summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/svga/svga_state_vs.c
blob: aa351bd1aa540f7339efc67f4e03cc13dff9b7d2 (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
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
/**********************************************************
 * Copyright 2008-2009 VMware, Inc.  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, sublicense, 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 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
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * 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.
 *
 **********************************************************/

#include "util/u_inlines.h"
#include "pipe/p_defines.h"
#include "util/u_math.h"
#include "util/u_memory.h"
#include "util/u_bitmask.h"
#include "translate/translate.h"
#include "tgsi/tgsi_ureg.h"

#include "svga_context.h"
#include "svga_state.h"
#include "svga_cmd.h"
#include "svga_shader.h"
#include "svga_tgsi.h"

#include "svga_hw_reg.h"


/**
 * If we fail to compile a vertex shader we'll use a dummy/fallback shader
 * that simply emits a (0,0,0,1) vertex position.
 */
static const struct tgsi_token *
get_dummy_vertex_shader(void)
{
   static const float zero[4] = { 0.0, 0.0, 0.0, 1.0 };
   struct ureg_program *ureg;
   const struct tgsi_token *tokens;
   struct ureg_src src;
   struct ureg_dst dst;
   unsigned num_tokens;

   ureg = ureg_create(PIPE_SHADER_VERTEX);
   if (!ureg)
      return NULL;

   dst = ureg_DECL_output(ureg, TGSI_SEMANTIC_POSITION, 0);
   src = ureg_DECL_immediate(ureg, zero, 4);
   ureg_MOV(ureg, dst, src);
   ureg_END(ureg);

   tokens = ureg_get_tokens(ureg, &num_tokens);

   ureg_destroy(ureg);

   return tokens;
}


static struct svga_shader_variant *
translate_vertex_program(struct svga_context *svga,
                         const struct svga_vertex_shader *vs,
                         const struct svga_compile_key *key)
{
   if (svga_have_vgpu10(svga)) {
      return svga_tgsi_vgpu10_translate(svga, &vs->base, key,
                                        PIPE_SHADER_VERTEX);
   }
   else {
      return svga_tgsi_vgpu9_translate(svga, &vs->base, key,
                                       PIPE_SHADER_VERTEX);
   }
}


/**
 * Replace the given shader's instruction with a simple / dummy shader.
 * We use this when normal shader translation fails.
 */
static struct svga_shader_variant *
get_compiled_dummy_vertex_shader(struct svga_context *svga,
                                 struct svga_vertex_shader *vs,
                                 const struct svga_compile_key *key)
{
   const struct tgsi_token *dummy = get_dummy_vertex_shader();
   struct svga_shader_variant *variant;

   if (!dummy) {
      return NULL;
   }

   FREE((void *) vs->base.tokens);
   vs->base.tokens = dummy;

   variant = translate_vertex_program(svga, vs, key);
   return variant;
}


/**
 * Translate TGSI shader into an svga shader variant.
 */
static enum pipe_error
compile_vs(struct svga_context *svga,
           struct svga_vertex_shader *vs,
           const struct svga_compile_key *key,
           struct svga_shader_variant **out_variant)
{
   struct svga_shader_variant *variant;
   enum pipe_error ret = PIPE_ERROR;

   variant = translate_vertex_program(svga, vs, key);
   if (variant == NULL) {
      debug_printf("Failed to compile vertex shader,"
                   " using dummy shader instead.\n");
      variant = get_compiled_dummy_vertex_shader(svga, vs, key);
   }
   else if (svga_shader_too_large(svga, variant)) {
      /* too big, use dummy shader */
      debug_printf("Shader too large (%u bytes),"
                   " using dummy shader instead.\n",
                   (unsigned) (variant->nr_tokens
                               * sizeof(variant->tokens[0])));
      /* Free the too-large variant */
      svga_destroy_shader_variant(svga, SVGA3D_SHADERTYPE_VS, variant);
      /* Use simple pass-through shader instead */
      variant = get_compiled_dummy_vertex_shader(svga, vs, key);
   }

   if (!variant) {
      return PIPE_ERROR;
   }

   ret = svga_define_shader(svga, SVGA3D_SHADERTYPE_VS, variant);
   if (ret != PIPE_OK) {
      svga_destroy_shader_variant(svga, SVGA3D_SHADERTYPE_VS, variant);
      return ret;
   }

   *out_variant = variant;

   return PIPE_OK;
}


/* SVGA_NEW_PRESCALE, SVGA_NEW_RAST, SVGA_NEW_FS
 */
static void
make_vs_key(struct svga_context *svga, struct svga_compile_key *key)
{
   const enum pipe_shader_type shader = PIPE_SHADER_VERTEX;

   memset(key, 0, sizeof *key);

   if (svga->state.sw.need_swtnl && svga_have_vgpu10(svga)) {
      /* Set both of these flags, to match compile_passthrough_vs() */
      key->vs.passthrough = 1;
      key->vs.undo_viewport = 1;
      return;
   }

