aboutsummaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/i965/brw_pipe_vertex.c
blob: d6a840857ec0fd7b20e853847ccc683e624b210c (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
#include "brw_context.h"
#include "brw_defines.h"
#include "brw_structs.h"

#include "util/u_memory.h"
#include "util/u_format.h"


static unsigned brw_translate_surface_format( unsigned id )
{
   switch (id) {
   case PIPE_FORMAT_R64_FLOAT:
      return BRW_SURFACEFORMAT_R64_FLOAT;
   case PIPE_FORMAT_R64G64_FLOAT:
      return BRW_SURFACEFORMAT_R64G64_FLOAT;
   case PIPE_FORMAT_R64G64B64_FLOAT:
      return BRW_SURFACEFORMAT_R64G64B64_FLOAT;
   case PIPE_FORMAT_R64G64B64A64_FLOAT:
      return BRW_SURFACEFORMAT_R64G64B64A64_FLOAT;

   case PIPE_FORMAT_R32_FLOAT:
      return BRW_SURFACEFORMAT_R32_FLOAT;
   case PIPE_FORMAT_R32G32_FLOAT:
      return BRW_SURFACEFORMAT_R32G32_FLOAT;
   case PIPE_FORMAT_R32G32B32_FLOAT:
      return BRW_SURFACEFORMAT_R32G32B32_FLOAT;
   case PIPE_FORMAT_R32G32B32A32_FLOAT:
      return BRW_SURFACEFORMAT_R32G32B32A32_FLOAT;

   case PIPE_FORMAT_R32_UNORM:
      return BRW_SURFACEFORMAT_R32_UNORM;
   case PIPE_FORMAT_R32G32_UNORM:
      return BRW_SURFACEFORMAT_R32G32_UNORM;
   case PIPE_FORMAT_R32G32B32_UNORM:
      return BRW_SURFACEFORMAT_R32G32B32_UNORM;
   case PIPE_FORMAT_R32G32B32A32_UNORM:
      return BRW_SURFACEFORMAT_R32G32B32A32_UNORM;

   case PIPE_FORMAT_R32_USCALED:
      return BRW_SURFACEFORMAT_R32_USCALED;
   case PIPE_FORMAT_R32G32_USCALED:
      return BRW_SURFACEFORMAT_R32G32_USCALED;
   case PIPE_FORMAT_R32G32B32_USCALED:
      return BRW_SURFACEFORMAT_R32G32B32_USCALED;
   case PIPE_FORMAT_R32G32B32A32_USCALED:
      return BRW_SURFACEFORMAT_R32G32B32A32_USCALED;

   case PIPE_FORMAT_R32_SNORM:
      return BRW_SURFACEFORMAT_R32_SNORM;
   case PIPE_FORMAT_R32G32_SNORM:
      return BRW_SURFACEFORMAT_R32G32_SNORM;
   case PIPE_FORMAT_R32G32B32_SNORM:
      return BRW_SURFACEFORMAT_R32G32B32_SNORM;
   case PIPE_FORMAT_R32G32B32A32_SNORM:
      return BRW_SURFACEFORMAT_R32G32B32A32_SNORM;

   case PIPE_FORMAT_R32_SSCALED:
      return BRW_SURFACEFORMAT_R32_SSCALED;
   case PIPE_FORMAT_R32G32_SSCALED:
      return BRW_SURFACEFORMAT_R32G32_SSCALED;
   case PIPE_FORMAT_R32G32B32_SSCALED:
      return BRW_SURFACEFORMAT_R32G32B32_SSCALED;
   case PIPE_FORMAT_R32G32B32A32_SSCALED:
      return BRW_SURFACEFORMAT_R32G32B32A32_SSCALED;

   case PIPE_FORMAT_R16_UNORM:
      return BRW_SURFACEFORMAT_R16_UNORM;
   case PIPE_FORMAT_R16G16_UNORM:
      return BRW_SURFACEFORMAT_R16G16_UNORM;
   case PIPE_FORMAT_R16G16B16_UNORM:
      return BRW_SURFACEFORMAT_R16G16B16_UNORM;
   case PIPE_FORMAT_R16G16B16A16_UNORM:
      return BRW_SURFACEFORMAT_R16G16B16A16_UNORM;

   case PIPE_FORMAT_R16_USCALED:
      return BRW_SURFACEFORMAT_R16_USCALED;
   case PIPE_FORMAT_R16G16_USCALED:
      return BRW_SURFACEFORMAT_R16G16_USCALED;
   case PIPE_FORMAT_R16G16B16_USCALED:
      return BRW_SURFACEFORMAT_R16G16B16_USCALED;
   case PIPE_FORMAT_R16G16B16A16_USCALED:
      return BRW_SURFACEFORMAT_R16G16B16A16_USCALED;

