aboutsummaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/svga/svga_state_rss.c
blob: 786839f705189f0a056391142d256294f1c86e35 (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
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
/**********************************************************
 * 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 "pipe/p_defines.h"
#include "util/u_bitmask.h"
#include "util/u_format.h"
#include "util/u_inlines.h"
#include "util/u_memory.h"
#include "util/u_math.h"
#include "util/u_memory.h"

#include "svga_context.h"
#include "svga_screen.h"
#include "svga_state.h"
#include "svga_cmd.h"
#include "svga_format.h"
#include "svga_shader.h"


struct rs_queue {
   unsigned rs_count;
   SVGA3dRenderState rs[SVGA3D_RS_MAX];
};


#define EMIT_RS(svga, value, token)                       \
do {                                                            \
   STATIC_ASSERT(SVGA3D_RS_##token < ARRAY_SIZE(svga->state.hw_draw.rs)); \
   if (svga->state.hw_draw.rs[SVGA3D_RS_##token] != value) {    \
      svga_queue_rs(&queue, SVGA3D_RS_##token, value);          \
      svga->state.hw_draw.rs[SVGA3D_RS_##token] = value;        \
   }                                                            \
} while (0)

#define EMIT_RS_FLOAT(svga, fvalue, token)                \
do {                                                            \
   unsigned value = fui(fvalue);                                \
   STATIC_ASSERT(SVGA3D_RS_##token < ARRAY_SIZE(svga->state.hw_draw.rs)); \
   if (svga->state.hw_draw.rs[SVGA3D_RS_##token] != value) {    \
      svga_queue_rs(&queue, SVGA3D_RS_##token, value);          \
      svga->state.hw_draw.rs[SVGA3D_RS_##token] = value;        \
   }                                                            \
} while (0)


static inline void
svga_queue_rs(struct rs_queue *q, unsigned rss, unsigned value)
{
   assert(q->rs_count < ARRAY_SIZE(q->rs));
   q->rs[q->rs_count].state = rss;
   q->rs[q->rs_count].uintValue = value;
   q->rs_count++;
}


static unsigned
translate_fill_mode(unsigned fill)
{
   switch (fill) {
   case PIPE_POLYGON_MODE_POINT:
      return SVGA3D_FILLMODE_POINT;
   case PIPE_POLYGON_MODE_LINE:
      return SVGA3D_FILLMODE_LINE;
   case PIPE_POLYGON_MODE_FILL:
      return SVGA3D_FILLMODE_FILL;
   default:
      assert(!"Bad fill mode");
      return SVGA3D_FILLMODE_FILL;
   }
}


/* Compare old and new render states and emit differences between them
 * to hardware.  Simplest implementation would be to emit the whole of
 * the "to" state.
 */
static enum pipe_error
emit_rss_vgpu9(struct svga_context *svga, unsigned dirty)
{
   struct svga_screen *screen = svga_screen(svga->pipe.screen);
   struct rs_queue queue;
   float point_size_min;

   queue.rs_count = 0;

   if (dirty & (SVGA_NEW_BLEND | SVGA_NEW_BLEND_COLOR)) {
      const struct svga_blend_state *curr = svga->curr.blend;

      EMIT_RS(svga, curr->rt[0].writemask, COLORWRITEENABLE);
      EMIT_RS(svga, curr->rt[0].blend_enable, BLENDENABLE);

      if (curr->rt[0].blend_enable) {
         EMIT_RS(svga, curr->rt[0].srcblend, SRCBLEND);
         EMIT_RS(svga, curr->rt[0].dstblend, DSTBLEND);
         EMIT_RS(svga, curr->rt[0].blendeq, BLENDEQUATION);

