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
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
|
/**************************************************************************
*
* Copyright 2010 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, 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.
*
**************************************************************************/
#include "util/u_math.h"
#include "util/u_memory.h"
#include "draw/draw_context.h"
#include "draw/draw_vbuf.h"
#include "draw/draw_vertex.h"
#include "draw/draw_pt.h"
#include "draw/draw_vs.h"
#include "draw/draw_gs.h"
#include "draw/draw_llvm.h"
struct llvm_middle_end {
struct draw_pt_middle_end base;
struct draw_context *draw;
struct pt_emit *emit;
struct pt_fetch *fetch;
struct pt_post_vs *post_vs;
unsigned vertex_data_offset;
unsigned vertex_size;
unsigned prim;
unsigned opt;
struct draw_llvm *llvm;
struct draw_llvm_variant *variants;
struct draw_llvm_variant *current_variant;
int nr_variants;
};
static void
llvm_middle_end_prepare( struct draw_pt_middle_end *middle,
unsigned prim,
unsigned opt,
unsigned *max_vertices )
{
struct llvm_middle_end *fpme = (struct llvm_middle_end *)middle;
struct draw_context *draw = fpme->draw;
struct draw_vertex_shader *vs = draw->vs.vertex_shader;
struct draw_geometry_shader *gs = draw->gs.geometry_shader;
struct draw_llvm_variant_key key;
struct draw_llvm_variant *variant = NULL;
unsigned i;
unsigned instance_id_index = ~0;
/* Add one to num_outputs because the pipeline occasionally tags on
* an additional texcoord, eg for AA lines.
*/
unsigned nr = MAX2( vs->info.num_inputs,
vs->info.num_outputs + 1 );
/* Scan for instanceID system value.
*/
for (i = 0; i < vs->info.num_inputs; i++) {
if (vs->info.input_semantic_name[i] == TGSI_SEMANTIC_INSTANCEID) {
instance_id_index = i;
break;
}
}
fpme->prim = prim;
fpme->opt = opt;
/* Always leave room for the vertex header whether we need it or
* not. It's hard to get rid of it in particular because of the
* viewport code in draw_pt_post_vs.c.
*/
fpme->vertex_size = sizeof(struct vertex_header) + nr * 4 * sizeof(float);
draw_pt_fetch_prepare( fpme->fetch,
vs->info.num_inputs,
fpme->vertex_size,
instance_id_index );
if (opt & PT_SHADE) {
vs->prepare(vs, draw);
draw_geometry_shader_prepare(gs, draw);
}
/* XXX: it's not really gl rasterization rules we care about here,
* but gl vs dx9 clip spaces.
*/
draw_pt_post_vs_prepare( fpme->post_vs,
(boolean)draw->bypass_clipping,
(boolean)(draw->identity_viewport),
(boolean)draw->rasterizer->gl_rasterization_rules,
(draw->vs.edgeflag_output ? true : false) );
if (!(opt & PT_PIPELINE)) {
draw_pt_emit_prepare( fpme->emit,
prim,
max_vertices );
*max_vertices = MAX2( *max_vertices,
DRAW_PIPE_MAX_VERTICES );
}
else {
*max_vertices = DRAW_PIPE_MAX_VERTICES;
}
/* return even number */
*max_vertices = *max_vertices & ~1;
draw_llvm_make_variant_key(fpme->llvm, &key);
variant = fpme->variants;
while(variant) {
if(memcmp(&variant->key, &key, sizeof key) == 0)
break;
variant = variant->next;
}
if (!variant) {
variant = draw_llvm_prepare(fpme->llvm, nr);
variant->next = fpme->variants;
fpme->variants = variant;
++fpme->nr_variants;
}
fpme->current_variant = variant;
/*XXX we only support one constant buffer */
fpme->llvm->jit_context.vs_constants =
draw->pt.user.vs_constants[0];
fpme->llvm->jit_context.gs_constants =
draw->pt.user.gs_constants[0];
}
static void llvm_middle_end_run( struct draw_pt_middle_end *middle,
const unsigned *fetch_elts,
unsigned fetch_count,
const ushort *draw_elts,
unsigned draw_count )
{
struct llvm_middle_end *fpme = (struct llvm_middle_end *)middle;
struct draw_context *draw = fpme->draw;
unsigned opt = fpme->opt;
unsigned alloc_count = align( fetch_count, 4 );
struct vertex_header *pipeline_verts =
(struct vertex_header *)MALLOC(fpme->vertex_size * alloc_count);
if (!pipeline_verts) {
/* Not much we can do here - just skip the rendering.
*/
assert(0);
return;
}
fpme->current_variant->jit_func_elts( &fpme->llvm->jit_context,
pipeline_verts,
(const char **)draw->pt.user.vbuffer,
fetch_elts,
fetch_count,
fpme->vertex_size,
draw->pt.vertex_buffer );
if (draw_pt_post_vs_run( fpme->post_vs,
pipeline_verts,
fetch_count,
fpme->vertex_size ))
{
opt |= PT_PIPELINE;
}
/* Do we need to run the pipeline?