   case PIPE_FORMAT_R16_SNORM:
      return BRW_SURFACEFORMAT_R16_SNORM;
   case PIPE_FORMAT_R16G16_SNORM:
      return BRW_SURFACEFORMAT_R16G16_SNORM;
   case PIPE_FORMAT_R16G16B16_SNORM:
      return BRW_SURFACEFORMAT_R16G16B16_SNORM;
   case PIPE_FORMAT_R16G16B16A16_SNORM:
      return BRW_SURFACEFORMAT_R16G16B16A16_SNORM;

   case PIPE_FORMAT_R16_SSCALED:
      return BRW_SURFACEFORMAT_R16_SSCALED;
   case PIPE_FORMAT_R16G16_SSCALED:
      return BRW_SURFACEFORMAT_R16G16_SSCALED;
   case PIPE_FORMAT_R16G16B16_SSCALED:
      return BRW_SURFACEFORMAT_R16G16B16_SSCALED;
   case PIPE_FORMAT_R16G16B16A16_SSCALED:
      return BRW_SURFACEFORMAT_R16G16B16A16_SSCALED;

   case PIPE_FORMAT_R8_UNORM:
      return BRW_SURFACEFORMAT_R8_UNORM;
   case PIPE_FORMAT_R8G8_UNORM:
      return BRW_SURFACEFORMAT_R8G8_UNORM;
   case PIPE_FORMAT_R8G8B8_UNORM:
      return BRW_SURFACEFORMAT_R8G8B8_UNORM;
   case PIPE_FORMAT_R8G8B8A8_UNORM:
      return BRW_SURFACEFORMAT_R8G8B8A8_UNORM;

   case PIPE_FORMAT_R8_USCALED:
      return BRW_SURFACEFORMAT_R8_USCALED;
   case PIPE_FORMAT_R8G8_USCALED:
      return BRW_SURFACEFORMAT_R8G8_USCALED;
   case PIPE_FORMAT_R8G8B8_USCALED:
      return BRW_SURFACEFORMAT_R8G8B8_USCALED;
   case PIPE_FORMAT_R8G8B8A8_USCALED:
      return BRW_SURFACEFORMAT_R8G8B8A8_USCALED;

   case PIPE_FORMAT_R8_SNORM:
      return BRW_SURFACEFORMAT_R8_SNORM;
   case PIPE_FORMAT_R8G8_SNORM:
      return BRW_SURFACEFORMAT_R8G8_SNORM;
   case PIPE_FORMAT_R8G8B8_SNORM:
      return BRW_SURFACEFORMAT_R8G8B8_SNORM;
   case PIPE_FORMAT_R8G8B8A8_SNORM:
      return BRW_SURFACEFORMAT_R8G8B8A8_SNORM;

   case PIPE_FORMAT_R8_SSCALED:
      return BRW_SURFACEFORMAT_R8_SSCALED;
   case PIPE_FORMAT_R8G8_SSCALED:
      return BRW_SURFACEFORMAT_R8G8_SSCALED;
   case PIPE_FORMAT_R8G8B8_SSCALED:
      return BRW_SURFACEFORMAT_R8G8B8_SSCALED;
   case PIPE_FORMAT_R8G8B8A8_SSCALED:
      return BRW_SURFACEFORMAT_R8G8B8A8_SSCALED;

   default:
      assert(0);
      return 0;
   }
}

static void brw_translate_vertex_elements(struct brw_context *brw,
                                          struct brw_vertex_element_packet *brw_velems,
                                          const struct pipe_vertex_element *attribs,
                                          unsigned count)
{
   unsigned i;

   /* If the VS doesn't read any inputs (calculating vertex position from
    * a state variable for some reason, for example), emit a single pad
    * VERTEX_ELEMENT struct and bail.
    *
    * The stale VB state stays in place, but they don't do anything unless
    * a VE loads from them.
    */
   brw_velems->header.opcode = CMD_VERTEX_ELEMENT;

   if (count == 0) {
      brw_velems->header.length = 1;
      brw_velems->ve[0].ve0.src_offset = 0;
      brw_velems->ve[0].ve0.src_format = BRW_SURFACEFORMAT_R32G32B32A32_FLOAT;
      brw_velems->ve[0].ve0.valid = 1;
      brw_velems->ve[0].ve0.vertex_buffer_index = 0;
      brw_velems->ve[0].ve1.dst_offset = 0;
      brw_velems->ve[0].ve1.vfcomponent0 = BRW_VE1_COMPONENT_STORE_0;
      brw_velems->ve[0].ve1.vfcomponent1 = BRW_VE1_COMPONENT_STORE_0;
      brw_velems->ve[0].ve1.vfcomponent2 = BRW_VE1_COMPONENT_STORE_0;
      brw_velems->ve[0].ve1.vfcomponent3 = BRW_VE1_COMPONENT_STORE_1_FLT;
      return;
   }


