AMD-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Melissa Wen <mwen@igalia.com>
To: airlied@gmail.com, alexander.deucher@amd.com,
	christian.koenig@amd.com, harry.wentland@amd.com,
	simona@ffwll.ch, siqueira@igalia.com, sunpeng.li@amd.com
Cc: Krunoslav Kovac <Krunoslav.Kovac@amd.com>,
	"Dr . David Alan Gilbert" <linux@treblig.org>,
	Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>,
	Alex Hung <alex.hung@amd.com>,
	Aurabindo Pillai <Aurabindo.Pillai@amd.com>,
	Matthew Schwartz <matthew.schwartz@linux.dev>,
	pekka.paalanen@collabora.com, amd-gfx@lists.freedesktop.org,
	dri-devel@lists.freedesktop.org, kernel-dev@igalia.com
Subject: [PATCH v2 4/5] drm/amd/display: support up to 256 samples per region in degamma/blend LUT
Date: Wed,  6 May 2026 16:11:51 -0300	[thread overview]
Message-ID: <20260506191606.15022-5-mwen@igalia.com> (raw)
In-Reply-To: <20260506191606.15022-1-mwen@igalia.com>

cm3_helper_translate_curve_to_degamma_hw_format() reads one tf_pts entry
per HW LUT point, limiting the number of samples per region to
NUMBER_SW_SEGMENTS (16, at seg_distr[k] = 4) - higher seg_distr[k]
underflows the increment to 0. But the next patch introduces a halving
distribution for PQ/sRGB EOTFs that requires up to 128 samples in its
upper region (seg_distr[k] = 7).

As preparation, extend the loop index by 4 bits and linearly interpolate
adjacent tf_pts entries with the new interp_tf_pts() helper, where the 4
least significant bits are weight in 1/16 increments. This raises the
cap to 256 samples per region (seg_distr[k] = 8). seg_distr[k] <= 4
paths remain unchanged: the 4 least significant bits remain zero and
interp_tf_pts() reduces to a direct lookup.

Co-developed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Melissa Wen <mwen@igalia.com>
---
 .../amd/display/dc/dcn30/dcn30_cm_common.c    | 32 +++++++++++++++----
 1 file changed, 25 insertions(+), 7 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 0949b1dffc63..70b7bc3494a2 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
@@ -305,6 +305,22 @@ bool cm3_helper_translate_curve_to_hw_format(struct dc_context *ctx,
 
 #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.
+ */
+static struct fixed31_32 interp_tf_pts(const struct fixed31_32 *output_tf_channel, int i)
+{
+	struct fixed31_32 in_plus_one, in, value;
+	uint32_t t = i & 0xf;
+
+	in_plus_one = output_tf_channel[(i >> 4) + 1];
+	in = output_tf_channel[i >> 4];
+	value = dc_fixpt_sub(in_plus_one, in);
+	value = dc_fixpt_shr(dc_fixpt_mul_int(value, t), 4);
+	value = dc_fixpt_add(in, value);
+
+	return value;
+}
 
 bool cm3_helper_translate_curve_to_degamma_hw_format(
 				const struct dc_transfer_func *output_tf,
@@ -348,18 +364,20 @@ bool cm3_helper_translate_curve_to_degamma_hw_format(
 
 	j = 0;
 	for (k = 0; k < (region_end - region_start); k++) {
-		increment = NUMBER_SW_SEGMENTS / (1 << seg_distr[k]);
+		increment = (NUMBER_SW_SEGMENTS << 4) / (1 << seg_distr[k]);
 		start_index = (region_start + k + MAX_LOW_POINT) *
 				NUMBER_SW_SEGMENTS;
-		for (i = start_index; i < start_index + NUMBER_SW_SEGMENTS;
-				i += increment) {
+		for (i = (start_index << 4);
+		     i < (start_index << 4) + (NUMBER_SW_SEGMENTS << 4);
+		     i += increment) {
 			if (j == hw_points - 1)
 				break;
-			if (i >= TRANSFER_FUNC_POINTS)
+			if ((i >> 4) + 1 >= TRANSFER_FUNC_POINTS)
 				return false;
-			rgb_resulted[j].red = output_tf->tf_pts.red[i];
-			rgb_resulted[j].green = output_tf->tf_pts.green[i];
-			rgb_resulted[j].blue = output_tf->tf_pts.blue[i];
+
+			rgb_resulted[j].red = interp_tf_pts(output_tf->tf_pts.red, i);
+			rgb_resulted[j].green = interp_tf_pts(output_tf->tf_pts.green, i);
+			rgb_resulted[j].blue = interp_tf_pts(output_tf->tf_pts.blue, i);
 			j++;
 		}
 	}
-- 
2.53.0


  parent reply	other threads:[~2026-05-06 19:16 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-06 19:11 [PATCH v2 0/5] better LUT segmentation for EOTFs Melissa Wen
2026-05-06 19:11 ` [PATCH v2 1/5] drm/amd/display: use GAMCOR for degamma private props in subsampled format Melissa Wen
2026-05-06 19:11 ` [PATCH v2 2/5] Revert "drm/amd/display: Remove unused cm3_helper_translate_curve_to_degamma_hw_format" Melissa Wen
2026-05-06 19:11 ` [PATCH v2 3/5] drm/amd/display: use a separate helper to translate degamma curves Melissa Wen
2026-05-06 19:11 ` Melissa Wen [this message]
2026-05-06 19:11 ` [PATCH v2 5/5] drm/amd/display: use halving distribution for PQ/sRGB linearizing LUT Melissa Wen
2026-05-08 17:45 ` [PATCH v2 0/5] better LUT segmentation for EOTFs Matthew Schwartz

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260506191606.15022-5-mwen@igalia.com \
    --to=mwen@igalia.com \
    --cc=Aurabindo.Pillai@amd.com \
    --cc=Bhawanpreet.Lakha@amd.com \
    --cc=Krunoslav.Kovac@amd.com \
    --cc=airlied@gmail.com \
    --cc=alex.hung@amd.com \
    --cc=alexander.deucher@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=christian.koenig@amd.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=harry.wentland@amd.com \
    --cc=kernel-dev@igalia.com \
    --cc=linux@treblig.org \
    --cc=matthew.schwartz@linux.dev \
    --cc=pekka.paalanen@collabora.com \
    --cc=simona@ffwll.ch \
    --cc=siqueira@igalia.com \
    --cc=sunpeng.li@amd.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox