sticking it into pipe wm state for
+		 * keeping it simple - anyway this is related to wm.
+		 * Proper way in ideal universe would be of course not
+		 * to lose parent atomic state object from child crtc_state,
+		 * and stick to OOP programming principles, which had been
+		 * scientifically proven to work.

Seems like random ramblings. I'd just nuke the comment.

Quoted 180 lines
drivers/gpu/drm/i915/intel_pm.c +100 -5
+		 */
+		pipe_wm->can_sagv = intel_can_enable_sagv(new_bw_state);
+	}
+
 	if (intel_can_enable_sagv(new_bw_state) != intel_can_enable_sagv(old_bw_state)) {
 		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
 		if (ret)
@@ -4664,12 +4687,39 @@ skl_plane_wm_level(const struct intel_crtc_state *crtc_state,
 		   int level,
 		   int color_plane)
 {
-	const struct skl_plane_wm *wm =
-		&crtc_state->wm.skl.optimal.planes[plane_id];
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (!level) {
+		if (pipe_wm->can_sagv)
+			return color_plane == 0 ? &wm->sagv_wm0 : &wm->uv_sagv_wm0;
+	}
 
 	return color_plane == 0 ? &wm->wm[level] : &wm->uv_wm[level];
 }
 
+static bool
+tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *plane_alloc =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (skl_ddb_entry_size(plane_alloc) < wm->sagv_wm0.min_ddb_alloc)
+			return false;
+	}
+
+	return true;
+}
+
 static int
 skl_allocate_pipe_ddb(struct intel_crtc_state *crtc_state)
 {
@@ -5251,10 +5301,17 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
 static void
 skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      const struct skl_wm_params *wm_params,
-		      struct skl_wm_level *levels)
+		      struct skl_plane_wm *plane_wm,
+		      int color_plane)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
 	int level, max_level = ilk_wm_max_level(dev_priv);
+	/*
+	 * Check which kind of plane is it and based on that calculate
+	 * correspondent WM levels.
+	 */
+	struct skl_wm_level *levels = color_plane == 1 ?
+				      plane_wm->uv_wm : plane_wm->wm;
 	struct skl_wm_level *result_prev = &levels[0];
 
 	for (level = 0; level <= max_level; level++) {
@@ -5268,6 +5325,40 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 	}
 }
 
+static void skl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm,
+				int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = color_plane == 1 ?
+				&plane_wm->uv_sagv_wm0 : &plane_wm->sagv_wm0;
+	struct skl_wm_level *levels = color_plane == 1 ?
+				plane_wm->uv_wm : plane_wm->wm;
+
+	/*
+	 * For Gen12 if it is an L0 we need to also
+	 * consider sagv_block_time when calculating
+	 * L0 watermark - we will need that when making
+	 * a decision whether enable SAGV or not.
+	 * For older gens we agreed to copy L0 value for
+	 * compatibility.
+	 */
+	if ((INTEL_GEN(dev_priv) >= 12)) {
+		u32 latency = dev_priv->wm.skl_latency[0];
+
+		latency += dev_priv->sagv_block_time_us;
+		skl_compute_plane_wm(crtc_state, 0, latency,
+				     wm_params, &levels[0],
+				     sagv_wm);
+		DRM_DEBUG_KMS("%d L0 blocks required for SAGV vs %d for non-SAGV\n",
+			      sagv_wm->min_ddb_alloc, levels[0].min_ddb_alloc);
+	} else {
+		/* Since all members are POD */
+		*sagv_wm = levels[0];
+	}
+}
+
 static void skl_compute_transition_wm(const struct intel_crtc_state *crtc_state,
 				      const struct skl_wm_params *wp,
 				      struct skl_plane_wm *wm)
@@ -5344,7 +5435,8 @@ static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 0);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 0);
 	skl_compute_transition_wm(crtc_state, &wm_params, wm);
 
 	return 0;
@@ -5366,7 +5458,8 @@ static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
-	skl_compute_wm_levels(crtc_state, &wm_params, wm->uv_wm);
+	skl_compute_wm_levels(crtc_state, &wm_params, wm, 1);
+	skl_compute_sagv_wm(crtc_state, &wm_params, wm, 1);
 
 	return 0;
 }
@@ -5751,6 +5844,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].plane_res_l,
 				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.plane_res_l);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 lines %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_l,
+				    new_wm->sagv_wm0.plane_res_l);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5766,6 +5865,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[6].plane_res_b, new_wm->wm[7].plane_res_b,
 				    new_wm->trans_wm.plane_res_b);
 
