1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
|
From 62c45a6738804fc2ee76ae77281374b2c9af4b8a Mon Sep 17 00:00:00 2001
From: SivapiriyanKumarasamy <sivapiriyan.kumarasamy@amd.com>
Date: Tue, 11 Sep 2018 17:48:07 -0400
Subject: [PATCH 0541/2940] drm/amd/display: Add support for Freesync 2 HDR and
Content to Display Mapping
[Why]
Freesync 2 HDR and support for HDR content
outside the range of the HDR display
require implementation on Dal 3 to better match
Dal2.
[How]
Add support for Freesync HDR and mapping
of source content to display ranges for better
representation of HDR content.
Signed-off-by: SivapiriyanKumarasamy <sivapiriyan.kumarasamy@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
.../amd/display/amdgpu_dm/amdgpu_dm_color.c | 2 +-
.../amd/display/dc/dcn10/dcn10_cm_common.c | 2 +-
.../amd/display/modules/color/color_gamma.c | 175 +++++++++++++++++-
.../amd/display/modules/color/color_gamma.h | 11 +-
4 files changed, 186 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c
index be19e6861189..216e48cec716 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c
@@ -164,7 +164,7 @@ int amdgpu_dm_set_regamma_lut(struct dm_crtc_state *crtc)
*/
stream->out_transfer_func->type = TF_TYPE_DISTRIBUTED_POINTS;
ret = mod_color_calculate_regamma_params(stream->out_transfer_func,
- gamma, true, adev->asic_type <= CHIP_RAVEN);
+ gamma, true, adev->asic_type <= CHIP_RAVEN, NULL);
dc_gamma_release(&gamma);
if (!ret) {
stream->out_transfer_func->type = old_type;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
index 5d95a997fd9f..97c059934feb 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
@@ -268,7 +268,7 @@ bool cm_helper_translate_curve_to_hw_format(
memset(lut_params, 0, sizeof(struct pwl_params));
memset(seg_distr, 0, sizeof(seg_distr));
- if (output_tf->tf == TRANSFER_FUNCTION_PQ) {
+ if (output_tf->tf == TRANSFER_FUNCTION_PQ || output_tf->tf == TRANSFER_FUNCTION_GAMMA22) {
/* 32 segments
* segments are from 2^-25 to 2^7
*/
diff --git a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c
index cdcefd087487..2e215c9e5445 100644
--- a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c
+++ b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c
@@ -306,6 +306,18 @@ static struct fixed31_32 translate_from_linear_space(
a1);
}
+static struct fixed31_32 calculate_gamma22(struct fixed31_32 arg)
+{
+ struct fixed31_32 gamma = dc_fixpt_from_fraction(22, 10);
+
+ return translate_from_linear_space(arg,
+ dc_fixpt_zero,
+ dc_fixpt_zero,
+ dc_fixpt_zero,
+ dc_fixpt_zero,
+ gamma);
+}
+
static struct fixed31_32 translate_to_linear_space(
struct fixed31_32 arg,
struct fixed31_32 a0,
@@ -709,6 +721,160 @@ static void build_regamma(struct pwl_float_data_ex *rgb_regamma,
}
}
+static void hermite_spline_eetf(struct fixed31_32 input_x,
+ struct fixed31_32 max_display,
+ struct fixed31_32 min_display,
+ struct fixed31_32 max_content,
+ struct fixed31_32 *out_x)
+{
+ struct fixed31_32 min_lum_pq;
+ struct fixed31_32 max_lum_pq;
+ struct fixed31_32 max_content_pq;
+ struct fixed31_32 ks;
+ struct fixed31_32 E1;
+ struct fixed31_32 E2;
+ struct fixed31_32 E3;
+ struct fixed31_32 t;
+ struct fixed31_32 t2;
+ struct fixed31_32 t3;
