aboutsummaryrefslogtreecommitdiffstats
path: root/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
blob: 1fc802cfa657e23d9d3c513c9634241bc911480f (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
/*
 Copyright (C) Intel Corp.  2006.  All Rights Reserved.
 Intel funded Tungsten Graphics (http://www.tungstengraphics.com) 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 <keith@tungstengraphics.com>
  */
                   

#include "brw_context.h"
#include "brw_state.h"
#include "brw_defines.h"

#include "main/macros.h"



/* Samplers aren't strictly wm state from the hardware's perspective,
 * but that is the only situation in which we use them in this driver.
 */



/* The brw (and related graphics cores) do not support GL_CLAMP.  The
 * Intel drivers for "other operating systems" implement GL_CLAMP as
 * GL_CLAMP_TO_EDGE, so the same is done here.
 */
static GLuint translate_wrap_mode( GLenum wrap )
{
   switch( wrap ) {
   case GL_REPEAT: 
      return BRW_TEXCOORDMODE_WRAP;
   case GL_CLAMP:  
      return BRW_TEXCOORDMODE_CLAMP;
   case GL_CLAMP_TO_EDGE: 
      return BRW_TEXCOORDMODE_CLAMP; /* conform likes it this way */
   case GL_CLAMP_TO_BORDER: 
      return BRW_TEXCOORDMODE_CLAMP_BORDER;
   case GL_MIRRORED_REPEAT: 
      return BRW_TEXCOORDMODE_MIRROR;
   default: 
      return BRW_TEXCOORDMODE_WRAP;
   }
}

static drm_intel_bo *upload_default_color( struct brw_context *brw,
				     const GLfloat *color )
{
   struct brw_sampler_default_color sdc;

   COPY_4V(sdc.color, color); 
   
   return brw_cache_data(&brw->cache, BRW_SAMPLER_DEFAULT_COLOR,
			 &sdc, sizeof(sdc));
}


struct wm_sampler_key {
   int sampler_count;

   struct wm_sampler_entry {
      GLenum tex_target;
      GLenum wrap_r, wrap_s, wrap_t;
      float maxlod, minlod;
      float lod_bias;
      float max_aniso;
      GLenum minfilter, magfilter;
      GLenum comparemode, comparefunc;

      /** If target is cubemap, take context setting.
       */
      GLboolean seamless_cube_map;
   } sampler[BRW_MAX_TEX_UNIT];
};

/**
 * Sets the sampler state for a single unit based off of the sampler key
 * entry.
 */
static void brw_update_sampler_state(struct wm_sampler_entry *key,
				     drm_intel_bo *sdc_bo,
				     struct brw_sampler_state *sampler)
{
   memset(sampler, 0, sizeof(*sampler));

   switch (key->minfilter) {
   case GL_NEAREST:
      sampler->ss0.min_filter = BRW_MAPFILTER_NEAREST;
      sampler->ss0.mip_filter = BRW_MIPFILTER_NONE;
      break;
   case GL_LINEAR:
      sampler->ss0.min_filter = BRW_MAPFILTER_LINEAR;
      sampler->ss0.mip_filter = BRW_MIPFILTER_NONE;
      break;
   case GL_NEAREST_MIPMAP_NEAREST:
      sampler->ss0.min_filter = BRW_MAPFILTER_NEAREST;
      sampler->ss0.mip_filter = BRW_MIPFILTER_NEAREST;
      break;
   case GL_LINEAR_MIPMAP_NEAREST:
      sampler->ss0.min_filter = BRW_MAPFILTER_LINEAR;
      sampler->ss0.mip_filter = BRW_MIPFILTER_NEAREST;
      break;
   case GL_NEAREST_MIPMAP_LINEAR:
      sampler->ss0.min_filter = BRW_MAPFILTER_NEAREST;
      sampler->ss0.mip_filter = BRW_MIPFILTER_LINEAR;
      break;
   case GL_LINEAR_MIPMAP_LINEAR:
      sampler->ss0.min_filter = BRW_MAPFILTER_LINEAR;
      sampler->ss0.mip_filter = BRW_MIPFILTER_LINEAR;
      break;
   default:
      break;
   }