   /* SVGA_NEW_PRESCALE */
   key->vs.need_prescale = svga->state.hw_clear.prescale.enabled &&
                           (svga->curr.gs == NULL);

   /* SVGA_NEW_RAST */
   key->vs.allow_psiz = svga->curr.rast->templ.point_size_per_vertex;

   /* SVGA_NEW_FS */
   key->vs.fs_generic_inputs = svga->curr.fs->generic_inputs;

   svga_remap_generics(key->vs.fs_generic_inputs, key->generic_remap_table);

   /* SVGA_NEW_VELEMENT */
   key->vs.adjust_attrib_range = svga->curr.velems->adjust_attrib_range;
   key->vs.adjust_attrib_w_1 = svga->curr.velems->adjust_attrib_w_1;
   key->vs.attrib_is_pure_int = svga->curr.velems->attrib_is_pure_int;
   key->vs.adjust_attrib_itof = svga->curr.velems->adjust_attrib_itof;
   key->vs.adjust_attrib_utof = svga->curr.velems->adjust_attrib_utof;
   key->vs.attrib_is_bgra = svga->curr.velems->attrib_is_bgra;
   key->vs.attrib_puint_to_snorm = svga->curr.velems->attrib_puint_to_snorm;
   key->vs.attrib_puint_to_uscaled = svga->curr.velems->attrib_puint_to_uscaled;
   key->vs.attrib_puint_to_sscaled = svga->curr.velems->attrib_puint_to_sscaled;

   /* SVGA_NEW_TEXTURE_BINDING | SVGA_NEW_SAMPLER */
   svga_init_shader_key_common(svga, shader, key);

   /* SVGA_NEW_RAST */
   key->clip_plane_enable = svga->curr.rast->templ.clip_plane_enable;
}


/**
 * svga_reemit_vs_bindings - Reemit the vertex shader bindings
 */
enum pipe_error
svga_reemit_vs_bindings(struct svga_context *svga)
{
   enum pipe_error ret;
   struct svga_winsys_gb_shader *gbshader = NULL;
   SVGA3dShaderId shaderId = SVGA3D_INVALID_ID;

   assert(svga->rebind.flags.vs);
   assert(svga_have_gb_objects(svga));

   if (svga->state.hw_draw.vs) {
      gbshader = svga->state.hw_draw.vs->gb_shader;
      shaderId = svga->state.hw_draw.vs->id;
   }

   if (!svga_need_to_rebind_resources(svga)) {
      ret =  svga->swc->resource_rebind(svga->swc, NULL, gbshader,
                                        SVGA_RELOC_READ);
      goto out;
   }

   if (svga_have_vgpu10(svga))
      ret = SVGA3D_vgpu10_SetShader(svga->swc, SVGA3D_SHADERTYPE_VS,
                                    gbshader, shaderId);
   else
      ret = SVGA3D_SetGBShader(svga->swc, SVGA3D_SHADERTYPE_VS, gbshader);

 out:
   if (ret != PIPE_OK)
      return ret;

   svga->rebind.flags.vs = FALSE;
   return PIPE_OK;
}


/**
 * The current vertex shader is already executed by the 'draw'
 * module, so we just need to generate a simple vertex shader
 * to pass through all those VS outputs that will
 * be consumed by the fragment shader.
 * Used when we employ the 'draw' module.
 */
static enum pipe_error
compile_passthrough_vs(struct svga_context *svga,
                       struct svga_vertex_shader *vs,
                       struct svga_fragment_shader *fs,
                       struct svga_shader_variant **out_variant)
{
   struct svga_shader_variant *variant = NULL;
   unsigned num_inputs;
   unsigned i;
   unsigned num_elements;
   struct svga_vertex_shader new_vs;
   struct ureg_src src[PIPE_MAX_SHADER_INPUTS];
   struct ureg_dst dst[PIPE_MAX_SHADER_OUTPUTS];
   struct ureg_program *ureg;
   unsigned num_tokens;
   struct svga_compile_key key;
   enum pipe_error ret;

   assert(svga_have_vgpu10(svga));
   assert(fs);

   num_inputs = fs->base.info.num_inputs;

   ureg = ureg_create(PIPE_SHADER_VERTEX);
   if (!ureg)
      return PIPE_ERROR_OUT_OF_MEMORY;

   /* draw will always add position */
   dst[0] = ureg_DECL_output(ureg, TGSI_SEMANTIC_POSITION, 0);
   src[0] = ureg_DECL_vs_input(ureg, 0);
   num_elements = 1;