   /* Now emit vertex element (VEP) state packets.
    *
    */
   brw_velems->header.length = (1 + count * 2) - 2;
   for (i = 0; i < count; i++) {
      const struct pipe_vertex_element *input = &attribs[i];
      unsigned nr_components = util_format_get_nr_components(input->src_format);

      uint32_t format = brw_translate_surface_format( input->src_format );
      uint32_t comp0 = BRW_VE1_COMPONENT_STORE_SRC;
      uint32_t comp1 = BRW_VE1_COMPONENT_STORE_SRC;
      uint32_t comp2 = BRW_VE1_COMPONENT_STORE_SRC;
      uint32_t comp3 = BRW_VE1_COMPONENT_STORE_SRC;

      switch (nr_components) {
      case 0: comp0 = BRW_VE1_COMPONENT_STORE_0; /* fallthrough */
      case 1: comp1 = BRW_VE1_COMPONENT_STORE_0; /* fallthrough */
      case 2: comp2 = BRW_VE1_COMPONENT_STORE_0; /* fallthrough */
      case 3: comp3 = BRW_VE1_COMPONENT_STORE_1_FLT;
         break;
      }

      brw_velems->ve[i].ve0.src_offset = input->src_offset;
      brw_velems->ve[i].ve0.src_format = format;
      brw_velems->ve[i].ve0.valid = 1;
      brw_velems->ve[i].ve0.vertex_buffer_index = input->vertex_buffer_index;
      brw_velems->ve[i].ve1.vfcomponent0 = comp0;
      brw_velems->ve[i].ve1.vfcomponent1 = comp1;
      brw_velems->ve[i].ve1.vfcomponent2 = comp2;
      brw_velems->ve[i].ve1.vfcomponent3 = comp3;

      if (BRW_IS_IGDNG(brw))
         brw_velems->ve[i].ve1.dst_offset = 0;
      else
         brw_velems->ve[i].ve1.dst_offset = i * 4;
   }
}

static void* brw_create_vertex_elements_state( struct pipe_context *pipe,
                                               unsigned count,
                                               const struct pipe_vertex_element *attribs )
{
   /* note: for the brw_swtnl.c code (if ever we need draw fallback) we'd also need
      to store the original data */
   struct brw_context *brw = brw_context(pipe);
   struct brw_vertex_element_packet *velems;
   assert(count <= BRW_VEP_MAX);
   velems = (struct brw_vertex_element_packet *) MALLOC(sizeof(struct brw_vertex_element_packet));
   if (velems) {
      brw_translate_vertex_elements(brw, velems, attribs, count);
   }
   return velems;
}

static void brw_bind_vertex_elements_state(struct pipe_context *pipe,
                                           void *velems)
{
   struct brw_context *brw = brw_context(pipe);
   struct brw_vertex_element_packet *brw_velems = (struct brw_vertex_element_packet *) velems;

   brw->curr.velems = brw_velems;

   brw->state.dirty.mesa |= PIPE_NEW_VERTEX_ELEMENT;
}

static void brw_delete_vertex_elements_state(struct pipe_context *pipe, void *velems)
{
   FREE( velems );
}


static void brw_set_vertex_buffers(struct pipe_context *pipe,
                                   unsigned count,
                                   const struct pipe_vertex_buffer *buffers)
{
   struct brw_context *brw = brw_context(pipe);
   unsigned i;

   /* Check for no change */
   if (count == brw->curr.num_vertex_buffers &&
       memcmp(brw->curr.vertex_buffer,
              buffers,
              count * sizeof buffers[0]) == 0)
      return;

   /* Adjust refcounts */
   for (i = 0; i < count; i++) 
      pipe_buffer_reference(&brw->curr.vertex_buffer[i].buffer, 
                            buffers[i].buffer);

   for ( ; i < brw->curr.num_vertex_buffers; i++)
      pipe_buffer_reference(&brw->curr.vertex_buffer[i].buffer,
                            NULL);

   /* Copy remaining data */
   memcpy(brw->curr.vertex_buffer, buffers, count * sizeof buffers[0]);
   brw->curr.num_vertex_buffers = count;

   brw->state.dirty.mesa |= PIPE_NEW_VERTEX_BUFFER;
}


void 
brw_pipe_vertex_init( struct brw_context *brw )
{
   brw->base.set_vertex_buffers = brw_set_vertex_buffers;
   brw->base.create_vertex_elements_state = brw_create_vertex_elements_state;
   brw->base.bind_vertex_elements_state = brw_bind_vertex_elements_state;
   brw->base.delete_vertex_elements_state = brw_delete_vertex_elements_state;
}


void 
brw_pipe_vertex_cleanup( struct brw_context *brw )
{

   /* Release bound pipe vertex_buffers
    */

   /* Release some other stuff
    */
#if 0
   for (i = 0; i < PIPE_MAX_ATTRIBS; i++) {
      bo_reference(&brw->vb.inputs[i].bo, NULL);
      brw->vb.inputs[i].bo = NULL;
   }
#endif
}