summaryrefslogtreecommitdiff
path: root/src/gallium/drivers/cell/spu/spu_render.c
blob: 6df59abd36d68b0ea8d6c901c7393889b5229e1a (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
/**************************************************************************
 * 
 * Copyright 2008 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.
 * 
 **************************************************************************/


#include <stdio.h>
#include <libmisc.h>
#include <spu_mfcio.h>

#include "spu_main.h"
#include "spu_render.h"
#include "spu_tri.h"
#include "spu_tile.h"
#include "cell/common.h"



/**
 * Given a rendering command's bounding box (in pixels) compute the
 * location of the corresponding screen tile bounding box.
 */
static INLINE void
tile_bounding_box(const struct cell_command_render *render,
                  uint *txmin, uint *tymin,
                  uint *box_num_tiles, uint *box_width_tiles)
{
#if 0
   /* Debug: full-window bounding box */
   uint txmax = spu.fb.width_tiles - 1;
   uint tymax = spu.fb.height_tiles - 1;
   *txmin = 0;
   *tymin = 0;
   *box_num_tiles = spu.fb.width_tiles * spu.fb.height_tiles;
   *box_width_tiles = spu.fb.width_tiles;
   (void) render;
   (void) txmax;
   (void) tymax;
#else
   uint txmax, tymax, box_height_tiles;

   *txmin = (uint) render->xmin / TILE_SIZE;
   *tymin = (uint) render->ymin / TILE_SIZE;
   txmax = (uint) render->xmax / TILE_SIZE;
   tymax = (uint) render->ymax / TILE_SIZE;
   if (txmax >= spu.fb.width_tiles)
      txmax = spu.fb.width_tiles-1;
   if (tymax >= spu.fb.height_tiles)
      tymax = spu.fb.height_tiles-1;
   *box_width_tiles = txmax - *txmin + 1;
   box_height_tiles = tymax - *tymin + 1;
   *box_num_tiles = *box_width_tiles * box_height_tiles;
#endif
#if 0
   printf("SPU %u: bounds: %g, %g  ...  %g, %g\n", spu.init.id,
          render->xmin, render->ymin, render->xmax, render->ymax);
   printf("SPU %u: tiles:  %u, %u .. %u, %u\n",
           spu.init.id, *txmin, *tymin, txmax, tymax);
   ASSERT(render->xmin <= render->xmax);
   ASSERT(render->ymin <= render->ymax);
#endif
}


/** Check if the tile at (tx,ty) belongs to this SPU */
static INLINE boolean
my_tile(uint tx, uint ty)
{
   return (spu.fb.width_tiles * ty + tx) % spu.init.num_spus == spu.init.id;
}


/**
 * Start fetching non-clear color/Z tiles from main memory
 */
static INLINE void
get_cz_tiles(uint tx, uint ty)
{
   if (spu.read_depth) {
      if (spu.cur_ztile_status != TILE_STATUS_CLEAR) {
         //printf("SPU %u: getting Z tile %u, %u\n", spu.init.id, tx, ty);
         get_tile(tx, ty, &spu.ztile, TAG_READ_TILE_Z, 1);
         spu.cur_ztile_status = TILE_STATUS_GETTING;
      }
   }

   if (spu.cur_ctile_status != TILE_STATUS_CLEAR) {
      //printf("SPU %u: getting C tile %u, %u\n", spu.init.id, tx, ty);
      get_tile(tx, ty, &spu.ctile, TAG_READ_TILE_COLOR, 0);
      spu.cur_ctile_status = TILE_STATUS_GETTING;
   }
}


/**
 * Start putting dirty color/Z tiles back to main memory
 */
static INLINE void
put_cz_tiles(uint tx, uint ty)
{
   if (spu.cur_ztile_status == TILE_STATUS_DIRTY) {
      /* tile was modified and needs to be written back */
      //printf("SPU %u: put dirty Z tile %u, %u\n", spu.init.id, tx, ty);
      put_tile(tx, ty, &spu.ztile, TAG_WRITE_TILE_Z, 1);
      spu.cur_ztile_status = TILE_STATUS_DEFINED;
   }
   else if (spu.cur_ztile_status == TILE_STATUS_GETTING) {
      /* tile was never used */
      spu.cur_ztile_status = TILE_STATUS_DEFINED;
      //printf("SPU %u: put getting Z tile %u, %u\n", spu.init.id, tx, ty);
   }

   if (spu.cur_ctile_status == TILE_STATUS_DIRTY) {
      /* tile was modified and needs to be written back */
      //printf("SPU %u: put dirty C tile %u, %u\n", spu.init.id, tx, ty);
      put_tile(tx, ty, &spu.ctile, TAG_WRITE_TILE_COLOR, 0);
      spu.cur_ctile_status = TILE_STATUS_DEFINED;
   }
   else if (spu.cur_ctile_status == TILE_STATUS_GETTING) {
      /* tile was never used */
      spu.cur_ctile_status = TILE_STATUS_DEFINED;
      //printf("SPU %u: put getting C tile %u, %u\n", spu.init.id, tx, ty);
   }
}