*/
if (opt & PT_PIPELINE) {
draw_pipeline_run( fpme->draw,
fpme->prim,
pipeline_verts,
fetch_count,
fpme->vertex_size,
draw_elts,
draw_count );
}
else {
draw_pt_emit( fpme->emit,
(const float (*)[4])pipeline_verts->data,
fetch_count,
fpme->vertex_size,
draw_elts,
draw_count );
}
FREE(pipeline_verts);
}
static void llvm_middle_end_linear_run( struct draw_pt_middle_end *middle,
unsigned start,
unsigned count)
{
struct llvm_middle_end *fpme = (struct llvm_middle_end *)middle;
struct draw_context *draw = fpme->draw;
unsigned opt = fpme->opt;
unsigned alloc_count = align( count, 4 );
struct vertex_header *pipeline_verts =
(struct vertex_header *)MALLOC(fpme->vertex_size * alloc_count);
if (!pipeline_verts) {
/* Not much we can do here - just skip the rendering.
*/
assert(0);
return;
}
#if 0
debug_printf("#### Pipeline = %p (data = %p)\n",
pipeline_verts, pipeline_verts->data);
#endif
fpme->current_variant->jit_func( &fpme->llvm->jit_context,
pipeline_verts,
(const char **)draw->pt.user.vbuffer,
start,
count,
fpme->vertex_size,
draw->pt.vertex_buffer );
if (draw_pt_post_vs_run( fpme->post_vs,
pipeline_verts,
count,
fpme->vertex_size ))
{
opt |= PT_PIPELINE;
}
/* Do we need to run the pipeline?
*/
if (opt & PT_PIPELINE) {
draw_pipeline_run_linear( fpme->draw,
fpme->prim,
pipeline_verts,
count,
fpme->vertex_size);
}
else {
draw_pt_emit_linear( fpme->emit,
(const float (*)[4])pipeline_verts->data,
fpme->vertex_size,
count );
}
FREE(pipeline_verts);
}
static boolean
llvm_middle_end_linear_run_elts( struct draw_pt_middle_end *middle,
unsigned start,
unsigned count,
const ushort *draw_elts,
unsigned draw_count )
{
struct llvm_middle_end *fpme = (struct llvm_middle_end *)middle;
struct draw_context *draw = fpme->draw;
unsigned opt = fpme->opt;
unsigned alloc_count = align( count, 4 );
struct vertex_header *pipeline_verts =
(struct vertex_header *)MALLOC(fpme->vertex_size * alloc_count);
if (!pipeline_verts)
return FALSE;
fpme->current_variant->jit_func( &fpme->llvm->jit_context,
pipeline_verts,
(const char **)draw->pt.user.vbuffer,
start,
count,
fpme->vertex_size,
draw->pt.vertex_buffer );
if (draw_pt_post_vs_run( fpme->post_vs,
pipeline_verts,
count,
fpme->vertex_size ))
{
opt |= PT_PIPELINE;
}
/* Do we need to run the pipeline?
*/
if (opt & PT_PIPELINE) {
draw_pipeline_run( fpme->draw,
fpme->prim,
pipeline_verts,
count,
fpme->vertex_size,
draw_elts,
draw_count );
}
else {
draw_pt_emit( fpme->emit,
(const float (*)[4])pipeline_verts->data,
count,
fpme->vertex_size,
draw_elts,
draw_count );
}
FREE(pipeline_verts);
return TRUE;
}
static void llvm_middle_end_finish( struct draw_pt_middle_end *middle )
{
/* nothing to do */
}
static void llvm_middle_end_destroy( struct draw_pt_middle_end *middle )
{
struct llvm_middle_end *fpme = (struct llvm_middle_end *)middle;
struct draw_context *draw = fpme->draw;
struct draw_llvm_variant *variant = NULL;
variant = fpme->variants;
while(variant) {
struct draw_llvm_variant *next = variant->next;
if (variant->function_elts) {
if (variant->function_elts)
LLVMFreeMachineCodeForFunction(draw->engine,
variant->function_elts);
LLVMDeleteFunction(variant->function_elts);
}
if (variant->function) {
if (variant->function)
LLVMFreeMachineCodeForFunction(draw->engine,
variant->function);
LLVMDeleteFunction(variant->function);
}
FREE(variant);
variant = next;
}
if (fpme->fetch)
draw_pt_fetch_destroy( fpme->fetch );
if (fpme->emit)
draw_pt_emit_destroy( fpme->emit );
if (fpme->post_vs)
draw_pt_post_vs_destroy( fpme->post_vs );
if (fpme->llvm)
draw_llvm_destroy( fpme->llvm );
FREE(middle);
}
struct draw_pt_middle_end *draw_pt_fetch_pipeline_or_emit_llvm( struct draw_context *draw )
{
struct llvm_middle_end *fpme = 0;
if (!draw->engine)
return NULL;
fpme = CALLOC_STRUCT( llvm_middle_end );
if (!fpme)
goto fail;
fpme->base.prepare = llvm_middle_end_prepare;
fpme->base.run = llvm_middle_end_run;
fpme->base.run_linear = llvm_middle_end_linear_run;
fpme->base.run_linear_elts = llvm_middle_end_linear_run_elts;
fpme->base.finish = llvm_middle_end_finish;
fpme->base.destroy = llvm_middle_end_destroy;
fpme->draw = draw;
fpme->fetch = draw_pt_fetch_create( draw );
if (!fpme->fetch)
goto fail;
fpme->post_vs = draw_pt_post_vs_create( draw );
if (!fpme->post_vs)
goto fail;
fpme->emit = draw_pt_emit_create( draw );
if (!fpme->emit)
goto fail;
fpme->llvm = draw_llvm_create(draw);
if (!fpme->llvm)
goto fail;
fpme->variants = NULL;
fpme->current_variant = NULL;
fpme->nr_variants = 0;
return &fpme->base;
fail:
if (fpme)
llvm_middle_end_destroy( &fpme->base );
return NULL;
}
|