summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/i965/brw_clip_util.c
blob: 6f4775dbb7632da8383ae2a413ffad95a3283fe6 (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
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
/*
 Copyright (C) Intel Corp.  2006.  All Rights Reserved.
 Intel funded Tungsten Graphics to
 develop this 3D driver.

 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 (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 NONINFRINGEMENT.
 IN NO EVENT SHALL THE COPYRIGHT OWNER(S) 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.

 **********************************************************************/
 /*
  * Authors:
  *   Keith Whitwell <keithw@vmware.com>
  */


#include "main/glheader.h"
#include "main/macros.h"
#include "main/enums.h"
#include "program/program.h"

#include "intel_batchbuffer.h"

#include "brw_defines.h"
#include "brw_context.h"
#include "brw_eu.h"
#include "brw_clip.h"




struct brw_reg get_tmp( struct brw_clip_compile *c )
{
   struct brw_reg tmp = brw_vec4_grf(c->last_tmp, 0);

   if (++c->last_tmp > c->prog_data.total_grf)
      c->prog_data.total_grf = c->last_tmp;

   return tmp;
}

static void release_tmp( struct brw_clip_compile *c, struct brw_reg tmp )
{
   if (tmp.nr == c->last_tmp-1)
      c->last_tmp--;
}


static struct brw_reg make_plane_ud(GLuint x, GLuint y, GLuint z, GLuint w)
{
   return brw_imm_ud((w<<24) | (z<<16) | (y<<8) | x);
}


void brw_clip_init_planes( struct brw_clip_compile *c )
{
   struct brw_compile *p = &c->func;

   if (!c->key.nr_userclip) {
      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 0), make_plane_ud( 0,    0, 0xff, 1));
      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 1), make_plane_ud( 0,    0,    1, 1));
      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 2), make_plane_ud( 0, 0xff,    0, 1));
      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 3), make_plane_ud( 0,    1,    0, 1));
      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 4), make_plane_ud(0xff,  0,    0, 1));
      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 5), make_plane_ud( 1,    0,    0, 1));
   }
}



#define W 3

/* Project 'pos' to screen space (or back again), overwrite with results:
 */
void brw_clip_project_position(struct brw_clip_compile *c, struct brw_reg pos )
{
   struct brw_compile *p = &c->func;

   /* calc rhw
    */
   brw_math_invert(p, get_element(pos, W), get_element(pos, W));

   /* value.xyz *= value.rhw
    */
   brw_set_default_access_mode(p, BRW_ALIGN_16);
   brw_MUL(p, brw_writemask(pos, WRITEMASK_XYZ), pos, brw_swizzle1(pos, W));
   brw_set_default_access_mode(p, BRW_ALIGN_1);
}


static void brw_clip_project_vertex( struct brw_clip_compile *c,
				     struct brw_indirect vert_addr )
{
   struct brw_compile *p = &c->func;
   struct brw_reg tmp = get_tmp(c);
   GLuint hpos_offset = brw_varying_to_offset(&c->vue_map, VARYING_SLOT_POS);
   GLuint ndc_offset = brw_varying_to_offset(&c->vue_map,
                                             BRW_VARYING_SLOT_NDC);

   /* Fixup position.  Extract from the original vertex and re-project
    * to screen space:
    */
   brw_MOV(p, tmp, deref_4f(vert_addr, hpos_offset));
   brw_clip_project_position(c, tmp);
   brw_MOV(p, deref_4f(vert_addr, ndc_offset), tmp);
	
   release_tmp(c, tmp);
}




/* Interpolate between two vertices and put the result into a0.0.
 * Increment a0.0 accordingly.
 *
 * Beware that dest_ptr can be equal to v0_ptr!
 */
void brw_clip_interp_vertex( struct brw_clip_compile *c,
			     struct brw_indirect dest_ptr,
			     struct brw_indirect v0_ptr, /* from */
			     struct brw_indirect v1_ptr, /* to */
			     struct brw_reg t0,
			     bool force_edgeflag)
{
   struct brw_compile *p = &c->func;
   struct brw_reg t_nopersp, v0_ndc_copy;
   GLuint slot;