+ struct fixed31_32 two;
+ struct fixed31_32 three;
+ struct fixed31_32 temp1;
+ struct fixed31_32 temp2;
+ struct fixed31_32 a = dc_fixpt_from_fraction(15, 10);
+ struct fixed31_32 b = dc_fixpt_from_fraction(5, 10);
+ struct fixed31_32 epsilon = dc_fixpt_from_fraction(1, 1000000); // dc_fixpt_epsilon is a bit too small
+
+ if (dc_fixpt_eq(max_content, dc_fixpt_zero)) {
+ *out_x = dc_fixpt_zero;
+ return;
+ }
+
+ compute_pq(input_x, &E1);
+ compute_pq(dc_fixpt_div(min_display, max_content), &min_lum_pq);
+ compute_pq(dc_fixpt_div(max_display, max_content), &max_lum_pq);
+ compute_pq(dc_fixpt_one, &max_content_pq); // always 1? DAL2 code is weird
+ a = dc_fixpt_div(dc_fixpt_add(dc_fixpt_one, b), max_content_pq); // (1+b)/maxContent
+ ks = dc_fixpt_sub(dc_fixpt_mul(a, max_lum_pq), b); // a * max_lum_pq - b
+
+ if (dc_fixpt_lt(E1, ks))
+ E2 = E1;
+ else if (dc_fixpt_le(ks, E1) && dc_fixpt_le(E1, dc_fixpt_one)) {
+ if (dc_fixpt_lt(epsilon, dc_fixpt_sub(dc_fixpt_one, ks)))
+ // t = (E1 - ks) / (1 - ks)
+ t = dc_fixpt_div(dc_fixpt_sub(E1, ks),
+ dc_fixpt_sub(dc_fixpt_one, ks));
+ else
+ t = dc_fixpt_zero;
+
+ two = dc_fixpt_from_int(2);
+ three = dc_fixpt_from_int(3);
+
+ t2 = dc_fixpt_mul(t, t);
+ t3 = dc_fixpt_mul(t2, t);
+ temp1 = dc_fixpt_mul(two, t3);
+ temp2 = dc_fixpt_mul(three, t2);
+
+ // (2t^3 - 3t^2 + 1) * ks
+ E2 = dc_fixpt_mul(ks, dc_fixpt_add(dc_fixpt_one,
+ dc_fixpt_sub(temp1, temp2)));
+
+ // (-2t^3 + 3t^2) * max_lum_pq
+ E2 = dc_fixpt_add(E2, dc_fixpt_mul(max_lum_pq,
+ dc_fixpt_sub(temp2, temp1)));
+
+ temp1 = dc_fixpt_mul(two, t2);
+ temp2 = dc_fixpt_sub(dc_fixpt_one, ks);
+
+ // (t^3 - 2t^2 + t) * (1-ks)
+ E2 = dc_fixpt_add(E2, dc_fixpt_mul(temp2,
+ dc_fixpt_add(t, dc_fixpt_sub(t3, temp1))));
+ }
+
+ temp1 = dc_fixpt_sub(dc_fixpt_one, E2);
+ temp2 = dc_fixpt_mul(temp1, temp1);
+ temp2 = dc_fixpt_mul(temp2, temp2);
+ // temp2 = (1-E2)^4
+
+ E3 = dc_fixpt_add(E2, dc_fixpt_mul(min_lum_pq, temp2));
+ compute_de_pq(E3, out_x);
+
+ *out_x = dc_fixpt_div(*out_x, dc_fixpt_div(max_display, max_content));
+}
+
+static bool build_freesync_hdr(struct pwl_float_data_ex *rgb_regamma,
+ uint32_t hw_points_num,
+ const struct hw_x_point *coordinate_x,
+ const struct freesync_hdr_tf_params *fs_params)
+{
+ uint32_t i;
+ struct pwl_float_data_ex *rgb = rgb_regamma;
+ const struct hw_x_point *coord_x = coordinate_x;
+ struct fixed31_32 scaledX = dc_fixpt_zero;
+ struct fixed31_32 scaledX1 = dc_fixpt_zero;
+ struct fixed31_32 max_display = dc_fixpt_from_int(fs_params->max_display);
+ struct fixed31_32 min_display = dc_fixpt_from_fraction(fs_params->min_display, 10000);
+ struct fixed31_32 max_content = dc_fixpt_from_int(fs_params->max_content);
+ struct fixed31_32 min_content = dc_fixpt_from_fraction(fs_params->min_content, 10000);
+ struct fixed31_32 clip = dc_fixpt_one;
+ struct fixed31_32 output;
+ bool use_eetf = false;
+ struct fixed31_32 sdr_white_level = dc_fixpt_from_int(fs_params->sdr_white_level);
+
+ if (fs_params == NULL || fs_params->max_content == 0 ||
+ fs_params->max_display == 0)
+ return false;
+
+ if (fs_params->min_display > 1000) // cap at 0.1 at the bottom
+ min_display = dc_fixpt_from_fraction(1, 10);