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 blocks %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.plane_res_b,
+				    new_wm->sagv_wm0.plane_res_b);
+
 			drm_dbg_kms(&dev_priv->drm,
 				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d"
 				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d\n",
@@ -5780,6 +5885,12 @@ skl_print_wm_changes(struct intel_atomic_state *state)
 				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
 				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
 				    new_wm->trans_wm.min_ddb_alloc);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[PLANE:%d:%s] sagv wm0 min ddb %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->sagv_wm0.min_ddb_alloc,
+				    new_wm->sagv_wm0.min_ddb_alloc);
 		}
 	}
 }
@@ -6072,6 +6183,9 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			skl_wm_level_from_reg_val(val, &wm->wm[level]);
 		}
 
+		memcpy(&wm->sagv_wm0, &wm->wm[0],
+		       sizeof(struct skl_wm_level));
+
 		if (plane_id != PLANE_CURSOR)
 			val = I915_READ(PLANE_WM_TRANS(pipe, plane_id));
 		else
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 2 lines
Ville Syrjälä
Intel
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 2 lines
Ville Syrjälä
Intel

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

#18 lore raw

PATCH v27 4/6 drm/i915: Added required new PCode commands

In reply to Stanislav Lisovskiy

We need a new PCode request commands and reply codes to be added as a prepartion patch for QGV points restricting for new SAGV support.

v2: - Extracted those changes into separate patch
(Ville Syrjälä)

v3: - Moved new PCode masks to another place from
PCode commands(Ville)

v4: - Moved new PCode masks to correspondent PCode
command, with identation(Ville)
- Changed naming to ICL_ instead of GEN11_ to fit more nicely into existing definition style.

---
 drivers/gpu/drm/i915/i915_reg.h       | 4 ++++
 drivers/gpu/drm/i915/intel_sideband.c | 2 ++
 2 files changed, 6 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +4 -0
index fd9f2904d93c..f23a18ee28f9 100644
@@ -9064,6 +9064,7 @@ enum {
 #define     GEN7_PCODE_ILLEGAL_DATA		0x3
 #define     GEN11_PCODE_ILLEGAL_SUBCOMMAND	0x4
 #define     GEN11_PCODE_LOCKED			0x6
+#define     GEN11_PCODE_REJECTED		0x11
 #define     GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE 0x10
 #define   GEN6_PCODE_WRITE_RC6VIDS		0x4
 #define   GEN6_PCODE_READ_RC6VIDS		0x5
@@ -9085,6 +9086,9 @@ enum {
 #define   ICL_PCODE_MEM_SUBSYSYSTEM_INFO	0xd
 #define     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO	(0x0 << 8)
 #define     ICL_PCODE_MEM_SS_READ_QGV_POINT_INFO(point)	(((point) << 16) | (0x1 << 8))
+#define   ICL_PCODE_SAGV_DE_MEM_SS_CONFIG	0xe
+#define     ICL_PCODE_POINTS_RESTRICTED		0x0
+#define     ICL_PCODE_POINTS_RESTRICTED_MASK	0x1
 #define   GEN6_PCODE_READ_D_COMP		0x10
 #define   GEN6_PCODE_WRITE_D_COMP		0x11
 #define   ICL_PCODE_EXIT_TCCOLD			0x12
drivers/gpu/drm/i915/intel_sideband.c +2 -0
index d5129c1dd452..916ccd1c0e96 100644
@@ -371,6 +371,8 @@ static int gen7_check_mailbox_status(u32 mbox)
 		return -ENXIO;
 	case GEN11_PCODE_LOCKED:
 		return -EBUSY;
+	case GEN11_PCODE_REJECTED:
+		return -EACCES;
 	case GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE:
 		return -EOVERFLOW;
 	default:
Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

Re: [Intel-gfx] [PATCH v27 4/6] drm/i915: Added required new PCode commands

Quoted 10 lines · Stanislav Lisovskiy

We need a new PCode request commands and reply codes to be added as a prepartion patch for QGV points restricting for new SAGV support.

v2: - Extracted those changes into separate patch
(Ville Syrjälä)

v3: - Moved new PCode masks to another place from
PCode commands(Ville)

Quoted 7 lines

v4: - Moved new PCode masks to correspondent PCode
command, with identation(Ville)
- Changed naming to ICL_ instead of GEN11_ to fit more nicely into existing definition style.

Thanks. Pushed this one.