         EMIT_RS(svga, curr->rt[0].separate_alpha_blend_enable,
                  SEPARATEALPHABLENDENABLE);

         if (curr->rt[0].separate_alpha_blend_enable) {
            EMIT_RS(svga, curr->rt[0].srcblend_alpha, SRCBLENDALPHA);
            EMIT_RS(svga, curr->rt[0].dstblend_alpha, DSTBLENDALPHA);
            EMIT_RS(svga, curr->rt[0].blendeq_alpha, BLENDEQUATIONALPHA);
         }
      }
   }

   if (dirty & SVGA_NEW_BLEND_COLOR) {
      uint32 color;
      uint32 r = float_to_ubyte(svga->curr.blend_color.color[0]);
      uint32 g = float_to_ubyte(svga->curr.blend_color.color[1]);
      uint32 b = float_to_ubyte(svga->curr.blend_color.color[2]);
      uint32 a = float_to_ubyte(svga->curr.blend_color.color[3]);

      color = (a << 24) | (r << 16) | (g << 8) | b;

      EMIT_RS(svga, color, BLENDCOLOR);
   }

   if (dirty & (SVGA_NEW_DEPTH_STENCIL_ALPHA | SVGA_NEW_RAST)) {
      const struct svga_depth_stencil_state *curr = svga->curr.depth;
      const struct svga_rasterizer_state *rast = svga->curr.rast;

      if (!curr->stencil[0].enabled) {
         /* Stencil disabled
          */
         EMIT_RS(svga, FALSE, STENCILENABLE);
         EMIT_RS(svga, FALSE, STENCILENABLE2SIDED);
      }
      else if (curr->stencil[0].enabled && !curr->stencil[1].enabled) {
         /* Regular stencil
          */
         EMIT_RS(svga, TRUE, STENCILENABLE);
         EMIT_RS(svga, FALSE, STENCILENABLE2SIDED);

         EMIT_RS(svga, curr->stencil[0].func,  STENCILFUNC);
         EMIT_RS(svga, curr->stencil[0].fail,  STENCILFAIL);
         EMIT_RS(svga, curr->stencil[0].zfail, STENCILZFAIL);
         EMIT_RS(svga, curr->stencil[0].pass,  STENCILPASS);

         EMIT_RS(svga, curr->stencil_mask, STENCILMASK);
         EMIT_RS(svga, curr->stencil_writemask, STENCILWRITEMASK);
      }
      else {
         int cw, ccw;

         /* Hardware frontwinding is always CW, so if ours is also CW,
          * then our definition of front face agrees with hardware.
          * Otherwise need to flip.
          */
         if (rast->templ.front_ccw) {
            ccw = 0;
            cw = 1;
         }
         else {
            ccw = 1;
            cw = 0;
         }

         /* Twoside stencil
          */
         EMIT_RS(svga, TRUE, STENCILENABLE);
         EMIT_RS(svga, TRUE, STENCILENABLE2SIDED);

         EMIT_RS(svga, curr->stencil[cw].func,  STENCILFUNC);
         EMIT_RS(svga, curr->stencil[cw].fail,  STENCILFAIL);
         EMIT_RS(svga, curr->stencil[cw].zfail, STENCILZFAIL);
         EMIT_RS(svga, curr->stencil[cw].pass,  STENCILPASS);

         EMIT_RS(svga, curr->stencil[ccw].func,  CCWSTENCILFUNC);
         EMIT_RS(svga, curr->stencil[ccw].fail,  CCWSTENCILFAIL);
         EMIT_RS(svga, curr->stencil[ccw].zfail, CCWSTENCILZFAIL);
         EMIT_RS(svga, curr->stencil[ccw].pass,  CCWSTENCILPASS);

         EMIT_RS(svga, curr->stencil_mask, STENCILMASK);
         EMIT_RS(svga, curr->stencil_writemask, STENCILWRITEMASK);
      }

      EMIT_RS(svga, curr->zenable, ZENABLE);
      if (curr->zenable) {
         EMIT_RS(svga, curr->zfunc, ZFUNC);
         EMIT_RS(svga, curr->zwriteenable, ZWRITEENABLE);
      }

