drm/i915: Replace wm.max_levels with wm.num_levels and use it everywhere

Replaces wm.max_level with wm.num_levels, since that generally
results in nicer looking code (for-loops can be in standard
form etc.).

Also get rid of the two different wrappers we have for this
(ilk_wm_max_level() and intel_wm_num_levels()). They don't
really do anything for us other than potentially slow things
down if the compiler actually emits the function calls every
time (num_planes*num_wm_levels*higher_level_wm_function_calls
could be a big number). The watermark code already shows up
far too prominently in cpu profiles. Though I must admit that
I didn't look at the generated code this time.

v2: Fix the ilk_wm_merge() off-by-one (Jani)
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230209222504.31478-1-ville.syrjala@linux.intel.com

drivers/gpu/drm/i915/display/intel_display_core.h

@@ -243,7 +243,7 @@ struct intel_wm {
 		struct g4x_wm_values g4x;
 	};
 
-	u8 max_level;
+	u8 num_levels;
 
 	/*
 	 * Should be held around atomic WM register writing; also

drivers/gpu/drm/i915/display/intel_display_debugfs.c

@@ -1286,13 +1286,10 @@ static void wm_latency_show(struct seq_file *m, const u16 wm[8])
 {
 	struct drm_i915_private *dev_priv = m->private;
 	int level;
-	int num_levels;
-
-	num_levels = ilk_wm_max_level(dev_priv) + 1;
 
 	drm_modeset_lock_all(&dev_priv->drm);
 
-	for (level = 0; level < num_levels; level++) {
+	for (level = 0; level < dev_priv->display.wm.num_levels; level++) {
 		unsigned int latency = wm[level];
 
 		/*
@@ -1395,13 +1392,10 @@ static ssize_t wm_latency_write(struct file *file, const char __user *ubuf,
 	struct seq_file *m = file->private_data;
 	struct drm_i915_private *dev_priv = m->private;
 	u16 new[8] = { 0 };
-	int num_levels;
 	int level;
 	int ret;
 	char tmp[32];
 
-	num_levels = ilk_wm_max_level(dev_priv) + 1;
-
 	if (len >= sizeof(tmp))
 		return -EINVAL;
 
@@ -1413,12 +1407,12 @@ static ssize_t wm_latency_write(struct file *file, const char __user *ubuf,
 	ret = sscanf(tmp, "%hu %hu %hu %hu %hu %hu %hu %hu",
 		     &new[0], &new[1], &new[2], &new[3],
 		     &new[4], &new[5], &new[6], &new[7]);
-	if (ret != num_levels)
+	if (ret != dev_priv->display.wm.num_levels)
 		return -EINVAL;
 
 	drm_modeset_lock_all(&dev_priv->drm);
 
-	for (level = 0; level < num_levels; level++)
+	for (level = 0; level < dev_priv->display.wm.num_levels; level++)
 		wm[level] = new[level];
 
 	drm_modeset_unlock_all(&dev_priv->drm);

drivers/gpu/drm/i915/display/skl_watermark.c

@@ -359,7 +359,7 @@ static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
 			continue;
 
 		/* Find the highest enabled wm level for this plane */
-		for (level = ilk_wm_max_level(i915);
+		for (level = i915->display.wm.num_levels - 1;
 		     !wm->wm[level].enable; --level)
 		     { }
 
@@ -710,10 +710,10 @@ skl_cursor_allocation(const struct intel_crtc_state *crtc_state,
 {
 	struct intel_plane *plane = to_intel_plane(crtc_state->uapi.crtc->cursor);
 	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
-	int level, max_level = ilk_wm_max_level(i915);
 	struct skl_wm_level wm = {};
 	int ret, min_ddb_alloc = 0;
 	struct skl_wm_params wp;
+	int level;
 
 	ret = skl_compute_wm_params(crtc_state, 256,
 				    drm_format_info(DRM_FORMAT_ARGB8888),
@@ -722,7 +722,7 @@ skl_cursor_allocation(const struct intel_crtc_state *crtc_state,
 				    crtc_state->pixel_rate, &wp, 0);
 	drm_WARN_ON(&i915->drm, ret);
 
-	for (level = 0; level <= max_level; level++) {
+	for (level = 0; level < i915->display.wm.num_levels; level++) {
 		unsigned int latency = i915->display.wm.skl_latency[level];
 
