bumpmap sample is correct now
[mesa.git] / src / mesa / math / m_xform.c
1 /*
2 * Mesa 3-D graphics library
3 * Version: 5.1
4 *
5 * Copyright (C) 1999-2003 Brian Paul 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 "Software"),
9 * to deal in the Software without restriction, including without limitation
10 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
11 * and/or sell copies of the Software, and to permit persons to whom the
12 * Software is furnished to do so, subject to the following conditions:
13 *
14 * The above copyright notice and this permission notice shall be included
15 * in all copies or substantial portions of the Software.
16 *
17 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
18 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
19 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
20 * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
21 * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
22 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
23 */
24
25
26 /*
27 * Matrix/vertex/vector transformation stuff
28 *
29 *
30 * NOTES:
31 * 1. 4x4 transformation matrices are stored in memory in column major order.
32 * 2. Points/vertices are to be thought of as column vectors.
33 * 3. Transformation of a point p by a matrix M is: p' = M * p
34 */
35
36 #include "glheader.h"
37 #include "macros.h"
38
39 #include "m_eval.h"
40 #include "m_matrix.h"
41 #include "m_translate.h"
42 #include "m_xform.h"
43 #include "mathmod.h"
44
45
46 #ifdef DEBUG_MATH
47 #include "m_debug.h"
48 #endif
49
50 #ifdef USE_X86_ASM
51 #include "x86/common_x86_asm.h"
52 #endif
53
54 #ifdef USE_X86_64_ASM
55 #include "x86-64/x86-64.h"
56 #endif
57
58 #ifdef USE_SPARC_ASM
59 #include "sparc/sparc.h"
60 #endif
61
62 #ifdef USE_PPC_ASM
63 #include "ppc/common_ppc_features.h"
64 #endif
65
66 clip_func _mesa_clip_tab[5];
67 clip_func _mesa_clip_np_tab[5];
68 dotprod_func _mesa_dotprod_tab[5];
69 vec_copy_func _mesa_copy_tab[0x10];
70 normal_func _mesa_normal_tab[0xf];
71 transform_func *_mesa_transform_tab[5];
72
73
74 /* Raw data format used for:
75 * - Object-to-eye transform prior to culling, although this too
76 * could be culled under some circumstances.
77 * - Eye-to-clip transform (via the function above).
78 * - Cliptesting
79 * - And everything else too, if culling happens to be disabled.
80 *
81 * GH: It's used for everything now, as clipping/culling is done
82 * elsewhere (most often by the driver itself).
83 */
84 #define TAG(x) x
85 #define TAG2(x,y) x##y
86 #define STRIDE_LOOP for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) )
87 #define LOOP for ( i = 0 ; i < n ; i++ )
88 #define ARGS
89 #include "m_xform_tmp.h"
90 #include "m_clip_tmp.h"
91 #include "m_norm_tmp.h"
92 #include "m_dotprod_tmp.h"
93 #include "m_copy_tmp.h"
94 #undef TAG
95 #undef TAG2
96 #undef LOOP
97 #undef ARGS
98
99
100
101
102 GLvector4f *_mesa_project_points( GLvector4f *proj_vec,
103 const GLvector4f *clip_vec )
104 {
105 const GLuint stride = clip_vec->stride;
106 const GLfloat *from = (GLfloat *)clip_vec->start;
107 const GLuint count = clip_vec->count;
108 GLfloat (*vProj)[4] = (GLfloat (*)[4])proj_vec->start;
109 GLuint i;
110
111 for (i = 0 ; i < count ; i++, STRIDE_F(from, stride))
112 {
113 GLfloat oow = 1.0F / from[3];
114 vProj[i][3] = oow;
115 vProj[i][0] = from[0] * oow;
116 vProj[i][1] = from[1] * oow;
117 vProj[i][2] = from[2] * oow;
118 }
119
120 proj_vec->flags |= VEC_SIZE_4;
121 proj_vec->size = 3;
122 proj_vec->count = clip_vec->count;
123 return proj_vec;
124 }
125
126
127
128
129
130
131 /*
132 * Transform a 4-element row vector (1x4 matrix) by a 4x4 matrix. This
133 * function is used for transforming clipping plane equations and spotlight
134 * directions.
135 * Mathematically, u = v * m.