   /* Set Anisotropy: 
    */
   if (key->max_aniso > 1.0) {
      sampler->ss0.min_filter = BRW_MAPFILTER_ANISOTROPIC; 
      sampler->ss0.mag_filter = BRW_MAPFILTER_ANISOTROPIC;

      if (key->max_aniso > 2.0) {
	 sampler->ss3.max_aniso = MIN2((key->max_aniso - 2) / 2,
				       BRW_ANISORATIO_16);
      }
   }
   else {
      switch (key->magfilter) {
      case GL_NEAREST:
	 sampler->ss0.mag_filter = BRW_MAPFILTER_NEAREST;
	 break;
      case GL_LINEAR:
	 sampler->ss0.mag_filter = BRW_MAPFILTER_LINEAR;
	 break;
      default:
	 break;
      }  
   }

   sampler->ss1.r_wrap_mode = translate_wrap_mode(key->wrap_r);
   sampler->ss1.s_wrap_mode = translate_wrap_mode(key->wrap_s);
   sampler->ss1.t_wrap_mode = translate_wrap_mode(key->wrap_t);

   /* Cube-maps on 965 and later must use the same wrap mode for all 3
    * coordinate dimensions.  Futher, only CUBE and CLAMP are valid.
    */
   if (key->tex_target == GL_TEXTURE_CUBE_MAP) {
      if (key->seamless_cube_map &&
	  (key->minfilter != GL_NEAREST || key->magfilter != GL_NEAREST)) {
	 sampler->ss1.r_wrap_mode = BRW_TEXCOORDMODE_CUBE;
	 sampler->ss1.s_wrap_mode = BRW_TEXCOORDMODE_CUBE;
	 sampler->ss1.t_wrap_mode = BRW_TEXCOORDMODE_CUBE;
      } else {
	 sampler->ss1.r_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
	 sampler->ss1.s_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
	 sampler->ss1.t_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
      }
   } else if (key->tex_target == GL_TEXTURE_1D) {
      /* There's a bug in 1D texture sampling - it actually pays
       * attention to the wrap_t value, though it should not.
       * Override the wrap_t value here to GL_REPEAT to keep
       * any nonexistent border pixels from floating in.
       */
      sampler->ss1.t_wrap_mode = BRW_TEXCOORDMODE_WRAP;
   }


   /* Set shadow function: 
    */
   if (key->comparemode == GL_COMPARE_R_TO_TEXTURE_ARB) {
      /* Shadowing is "enabled" by emitting a particular sampler
       * message (sample_c).  So need to recompile WM program when
       * shadow comparison is enabled on each/any texture unit.
       */
      sampler->ss0.shadow_function =
	 intel_translate_shadow_compare_func(key->comparefunc);
   }

   /* Set LOD bias: 
    */
   sampler->ss0.lod_bias = S_FIXED(CLAMP(key->lod_bias, -16, 15), 6);

   sampler->ss0.lod_preclamp = 1; /* OpenGL mode */
   sampler->ss0.default_color_mode = 0; /* OpenGL/DX10 mode */

   /* Set BaseMipLevel, MaxLOD, MinLOD: 
    *
    * XXX: I don't think that using firstLevel, lastLevel works,
    * because we always setup the surface state as if firstLevel ==
    * level zero.  Probably have to subtract firstLevel from each of
    * these:
    */
   sampler->ss0.base_level = U_FIXED(0, 1);

   sampler->ss1.max_lod = U_FIXED(CLAMP(key->maxlod, 0, 13), 6);
   sampler->ss1.min_lod = U_FIXED(CLAMP(key->minlod, 0, 13), 6);
   
   sampler->ss2.default_color_pointer = sdc_bo->offset >> 5; /* reloc */
}


/** Sets up the cache key for sampler state for all texture units */
static void
brw_wm_sampler_populate_key(struct brw_context *brw,
			    struct wm_sampler_key *key)
{
   GLcontext *ctx = &brw->intel.ctx;
   int unit;
   char *last_entry_end = ((char*)&key->sampler_count) + 
      sizeof(key->sampler_count);

   key->sampler_count = 0;

   for (unit = 0; unit < BRW_MAX_TEX_UNIT; unit++) {
      if (ctx->Texture.Unit[unit]._ReallyEnabled) {
	 struct wm_sampler_entry *entry = &key->sampler[unit];
	 struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
	 struct gl_texture_object *texObj = texUnit->_Current;
	 struct intel_texture_object *intelObj = intel_texture_object(texObj);
	 struct gl_texture_image *firstImage =
	    texObj->Image[0][intelObj->firstLevel];

	 memset(last_entry_end, 0, 
		(char*)entry - last_entry_end + sizeof(*entry));
	 last_entry_end = ((char*)entry) + sizeof(*entry);

         entry->tex_target = texObj->Target;