      EMIT_RS(svga, curr->alphatestenable, ALPHATESTENABLE);
      if (curr->alphatestenable) {
         EMIT_RS(svga, curr->alphafunc, ALPHAFUNC);
         EMIT_RS_FLOAT(svga, curr->alpharef, ALPHAREF);
      }
   }

   if (dirty & SVGA_NEW_STENCIL_REF) {
      EMIT_RS(svga, svga->curr.stencil_ref.ref_value[0], STENCILREF);
   }

   if (dirty & (SVGA_NEW_RAST | SVGA_NEW_NEED_PIPELINE)) {
      const struct svga_rasterizer_state *curr = svga->curr.rast;
      unsigned cullmode = curr->cullmode;

      /* Shademode: still need to rearrange index list to move
       * flat-shading PV first vertex.
       */
      EMIT_RS(svga, curr->shademode, SHADEMODE);

      EMIT_RS(svga, translate_fill_mode(curr->hw_fillmode), FILLMODE);

      /* Don't do culling while the software pipeline is active.  It
       * does it for us, and additionally introduces potentially
       * back-facing triangles.
       */
      if (svga->state.sw.need_pipeline)
         cullmode = SVGA3D_FACE_NONE;

      point_size_min = util_get_min_point_size(&curr->templ);

      EMIT_RS(svga, cullmode, CULLMODE);
      EMIT_RS(svga, curr->scissortestenable, SCISSORTESTENABLE);
      EMIT_RS(svga, curr->multisampleantialias, MULTISAMPLEANTIALIAS);
      EMIT_RS(svga, curr->lastpixel, LASTPIXEL);
      EMIT_RS_FLOAT(svga, curr->pointsize, POINTSIZE);
      EMIT_RS_FLOAT(svga, point_size_min, POINTSIZEMIN);
      EMIT_RS_FLOAT(svga, screen->maxPointSize, POINTSIZEMAX);
      EMIT_RS(svga, curr->pointsprite, POINTSPRITEENABLE);

      /* Emit line state, when the device understands it */
      if (screen->haveLineStipple)
         EMIT_RS(svga, curr->linepattern, LINEPATTERN);
      if (screen->haveLineSmooth)
         EMIT_RS(svga, curr->antialiasedlineenable, ANTIALIASEDLINEENABLE);
      if (screen->maxLineWidth > 1.0F)
         EMIT_RS_FLOAT(svga, curr->linewidth, LINEWIDTH);
   }

   if (dirty & (SVGA_NEW_RAST |
                SVGA_NEW_FRAME_BUFFER |
                SVGA_NEW_NEED_PIPELINE)) {
      const struct svga_rasterizer_state *curr = svga->curr.rast;
      float slope = 0.0;
      float bias  = 0.0;

      /* Need to modify depth bias according to bound depthbuffer
       * format.  Don't do hardware depthbias while the software
       * pipeline is active.
       */
      if (!svga->state.sw.need_pipeline &&
          svga->curr.framebuffer.zsbuf)
      {
         slope = curr->slopescaledepthbias;
         bias  = svga->curr.depthscale * curr->depthbias;
      }

      EMIT_RS_FLOAT(svga, slope, SLOPESCALEDEPTHBIAS);
      EMIT_RS_FLOAT(svga, bias, DEPTHBIAS);
   }

   if (dirty & SVGA_NEW_FRAME_BUFFER) {
      /* XXX: we only look at the first color buffer's sRGB state */
      float gamma = 1.0f;
      if (svga->curr.framebuffer.cbufs[0] &&
          util_format_is_srgb(svga->curr.framebuffer.cbufs[0]->format)) {
         gamma = 2.2f;
      }
      EMIT_RS_FLOAT(svga, gamma, OUTPUTGAMMA);
   }

   if (dirty & SVGA_NEW_RAST) {
      /* bitmask of the enabled clip planes */
      unsigned enabled = svga->curr.rast->templ.clip_plane_enable;
      EMIT_RS(svga, enabled, CLIPPLANEENABLE);
   }

   if (queue.rs_count) {
      SVGA3dRenderState *rs;

      if (SVGA3D_BeginSetRenderState(svga->swc, &rs, queue.rs_count)
          != PIPE_OK) {
         /* XXX: need to poison cached hardware state on failure to ensure
          * dirty state gets re-emitted.  Fix this by re-instating partial
          * FIFOCommit command and only updating cached hw state once the
          * initial allocation has succeeded.
          */
         memset(svga->state.hw_draw.rs, 0xcd, sizeof(svga->state.hw_draw.rs));

         return PIPE_ERROR_OUT_OF_MEMORY;
      }

      memcpy(rs, queue.rs, queue.rs_count * sizeof queue.rs[0]);