136 * Input: v - input vector
137 * m - transformation matrix
138 * Output: u - transformed vector
139 */
140 void _mesa_transform_vector( GLfloat u[4], const GLfloat v[4], const GLfloat m[16] )
141 {
142 GLfloat v0=v[0], v1=v[1], v2=v[2], v3=v[3];
143 #define M(row,col) m[row + col*4]
144 u[0] = v0 * M(0,0) + v1 * M(1,0) + v2 * M(2,0) + v3 * M(3,0);
145 u[1] = v0 * M(0,1) + v1 * M(1,1) + v2 * M(2,1) + v3 * M(3,1);
146 u[2] = v0 * M(0,2) + v1 * M(1,2) + v2 * M(2,2) + v3 * M(3,2);
147 u[3] = v0 * M(0,3) + v1 * M(1,3) + v2 * M(2,3) + v3 * M(3,3);
148 #undef M
149 }
150
151
152 /* Useful for one-off point transformations, as in clipping.
153 * Note that because the matrix isn't analysed we do too many
154 * multiplies, and that the result is always 4-clean.
155 */
156 void _mesa_transform_point_sz( GLfloat Q[4], const GLfloat M[16],
157 const GLfloat P[4], GLuint sz )
158 {
159 if (Q == P)
160 return;
161
162 if (sz == 4)
163 {
164 Q[0] = M[0] * P[0] + M[4] * P[1] + M[8] * P[2] + M[12] * P[3];
165 Q[1] = M[1] * P[0] + M[5] * P[1] + M[9] * P[2] + M[13] * P[3];
166 Q[2] = M[2] * P[0] + M[6] * P[1] + M[10] * P[2] + M[14] * P[3];
167 Q[3] = M[3] * P[0] + M[7] * P[1] + M[11] * P[2] + M[15] * P[3];
168 }
169 else if (sz == 3)
170 {
171 Q[0] = M[0] * P[0] + M[4] * P[1] + M[8] * P[2] + M[12];
172 Q[1] = M[1] * P[0] + M[5] * P[1] + M[9] * P[2] + M[13];
173 Q[2] = M[2] * P[0] + M[6] * P[1] + M[10] * P[2] + M[14];
174 Q[3] = M[3] * P[0] + M[7] * P[1] + M[11] * P[2] + M[15];
175 }
176 else if (sz == 2)
177 {
178 Q[0] = M[0] * P[0] + M[4] * P[1] + M[12];
179 Q[1] = M[1] * P[0] + M[5] * P[1] + M[13];
180 Q[2] = M[2] * P[0] + M[6] * P[1] + M[14];
181 Q[3] = M[3] * P[0] + M[7] * P[1] + M[15];
182 }
183 else if (sz == 1)
184 {
185 Q[0] = M[0] * P[0] + M[12];
186 Q[1] = M[1] * P[0] + M[13];
187 Q[2] = M[2] * P[0] + M[14];
188 Q[3] = M[3] * P[0] + M[15];
189 }
190 }
191
192
193 /*
194 * This is called only once. It initializes several tables with pointers
195 * to optimized transformation functions. This is where we can test for
196 * AMD 3Dnow! capability, Intel SSE, etc. and hook in the right code.
197 */
198 void
199 _math_init_transformation( void )
200 {
201 init_c_transformations();
202 init_c_norm_transform();
203 init_c_cliptest();
204 init_copy0();
205 init_dotprod();
206
207 #ifdef DEBUG_MATH
208 _math_test_all_transform_functions( "default" );
209 _math_test_all_normal_transform_functions( "default" );
210 _math_test_all_cliptest_functions( "default" );
211 #endif
212
213 #ifdef USE_X86_ASM
214 _mesa_init_all_x86_transform_asm();
215 #elif defined( USE_SPARC_ASM )
216 _mesa_init_all_sparc_transform_asm();
217 #elif defined( USE_PPC_ASM )
218 _mesa_init_all_ppc_transform_asm();
219 #elif defined( USE_X86_64_ASM )
220 _mesa_init_all_x86_64_transform_asm();
221 #endif
222 }
223
224 void
225 _math_init( void )
226 {
227 _math_init_transformation();
228 _math_init_translate();
229 _math_init_eval();
230 }