Quoted 42 lines
---
 drivers/gpu/drm/i915/i915_reg.h       | 4 ++++
 drivers/gpu/drm/i915/intel_sideband.c | 2 ++
 2 files changed, 6 insertions(+)
drivers/gpu/drm/i915/i915_reg.h +4 -0
index fd9f2904d93c..f23a18ee28f9 100644
@@ -9064,6 +9064,7 @@ enum {
 #define     GEN7_PCODE_ILLEGAL_DATA		0x3
 #define     GEN11_PCODE_ILLEGAL_SUBCOMMAND	0x4
 #define     GEN11_PCODE_LOCKED			0x6
+#define     GEN11_PCODE_REJECTED		0x11
 #define     GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE 0x10
 #define   GEN6_PCODE_WRITE_RC6VIDS		0x4
 #define   GEN6_PCODE_READ_RC6VIDS		0x5
@@ -9085,6 +9086,9 @@ enum {
 #define   ICL_PCODE_MEM_SUBSYSYSTEM_INFO	0xd
 #define     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO	(0x0 << 8)
 #define     ICL_PCODE_MEM_SS_READ_QGV_POINT_INFO(point)	(((point) << 16) | (0x1 << 8))
+#define   ICL_PCODE_SAGV_DE_MEM_SS_CONFIG	0xe
+#define     ICL_PCODE_POINTS_RESTRICTED		0x0
+#define     ICL_PCODE_POINTS_RESTRICTED_MASK	0x1
 #define   GEN6_PCODE_READ_D_COMP		0x10
 #define   GEN6_PCODE_WRITE_D_COMP		0x11
 #define   ICL_PCODE_EXIT_TCCOLD			0x12
drivers/gpu/drm/i915/intel_sideband.c +2 -0
index d5129c1dd452..916ccd1c0e96 100644
@@ -371,6 +371,8 @@ static int gen7_check_mailbox_status(u32 mbox)
 		return -ENXIO;
 	case GEN11_PCODE_LOCKED:
 		return -EBUSY;
+	case GEN11_PCODE_REJECTED:
+		return -EACCES;
 	case GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE:
 		return -EOVERFLOW;
 	default:
Signature · 1 line
2.24.1.485.gad05a3d8e5
Signature · 6 lines
Ville Syrjälä
Intel
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#20 lore raw

PATCH v27 5/6 drm/i915: Restrict qgv points which don't have enough bandwidth.

In reply to Stanislav Lisovskiy

According to BSpec 53998, we should try to restrict qgv points, which can't provide enough bandwidth for desired display configuration.

Currently we are just comparing against all of
those and take minimum(worst case).

v2: Fixed wrong PCode reply mask, removed hardcoded
values.

v3: Forbid simultaneous legacy SAGV PCode requests and
restricting qgv points. Put the actual restriction to commit function, added serialization(thanks to Ville) to prevent commit being applied out of order in case of nonblocking and/or nomodeset commits.

v4:
- Minor code refactoring, fixed few typos(thanks to James Ausmus)
- Change the naming of qgv point
masking/unmasking functions(James Ausmus).
- Simplify the masking/unmasking operation itself,
as we don't need to mask only single point per request(James Ausmus)
- Reject and stick to highest bandwidth point if SAGV can't be enabled(BSpec)

v5:
- Add new mailbox reply codes, which seems to happen during boot time for TGL and indicate that QGV setting is not yet available.

v6:
- Increase number of supported QGV points to be in sync with BSpec.

v7: - Rebased and resolved conflict to fix build failure.
- Fix NUM_QGV_POINTS to 8 and moved that to header file(James Ausmus)

v8: - Don't report an error if we can't restrict qgv points, as SAGV
can be disabled by BIOS, which is completely legal. So don't make CI panic. Instead if we detect that there is only 1 QGV point accessible just analyze if we can fit the required bandwidth requirements, but no need in restricting.

v9: - Fix wrong QGV transition if we have 0 planes and no SAGV
simultaneously.

v10: - Fix CDCLK corruption, because of global state getting serialized
without modeset, which caused copying of non-calculated cdclk to be copied to dev_priv(thanks to Ville for the hint).

v11: - Remove unneeded headers and spaces(Matthew Roper)
- Remove unneeded intel_qgv_info qi struct from bw check and zero out the needed one(Matthew Roper)
- Changed QGV error message to have more clear meaning(Matthew Roper)
- Use state->modeset_set instead of any_ms(Matthew Roper)
- Moved NUM_SAGV_POINTS from i915_reg.h to i915_drv.h where it's used
- Keep using crtc_state->hw.active instead of .enable(Matthew Roper)
- Moved unrelated changes to other patch(using latency as parameter for plane wm calculation, moved to SAGV refactoring patch)

v12: - Fix rebase conflict with own temporary SAGV/QGV fix.
- Remove unnecessary mask being zero check when unmasking qgv points as this is completely legal(Matt Roper)
- Check if we are setting the same mask as already being set in hardware to prevent error from PCode.
- Fix error message when restricting/unrestricting qgv points to "mask/unmask" which sounds more accurate(Matt Roper)
- Move sagv status setting to icl_get_bw_info from atomic check as this should be calculated only once.(Matt Roper)
- Edited comments for the case when we can't enable SAGV and use only 1 QGV point with highest bandwidth to be more understandable.(Matt Roper)

