summaryrefslogtreecommitdiff
path: root/src/gallium/drivers/svga/svga_state_fs.c
blob: ec2886348b9c24aaa9e8f8b7131bf8679ef94c1a (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
/**********************************************************
 * 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_inlines.h"
#include "pipe/p_defines.h"
#include "util/u_math.h"
#include "util/u_bitmask.h"

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

#include "svga_hw_reg.h"



static INLINE int compare_fs_keys( const struct svga_fs_compile_key *a,
                                   const struct svga_fs_compile_key *b )
{
   unsigned keysize_a = svga_fs_key_size( a );
   unsigned keysize_b = svga_fs_key_size( b );

   if (keysize_a != keysize_b) {
      return (int)(keysize_a - keysize_b);
   }
   return memcmp( a, b, keysize_a );
}


static struct svga_shader_result *search_fs_key( struct svga_fragment_shader *fs,
                                                 const struct svga_fs_compile_key *key )
{
   struct svga_shader_result *result = fs->base.results;

   assert(key);

   for ( ; result; result = result->next) {
      if (compare_fs_keys( key, &result->key.fkey ) == 0)
         return result;
   }
   
   return NULL;
}


static enum pipe_error compile_fs( struct svga_context *svga,
                                   struct svga_fragment_shader *fs,
                                   const struct svga_fs_compile_key *key,
                                   struct svga_shader_result **out_result )
{
   struct svga_shader_result *result;
   enum pipe_error ret = PIPE_ERROR;

   result = svga_translate_fragment_program( fs, key );
   if (result == NULL) {
      ret = PIPE_ERROR_OUT_OF_MEMORY;
      goto fail;
   }

   result->id = util_bitmask_add(svga->fs_bm);
   if(result->id == UTIL_BITMASK_INVALID_INDEX) {
      ret = PIPE_ERROR_OUT_OF_MEMORY;
      goto fail;
   }

   ret = SVGA3D_DefineShader(svga->swc, 
                             result->id,
                             SVGA3D_SHADERTYPE_PS,
                             result->tokens, 
                             result->nr_tokens * sizeof result->tokens[0]);
   if (ret)
      goto fail;

   *out_result = result;
   result->next = fs->base.results;
   fs->base.results = result;
   return PIPE_OK;

fail:
   if (result) {
      if (result->id != UTIL_BITMASK_INVALID_INDEX)
         util_bitmask_clear( svga->fs_bm, result->id );
      svga_destroy_shader_result( result );
   }
   return ret;
}

/* The blend workaround for simulating logicop xor behaviour requires
 * that the incoming fragment color be white.  This change achieves
 * that by hooking up a hard-wired fragment shader that just emits
 * color 1,1,1,1
 *   
 * This is a slightly incomplete solution as it assumes that the
 * actual bound shader has no other effects beyond generating a
 * fragment color.  In particular shaders containing TEXKIL and/or
 * depth-write will not have the correct behaviour, nor will those
 * expecting to use alphatest.
 *   
 * These are avoidable issues, but they are not much worse than the
 * unavoidable ones associated with this technique, so it's not clear
 * how much effort should be expended trying to resolve them - the
 * ultimate result will still not be correct in most cases.
 *
 * Shader below was generated with:
 *   SVGA_DEBUG=tgsi ./mesa/progs/fp/fp-tri white.txt
 */
static int emit_white_fs( struct svga_context *svga )
{
   int ret = PIPE_ERROR;

   /* ps_3_0
    * def c0, 1.000000, 0.000000, 0.000000, 1.000000
    * mov oC0, c0.x
    * end
    */
   static const unsigned white_tokens[] = {
      0xffff0300,
      0x05000051,
      0xa00f0000,
      0x3f800000,
      0x00000000,
      0x00000000,
      0x3f800000,
      0x02000001,
      0x800f0800,
      0xa0000000,
      0x0000ffff,
   };

   assert(SVGA3D_INVALID_ID == UTIL_BITMASK_INVALID_INDEX);
   svga->state.white_fs_id = util_bitmask_add(svga->fs_bm);
   if(svga->state.white_fs_id == SVGA3D_INVALID_ID)
      goto no_fs_id;

   ret = SVGA3D_DefineShader(svga->swc, 
                             svga->state.white_fs_id,
                             SVGA3D_SHADERTYPE_PS,
                             white_tokens, 
                             sizeof(white_tokens));
   if (ret)
      goto no_definition;

   return 0;

no_definition:
   util_bitmask_clear(svga->fs_bm, svga->state.white_fs_id);
   svga->state.white_fs_id = SVGA3D_INVALID_ID;
no_fs_id:
   return ret;
}