   /* Just copy the vertex header:
    */
   /*
    * After CLIP stage, only first 256 bits of the VUE are read
    * back on Ironlake, so needn't change it
    */
   brw_copy_indirect_to_indirect(p, dest_ptr, v0_ptr, 1);


   /* First handle the 3D and NDC interpolation, in case we
    * need noperspective interpolation. Doing it early has no
    * performance impact in any case.
    */

   /* Take a copy of the v0 NDC coordinates, in case dest == v0. */
   if (c->has_noperspective_shading) {
      GLuint offset = brw_varying_to_offset(&c->vue_map,
                                                 BRW_VARYING_SLOT_NDC);
      v0_ndc_copy = get_tmp(c);
      brw_MOV(p, v0_ndc_copy, deref_4f(v0_ptr, offset));
   }

   /* Compute the new 3D position
    *
    * dest_hpos = v0_hpos * (1 - t0) + v1_hpos * t0
    */
   {
      GLuint delta = brw_varying_to_offset(&c->vue_map, VARYING_SLOT_POS);
      struct brw_reg tmp = get_tmp(c);
      brw_MUL(p, vec4(brw_null_reg()), deref_4f(v1_ptr, delta), t0);
      brw_MAC(p, tmp, negate(deref_4f(v0_ptr, delta)), t0);
      brw_ADD(p, deref_4f(dest_ptr, delta), deref_4f(v0_ptr, delta), tmp);
      release_tmp(c, tmp);
   }

   /* Recreate the projected (NDC) coordinate in the new vertex header */
   brw_clip_project_vertex(c, dest_ptr);

   /* If we have noperspective attributes,
    * we need to compute the screen-space t
    */
   if (c->has_noperspective_shading) {
      GLuint delta = brw_varying_to_offset(&c->vue_map,
                                                BRW_VARYING_SLOT_NDC);
      struct brw_reg tmp = get_tmp(c);
      t_nopersp = get_tmp(c);

      /* t_nopersp = vec4(v1.xy, dest.xy) */
      brw_MOV(p, t_nopersp, deref_4f(v1_ptr, delta));
      brw_MOV(p, tmp, deref_4f(dest_ptr, delta));
      brw_set_default_access_mode(p, BRW_ALIGN_16);
      brw_MOV(p,
              brw_writemask(t_nopersp, WRITEMASK_ZW),
              brw_swizzle(tmp, 0, 1, 0, 1));

      /* t_nopersp = vec4(v1.xy, dest.xy) - v0.xyxy */
      brw_ADD(p, t_nopersp, t_nopersp,
              negate(brw_swizzle(v0_ndc_copy, 0, 1, 0, 1)));

      /* Add the absolute values of the X and Y deltas so that if
       * the points aren't in the same place on the screen we get
       * nonzero values to divide.
       *
       * After that, we have vert1 - vert0 in t_nopersp.x and
       * vertnew - vert0 in t_nopersp.y
       *
       * t_nopersp = vec2(|v1.x  -v0.x| + |v1.y  -v0.y|,
       *                  |dest.x-v0.x| + |dest.y-v0.y|)
       */
      brw_ADD(p,
              brw_writemask(t_nopersp, WRITEMASK_XY),
              brw_abs(brw_swizzle(t_nopersp, 0, 2, 0, 0)),
              brw_abs(brw_swizzle(t_nopersp, 1, 3, 0, 0)));
      brw_set_default_access_mode(p, BRW_ALIGN_1);

      /* If the points are in the same place, just substitute a
       * value to avoid divide-by-zero
       */
      brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_EQ,
              vec1(t_nopersp),
              brw_imm_f(0));
      brw_IF(p, BRW_EXECUTE_1);
      brw_MOV(p, t_nopersp, brw_imm_vf4(1, 0, 0, 0));
      brw_ENDIF(p);