v13: - Moved max_data_rate in bw check to closer scope(Ville Syrjälä)
- Changed comment for zero new_mask in qgv points masking function to better reflect reality(Ville Syrjälä)
- Simplified bit mask operation in qgv points masking function (Ville Syrjälä)
- Moved intel_qgv_points_mask closer to gen11 SAGV disabling, however this still can't be under modeset condition(Ville Syrjälä)
- Packed qgv_points_mask as u8 and moved closer to pipe_sagv_mask (Ville Syrjälä)
- Extracted PCode changes to separate patch.(Ville Syrjälä)
- Now treat num_planes 0 same as 1 to avoid confusion and returning max_bw as 0, which would prevent choosing QGV point having max bandwidth in case if SAGV is not allowed, as per BSpec(Ville Syrjälä)
- Do the actual qgv_points_mask swap in the same place as all other global state parts like cdclk are swapped. In the next patch, this all will be moved to bw state as global state, once new global state patch series from Ville lands

v14: - Now using global state to serialize access to qgv points
- Added global state locking back, otherwise we seem to read bw state in a wrong way.

v15: - Added TODO comment for near atomic global state locking in
bw code.

v16: - Fixed intel_atomic_bw_* functions to be intel_bw_* as discussed
with Jani Nikula.
- Take bw_state_changed flag into use.

v17: - Moved qgv point related manipulations next to SAGV code, as
those are semantically related(Ville Syrjälä)
- Renamed those into intel_sagv_(pre)|(post)_plane_update (Ville Syrjälä)

v18: - Move sagv related calls from commit tail into
intel_sagv_(pre)|(post)_plane_update(Ville Syrjälä)

v19: - Use intel_atomic_get_bw_(old)|(new)_state which is intended
for commit tail stage.

v20: - Return max bandwidth for 0 planes(Ville)
- Constify old_bw_state in bw_atomic_check(Ville)
- Removed some debugs(Ville)
- Added data rate to debug print when no QGV points(Ville)
- Removed some comments(Ville)

v21, v22, v23: - Fixed rebase conflict

v24: - Changed PCode mask to use ICL_ prefix

Cc Ville Syrjälä <[email protected]>
Cc James Ausmus <[email protected]>
---
 drivers/gpu/drm/i915/display/intel_bw.c       | 139 +++++++++++++-----
 drivers/gpu/drm/i915/display/intel_bw.h       |   9 ++
 .../drm/i915/display/intel_display_types.h    |   3 +
 drivers/gpu/drm/i915/intel_pm.c               |  66 ++++++++-
 drivers/gpu/drm/i915/intel_pm.h               |   2 +
 5 files changed, 181 insertions(+), 38 deletions(-)
drivers/gpu/drm/i915/display/intel_bw.c +104 -35
index 6e7cc3a4f1aa..a192fd02c463 100644
@@ -8,6 +8,9 @@
 #include "intel_bw.h"
 #include "intel_display_types.h"
 #include "intel_sideband.h"
+#include "intel_atomic.h"
+#include "intel_pm.h"
+
 
 /* Parameters for Qclk Geyserville (QGV) */
 struct intel_qgv_point {
@@ -113,6 +116,26 @@ static int icl_pcode_read_qgv_point_info(struct drm_i915_private *dev_priv,
 	return 0;
 }
 
+int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
+				  u32 points_mask)
+{
+	int ret;
+
+	/* bspec says to keep retrying for at least 1 ms */
+	ret = skl_pcode_request(dev_priv, ICL_PCODE_SAGV_DE_MEM_SS_CONFIG,
+				points_mask,
+				ICL_PCODE_POINTS_RESTRICTED_MASK,
+				ICL_PCODE_POINTS_RESTRICTED,
+				1);
+
+	if (ret < 0) {
+		DRM_ERROR("Failed to disable qgv points (%d)\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
 static int icl_get_qgv_points(struct drm_i915_private *dev_priv,
 			      struct intel_qgv_info *qi)
 {
@@ -240,6 +263,16 @@ static int icl_get_bw_info(struct drm_i915_private *dev_priv, const struct intel
 			break;
 	}
 
+	/*
+	 * In case if SAGV is disabled in BIOS, we always get 1
+	 * SAGV point, but we can't send PCode commands to restrict it
+	 * as it will fail and pointless anyway.
+	 */
+	if (qi.num_points == 1)
+		dev_priv->sagv_status = I915_SAGV_NOT_CONTROLLED;
+	else
+		dev_priv->sagv_status = I915_SAGV_ENABLED;
+
 	return 0;
 }
 
@@ -248,6 +281,11 @@ static unsigned int icl_max_bw(struct drm_i915_private *dev_priv,
 {
 	int i;
 
+	/*
+	 * Let's return max bw for 0 planes
+	 */
+	num_planes = max(1, num_planes);
+
 	for (i = 0; i < ARRAY_SIZE(dev_priv->max_bw); i++) {
 		const struct intel_bw_info *bi =
 			&dev_priv->max_bw[i];
@@ -277,34 +315,6 @@ void intel_bw_init_hw(struct drm_i915_private *dev_priv)
 		icl_get_bw_info(dev_priv, &icl_sa_info);
 }
 