	 entry->seamless_cube_map = (texObj->Target == GL_TEXTURE_CUBE_MAP)
	    ? ctx->Texture.CubeMapSeamless : GL_FALSE;

	 entry->wrap_r = texObj->WrapR;
	 entry->wrap_s = texObj->WrapS;
	 entry->wrap_t = texObj->WrapT;

	 entry->maxlod = texObj->MaxLod;
	 entry->minlod = texObj->MinLod;
	 entry->lod_bias = texUnit->LodBias + texObj->LodBias;
	 entry->max_aniso = texObj->MaxAnisotropy;
	 entry->minfilter = texObj->MinFilter;
	 entry->magfilter = texObj->MagFilter;
	 entry->comparemode = texObj->CompareMode;
         entry->comparefunc = texObj->CompareFunc;

	 drm_intel_bo_unreference(brw->wm.sdc_bo[unit]);
	 if (firstImage->_BaseFormat == GL_DEPTH_COMPONENT) {
	    float bordercolor[4] = {
	       texObj->BorderColor.f[0],
	       texObj->BorderColor.f[0],
	       texObj->BorderColor.f[0],
	       texObj->BorderColor.f[0]
	    };
	    /* GL specs that border color for depth textures is taken from the
	     * R channel, while the hardware uses A.  Spam R into all the
	     * channels for safety.
	     */
	    brw->wm.sdc_bo[unit] = upload_default_color(brw, bordercolor);
	 } else {
	    brw->wm.sdc_bo[unit] = upload_default_color(brw,
							texObj->BorderColor.f);
	 }
	 key->sampler_count = unit + 1;
      }
   }
   struct wm_sampler_entry *entry = &key->sampler[key->sampler_count];
   memset(last_entry_end, 0, (char*)entry - last_entry_end);
}

/* All samplers must be uploaded in a single contiguous array, which
 * complicates various things.  However, this is still too confusing -
 * FIXME: simplify all the different new texture state flags.
 */
static void upload_wm_samplers( struct brw_context *brw )
{
   GLcontext *ctx = &brw->intel.ctx;
   struct wm_sampler_key key;
   int i, sampler_key_size;

   brw_wm_sampler_populate_key(brw, &key);

   if (brw->wm.sampler_count != key.sampler_count) {
      brw->wm.sampler_count = key.sampler_count;
      brw->state.dirty.cache |= CACHE_NEW_SAMPLER;
   }

   drm_intel_bo_unreference(brw->wm.sampler_bo);
   brw->wm.sampler_bo = NULL;
   if (brw->wm.sampler_count == 0)
      return;

   /* Only include the populated portion of the key in the search. */
   sampler_key_size = offsetof(struct wm_sampler_key,
			       sampler[key.sampler_count]);
   brw->wm.sampler_bo = brw_search_cache(&brw->cache, BRW_SAMPLER,
					 &key, sampler_key_size,
					 brw->wm.sdc_bo, key.sampler_count,
					 NULL);

   /* If we didnt find it in the cache, compute the state and put it in the
    * cache.
    */
   if (brw->wm.sampler_bo == NULL) {
      struct brw_sampler_state sampler[BRW_MAX_TEX_UNIT];

      memset(sampler, 0, sizeof(sampler));
      for (i = 0; i < key.sampler_count; i++) {
	 if (brw->wm.sdc_bo[i] == NULL)
	    continue;

	 brw_update_sampler_state(&key.sampler[i], brw->wm.sdc_bo[i],
				  &sampler[i]);
      }

      brw->wm.sampler_bo = brw_upload_cache(&brw->cache, BRW_SAMPLER,
					    &key, sampler_key_size,
					    brw->wm.sdc_bo, key.sampler_count,
					    &sampler, sizeof(sampler));

      /* Emit SDC relocations */
      for (i = 0; i < BRW_MAX_TEX_UNIT; i++) {
	 if (!ctx->Texture.Unit[i]._ReallyEnabled)
	    continue;

	 drm_intel_bo_emit_reloc(brw->wm.sampler_bo,
				 i * sizeof(struct brw_sampler_state) +
				 offsetof(struct brw_sampler_state, ss2),
				 brw->wm.sdc_bo[i], 0,
				 I915_GEM_DOMAIN_SAMPLER, 0);
      }
   }
}

const struct brw_tracked_state brw_wm_samplers = {
   .dirty = {
      .mesa = _NEW_TEXTURE,
      .brw = 0,
      .cache = 0
   },
   .prepare = upload_wm_samplers,
};