      SVGA_FIFOCommitAll(svga->swc);
   }

   return PIPE_OK;
}


/** Returns a non-culling rasterizer state object to be used with
 *  point sprite.
 */
static struct svga_rasterizer_state *
get_no_cull_rasterizer_state(struct svga_context *svga)
{
   const struct svga_rasterizer_state *r = svga->curr.rast;
   unsigned int aa_point = r->templ.point_smooth;

   if (!svga->rasterizer_no_cull[aa_point]) {
      struct pipe_rasterizer_state rast;

      memset(&rast, 0, sizeof(rast));
      rast.flatshade = 1;
      rast.front_ccw = 1;
      rast.point_smooth = r->templ.point_smooth;

      /* All rasterizer states have the same half_pixel_center,
       * bottom_edge_rule and clip_halfz values since they are
       * constant for a context. If we ever implement
       * GL_ARB_clip_control, the clip_halfz field would have to be observed.
       */
      rast.half_pixel_center = r->templ.half_pixel_center;
      rast.bottom_edge_rule = r->templ.bottom_edge_rule;
      rast.clip_halfz = r->templ.clip_halfz;

      svga->rasterizer_no_cull[aa_point] =
               svga->pipe.create_rasterizer_state(&svga->pipe, &rast);
   }
   return svga->rasterizer_no_cull[aa_point];
}


/** Returns a depth stencil state object with depth and stencil test disabled.
 */
static struct svga_depth_stencil_state *
get_no_depth_stencil_test_state(struct svga_context *svga)
{
   if (!svga->depthstencil_disable) {
      struct pipe_depth_stencil_alpha_state ds = {{0}};
      svga->depthstencil_disable =
         svga->pipe.create_depth_stencil_alpha_state(&svga->pipe, &ds);
   }
   return svga->depthstencil_disable;
}


static enum pipe_error
emit_rss_vgpu10(struct svga_context *svga, unsigned dirty)
{
   enum pipe_error ret = PIPE_OK;

   svga_hwtnl_flush_retry(svga);

   if (dirty & (SVGA_NEW_BLEND | SVGA_NEW_BLEND_COLOR)) {
      const struct svga_blend_state *curr;
      float blend_factor[4];

      if (svga_has_any_integer_cbufs(svga)) {
         /* Blending is not supported in integer-valued render targets. */
         curr = svga->noop_blend;
         blend_factor[0] =
         blend_factor[1] =
         blend_factor[2] =
         blend_factor[3] = 0;
      }
      else {
         curr = svga->curr.blend;

         if (curr->blend_color_alpha) {
            blend_factor[0] =
            blend_factor[1] =
            blend_factor[2] =
            blend_factor[3] = svga->curr.blend_color.color[3];
         }
         else {
            blend_factor[0] = svga->curr.blend_color.color[0];
            blend_factor[1] = svga->curr.blend_color.color[1];
            blend_factor[2] = svga->curr.blend_color.color[2];
            blend_factor[3] = svga->curr.blend_color.color[3];
         }
      }

      /* Set/bind the blend state object */
      if (svga->state.hw_draw.blend_id != curr->id ||
          svga->state.hw_draw.blend_factor[0] != blend_factor[0] ||
          svga->state.hw_draw.blend_factor[1] != blend_factor[1] ||
          svga->state.hw_draw.blend_factor[2] != blend_factor[2] ||
          svga->state.hw_draw.blend_factor[3] != blend_factor[3] ||
          svga->state.hw_draw.blend_sample_mask != svga->curr.sample_mask) {
         ret = SVGA3D_vgpu10_SetBlendState(svga->swc, curr->id,
                                           blend_factor,
                                           svga->curr.sample_mask);
         if (ret != PIPE_OK)
            return ret;