-static unsigned int intel_max_data_rate(struct drm_i915_private *dev_priv,
-					int num_planes)
-{
-	if (INTEL_GEN(dev_priv) >= 11) {
-		/*
-		 * Any bw group has same amount of QGV points
-		 */
-		const struct intel_bw_info *bi =
-			&dev_priv->max_bw[0];
-		unsigned int min_bw = UINT_MAX;
-		int i;
-
-		/*
-		 * FIXME with SAGV disabled maybe we can assume
-		 * point 1 will always be used? Seems to match
-		 * the behaviour observed in the wild.
-		 */
-		for (i = 0; i < bi->num_qgv_points; i++) {
-			unsigned int bw = icl_max_bw(dev_priv, num_planes, i);
-
-			min_bw = min(bw, min_bw);
-		}
-		return min_bw;
-	} else {
-		return UINT_MAX;
-	}
-}
-
 static unsigned int intel_bw_crtc_num_active_planes(const struct intel_crtc_state *crtc_state)
 {
 	/*
@@ -415,10 +425,15 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	struct intel_crtc_state *new_crtc_state, *old_crtc_state;
 	struct intel_bw_state *new_bw_state = NULL;
-	unsigned int data_rate, max_data_rate;
+	const struct intel_bw_state *old_bw_state = NULL;
+	unsigned int data_rate;
 	unsigned int num_active_planes;
 	struct intel_crtc *crtc;
 	int i, ret;
+	u32 allowed_points = 0;
+	unsigned int max_bw_point = 0, max_bw = 0;
+	unsigned int num_qgv_points = dev_priv->max_bw[0].num_qgv_points;
+	u32 mask = (1 << num_qgv_points) - 1;
 
 	/* FIXME earlier gens need some checks too */
 	if (INTEL_GEN(dev_priv) < 11)
@@ -465,19 +480,73 @@ int intel_bw_atomic_check(struct intel_atomic_state *state)
 		return ret;
 
 	data_rate = intel_bw_data_rate(dev_priv, new_bw_state);
+	data_rate = DIV_ROUND_UP(data_rate, 1000);
+
 	num_active_planes = intel_bw_num_active_planes(dev_priv, new_bw_state);
 
-	max_data_rate = intel_max_data_rate(dev_priv, num_active_planes);
+	for (i = 0; i < num_qgv_points; i++) {
+		unsigned int max_data_rate;
 
-	data_rate = DIV_ROUND_UP(data_rate, 1000);
+		max_data_rate = icl_max_bw(dev_priv, num_active_planes, i);
+		/*
+		 * We need to know which qgv point gives us
+		 * maximum bandwidth in order to disable SAGV
+		 * if we find that we exceed SAGV block time
+		 * with watermarks. By that moment we already
+		 * have those, as it is calculated earlier in
+		 * intel_atomic_check,
+		 */
+		if (max_data_rate > max_bw) {
+			max_bw_point = i;
+			max_bw = max_data_rate;
+		}
+		if (max_data_rate >= data_rate)
+			allowed_points |= BIT(i);
+		DRM_DEBUG_KMS("QGV point %d: max bw %d required %d\n",
+			      i, max_data_rate, data_rate);
+	}
 
-	if (data_rate > max_data_rate) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "Bandwidth %u MB/s exceeds max available %d MB/s (%d active planes)\n",
-			    data_rate, max_data_rate, num_active_planes);
+	/*
+	 * BSpec states that we always should have at least one allowed point
+	 * left, so if we couldn't - simply reject the configuration for obvious
+	 * reasons.
+	 */
+	if (allowed_points == 0) {
+		DRM_DEBUG_KMS("No QGV points provide sufficient memory"
+			      " bandwidth %d for display configuration.\n", data_rate);
 		return -EINVAL;
 	}
 
+	/*
+	 * Leave only single point with highest bandwidth, if
+	 * we can't enable SAGV due to the increased memory latency it may
+	 * cause.
+	 */
+	if (!intel_can_enable_sagv(new_bw_state)) {
+		allowed_points = BIT(max_bw_point);
+		DRM_DEBUG_KMS("No SAGV, using single QGV point %d\n",
+			      max_bw_point);
+	}
+	/*
+	 * We store the ones which need to be masked as that is what PCode
+	 * actually accepts as a parameter.
+	 */
+	new_bw_state->qgv_points_mask = (~allowed_points) & mask;
+
+	old_bw_state = intel_atomic_get_old_bw_state(state);
+	if (!old_bw_state)
+		return -EINVAL;
+
+	/*
+	 * If the actual mask had changed we need to make sure that
+	 * the commits are serialized(in case this is a nomodeset, nonblocking)
+	 */
+	if (new_bw_state->qgv_points_mask != old_bw_state->qgv_points_mask) {
+		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
+		if (ret)
+			return ret;
+	}
+
 	return 0;
 }
 
drivers/gpu/drm/i915/display/intel_bw.h +9 -0
index 898b4a85ccab..bbcaaa73ec1b 100644
@@ -24,6 +24,13 @@ struct intel_bw_state {
 	 */
 	u8 pipe_sagv_reject;
 
