summaryrefslogtreecommitdiffstats
path: root/gfx/wr/webrender/res/ps_quad.glsl
blob: 94c80a93f70f9d9d1bf05acdd5a69d6635639256 (plain)
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
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

/// The common infrastructure for ps_quad_* shaders.
///
/// # Memory layout
///
/// The diagram below shows the the various pieces of data fectched in the vertex shader:
///
///```ascii
///                                       (int gpu buffer)
///                                       +---------------+    (sGpuCache)
///  (instance-step vertex attr)          |  Int header   |   +-----------+
/// +-----------------------------+       |               |   | Transform |
/// |    Quad instance (uvec4)    |  +--> | transform id +--> +-----------+
/// |                             |  |    | z id          |
/// | x: int prim address        +---+    +---------------+   (float gpu buffer)
/// | y: float prim address      +--------------------------> +-----------+--------------+-+-+
/// | z: quad flags               |      (sGpuCache)          | Quad Prim | Quad Segment | | |
/// |    edge flags               |   +--------------------+  |           |              | | |
/// |    part index               |   |     Picture task   |  | bounds    | rect         | | |
/// |    segment index            |   |                    |  | clip      | uv rect      | | |
/// | w: picture task address    +--> | task rect          |  | color     |              | | |
/// +-----------------------------+   | device pixel scale |  +-----------+--------------+-+-+
///                                   | content origin     |
///                                   +--------------------+
///```

#define WR_FEATURE_TEXTURE_2D

#include shared,rect,transform,render_task,gpu_buffer

flat varying mediump vec4 v_color;
flat varying mediump vec4 v_uv_sample_bounds;
// x: (in ps_quad_textured) has edge flags
// y: has uv rect
// z: (in ps_quad_textured) sample as mask
flat varying lowp ivec4 v_flags;
varying highp vec2 v_uv;

#ifdef WR_VERTEX_SHADER

#define EDGE_AA_LEFT    1
#define EDGE_AA_TOP     2
#define EDGE_AA_RIGHT   4
#define EDGE_AA_BOTTOM  8

#define PART_CENTER     0
#define PART_LEFT       1
#define PART_TOP        2
#define PART_RIGHT      3
#define PART_BOTTOM     4
#define PART_ALL        5

#define QF_IS_OPAQUE            1
#define QF_APPLY_DEVICE_CLIP    2
#define QF_IGNORE_DEVICE_SCALE  4
#define QF_USE_AA_SEGMENTS      8
#define QF_SAMPLE_AS_MASK       16

#define INVALID_SEGMENT_INDEX   0xff

#define AA_PIXEL_RADIUS 2.0

PER_INSTANCE in ivec4 aData;

struct PrimitiveInfo {
    vec2 local_pos;

    RectWithEndpoint local_prim_rect;
    RectWithEndpoint local_clip_rect;

    int edge_flags;
    int quad_flags;
};

struct QuadSegment {
    RectWithEndpoint rect;
    vec4 uv_rect;
};

struct QuadPrimitive {
    RectWithEndpoint bounds;
    RectWithEndpoint clip;
    vec4 color;
};

QuadSegment fetch_segment(int base, int index) {
    QuadSegment seg;

    vec4 texels[2] = fetch_from_gpu_buffer_2f(base + 3 + index * 2);

    seg.rect = RectWithEndpoint(texels[0].xy, texels[0].zw);
    seg.uv_rect = texels[1];

    return seg;
}

QuadPrimitive fetch_primitive(int index) {
    QuadPrimitive prim;

    vec4 texels[3] = fetch_from_gpu_buffer_3f(index);

    prim.bounds = RectWithEndpoint(texels[0].xy, texels[0].zw);
    prim.clip = RectWithEndpoint(texels[1].xy, texels[1].zw);
    prim.color = texels[2];

    return prim;
}

struct QuadHeader {
    int transform_id;
    int z_id;
};

QuadHeader fetch_header(int address) {
    ivec4 header = fetch_from_gpu_buffer_1i(address);

    QuadHeader qh = QuadHeader(
        header.x,
        header.y
    );

    return qh;
}

struct QuadInstance {
    // x
    int prim_address_i;

