summaryrefslogtreecommitdiff
path: root/src/mesa/state_tracker/st_atom_shader.c
blob: 4c106eb89b0bff180014946737c36bc34559ed82 (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
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
/**************************************************************************
 * 
 * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
 * All Rights Reserved.
 * 
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the
 * "Software"), to deal in the Software without restriction, including
 * without limitation the rights to use, copy, modify, merge, publish,
 * distribute, sub license, and/or sell copies of the Software, and to
 * permit persons to whom the Software is furnished to do so, subject to
 * the following conditions:
 * 
 * The above copyright notice and this permission notice (including the
 * next paragraph) shall be included in all copies or substantial portions
 * of the Software.
 * 
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
 * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
 * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 * 
 **************************************************************************/

/**
 * State validation for vertex/fragment shaders.
 * Note that we have to delay most vertex/fragment shader translation
 * until rendering time since the linkage between the vertex outputs and
 * fragment inputs can vary depending on the pairing of shaders.
 *
 * Authors:
 *   Brian Paul
 */



#include "main/imports.h"
#include "main/mtypes.h"
#include "main/macros.h"
#include "shader/program.h"

#include "pipe/p_context.h"
#include "pipe/p_shader_tokens.h"

#include "util/u_simple_shaders.h"

#include "cso_cache/cso_context.h"

#include "st_context.h"
#include "st_atom.h"
#include "st_program.h"
#include "st_atom_shader.h"


/**
 * This represents a vertex program, especially translated to match
 * the inputs of a particular fragment shader.
 */
struct translated_vertex_program
{
   struct st_vertex_program *master;

   /** The fragment shader "signature" this vertex shader is meant for: */
   GLbitfield frag_inputs;

   /** Compared against master vertex program's serialNo: */
   GLuint serialNo;

   /** Maps VERT_RESULT_x to slot */
   GLuint output_to_slot[VERT_RESULT_MAX];
   ubyte output_to_semantic_name[VERT_RESULT_MAX];
   ubyte output_to_semantic_index[VERT_RESULT_MAX];

   /** Pointer to the translated vertex program */
   struct st_vertex_program *vp;

   struct translated_vertex_program *next;  /**< next in linked list */
};



/**
 * Given a vertex program output attribute, return the corresponding
 * fragment program input attribute.
 * \return -1 for vertex outputs that have no corresponding fragment input
 */
static GLint
vp_out_to_fp_in(GLuint vertResult)
{
   if (vertResult >= VERT_RESULT_TEX0 &&
       vertResult < VERT_RESULT_TEX0 + MAX_TEXTURE_COORD_UNITS)
      return FRAG_ATTRIB_TEX0 + (vertResult - VERT_RESULT_TEX0);

   if (vertResult >= VERT_RESULT_VAR0 &&
       vertResult < VERT_RESULT_VAR0 + MAX_VARYING)
      return FRAG_ATTRIB_VAR0 + (vertResult - VERT_RESULT_VAR0);

   switch (vertResult) {
   case VERT_RESULT_HPOS:
      return FRAG_ATTRIB_WPOS;
   case VERT_RESULT_COL0:
      return FRAG_ATTRIB_COL0;
   case VERT_RESULT_COL1:
      return FRAG_ATTRIB_COL1;
   case VERT_RESULT_FOGC:
      return FRAG_ATTRIB_FOGC;
   default:
      /* Back-face colors, edge flags, etc */
      return -1;
   }
}


/**
 * Find a translated vertex program that corresponds to stvp and
 * has outputs matched to stfp's inputs.
 * This performs vertex and fragment translation (to TGSI) when needed.
 */
static struct translated_vertex_program *
find_translated_vp(struct st_context *st,
                   struct st_vertex_program *stvp,
                   struct st_fragment_program *stfp)
{
   static const GLuint UNUSED = ~0;
   struct translated_vertex_program *xvp;
   const GLbitfield fragInputsRead = stfp->Base.Base.InputsRead;

   /*
    * Translate fragment program if needed.
    */
   if (!stfp->state.tokens) {
      GLuint inAttr, numIn = 0;

      for (inAttr = 0; inAttr < FRAG_ATTRIB_MAX; inAttr++) {
         if (fragInputsRead & (1 << inAttr)) {
            stfp->input_to_slot[inAttr] = numIn;
            numIn++;
         }
         else {
            stfp->input_to_slot[inAttr] = UNUSED;
         }
      }

      stfp->num_input_slots = numIn;

      assert(stfp->Base.Base.NumInstructions > 0);

      st_translate_fragment_program(st, stfp, stfp->input_to_slot);
   }


   /* See if we've got a translated vertex program whose outputs match
    * the fragment program's inputs.
    * XXX This could be a hash lookup, using InputsRead as the key.
    */
   for (xvp = stfp->vertex_programs; xvp; xvp = xvp->next) {
      if (xvp->master == stvp && xvp->frag_inputs == fragInputsRead) {
         break;
      }
   }