      /* Now compute t_nopersp = t_nopersp.y/t_nopersp.x and broadcast it. */
      brw_math_invert(p, get_element(t_nopersp, 0), get_element(t_nopersp, 0));
      brw_MUL(p, vec1(t_nopersp), vec1(t_nopersp),
            vec1(suboffset(t_nopersp, 1)));
      brw_set_default_access_mode(p, BRW_ALIGN_16);
      brw_MOV(p, t_nopersp, brw_swizzle(t_nopersp, 0, 0, 0, 0));
      brw_set_default_access_mode(p, BRW_ALIGN_1);

      release_tmp(c, tmp);
      release_tmp(c, v0_ndc_copy);
   }

   /* Now we can iterate over each attribute
    * (could be done in pairs?)
    */
   for (slot = 0; slot < c->vue_map.num_slots; slot++) {
      int varying = c->vue_map.slot_to_varying[slot];
      GLuint delta = brw_vue_slot_to_offset(slot);

      /* HPOS, NDC already handled above */
      if (varying == VARYING_SLOT_POS || varying == BRW_VARYING_SLOT_NDC)
         continue;


      if (varying == VARYING_SLOT_EDGE) {
	 if (force_edgeflag)
	    brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(1));
	 else
	    brw_MOV(p, deref_4f(dest_ptr, delta), deref_4f(v0_ptr, delta));
      } else if (varying == VARYING_SLOT_PSIZ) {
         /* PSIZ doesn't need interpolation because it isn't used by the
          * fragment shader.
          */
      } else if (varying < VARYING_SLOT_MAX) {
	 /* This is a true vertex result (and not a special value for the VUE
	  * header), so interpolate:
	  *
	  *        New = attr0 + t*attr1 - t*attr0
          *
          * Unless the attribute is flat shaded -- in which case just copy
          * from one of the sources (doesn't matter which; already copied from pv)
	  */
         GLuint interp = c->key.interpolation_mode.mode[slot];

         if (interp != INTERP_QUALIFIER_FLAT) {
            struct brw_reg tmp = get_tmp(c);
            struct brw_reg t =
               interp == INTERP_QUALIFIER_NOPERSPECTIVE ? t_nopersp : t0;

            brw_MUL(p,
                  vec4(brw_null_reg()),
                  deref_4f(v1_ptr, delta),
                  t);

            brw_MAC(p,
                  tmp,
                  negate(deref_4f(v0_ptr, delta)),
                  t);

            brw_ADD(p,
                  deref_4f(dest_ptr, delta),
                  deref_4f(v0_ptr, delta),
                  tmp);

            release_tmp(c, tmp);
         }
         else {
            brw_MOV(p,
                  deref_4f(dest_ptr, delta),
                  deref_4f(v0_ptr, delta));
         }
      }
   }

   if (c->vue_map.num_slots % 2) {
      GLuint delta = brw_vue_slot_to_offset(c->vue_map.num_slots);

      brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(0));
   }

   if (c->has_noperspective_shading)
      release_tmp(c, t_nopersp);
}

void brw_clip_emit_vue(struct brw_clip_compile *c,
		       struct brw_indirect vert,
                       enum brw_urb_write_flags flags,
		       GLuint header)
{
   struct brw_compile *p = &c->func;
   bool allocate = flags & BRW_URB_WRITE_ALLOCATE;

   brw_clip_ff_sync(c);

   /* Any URB entry that is allocated must subsequently be used or discarded,
    * so it doesn't make sense to mark EOT and ALLOCATE at the same time.
    */
   assert(!(allocate && (flags & BRW_URB_WRITE_EOT)));

   /* Copy the vertex from vertn into m1..mN+1:
    */
   brw_copy_from_indirect(p, brw_message_reg(1), vert, c->nr_regs);

   /* Overwrite PrimType and PrimStart in the message header, for
    * each vertex in turn:
    */
   brw_MOV(p, get_element_ud(c->reg.R0, 2), brw_imm_ud(header));