    // y
    int prim_address_f;

    // z
    int quad_flags;
    int edge_flags;
    int part_index;
    int segment_index;

    // w
    int picture_task_address;
};

QuadInstance decode_instance() {
    QuadInstance qi = QuadInstance(
        aData.x,

        aData.y,

        (aData.z >> 24) & 0xff,
        (aData.z >> 16) & 0xff,
        (aData.z >>  8) & 0xff,
        (aData.z >>  0) & 0xff,

        aData.w
    );

    return qi;
}

struct VertexInfo {
    vec2 local_pos;
};

VertexInfo write_vertex(vec2 local_pos,
                        float z,
                        Transform transform,
                        vec2 content_origin,
                        RectWithEndpoint task_rect,
                        float device_pixel_scale,
                        int quad_flags) {
    VertexInfo vi;

    // Transform the current vertex to world space.
    vec4 world_pos = transform.m * vec4(local_pos, 0.0, 1.0);

    // Convert the world positions to device pixel space.
    vec2 device_pos = world_pos.xy * device_pixel_scale;

    if ((quad_flags & QF_APPLY_DEVICE_CLIP) != 0) {
        RectWithEndpoint device_clip_rect = RectWithEndpoint(
            content_origin,
            content_origin + task_rect.p1 - task_rect.p0
        );

        // Clip to task rect
        device_pos = rect_clamp(device_clip_rect, device_pos);

        vi.local_pos = (transform.inv_m * vec4(device_pos / device_pixel_scale, 0.0, 1.0)).xy;
    } else {
        vi.local_pos = local_pos;
    }

    // Apply offsets for the render task to get correct screen location.
    vec2 final_offset = -content_origin + task_rect.p0;

    gl_Position = uTransform * vec4(device_pos + final_offset * world_pos.w, z * world_pos.w, world_pos.w);

    return vi;
}

float edge_aa_offset(int edge, int flags) {
    return ((flags & edge) != 0) ? AA_PIXEL_RADIUS : 0.0;
}

PrimitiveInfo ps_quad_main(void) {
    QuadInstance qi = decode_instance();

    QuadHeader qh = fetch_header(qi.prim_address_i);
    Transform transform = fetch_transform(qh.transform_id);
    PictureTask task = fetch_picture_task(qi.picture_task_address);
    QuadPrimitive prim = fetch_primitive(qi.prim_address_f);
    float z = float(qh.z_id);

    QuadSegment seg;
    if (qi.segment_index == INVALID_SEGMENT_INDEX) {
        seg.rect = prim.bounds;
        seg.uv_rect = vec4(0.0);
    } else {
        seg = fetch_segment(qi.prim_address_f, qi.segment_index);
    }

    // The local space rect that we will draw, which is effectively:
    //  - The tile within the primitive we will draw
    //  - Intersected with any local-space clip rect(s)
    //  - Expanded for AA edges where appropriate
    RectWithEndpoint local_coverage_rect = seg.rect;

    // Apply local clip rect
    local_coverage_rect.p0 = max(local_coverage_rect.p0, prim.clip.p0);
    local_coverage_rect.p1 = min(local_coverage_rect.p1, prim.clip.p1);
    local_coverage_rect.p1 = max(local_coverage_rect.p0, local_coverage_rect.p1);