 		skl_compute_plane_wm(crtc_state, plane, level, latency, &wp, &wm, &wm);
@@ -1492,7 +1492,7 @@ skl_crtc_allocate_plane_ddb(struct intel_atomic_state *state,
 	 * Find the highest watermark level for which we can satisfy the block
 	 * requirement of active planes.
 	 */
-	for (level = ilk_wm_max_level(i915); level >= 0; level--) {
+	for (level = i915->display.wm.num_levels - 1; level >= 0; level--) {
 		blocks = 0;
 		for_each_plane_id_on_crtc(crtc, plane_id) {
 			const struct skl_plane_wm *wm =
@@ -1568,7 +1568,7 @@ skl_crtc_allocate_plane_ddb(struct intel_atomic_state *state,
 	 * all levels as "enabled."  Go back now and disable the ones
 	 * that aren't actually possible.
 	 */
-	for (level++; level <= ilk_wm_max_level(i915); level++) {
+	for (level++; level < i915->display.wm.num_levels; level++) {
 		for_each_plane_id_on_crtc(crtc, plane_id) {
 			const struct skl_ddb_entry *ddb =
 				&crtc_state->wm.skl.plane_ddb[plane_id];
@@ -1967,10 +1967,10 @@ skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
 		      struct skl_wm_level *levels)
 {
 	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
-	int level, max_level = ilk_wm_max_level(i915);
 	struct skl_wm_level *result_prev = &levels[0];
+	int level;
 
-	for (level = 0; level <= max_level; level++) {
+	for (level = 0; level < i915->display.wm.num_levels; level++) {
 		struct skl_wm_level *result = &levels[level];
 		unsigned int latency = i915->display.wm.skl_latency[level];
 
@@ -2248,7 +2248,6 @@ void skl_write_plane_wm(struct intel_plane *plane,
 			const struct intel_crtc_state *crtc_state)
 {
 	struct drm_i915_private *i915 = to_i915(plane->base.dev);
-	int level, max_level = ilk_wm_max_level(i915);
 	enum plane_id plane_id = plane->id;
 	enum pipe pipe = plane->pipe;
 	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
@@ -2256,8 +2255,9 @@ void skl_write_plane_wm(struct intel_plane *plane,
 		&crtc_state->wm.skl.plane_ddb[plane_id];
 	const struct skl_ddb_entry *ddb_y =
 		&crtc_state->wm.skl.plane_ddb_y[plane_id];
+	int level;
 
-	for (level = 0; level <= max_level; level++)
+	for (level = 0; level < i915->display.wm.num_levels; level++)
 		skl_write_wm_level(i915, PLANE_WM(pipe, plane_id, level),
 				   skl_plane_wm_level(pipe_wm, plane_id, level));
 
@@ -2285,14 +2285,14 @@ void skl_write_cursor_wm(struct intel_plane *plane,
 			 const struct intel_crtc_state *crtc_state)
 {
 	struct drm_i915_private *i915 = to_i915(plane->base.dev);
-	int level, max_level = ilk_wm_max_level(i915);
 	enum plane_id plane_id = plane->id;
 	enum pipe pipe = plane->pipe;
 	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
 	const struct skl_ddb_entry *ddb =
 		&crtc_state->wm.skl.plane_ddb[plane_id];
+	int level;
 
-	for (level = 0; level <= max_level; level++)
+	for (level = 0; level < i915->display.wm.num_levels; level++)
 		skl_write_wm_level(i915, CUR_WM(pipe, level),
 				   skl_plane_wm_level(pipe_wm, plane_id, level));
 
@@ -2324,9 +2324,9 @@ static bool skl_plane_wm_equals(struct drm_i915_private *i915,
 				const struct skl_plane_wm *wm1,
 				const struct skl_plane_wm *wm2)
 {
-	int level, max_level = ilk_wm_max_level(i915);
+	int level;
 
-	for (level = 0; level <= max_level; level++) {
+	for (level = 0; level < i915->display.wm.num_levels; level++) {
 		/*
 		 * We don't check uv_wm as the hardware doesn't actually
 		 * use it. It only gets used for calculating the required
@@ -2676,9 +2676,9 @@ static bool skl_plane_selected_wm_equals(struct intel_plane *plane,
 					 const struct skl_pipe_wm *new_pipe_wm)
 {
 	struct drm_i915_private *i915 = to_i915(plane->base.dev);
-	int level, max_level = ilk_wm_max_level(i915);
+	int level;
 
-	for (level = 0; level <= max_level; level++) {
+	for (level = 0; level < i915->display.wm.num_levels; level++) {
 		/*
 		 * We don't check uv_wm as the hardware doesn't actually
 		 * use it. It only gets used for calculating the required
@@ -2814,16 +2814,14 @@ static void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
 {
 	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
 	enum pipe pipe = crtc->pipe;
-	int level, max_level;
 	enum plane_id plane_id;
+	int level;
 	u32 val;
 
-	max_level = ilk_wm_max_level(i915);
-
 	for_each_plane_id_on_crtc(crtc, plane_id) {
 		struct skl_plane_wm *wm = &out->planes[plane_id];
 
