Merge branch 'gallium-vertex-linear' into gallium-0.1
[mesa.git] / src / gallium / auxiliary / draw / draw_vs_exec.c
1 /**************************************************************************
2 *
3 * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
4 * All Rights Reserved.
5 *
6 * Permission is hereby granted, free of charge, to any person obtaining a
7 * copy of this software and associated documentation files (the
8 * "Software"), to deal in the Software without restriction, including
9 * without limitation the rights to use, copy, modify, merge, publish,
10 * distribute, sub license, and/or sell copies of the Software, and to
11 * permit persons to whom the Software is furnished to do so, subject to
12 * the following conditions:
13 *
14 * The above copyright notice and this permission notice (including the
15 * next paragraph) shall be included in all copies or substantial portions
16 * of the Software.
17 *
18 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
19 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
20 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
21 * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
22 * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
23 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
24 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
25 *
26 **************************************************************************/
27
28 /*
29 * Authors:
30 * Keith Whitwell <keith@tungstengraphics.com>
31 * Brian Paul
32 */
33
34 #include "pipe/p_util.h"
35 #include "pipe/p_shader_tokens.h"
36
37 #include "draw_private.h"
38 #include "draw_context.h"
39 #include "draw_vs.h"
40
41 #include "tgsi/util/tgsi_parse.h"
42 #include "tgsi/util/tgsi_scan.h"
43
44
45 struct exec_vertex_shader {
46 struct draw_vertex_shader base;
47 struct tgsi_exec_machine *machine;
48 };
49
50 static struct exec_vertex_shader *exec_vertex_shader( struct draw_vertex_shader *vs )
51 {
52 return (struct exec_vertex_shader *)vs;
53 }
54
55
56 /* Not required for run_linear.
57 */
58 static void
59 vs_exec_prepare( struct draw_vertex_shader *shader,
60 struct draw_context *draw )
61 {
62 struct exec_vertex_shader *evs = exec_vertex_shader(shader);
63
64 /* specify the vertex program to interpret/execute */
65 tgsi_exec_machine_bind_shader(evs->machine,
66 shader->state.tokens,
67 PIPE_MAX_SAMPLERS,
68 NULL /*samplers*/ );
69
70 }
71
72
73
74
75 /* Simplified vertex shader interface for the pt paths. Given the
76 * complexity of code-generating all the above operations together,
77 * it's time to try doing all the other stuff separately.
78 */
79 static void
80 vs_exec_run_linear( struct draw_vertex_shader *shader,
81 const float (*input)[4],
82 float (*output)[4],
83 const float (*constants)[4],
84 unsigned count,
85 unsigned input_stride,
86 unsigned output_stride )
87 {
88 struct exec_vertex_shader *evs = exec_vertex_shader(shader);
89 struct tgsi_exec_machine *machine = evs->machine;
90 unsigned int i, j;
91 unsigned slot;
92
93 machine->Consts = constants;
94
95 for (i = 0; i < count; i += MAX_TGSI_VERTICES) {
96 unsigned int max_vertices = MIN2(MAX_TGSI_VERTICES, count - i);
97
98 /* Swizzle inputs.
99 */
100 for (j = 0; j < max_vertices; j++) {
101 #if 0
102 debug_printf("%d) Input vert:\n", i + j);
103 for (slot = 0; slot < shader->info.num_inputs; slot++) {
104 debug_printf("\t%d: %f %f %f %f\n", slot,
105 input[slot][0],
106 input[slot][1],
107 input[slot][2],
108 input[slot][3]);
109 }
110 #endif
111
112 for (slot = 0; slot < shader->info.num_inputs; slot++) {
113 machine->Inputs[slot].xyzw[0].f[j] = input[slot][0];
114 machine->Inputs[slot].xyzw[1].f[j] = input[slot][1];
115 machine->Inputs[slot].xyzw[2].f[j] = input[slot][2];
116 machine->Inputs[slot].xyzw[3].f[j] = input[slot][3];
117 }
118
119 input = (const float (*)[4])((const char *)input + input_stride);
120 }
121
122 /* run interpreter */
123 tgsi_exec_machine_run( machine );
124
125 /* Unswizzle all output results.
126 */
127 for (j = 0; j < max_vertices; j++) {
128 for (slot = 0; slot < shader->info.num_outputs; slot++) {
129 output[slot][0] = machine->Outputs[slot].xyzw[0].f[j];
130 output[slot][1] = machine->Outputs[slot].xyzw[1].f[j];
131 output[slot][2] = machine->Outputs[slot].xyzw[2].f[j];
132 output[slot][3] = machine->Outputs[slot].xyzw[3].f[j];
133
134 }
135
136 #if 0
137 debug_printf("%d) Post xform vert:\n", i + j);
138 for (slot = 0; slot < shader->info.num_outputs; slot++) {
139 debug_printf("\t%d: %f %f %f %f\n", slot,
140 output[slot][0],
141 output[slot][1],
142 output[slot][2],
143 output[slot][3]);
144 }
145 #endif
146
147 output = (float (*)[4])((char *)output + output_stride);
148 }
149
150 }
151 }
152
153
154
155
156 static void
157 vs_exec_delete( struct draw_vertex_shader *dvs )
158 {
159 FREE((void*) dvs->state.tokens);
160 FREE( dvs );
161 }
162
163
164 struct draw_vertex_shader *
165 draw_create_vs_exec(struct draw_context *draw,
166 const struct pipe_shader_state *state)
167 {
168 struct exec_vertex_shader *vs = CALLOC_STRUCT( exec_vertex_shader );
169
170 if (vs == NULL)
171 return NULL;
172
173 /* we make a private copy of the tokens */
174 vs->base.state.tokens = tgsi_dup_tokens(state->tokens);
175 if (!vs->base.state.tokens) {
176 FREE(vs);
177 return NULL;
178 }
179
180 tgsi_scan_shader(state->tokens, &vs->base.info);
181
182 vs->base.draw = draw;
183 vs->base.prepare = vs_exec_prepare;
184 vs->base.run_linear = vs_exec_run_linear;
185 vs->base.delete = vs_exec_delete;
186 vs->base.create_varient = draw_vs_varient_generic;
187 vs->machine = &draw->vs.machine;
188
189 return &vs->base;
190 }