drm/amd/display: use halving distribution for PQ/sRGB linearizing LUT

When linearizing, the input is an encoded signal bounded to [0,1] and
PQ/sRGB EOTFs are steepest near 1, requiring more precision near the
bright end.

Take the 8-bit sRGB case as a reference: 256 possible inputs and 256 HW
LUT points line up, so the LUT acts as plain indexing. Float
representations don't land perfectly, but LERP-ing between two HW
entries, when input is within a small epsilon of one of them, doesn't
materially change the result.

Replace the uniform 12-region distribution (16 points each,
192 total, range [2^-12, 1]) with a 9-region halving distribution for
the PQ/sRGB pre-defined EOTF: 128 points in the top region [0.5, 1], 64
in the next, 32 in the next, and so on, down to 1 point in each of the
two darkest regions. Total samples grow from 192 to 256, with uniform
1/256 spacing across [0, 1]. The dark tail below 2^-9 is no longer
sampled separately, which is acceptable for PQ/sRGB.

Suggested-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Tested-by: Matthew Schwartz <matthew.schwartz@linux.dev>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Melissa Wen <mwen@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 24 insertions, 9 deletions

diff --git a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c
index 70b7bc3494a2..66fe7f313ea3 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn30/dcn30_cm_common.c
@@ -303,8 +303,6 @@ bool cm3_helper_translate_curve_to_hw_format(struct dc_context *ctx,
 	return true;
 }
 
-#define NUM_DEGAMMA_REGIONS    12
-
 /* Linear interpolation of tf_pts entries, where (i >> 4) is the integer tf_pts
  * index, (i & 0xf) is the 1/16 sub-position.
  */
@@ -345,17 +343,34 @@ bool cm3_helper_translate_curve_to_degamma_hw_format(
 	memset(lut_params, 0, sizeof(struct pwl_params));
 	memset(seg_distr, 0, sizeof(seg_distr));
 
-	region_start = -NUM_DEGAMMA_REGIONS;
-	region_end   = 0;
+	if (output_tf->tf == TRANSFER_FUNCTION_PQ ||
+	    output_tf->tf == TRANSFER_FUNCTION_SRGB) {
+		/* 9 segments
+		 * segments are from 2^-9 to 0
+		 */
+		const uint8_t SEG_COUNT = 9;
+		seg_distr[0] = 0; // Since we only have one point in darkest region
+		for (k = 1; k < SEG_COUNT; k++)
+			seg_distr[k] = k - 1; // 2^(k-1) points per region; halves as k decreases
+
+		region_start = -SEG_COUNT;
+		region_end = 0;
+	} else {
+		/* 12 segments
+		 * segments are from 2^-12 to 2^0
+		 * There are less than 256 points, for optimization
+		 */
+		const uint8_t SEG_COUNT = 12;
+
+		for (i = 0; i < SEG_COUNT; i++)
+			seg_distr[i] = 4;
 
+		region_start = -SEG_COUNT;
+		region_end = 0;
+	}
 
 	for (i = region_end - region_start; i < MAX_REGIONS_NUMBER ; i++)
 		seg_distr[i] = -1;
-	/* 12 segments
-	 * segments are from 2^-12 to 0
-	 */
-	for (i = 0; i < NUM_DEGAMMA_REGIONS ; i++)
-		seg_distr[i] = 4;
 
 	for (k = 0; k < MAX_REGIONS_NUMBER; k++) {
 		if (seg_distr[k] != -1)