-		for (level = 0; level <= max_level; level++) {
+		for (level = 0; level < i915->display.wm.num_levels; level++) {
 			if (plane_id != PLANE_CURSOR)
 				val = intel_de_read(i915, PLANE_WM(pipe, plane_id, level));
 			else
@@ -3006,9 +3004,9 @@ void intel_wm_state_verify(struct intel_crtc *crtc,
 		struct skl_pipe_wm wm;
 	} *hw;
 	const struct skl_pipe_wm *sw_wm = &new_crtc_state->wm.skl.optimal;
-	int level, max_level = ilk_wm_max_level(i915);
 	struct intel_plane *plane;
 	u8 hw_enabled_slices;
+	int level;
 
 	if (DISPLAY_VER(i915) < 9 || !new_crtc_state->hw.active)
 		return;
@@ -3035,7 +3033,7 @@ void intel_wm_state_verify(struct intel_crtc *crtc,
 		const struct skl_wm_level *hw_wm_level, *sw_wm_level;
 
 		/* Watermarks */
-		for (level = 0; level <= max_level; level++) {
+		for (level = 0; level < i915->display.wm.num_levels; level++) {
 			hw_wm_level = &hw->wm.planes[plane->id].wm[level];
 			sw_wm_level = skl_plane_wm_level(sw_wm, plane->id, level);
 
@@ -3157,7 +3155,7 @@ void skl_watermark_ipc_init(struct drm_i915_private *i915)
 
 static void
 adjust_wm_latency(struct drm_i915_private *i915,
-		  u16 wm[], int max_level, int read_latency)
+		  u16 wm[], int num_levels, int read_latency)
 {
 	bool wm_lv_0_adjust_needed = i915->dram_info.wm_lv_0_adjust_needed;
 	int i, level;
@@ -3167,12 +3165,12 @@ adjust_wm_latency(struct drm_i915_private *i915,
 	 * need to be disabled. We make sure to sanitize the values out
 	 * of the punit to satisfy this requirement.
 	 */
-	for (level = 1; level <= max_level; level++) {
+	for (level = 1; level < num_levels; level++) {
 		if (wm[level] == 0) {
-			for (i = level + 1; i <= max_level; i++)
+			for (i = level + 1; i < num_levels; i++)
 				wm[i] = 0;
 
-			max_level = level - 1;
+			num_levels = level;
 			break;
 		}
 	}
@@ -3185,7 +3183,7 @@ adjust_wm_latency(struct drm_i915_private *i915,
 	 * from the punit when level 0 response data is 0us.
 	 */
 	if (wm[0] == 0) {
-		for (level = 0; level <= max_level; level++)
+		for (level = 0; level < num_levels; level++)
 			wm[level] += read_latency;
 	}
 
@@ -3201,7 +3199,7 @@ adjust_wm_latency(struct drm_i915_private *i915,
 
 static void mtl_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
 {
-	int max_level = ilk_wm_max_level(i915);
+	int num_levels = i915->display.wm.num_levels;
 	u32 val;
 
 	val = intel_de_read(i915, MTL_LATENCY_LP0_LP1);
@@ -3216,12 +3214,12 @@ static void mtl_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
 	wm[4] = REG_FIELD_GET(MTL_LATENCY_LEVEL_EVEN_MASK, val);
 	wm[5] = REG_FIELD_GET(MTL_LATENCY_LEVEL_ODD_MASK, val);
 
-	adjust_wm_latency(i915, wm, max_level, 6);
+	adjust_wm_latency(i915, wm, num_levels, 6);
 }
 
 static void skl_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
 {
-	int max_level = ilk_wm_max_level(i915);
+	int num_levels = i915->display.wm.num_levels;
 	int read_latency = DISPLAY_VER(i915) >= 12 ? 3 : 2;
 	int mult = IS_DG2(i915) ? 2 : 1;
 	u32 val;
@@ -3253,15 +3251,15 @@ static void skl_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
 	wm[6] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_2_6_MASK, val) * mult;
 	wm[7] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_3_7_MASK, val) * mult;
 
-	adjust_wm_latency(i915, wm, max_level, read_latency);
+	adjust_wm_latency(i915, wm, num_levels, read_latency);
 }
 
 static void skl_setup_wm_latency(struct drm_i915_private *i915)
 {
 	if (HAS_HW_SAGV_WM(i915))
-		i915->display.wm.max_level = 5;
+		i915->display.wm.num_levels = 6;
 	else
-		i915->display.wm.max_level = 7;
+		i915->display.wm.num_levels = 8;
 
 	if (DISPLAY_VER(i915) >= 14)
 		mtl_read_wm_latency(i915, i915->display.wm.skl_latency);

drivers/gpu/drm/i915/intel_pm.h

@@ -14,7 +14,6 @@ struct intel_plane_state;
 
 void intel_init_clock_gating(struct drm_i915_private *dev_priv);
 void intel_suspend_hw(struct drm_i915_private *dev_priv);
-int ilk_wm_max_level(const struct drm_i915_private *dev_priv);
 void intel_init_pm(struct drm_i915_private *dev_priv);
 void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv);
 void intel_pm_setup(struct drm_i915_private *dev_priv);