/**
 * Wait for 'put' of color/z tiles to complete.
 */
static INLINE void
wait_put_cz_tiles(void)
{
   wait_on_mask(1 << TAG_WRITE_TILE_COLOR);
   if (spu.read_depth) {
      wait_on_mask(1 << TAG_WRITE_TILE_Z);
   }
}


/**
 * Render primitives
 * \param pos_incr  returns value indicating how may words to skip after
 *                  this command in the batch buffer
 */
void
cmd_render(const struct cell_command_render *render, uint *pos_incr)
{
   /* we'll DMA into these buffers */
   ubyte vertex_data[CELL_BUFFER_SIZE] ALIGN16_ATTRIB;
   const uint vertex_size = render->vertex_size; /* in bytes */
   /*const*/ uint total_vertex_bytes = render->num_verts * vertex_size;
   uint index_bytes;
   const ubyte *vertices;
   const ushort *indexes;
   uint i, j;


   if (Debug) {
      printf("SPU %u: RENDER prim %u, num_vert=%u  num_ind=%u  "
             "inline_vert=%u\n",
             spu.init.id,
             render->prim_type,
             render->num_verts,
             render->num_indexes,
             render->inline_verts);

      /*
      printf("       bound: %g, %g .. %g, %g\n",
             render->xmin, render->ymin, render->xmax, render->ymax);
      */
   }

   ASSERT(sizeof(*render) % 4 == 0);
   ASSERT(total_vertex_bytes % 16 == 0);
   ASSERT(render->prim_type == PIPE_PRIM_TRIANGLES);
   ASSERT(render->num_indexes % 3 == 0);


   /* indexes are right after the render command in the batch buffer */
   indexes = (const ushort *) (render + 1);
   index_bytes = ROUNDUP8(render->num_indexes * 2);
   *pos_incr = index_bytes / 8 + sizeof(*render) / 8;


   if (render->inline_verts) {
      /* Vertices are after indexes in batch buffer at next 16-byte addr */
      vertices = (const ubyte *) render + (*pos_incr * 8);
      vertices = (const ubyte *) align_pointer((void *) vertices, 16);
      ASSERT_ALIGN16(vertices);
      *pos_incr = ((vertices + total_vertex_bytes) - (ubyte *) render) / 8;
   }
   else {
      /* Begin DMA fetch of vertex buffer */
      ubyte *src = spu.init.buffers[render->vertex_buf];
      ubyte *dest = vertex_data;

      /* skip vertex data we won't use */
#if 01
      src += render->min_index * vertex_size;
      dest += render->min_index * vertex_size;
      total_vertex_bytes -= render->min_index * vertex_size;
#endif
      ASSERT(total_vertex_bytes % 16 == 0);
      ASSERT_ALIGN16(dest);
      ASSERT_ALIGN16(src);

      mfc_get(dest,   /* in vertex_data[] array */
              (unsigned int) src,  /* src in main memory */
              total_vertex_bytes,  /* size */
              TAG_VERTEX_BUFFER,
              0, /* tid */
              0  /* rid */);

      vertices = vertex_data;

      wait_on_mask(1 << TAG_VERTEX_BUFFER);
   }


   /**
    ** find tiles which intersect the prim bounding box
    **/
   uint txmin, tymin, box_width_tiles, box_num_tiles;
   tile_bounding_box(render, &txmin, &tymin,
                     &box_num_tiles, &box_width_tiles);


   /* make sure any pending clears have completed */
   wait_on_mask(1 << TAG_SURFACE_CLEAR); /* XXX temporary */


   /**
    ** loop over tiles, rendering tris
    **/
   for (i = 0; i < box_num_tiles; i++) {
      const uint tx = txmin + i % box_width_tiles;
      const uint ty = tymin + i / box_width_tiles;

      ASSERT(tx < spu.fb.width_tiles);
      ASSERT(ty < spu.fb.height_tiles);

      if (!my_tile(tx, ty))
         continue;

      spu.cur_ctile_status = spu.ctile_status[ty][tx];
      spu.cur_ztile_status = spu.ztile_status[ty][tx];

      get_cz_tiles(tx, ty);

      uint drawn = 0;

      /* loop over tris */
      for (j = 0; j < render->num_indexes; j += 3) {
         const float *v0, *v1, *v2;

         v0 = (const float *) (vertices + indexes[j+0] * vertex_size);
         v1 = (const float *) (vertices + indexes[j+1] * vertex_size);
         v2 = (const float *) (vertices + indexes[j+2] * vertex_size);

         drawn += tri_draw(v0, v1, v2, tx, ty);
      }

      //printf("SPU %u: drew %u of %u\n", spu.init.id, drawn, render->num_indexes/3);

      /* write color/z tiles back to main framebuffer, if dirtied */
      put_cz_tiles(tx, ty);

      wait_put_cz_tiles(); /* XXX seems unnecessary... */

      spu.ctile_status[ty][tx] = spu.cur_ctile_status;
      spu.ztile_status[ty][tx] = spu.cur_ztile_status;
   }

   if (Debug)
      printf("SPU %u: RENDER done\n",
             spu.init.id);
}