   /**
    * swtnl backend redefines the input layout based on the
    * fragment shader's inputs. So we only need to passthrough
    * those inputs that will be consumed by the fragment shader.
    * Note: DX10 requires the number of vertex elements
    * specified in the input layout to be no less than the
    * number of inputs to the vertex shader.
    */
   for (i = 0; i < num_inputs; i++) {
      switch (fs->base.info.input_semantic_name[i]) {
      case TGSI_SEMANTIC_COLOR:
      case TGSI_SEMANTIC_GENERIC:
      case TGSI_SEMANTIC_FOG:
         dst[num_elements] = ureg_DECL_output(ureg,
                                fs->base.info.input_semantic_name[i],
                                fs->base.info.input_semantic_index[i]);
         src[num_elements] = ureg_DECL_vs_input(ureg, num_elements);
         num_elements++;
         break;
      default:
         break;
      }
   }

   for (i = 0; i < num_elements; i++) {
      ureg_MOV(ureg, dst[i], src[i]);
   }

   ureg_END(ureg);

   memset(&new_vs, 0, sizeof(new_vs));
   new_vs.base.tokens = ureg_get_tokens(ureg, &num_tokens);
   tgsi_scan_shader(new_vs.base.tokens, &new_vs.base.info);

   memset(&key, 0, sizeof(key));
   key.vs.undo_viewport = 1;

   ret = compile_vs(svga, &new_vs, &key, &variant);
   if (ret != PIPE_OK)
      return ret;

   ureg_free_tokens(new_vs.base.tokens);
   ureg_destroy(ureg);

   /* Overwrite the variant key to indicate it's a pass-through VS */
   memset(&variant->key, 0, sizeof(variant->key));
   variant->key.vs.passthrough = 1;
   variant->key.vs.undo_viewport = 1;

   *out_variant = variant;

   return PIPE_OK;
}


static enum pipe_error
emit_hw_vs(struct svga_context *svga, unsigned dirty)
{
   struct svga_shader_variant *variant;
   struct svga_vertex_shader *vs = svga->curr.vs;
   struct svga_fragment_shader *fs = svga->curr.fs;
   enum pipe_error ret = PIPE_OK;
   struct svga_compile_key key;

   SVGA_STATS_TIME_PUSH(svga_sws(svga), SVGA_STATS_TIME_EMITVS);

   /* If there is an active geometry shader, and it has stream output
    * defined, then we will skip the stream output from the vertex shader
    */
   if (!svga_have_gs_streamout(svga)) {
      /* No GS stream out */
      if (svga_have_vs_streamout(svga)) {
         /* Set VS stream out */
         svga_set_stream_output(svga, vs->base.stream_output);
      }
      else {
         /* turn off stream out */
         svga_set_stream_output(svga, NULL);
      }
   }

   /* SVGA_NEW_NEED_SWTNL */
   if (svga->state.sw.need_swtnl && !svga_have_vgpu10(svga)) {
      /* No vertex shader is needed */
      variant = NULL;
   }
   else {
      make_vs_key(svga, &key);

      /* See if we already have a VS variant that matches the key */
      variant = svga_search_shader_key(&vs->base, &key);

      if (!variant) {
         /* Create VS variant now */
         if (key.vs.passthrough) {
            ret = compile_passthrough_vs(svga, vs, fs, &variant);
         }
         else {
            ret = compile_vs(svga, vs, &key, &variant);
         }
         if (ret != PIPE_OK)
            goto done;

         /* insert the new variant at head of linked list */
         assert(variant);
         variant->next = vs->base.variants;
         vs->base.variants = variant;
      }
   }

   if (variant != svga->state.hw_draw.vs) {
      /* Bind the new variant */
      if (variant) {
         ret = svga_set_shader(svga, SVGA3D_SHADERTYPE_VS, variant);
         if (ret != PIPE_OK)
            goto done;
         svga->rebind.flags.vs = FALSE;
      }

      svga->dirty |= SVGA_NEW_VS_VARIANT;
      svga->state.hw_draw.vs = variant;
   }

done:
   SVGA_STATS_TIME_POP(svga_sws(svga));
   return ret;
}

struct svga_tracked_state svga_hw_vs = 
{
   "vertex shader (hwtnl)",
   (SVGA_NEW_VS |
    SVGA_NEW_FS |
    SVGA_NEW_TEXTURE_BINDING |
    SVGA_NEW_SAMPLER |
    SVGA_NEW_RAST |
    SVGA_NEW_PRESCALE |
    SVGA_NEW_VELEMENT |
    SVGA_NEW_NEED_SWTNL),
   emit_hw_vs
};