+	/*
+	 * Current QGV points mask, which restricts
+	 * some particular SAGV states, not to confuse
+	 * with pipe_sagv_mask.
+	 */
+	u8 qgv_points_mask;
+
 	unsigned int data_rate[I915_MAX_PIPES];
 	u8 num_active_planes[I915_MAX_PIPES];
 
@@ -47,5 +54,7 @@ int intel_bw_init(struct drm_i915_private *dev_priv);
 int intel_bw_atomic_check(struct intel_atomic_state *state);
 void intel_bw_crtc_update(struct intel_bw_state *bw_state,
 			  const struct intel_crtc_state *crtc_state);
+int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
+				  u32 points_mask);
 
 #endif /* __INTEL_BW_H__ */
drivers/gpu/drm/i915/display/intel_display_types.h +3 -0
index 32cbbf7dddc6..a16aeb235654 100644
@@ -693,6 +693,9 @@ struct skl_plane_wm {
 	bool is_planar;
 };
 
+/* BSpec precisely defines this */
+#define NUM_SAGV_POINTS 8
+
 struct skl_pipe_wm {
 	struct skl_plane_wm planes[I915_MAX_PLANES];
 	bool can_sagv;
drivers/gpu/drm/i915/intel_pm.c +63 -3

93 lines not on this page · raw message

drivers/gpu/drm/i915/intel_pm.h +2 -0
index fd1dc422e6c5..4ae91ad5d5b3 100644
@@ -42,6 +42,8 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 			      struct skl_pipe_wm *out);
 void g4x_wm_sanitize(struct drm_i915_private *dev_priv);
 void vlv_wm_sanitize(struct drm_i915_private *dev_priv);
+void intel_sagv_pre_plane_update(struct intel_atomic_state *state);
+void intel_sagv_post_plane_update(struct intel_atomic_state *state);
 bool intel_can_enable_sagv(const struct intel_bw_state *bw_state);
 int intel_enable_sagv(struct drm_i915_private *dev_priv);
 int intel_disable_sagv(struct drm_i915_private *dev_priv);
Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

PATCH v27 6/6 drm/i915: Enable SAGV support for Gen12

In reply to Stanislav Lisovskiy

Flip the switch and enable SAGV support
for Gen12 also.

---
 drivers/gpu/drm/i915/intel_pm.c | 4 ----
 1 file changed, 4 deletions(-)
drivers/gpu/drm/i915/intel_pm.c +0 -4
index 5d0aab515e2a..a12f1d0a0be2 100644
@@ -3638,10 +3638,6 @@ static bool skl_needs_memory_bw_wa(struct drm_i915_private *dev_priv)
 static bool
 intel_has_sagv(struct drm_i915_private *dev_priv)
 {
-	/* HACK! */
-	if (IS_GEN(dev_priv, 12))
-		return false;
-
 	return (IS_GEN9_BC(dev_priv) || INTEL_GEN(dev_priv) >= 10) &&
 		dev_priv->sagv_status != I915_SAGV_NOT_CONTROLLED;
 }
Signature · 6 lines
2.24.1.485.gad05a3d8e5

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

[Intel-gfx] ✓ Fi.CI.BAT: success for SAGV support for Gen12+ (rev35)

In reply to Stanislav Lisovskiy
== Series Details ==

== Series Details ==

Series: SAGV support for Gen12+ (rev35)
URL : https://patchwork.freedesktop.org/series/75129/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_8430 -> Patchwork_17581
====================================================

Summary
-------

**SUCCESS**

No regressions found.

External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/index.html

Changes
-------

No changes found

Participating hosts (50 -> 42)
------------------------------

Missing (8): fi-ilk-m540 fi-hsw-4200u fi-byt-squawks fi-bsw-cyan fi-bwr-2160 fi-ctg-p8600 fi-byt-clapper fi-bdw-samus

Build changes
-------------

* CI: CI-20190529 -> None
* Linux: CI_DRM_8430 -> Patchwork_17581

CI-20190529: 20190529
CI_DRM_8430: 2daa6f8cad645f49a898158190a20a893b4aabe3 @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_5632: e630cb8cd2ec01d6d5358eb2a3f6ea70498b8183 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
Patchwork_17581: c9b00f24365bb8afb5caae9aa834993b996bc66a @ git://anongit.freedesktop.org/gfx-ci/linux

== Linux commits ==

c9b00f24365b drm/i915: Enable SAGV support for Gen12
f4e7158266f6 drm/i915: Restrict qgv points which don't have enough bandwidth. 5dd210b03447 drm/i915: Add TGL+ SAGV support
9ac87f34c218 drm/i915: Separate icl and skl SAGV checking c2e36d77edc9 drm/i915: Introduce skl_plane_wm_level accessor.

