s/Tungsten Graphics/VMware/
[mesa.git] / src / gallium / drivers / softpipe / sp_quad_fs.c
1 /**************************************************************************
2 *
3 * Copyright 2007 VMware, Inc.
4 * All Rights Reserved.
5 * Copyright 2008 VMware, Inc. All rights reserved.
6 *
7 * Permission is hereby granted, free of charge, to any person obtaining a
8 * copy of this software and associated documentation files (the
9 * "Software"), to deal in the Software without restriction, including
10 * without limitation the rights to use, copy, modify, merge, publish,
11 * distribute, sub license, and/or sell copies of the Software, and to
12 * permit persons to whom the Software is furnished to do so, subject to
13 * the following conditions:
14 *
15 * The above copyright notice and this permission notice (including the
16 * next paragraph) shall be included in all copies or substantial portions
17 * of the Software.
18 *
19 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
20 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
21 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
22 * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
23 * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
24 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
25 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
26 *
27 **************************************************************************/
28
29 /* Vertices are just an array of floats, with all the attributes
30 * packed. We currently assume a layout like:
31 *
32 * attr[0][0..3] - window position
33 * attr[1..n][0..3] - remaining attributes.
34 *
35 * Attributes are assumed to be 4 floats wide but are packed so that
36 * all the enabled attributes run contiguously.
37 */
38
39 #include "util/u_math.h"
40 #include "util/u_memory.h"
41 #include "pipe/p_defines.h"
42 #include "pipe/p_shader_tokens.h"
43
44 #include "sp_context.h"
45 #include "sp_state.h"
46 #include "sp_quad.h"
47 #include "sp_quad_pipe.h"
48
49
50 struct quad_shade_stage
51 {
52 struct quad_stage stage; /**< base class */
53
54 /* no other fields at this time */
55 };
56
57
58 /** cast wrapper */
59 static INLINE struct quad_shade_stage *
60 quad_shade_stage(struct quad_stage *qs)
61 {
62 return (struct quad_shade_stage *) qs;
63 }
64
65
66 /**
67 * Execute fragment shader for the four fragments in the quad.
68 * \return TRUE if quad is alive, FALSE if all four pixels are killed
69 */
70 static INLINE boolean
71 shade_quad(struct quad_stage *qs, struct quad_header *quad)
72 {
73 struct softpipe_context *softpipe = qs->softpipe;
74 struct tgsi_exec_machine *machine = softpipe->fs_machine;
75
76 if (softpipe->active_statistics_queries) {
77 softpipe->pipeline_statistics.ps_invocations +=
78 util_bitcount(quad->inout.mask);
79 }
80
81 /* run shader */
82 machine->flatshade_color = softpipe->rasterizer->flatshade ? TRUE : FALSE;
83 return softpipe->fs_variant->run( softpipe->fs_variant, machine, quad );
84 }
85
86
87
88 static void
89 coverage_quad(struct quad_stage *qs, struct quad_header *quad)
90 {
91 struct softpipe_context *softpipe = qs->softpipe;
92 uint cbuf;
93
94 /* loop over colorbuffer outputs */
95 for (cbuf = 0; cbuf < softpipe->framebuffer.nr_cbufs; cbuf++) {
96 float (*quadColor)[4] = quad->output.color[cbuf];
97 unsigned j;
98 for (j = 0; j < TGSI_QUAD_SIZE; j++) {
99 assert(quad->input.coverage[j] >= 0.0);
100 assert(quad->input.coverage[j] <= 1.0);
101 quadColor[3][j] *= quad->input.coverage[j];
102 }
103 }
104 }
105
106
107 /**
108 * Shade/write an array of quads
109 * Called via quad_stage::run()
110 */
111 static void
112 shade_quads(struct quad_stage *qs,
113 struct quad_header *quads[],
114 unsigned nr)
115 {
116 struct softpipe_context *softpipe = qs->softpipe;
117 struct tgsi_exec_machine *machine = softpipe->fs_machine;
118 unsigned i, nr_quads = 0;
119
120 tgsi_exec_set_constant_buffers(machine, PIPE_MAX_CONSTANT_BUFFERS,
121 softpipe->mapped_constants[PIPE_SHADER_FRAGMENT],
122 softpipe->const_buffer_size[PIPE_SHADER_FRAGMENT]);
123
124 machine->InterpCoefs = quads[0]->coef;
125
126 for (i = 0; i < nr; i++) {
127 /* Only omit this quad from the output list if all the fragments
128 * are killed _AND_ it's not the first quad in the list.
129 * The first quad is special in the (optimized) depth-testing code:
130 * the quads' Z coordinates are step-wise interpolated with respect
131 * to the first quad in the list.
132 * For multi-pass algorithms we need to produce exactly the same
133 * Z values in each pass. If interpolation starts with different quads
134 * we can get different Z values for the same (x,y).
135 */
136 if (!shade_quad(qs, quads[i]) && i > 0)
137 continue; /* quad totally culled/killed */
138
139 if (/*do_coverage*/ 0)
140 coverage_quad( qs, quads[i] );
141
142 quads[nr_quads++] = quads[i];
143 }
144
145 if (nr_quads)
146 qs->next->run(qs->next, quads, nr_quads);
147 }
148
149
150 /**
151 * Per-primitive (or per-begin?) setup
152 */
153 static void
154 shade_begin(struct quad_stage *qs)
155 {
156 qs->next->begin(qs->next);
157 }
158
159
160 static void
161 shade_destroy(struct quad_stage *qs)
162 {
163 FREE( qs );
164 }
165
166
167 struct quad_stage *
168 sp_quad_shade_stage( struct softpipe_context *softpipe )
169 {
170 struct quad_shade_stage *qss = CALLOC_STRUCT(quad_shade_stage);
171 if (!qss)
172 goto fail;
173
174 qss->stage.softpipe = softpipe;
175 qss->stage.begin = shade_begin;
176 qss->stage.run = shade_quads;
177 qss->stage.destroy = shade_destroy;
178
179 return &qss->stage;
180
181 fail:
182 FREE(qss);
183 return NULL;
184 }