+ if (fs_params->max_display < 100) // cap at 100 at the top
+ max_display = dc_fixpt_from_int(100);
+
+ if (fs_params->min_content < fs_params->min_display)
+ use_eetf = true;
+ else
+ min_content = min_display;
+
+ if (fs_params->max_content > fs_params->max_display)
+ use_eetf = true;
+ else
+ max_content = max_display;
+
+ rgb += 32; // first 32 points have problems with fixed point, too small
+ coord_x += 32;
+ for (i = 32; i <= hw_points_num; i++) {
+ if (use_eetf) {
+ /*max content is equal 1 */
+ scaledX1 = dc_fixpt_div(coord_x->x,
+ dc_fixpt_div(max_content, sdr_white_level));
+ hermite_spline_eetf(scaledX1, max_display, min_display,
+ max_content, &scaledX);
+ } else
+ scaledX = dc_fixpt_div(coord_x->x,
+ dc_fixpt_div(max_display, sdr_white_level));
+
+ if (dc_fixpt_lt(scaledX, clip)) {
+ if (dc_fixpt_lt(scaledX, dc_fixpt_zero))
+ output = dc_fixpt_zero;
+ else
+ output = calculate_gamma22(scaledX);
+
+ rgb->r = output;
+ rgb->g = output;
+ rgb->b = output;
+ } else {
+ rgb->r = clip;
+ rgb->g = clip;
+ rgb->b = clip;
+ }
+
+ ++coord_x;
+ ++rgb;
+ }
+
+ return true;
+}
+
static void build_degamma(struct pwl_float_data_ex *curve,
uint32_t hw_points_num,
const struct hw_x_point *coordinate_x, bool is_2_4)
@@ -1356,7 +1522,8 @@ static bool map_regamma_hw_to_x_user(
#define _EXTRA_POINTS 3
bool mod_color_calculate_regamma_params(struct dc_transfer_func *output_tf,
- const struct dc_gamma *ramp, bool mapUserRamp, bool canRomBeUsed)
+ const struct dc_gamma *ramp, bool mapUserRamp, bool canRomBeUsed,
+ const struct freesync_hdr_tf_params *fs_params)
{
struct dc_transfer_func_distributed_points *tf_pts = &output_tf->tf_pts;
struct dividers dividers;
@@ -1424,6 +1591,12 @@ bool mod_color_calculate_regamma_params(struct dc_transfer_func *output_tf,
MAX_HW_POINTS,
coordinates_x,
output_tf->sdr_ref_white_level);
+ } else if (tf == TRANSFER_FUNCTION_GAMMA22 &&
+ fs_params != NULL) {
+ build_freesync_hdr(rgb_regamma,
+ MAX_HW_POINTS,
+ coordinates_x,
+ fs_params);
} else {
tf_pts->end_exponent = 0;
tf_pts->x_point_at_y1_red = 1;
diff --git a/drivers/gpu/drm/amd/display/modules/color/color_gamma.h b/drivers/gpu/drm/amd/display/modules/color/color_gamma.h
index 63ccb9c91224..a6e164df090a 100644
--- a/drivers/gpu/drm/amd/display/modules/color/color_gamma.h
+++ b/drivers/gpu/drm/amd/display/modules/color/color_gamma.h
@@ -73,12 +73,21 @@ struct regamma_lut {
};
};
+struct freesync_hdr_tf_params {
+ unsigned int sdr_white_level;
+ unsigned int min_content; // luminance in 1/10000 nits
+ unsigned int max_content; // luminance in nits
+ unsigned int min_display; // luminance in 1/10000 nits
+ unsigned int max_display; // luminance in nits
+};
+
void setup_x_points_distribution(void);
void precompute_pq(void);
void precompute_de_pq(void);
bool mod_color_calculate_regamma_params(struct dc_transfer_func *output_tf,
- const struct dc_gamma *ramp, bool mapUserRamp, bool canRomBeUsed);
+ const struct dc_gamma *ramp, bool mapUserRamp, bool canRomBeUsed,
+ const struct freesync_hdr_tf_params *fs_params);
bool mod_color_calculate_degamma_params(struct dc_transfer_func *output_tf,
const struct dc_gamma *ramp, bool mapUserRamp);
--
2.17.1
|