/* SVGA_NEW_TEXTURE_BINDING
 * SVGA_NEW_RAST
 * SVGA_NEW_NEED_SWTNL
 * SVGA_NEW_SAMPLER
 */
static int make_fs_key( const struct svga_context *svga,
                        struct svga_fs_compile_key *key )
{
   int i;
   int idx = 0;

   memset(key, 0, sizeof *key);

   /* Only need fragment shader fixup for twoside lighting if doing
    * hwtnl.  Otherwise the draw module does the whole job for us.
    *
    * SVGA_NEW_SWTNL
    */
   if (!svga->state.sw.need_swtnl) {
      /* SVGA_NEW_RAST
       */
      key->light_twoside = svga->curr.rast->templ.light_twoside;
      key->front_cw = (svga->curr.rast->templ.front_winding == 
                       PIPE_WINDING_CW);
   }

   
   /* XXX: want to limit this to the textures that the shader actually
    * refers to.
    *
    * SVGA_NEW_TEXTURE_BINDING | SVGA_NEW_SAMPLER
    */
   for (i = 0; i < svga->curr.num_textures; i++) {
      if (svga->curr.texture[i]) {
         assert(svga->curr.sampler[i]);
         key->tex[i].texture_target = svga->curr.texture[i]->target;
         if (!svga->curr.sampler[i]->normalized_coords) {
            key->tex[i].width_height_idx = idx++;
            key->tex[i].unnormalized = TRUE;
            ++key->num_unnormalized_coords;
         }
      }
   }
   key->num_textures = svga->curr.num_textures;

   idx = 0;
   for (i = 0; i < svga->curr.num_samplers; ++i) {
      if (svga->curr.sampler[i]) {
         key->tex[i].compare_mode = svga->curr.sampler[i]->compare_mode;
         key->tex[i].compare_func = svga->curr.sampler[i]->compare_func;
      }
   }

   return 0;
}



static int emit_hw_fs( struct svga_context *svga,
                       unsigned dirty )
{
   struct svga_shader_result *result = NULL;
   unsigned id = SVGA3D_INVALID_ID;
   int ret = 0;

   /* SVGA_NEW_BLEND
    */
   if (svga->curr.blend->need_white_fragments) {
      if (svga->state.white_fs_id == SVGA3D_INVALID_ID) {
         ret = emit_white_fs( svga );
         if (ret)
            return ret;
      }
      id = svga->state.white_fs_id;
   }
   else {
      struct svga_fragment_shader *fs = svga->curr.fs;
      struct svga_fs_compile_key key;

      /* SVGA_NEW_TEXTURE_BINDING
       * SVGA_NEW_RAST
       * SVGA_NEW_NEED_SWTNL
       * SVGA_NEW_SAMPLER
       */
      ret = make_fs_key( svga, &key );
      if (ret)
         return ret;

      result = search_fs_key( fs, &key );
      if (!result) {
         ret = compile_fs( svga, fs, &key, &result );
         if (ret)
            return ret;
      }

      assert (result);
      id = result->id;
   }

   assert(id != SVGA3D_INVALID_ID);

   if (result != svga->state.hw_draw.fs) {
      ret = SVGA3D_SetShader(svga->swc,
                             SVGA3D_SHADERTYPE_PS,
                             id );
      if (ret)
         return ret;

      svga->dirty |= SVGA_NEW_FS_RESULT;
      svga->state.hw_draw.fs = result;      
   }

   return 0;
}

struct svga_tracked_state svga_hw_fs = 
{
   "fragment shader (hwtnl)",
   (SVGA_NEW_FS |
    SVGA_NEW_TEXTURE_BINDING |
    SVGA_NEW_NEED_SWTNL |
    SVGA_NEW_RAST |
    SVGA_NEW_SAMPLER |
    SVGA_NEW_BLEND),
   emit_hw_fs
};