   /* No?  Allocate translated vp object now */
   if (!xvp) {
      xvp = ST_CALLOC_STRUCT(translated_vertex_program);
      xvp->frag_inputs = fragInputsRead;
      xvp->master = stvp;

      xvp->next = stfp->vertex_programs;
      stfp->vertex_programs = xvp;
   }

   /* See if we need to translate vertex program to TGSI form */
   if (xvp->serialNo != stvp->serialNo) {
      GLuint outAttr;
      const GLbitfield64 outputsWritten = stvp->Base.Base.OutputsWritten;
      GLuint numVpOuts = 0;
      GLboolean emitPntSize = GL_FALSE, emitBFC0 = GL_FALSE, emitBFC1 = GL_FALSE;
      GLbitfield usedGenerics = 0x0;
      GLbitfield usedOutputSlots = 0x0;

      /* Compute mapping of vertex program outputs to slots, which depends
       * on the fragment program's input->slot mapping.
       */
      for (outAttr = 0; outAttr < VERT_RESULT_MAX; outAttr++) {
         /* set defaults: */
         xvp->output_to_slot[outAttr] = UNUSED;
         xvp->output_to_semantic_name[outAttr] = TGSI_SEMANTIC_COUNT;
         xvp->output_to_semantic_index[outAttr] = 99;

         if (outAttr == VERT_RESULT_HPOS) {
            /* always put xformed position into slot zero */
            GLuint slot = 0;
            xvp->output_to_slot[VERT_RESULT_HPOS] = slot;
            xvp->output_to_semantic_name[outAttr] = TGSI_SEMANTIC_POSITION;
            xvp->output_to_semantic_index[outAttr] = 0;
            numVpOuts++;
            usedOutputSlots |= (1 << slot);
         }
         else if (outputsWritten & (1 << outAttr)) {
            /* see if the frag prog wants this vert output */
            GLint fpInAttrib = vp_out_to_fp_in(outAttr);
            if (fpInAttrib >= 0) {
               GLuint fpInSlot = stfp->input_to_slot[fpInAttrib];
               if (fpInSlot != ~0) {
                  /* match this vp output to the fp input */
                  GLuint vpOutSlot = stfp->input_map[fpInSlot];
                  xvp->output_to_slot[outAttr] = vpOutSlot;
                  xvp->output_to_semantic_name[outAttr] = stfp->input_semantic_name[fpInSlot];
                  xvp->output_to_semantic_index[outAttr] = stfp->input_semantic_index[fpInSlot];
                  numVpOuts++;
                  usedOutputSlots |= (1 << vpOutSlot);
               }
               else {
#if 0 /*debug*/
                  printf("VP output %d not used by FP\n", outAttr);
#endif
               }
            }
            else if (outAttr == VERT_RESULT_PSIZ)
               emitPntSize = GL_TRUE;
            else if (outAttr == VERT_RESULT_BFC0)
               emitBFC0 = GL_TRUE;
            else if (outAttr == VERT_RESULT_BFC1)
               emitBFC1 = GL_TRUE;
         }
#if 0 /*debug*/
         printf("assign vp output_to_slot[%d] = %d\n", outAttr, 
                xvp->output_to_slot[outAttr]);
#endif
      }

      /* must do these last */
      if (emitPntSize) {
         GLuint slot = numVpOuts++;
         xvp->output_to_slot[VERT_RESULT_PSIZ] = slot;
         xvp->output_to_semantic_name[VERT_RESULT_PSIZ] = TGSI_SEMANTIC_PSIZE;
         xvp->output_to_semantic_index[VERT_RESULT_PSIZ] = 0;
         usedOutputSlots |= (1 << slot);
      }
      if (emitBFC0) {
         GLuint slot = numVpOuts++;
         xvp->output_to_slot[VERT_RESULT_BFC0] = slot;
         xvp->output_to_semantic_name[VERT_RESULT_BFC0] = TGSI_SEMANTIC_COLOR;
         xvp->output_to_semantic_index[VERT_RESULT_BFC0] = 0;
         usedOutputSlots |= (1 << slot);
      }
      if (emitBFC1) {
         GLuint slot = numVpOuts++;
         xvp->output_to_slot[VERT_RESULT_BFC1] = slot;
         xvp->output_to_semantic_name[VERT_RESULT_BFC1] = TGSI_SEMANTIC_COLOR;
         xvp->output_to_semantic_index[VERT_RESULT_BFC1] = 1;
         usedOutputSlots |= (1 << slot);
      }