    switch (qi.part_index) {
        case PART_LEFT:
            local_coverage_rect.p1.x = local_coverage_rect.p0.x + AA_PIXEL_RADIUS;
#ifdef SWGL_ANTIALIAS
            swgl_antiAlias(EDGE_AA_LEFT);
#else
            local_coverage_rect.p0.x -= AA_PIXEL_RADIUS;
            local_coverage_rect.p0.y -= AA_PIXEL_RADIUS;
            local_coverage_rect.p1.y += AA_PIXEL_RADIUS;
#endif
            break;
        case PART_TOP:
            local_coverage_rect.p0.x = local_coverage_rect.p0.x + AA_PIXEL_RADIUS;
            local_coverage_rect.p1.x = local_coverage_rect.p1.x - AA_PIXEL_RADIUS;
            local_coverage_rect.p1.y = local_coverage_rect.p0.y + AA_PIXEL_RADIUS;
#ifdef SWGL_ANTIALIAS
            swgl_antiAlias(EDGE_AA_TOP);
#else
            local_coverage_rect.p0.y -= AA_PIXEL_RADIUS;
#endif
            break;
        case PART_RIGHT:
            local_coverage_rect.p0.x = local_coverage_rect.p1.x - AA_PIXEL_RADIUS;
#ifdef SWGL_ANTIALIAS
            swgl_antiAlias(EDGE_AA_RIGHT);
#else
            local_coverage_rect.p1.x += AA_PIXEL_RADIUS;
            local_coverage_rect.p0.y -= AA_PIXEL_RADIUS;
            local_coverage_rect.p1.y += AA_PIXEL_RADIUS;
#endif
            break;
        case PART_BOTTOM:
            local_coverage_rect.p0.x = local_coverage_rect.p0.x + AA_PIXEL_RADIUS;
            local_coverage_rect.p1.x = local_coverage_rect.p1.x - AA_PIXEL_RADIUS;
            local_coverage_rect.p0.y = local_coverage_rect.p1.y - AA_PIXEL_RADIUS;
#ifdef SWGL_ANTIALIAS
            swgl_antiAlias(EDGE_AA_BOTTOM);
#else
            local_coverage_rect.p1.y += AA_PIXEL_RADIUS;
#endif
            break;
        case PART_CENTER:
            local_coverage_rect.p0.x += edge_aa_offset(EDGE_AA_LEFT, qi.edge_flags);
            local_coverage_rect.p1.x -= edge_aa_offset(EDGE_AA_RIGHT, qi.edge_flags);
            local_coverage_rect.p0.y += edge_aa_offset(EDGE_AA_TOP, qi.edge_flags);
            local_coverage_rect.p1.y -= edge_aa_offset(EDGE_AA_BOTTOM, qi.edge_flags);
            break;
        case PART_ALL:
        default:
#ifdef SWGL_ANTIALIAS
            swgl_antiAlias(qi.edge_flags);
#else
            local_coverage_rect.p0.x -= edge_aa_offset(EDGE_AA_LEFT, qi.edge_flags);
            local_coverage_rect.p1.x += edge_aa_offset(EDGE_AA_RIGHT, qi.edge_flags);
            local_coverage_rect.p0.y -= edge_aa_offset(EDGE_AA_TOP, qi.edge_flags);
            local_coverage_rect.p1.y += edge_aa_offset(EDGE_AA_BOTTOM, qi.edge_flags);
#endif
            break;
    }

    vec2 local_pos = mix(local_coverage_rect.p0, local_coverage_rect.p1, aPosition);

    float device_pixel_scale = task.device_pixel_scale;
    if ((qi.quad_flags & QF_IGNORE_DEVICE_SCALE) != 0) {
        device_pixel_scale = 1.0f;
    }

    VertexInfo vi = write_vertex(
        local_pos,
        z,
        transform,
        task.content_origin,
        task.task_rect,
        device_pixel_scale,
        qi.quad_flags
    );

    if (seg.uv_rect.xy == seg.uv_rect.zw) {
        v_color = prim.color;
        v_flags.y = 0;
    } else {
        v_color = vec4(1.0);
        v_flags.y = 1;

        vec2 f = (vi.local_pos - seg.rect.p0) / (seg.rect.p1 - seg.rect.p0);

        vec2 uv = mix(
            seg.uv_rect.xy,
            seg.uv_rect.zw,
            f
        );

        vec2 texture_size = vec2(TEX_SIZE(sColor0));

        v_uv = uv / texture_size;

        v_uv_sample_bounds = vec4(
            seg.uv_rect.xy + vec2(0.5),
            seg.uv_rect.zw - vec2(0.5)
        ) / texture_size.xyxy;
    }

    return PrimitiveInfo(
        vi.local_pos,
        prim.bounds,
        prim.clip,
        qi.edge_flags,
        qi.quad_flags
    );
}
#endif