         svga->state.hw_draw.blend_id = curr->id;
         svga->state.hw_draw.blend_factor[0] = blend_factor[0];
         svga->state.hw_draw.blend_factor[1] = blend_factor[1];
         svga->state.hw_draw.blend_factor[2] = blend_factor[2];
         svga->state.hw_draw.blend_factor[3] = blend_factor[3];
         svga->state.hw_draw.blend_sample_mask = svga->curr.sample_mask;
      }
   }

   if (svga->disable_rasterizer) {
      if (!svga->state.hw_draw.rasterizer_discard) {
         struct svga_depth_stencil_state *ds;

         /* If rasterization is to be disabled, disable depth and stencil
          * testing as well.
          */
         ds = get_no_depth_stencil_test_state(svga);
         if (ds->id != svga->state.hw_draw.depth_stencil_id) {
            ret = SVGA3D_vgpu10_SetDepthStencilState(svga->swc, ds->id, 0);
            if (ret != PIPE_OK)
               return ret;

            svga->state.hw_draw.depth_stencil_id = ds->id;
            svga->state.hw_draw.stencil_ref = 0;
         }
         svga->state.hw_draw.rasterizer_discard = TRUE;
      }
   } else {
      if ((dirty & (SVGA_NEW_DEPTH_STENCIL_ALPHA | SVGA_NEW_STENCIL_REF)) ||
          svga->state.hw_draw.rasterizer_discard) {
         const struct svga_depth_stencil_state *curr = svga->curr.depth;
         unsigned curr_ref = svga->curr.stencil_ref.ref_value[0];

         if (curr->id != svga->state.hw_draw.depth_stencil_id ||
             curr_ref != svga->state.hw_draw.stencil_ref) {
            /* Set/bind the depth/stencil state object */
            ret = SVGA3D_vgpu10_SetDepthStencilState(svga->swc, curr->id,
                                                     curr_ref);
            if (ret != PIPE_OK)
               return ret;

            svga->state.hw_draw.depth_stencil_id = curr->id;
            svga->state.hw_draw.stencil_ref = curr_ref;
         }
      }

      if (dirty & (SVGA_NEW_REDUCED_PRIMITIVE | SVGA_NEW_RAST)) {
         const struct svga_rasterizer_state *rast;

         if (svga->curr.reduced_prim == PIPE_PRIM_POINTS &&
             svga->curr.gs && svga->curr.gs->wide_point) {

            /* If we are drawing a point sprite, we will need to
             * bind a non-culling rasterizer state object
             */
            rast = get_no_cull_rasterizer_state(svga);
         }
         else {
            rast = svga->curr.rast;
         }

         if (svga->state.hw_draw.rasterizer_id != rast->id) {
            /* Set/bind the rasterizer state object */
            ret = SVGA3D_vgpu10_SetRasterizerState(svga->swc, rast->id);
            if (ret != PIPE_OK)
               return ret;
            svga->state.hw_draw.rasterizer_id = rast->id;
         }
      }
      svga->state.hw_draw.rasterizer_discard = FALSE;
   }
   return PIPE_OK;
}


static enum pipe_error
emit_rss(struct svga_context *svga, unsigned dirty)
{
   if (svga_have_vgpu10(svga)) {
      return emit_rss_vgpu10(svga, dirty);
   }
   else {
      return emit_rss_vgpu9(svga, dirty);
   }
}


struct svga_tracked_state svga_hw_rss =
{
   "hw rss state",

   (SVGA_NEW_BLEND |
    SVGA_NEW_BLEND_COLOR |
    SVGA_NEW_DEPTH_STENCIL_ALPHA |
    SVGA_NEW_STENCIL_REF |
    SVGA_NEW_RAST |
    SVGA_NEW_FRAME_BUFFER |
    SVGA_NEW_NEED_PIPELINE |
    SVGA_NEW_FS |
    SVGA_NEW_REDUCED_PRIMITIVE),

   emit_rss
};