      /* build usedGenerics mask */
      usedGenerics = 0x0;
      for (outAttr = 0; outAttr < VERT_RESULT_MAX; outAttr++) {
         if (xvp->output_to_semantic_name[outAttr] == TGSI_SEMANTIC_GENERIC) {
            usedGenerics |= (1 << xvp->output_to_semantic_index[outAttr]);
         }
      }

      /* For each vertex program output that doesn't match up to a fragment
       * program input, map the vertex program output to a free slot and
       * free generic attribute.
       */
      for (outAttr = 0; outAttr < VERT_RESULT_MAX; outAttr++) {
         if (outputsWritten & (1 << outAttr)) {
            if (xvp->output_to_slot[outAttr] == UNUSED) {
               GLint freeGeneric = _mesa_ffs(~usedGenerics) - 1;
               GLint freeSlot = _mesa_ffs(~usedOutputSlots) - 1;
               usedGenerics |= (1 << freeGeneric);
               usedOutputSlots |= (1 << freeSlot);
               xvp->output_to_slot[outAttr] = freeSlot;
               xvp->output_to_semantic_name[outAttr] = TGSI_SEMANTIC_GENERIC;
               xvp->output_to_semantic_index[outAttr] = freeGeneric;
            }
         }

#if 0 /*debug*/
         printf("vp output_to_slot[%d] = %d\n", outAttr, 
                xvp->output_to_slot[outAttr]);
#endif
      }

      st_translate_vertex_program(st, stvp, xvp->output_to_slot,
                                  xvp->output_to_semantic_name,
                                  xvp->output_to_semantic_index);

      xvp->vp = stvp;

      /* translated VP is up to date now */
      xvp->serialNo = stvp->serialNo;
   }

   return xvp;
}


void
st_free_translated_vertex_programs(struct st_context *st,
                                   struct translated_vertex_program *xvp)
{
   struct translated_vertex_program *next;

   while (xvp) {
      next = xvp->next;
      _mesa_free(xvp);
      xvp = next;
   }
}


static void *
get_passthrough_fs(struct st_context *st)
{
   if (!st->passthrough_fs) {
      st->passthrough_fs =
         util_make_fragment_passthrough_shader(st->pipe);
   }

   return st->passthrough_fs;
}


static void
update_linkage( struct st_context *st )
{
   struct st_vertex_program *stvp;
   struct st_fragment_program *stfp;
   struct translated_vertex_program *xvp;

   /* find active shader and params -- Should be covered by
    * ST_NEW_VERTEX_PROGRAM
    */
   assert(st->ctx->VertexProgram._Current);
   stvp = st_vertex_program(st->ctx->VertexProgram._Current);
   assert(stvp->Base.Base.Target == GL_VERTEX_PROGRAM_ARB);

   assert(st->ctx->FragmentProgram._Current);
   stfp = st_fragment_program(st->ctx->FragmentProgram._Current);
   assert(stfp->Base.Base.Target == GL_FRAGMENT_PROGRAM_ARB);

   xvp = find_translated_vp(st, stvp, stfp);

   st_reference_vertprog(st, &st->vp, stvp);
   st_reference_fragprog(st, &st->fp, stfp);

   cso_set_vertex_shader_handle(st->cso_context, stvp->driver_shader);

   if (st->missing_textures) {
      /* use a pass-through frag shader that uses no textures */
      void *fs = get_passthrough_fs(st);
      cso_set_fragment_shader_handle(st->cso_context, fs);
   }
   else {
      cso_set_fragment_shader_handle(st->cso_context, stfp->driver_shader);
   }

   st->vertex_result_to_slot = xvp->output_to_slot;
}


const struct st_tracked_state st_update_shader = {
   "st_update_shader",					/* name */
   {							/* dirty */
      0,						/* mesa */
      ST_NEW_VERTEX_PROGRAM | ST_NEW_FRAGMENT_PROGRAM	/* st */
   },
   update_linkage					/* update */
};