   /* Send each vertex as a seperate write to the urb.  This
    * is different to the concept in brw_sf_emit.c, where
    * subsequent writes are used to build up a single urb
    * entry.  Each of these writes instantiates a seperate
    * urb entry - (I think... what about 'allocate'?)
    */
   brw_urb_WRITE(p,
		 allocate ? c->reg.R0 : retype(brw_null_reg(), BRW_REGISTER_TYPE_UD),
		 0,
		 c->reg.R0,
                 flags,
		 c->nr_regs + 1, /* msg length */
		 allocate ? 1 : 0, /* response_length */
		 0,		/* urb offset */
		 BRW_URB_SWIZZLE_NONE);
}



void brw_clip_kill_thread(struct brw_clip_compile *c)
{
   struct brw_compile *p = &c->func;

   brw_clip_ff_sync(c);
   /* Send an empty message to kill the thread and release any
    * allocated urb entry:
    */
   brw_urb_WRITE(p,
		 retype(brw_null_reg(), BRW_REGISTER_TYPE_UD),
		 0,
		 c->reg.R0,
                 BRW_URB_WRITE_UNUSED | BRW_URB_WRITE_EOT_COMPLETE,
		 1, 		/* msg len */
		 0, 		/* response len */
		 0,
		 BRW_URB_SWIZZLE_NONE);
}




struct brw_reg brw_clip_plane0_address( struct brw_clip_compile *c )
{
   return brw_address(c->reg.fixed_planes);
}


struct brw_reg brw_clip_plane_stride( struct brw_clip_compile *c )
{
   if (c->key.nr_userclip) {
      return brw_imm_uw(16);
   }
   else {
      return brw_imm_uw(4);
   }
}


/* Distribute flatshaded attributes from provoking vertex prior to
 * clipping.
 */
void brw_clip_copy_flatshaded_attributes( struct brw_clip_compile *c,
			   GLuint to, GLuint from )
{
   struct brw_compile *p = &c->func;

   for (int i = 0; i < c->vue_map.num_slots; i++) {
      if (c->key.interpolation_mode.mode[i] == INTERP_QUALIFIER_FLAT) {
         brw_MOV(p,
                 byte_offset(c->reg.vertex[to], brw_vue_slot_to_offset(i)),
                 byte_offset(c->reg.vertex[from], brw_vue_slot_to_offset(i)));
      }
   }
}



void brw_clip_init_clipmask( struct brw_clip_compile *c )
{
   struct brw_compile *p = &c->func;
   struct brw_reg incoming = get_element_ud(c->reg.R0, 2);
   struct brw_context *brw = p->brw;

   /* Shift so that lowest outcode bit is rightmost:
    */
   brw_SHR(p, c->reg.planemask, incoming, brw_imm_ud(26));

   if (c->key.nr_userclip) {
      struct brw_reg tmp = retype(vec1(get_tmp(c)), BRW_REGISTER_TYPE_UD);

      /* Rearrange userclip outcodes so that they come directly after
       * the fixed plane bits.
       */
      if (brw->gen == 5 || brw->is_g4x)
         brw_AND(p, tmp, incoming, brw_imm_ud(0xff<<14));
      else
         brw_AND(p, tmp, incoming, brw_imm_ud(0x3f<<14));

      brw_SHR(p, tmp, tmp, brw_imm_ud(8));
      brw_OR(p, c->reg.planemask, c->reg.planemask, tmp);

      release_tmp(c, tmp);
   }
}

void brw_clip_ff_sync(struct brw_clip_compile *c)
{
    struct brw_compile *p = &c->func;
    struct brw_context *brw = p->brw;

    if (brw->gen == 5) {
        brw_AND(p, brw_null_reg(), c->reg.ff_sync, brw_imm_ud(0x1));
        brw_inst_set_cond_modifier(brw, brw_last_inst, BRW_CONDITIONAL_Z);
        brw_IF(p, BRW_EXECUTE_1);
        {
            brw_OR(p, c->reg.ff_sync, c->reg.ff_sync, brw_imm_ud(0x1));
            brw_ff_sync(p,
			c->reg.R0,
			0,
			c->reg.R0,
			1, /* allocate */
			1, /* response length */
			0 /* eot */);
        }
        brw_ENDIF(p);
        brw_set_default_predicate_control(p, BRW_PREDICATE_NONE);
    }
}

void brw_clip_init_ff_sync(struct brw_clip_compile *c)
{
    struct brw_context *brw = c->func.brw;

    if (brw->gen == 5) {
	struct brw_compile *p = &c->func;

        brw_MOV(p, c->reg.ff_sync, brw_imm_ud(0));
    }
}