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/index.html
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#23 lore raw

[Intel-gfx] ✓ Fi.CI.IGT: success for SAGV support for Gen12+ (rev35)

In reply to Stanislav Lisovskiy
== Series Details ==

== Series Details ==

Series: SAGV support for Gen12+ (rev35)
URL : https://patchwork.freedesktop.org/series/75129/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_8430_full -> Patchwork_17581_full
====================================================

Summary
-------

**SUCCESS**

No regressions found.

Known issues
------------

Here are the changes found in Patchwork_17581_full that come from known issues:

### IGT changes ###

#### Issues hit ####

* igt@gem_workarounds@suspend-resume:
- shard-apl: [PASS][1] -> [DMESG-WARN][2] ([i915#180])
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl4/igt@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl6/igt@[email protected]

* igt@kms_cursor_edge_walk@pipe-a-64x64-top-edge:
- shard-apl: [PASS][3] -> [FAIL][4] ([i915#70] / [i915#95])
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl3/igt@[email protected]
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl3/igt@[email protected]
- shard-kbl: [PASS][5] -> [FAIL][6] ([i915#70])
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-kbl6/igt@[email protected]
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-kbl7/igt@[email protected]

* igt@kms_cursor_legacy@all-pipes-single-move:
- shard-hsw: [PASS][7] -> [INCOMPLETE][8] ([i915#61])
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-hsw7/igt@[email protected]
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-hsw8/igt@[email protected]

* igt@kms_frontbuffer_tracking@psr-1p-primscrn-indfb-pgflip-blt:
- shard-skl: [PASS][9] -> [FAIL][10] ([i915#49])
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-skl2/igt@[email protected]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-skl2/igt@[email protected]

* igt@kms_pipe_crc_basic@suspend-read-crc-pipe-c:
- shard-skl: [PASS][11] -> [INCOMPLETE][12] ([i915#69])
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-skl3/igt@[email protected]
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-skl10/igt@[email protected]

* igt@kms_psr@psr2_cursor_render:
- shard-iclb: [PASS][13] -> [SKIP][14] ([fdo#109441]) +2 similar issues
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-iclb2/igt@kms_psr@psr2_cursor_render.html
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-iclb4/igt@kms_psr@psr2_cursor_render.html

#### Possible fixes ####

* igt@kms_cursor_crc@pipe-a-cursor-suspend:
- shard-apl: [DMESG-WARN][15] ([i915#180] / [i915#95]) -> [PASS][16]
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl8/igt@[email protected]
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl3/igt@[email protected]

* igt@kms_cursor_legacy@all-pipes-torture-move:
- shard-hsw: [DMESG-WARN][17] ([i915#128]) -> [PASS][18]
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-hsw1/igt@[email protected]
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-hsw4/igt@[email protected]

* {igt@kms_flip@flip-vs-expired-vblank-interruptible@c-dp1}:
- shard-apl: [FAIL][19] ([i915#79]) -> [PASS][20]
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl7/igt@kms_flip@[email protected]
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl8/igt@kms_flip@[email protected]

* {igt@kms_flip@flip-vs-expired-vblank@b-edp1}:
- shard-skl: [FAIL][21] ([i915#79]) -> [PASS][22]
[21]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-skl2/igt@kms_flip@[email protected]
[22]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-skl2/igt@kms_flip@[email protected]

* {igt@kms_flip@plain-flip-ts-check-interruptible@a-edp1}:
- shard-skl: [FAIL][23] ([i915#34]) -> [PASS][24]
[23]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-skl1/igt@kms_flip@[email protected]
[24]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-skl5/igt@kms_flip@[email protected]

* igt@kms_hdr@bpc-switch-dpms:
- shard-skl: [FAIL][25] ([i915#1188]) -> [PASS][26]
[25]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-skl6/igt@[email protected]
[26]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-skl1/igt@[email protected]

* igt@kms_hdr@bpc-switch-suspend:
- shard-kbl: [INCOMPLETE][27] ([i915#155]) -> [PASS][28]
[27]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-kbl3/igt@[email protected]
[28]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-kbl7/igt@[email protected]

* igt@kms_plane_alpha_blend@pipe-c-constant-alpha-min:
- shard-skl: [FAIL][29] ([fdo#108145] / [i915#265]) -> [PASS][30]
[29]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-skl8/igt@[email protected]
[30]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-skl2/igt@[email protected]

* igt@kms_psr2_su@frontbuffer:
- shard-iclb: [SKIP][31] ([fdo#109642] / [fdo#111068]) -> [PASS][32]
[31]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-iclb3/igt@[email protected]
[32]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-iclb2/igt@[email protected]

* igt@kms_psr@psr2_cursor_plane_onoff:
- shard-iclb: [SKIP][33] ([fdo#109441]) -> [PASS][34] +2 similar issues
[33]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-iclb4/igt@kms_psr@psr2_cursor_plane_onoff.html
[34]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-iclb2/igt@kms_psr@psr2_cursor_plane_onoff.html

* igt@kms_vblank@pipe-a-ts-continuation-suspend:
- shard-apl: [DMESG-WARN][35] ([i915#180]) -> [PASS][36] +2 similar issues
[35]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl6/igt@[email protected]
[36]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl7/igt@[email protected]

#### Warnings ####

* igt@i915_pm_dc@dc6-psr:
- shard-tglb: [SKIP][37] ([i915#468]) -> [FAIL][38] ([i915#454])
[37]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-tglb2/igt@[email protected]
[38]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-tglb3/igt@[email protected]

* igt@kms_content_protection@atomic:
- shard-apl: [FAIL][39] ([fdo#110321] / [fdo#110336]) -> [TIMEOUT][40] ([i915#1319])
[39]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl1/igt@[email protected]
[40]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl1/igt@[email protected]

* igt@kms_content_protection@atomic-dpms:
- shard-apl: [TIMEOUT][41] ([i915#1319]) -> [FAIL][42] ([fdo#110321] / [fdo#110336])
[41]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl7/igt@[email protected]
[42]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl8/igt@[email protected]

* igt@kms_content_protection@lic:
- shard-kbl: [TIMEOUT][43] ([i915#1319]) -> [FAIL][44] ([fdo#110321] / [i915#93] / [i915#95])
[43]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-kbl4/igt@[email protected]
[44]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-kbl2/igt@[email protected]

* igt@kms_content_protection@srm:
- shard-apl: [TIMEOUT][45] ([i915#1319]) -> [FAIL][46] ([fdo#110321])
[45]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_8430/shard-apl1/igt@[email protected]
[46]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/shard-apl1/igt@[email protected]

{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).

[fdo#108145]: https://bugs.freedesktop.org/show_bug.cgi?id=108145
[fdo#109441]: https://bugs.freedesktop.org/show_bug.cgi?id=109441
[fdo#109642]: https://bugs.freedesktop.org/show_bug.cgi?id=109642
[fdo#110321]: https://bugs.freedesktop.org/show_bug.cgi?id=110321
[fdo#110336]: https://bugs.freedesktop.org/show_bug.cgi?id=110336
[fdo#111068]: https://bugs.freedesktop.org/show_bug.cgi?id=111068
[i915#1188]: https://gitlab.freedesktop.org/drm/intel/issues/1188
[i915#128]: https://gitlab.freedesktop.org/drm/intel/issues/128
[i915#1319]: https://gitlab.freedesktop.org/drm/intel/issues/1319
[i915#155]: https://gitlab.freedesktop.org/drm/intel/issues/155
[i915#180]: https://gitlab.freedesktop.org/drm/intel/issues/180
[i915#265]: https://gitlab.freedesktop.org/drm/intel/issues/265
[i915#34]: https://gitlab.freedesktop.org/drm/intel/issues/34
[i915#454]: https://gitlab.freedesktop.org/drm/intel/issues/454
[i915#468]: https://gitlab.freedesktop.org/drm/intel/issues/468
[i915#49]: https://gitlab.freedesktop.org/drm/intel/issues/49
[i915#61]: https://gitlab.freedesktop.org/drm/intel/issues/61
[i915#69]: https://gitlab.freedesktop.org/drm/intel/issues/69
[i915#70]: https://gitlab.freedesktop.org/drm/intel/issues/70
[i915#79]: https://gitlab.freedesktop.org/drm/intel/issues/79
[i915#93]: https://gitlab.freedesktop.org/drm/intel/issues/93
[i915#95]: https://gitlab.freedesktop.org/drm/intel/issues/95

Participating hosts (11 -> 11)
------------------------------

No changes in participating hosts

Build changes
-------------

* CI: CI-20190529 -> None
* Linux: CI_DRM_8430 -> Patchwork_17581

CI-20190529: 20190529
CI_DRM_8430: 2daa6f8cad645f49a898158190a20a893b4aabe3 @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_5632: e630cb8cd2ec01d6d5358eb2a3f6ea70498b8183 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
Patchwork_17581: c9b00f24365bb8afb5caae9aa834993b996bc66a @ git://anongit.freedesktop.org/gfx-ci/linux
piglit_4509: fdc5a4ca11124ab8413c7988896eec4c97336694 @ git://anongit.freedesktop.org/piglit

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_17581/index.html
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx