summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/i965/brw_sf_state.c
blob: c99918724b36b8964b021c74cf82563bab12a34e (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
/*
 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"
#include "intel_fbo.h"

static void upload_sf_vp(struct brw_context *brw)
{
   GLcontext *ctx = &brw->intel.ctx;
   const GLfloat depth_scale = 1.0F / ctx->DrawBuffer->_DepthMaxF;
   struct brw_sf_viewport sfv;
   GLfloat y_scale, y_bias;
   const GLboolean render_to_fbo = (ctx->DrawBuffer->Name != 0);
   const GLfloat *v = ctx->Viewport._WindowMap.m;

   memset(&sfv, 0, sizeof(sfv));

   if (render_to_fbo) {
      y_scale = 1.0;
      y_bias = 0;
   }
   else {
      y_scale = -1.0;
      y_bias = ctx->DrawBuffer->Height;
   }

   /* _NEW_VIEWPORT */

   sfv.viewport.m00 = v[MAT_SX];
   sfv.viewport.m11 = v[MAT_SY] * y_scale;
   sfv.viewport.m22 = v[MAT_SZ] * depth_scale;
   sfv.viewport.m30 = v[MAT_TX];
   sfv.viewport.m31 = v[MAT_TY] * y_scale + y_bias;
   sfv.viewport.m32 = v[MAT_TZ] * depth_scale;

   /* _NEW_SCISSOR | _NEW_BUFFERS | _NEW_VIEWPORT
    * for DrawBuffer->_[XY]{min,max}
    */

   /* The scissor only needs to handle the intersection of drawable and
    * scissor rect.  Clipping to the boundaries of static shared buffers
    * for front/back/depth is covered by looping over cliprects in brw_draw.c.
    *
    * Note that the hardware's coordinates are inclusive, while Mesa's min is
    * inclusive but max is exclusive.
    */
   if (render_to_fbo) {
      /* texmemory: Y=0=bottom */
      sfv.scissor.xmin = ctx->DrawBuffer->_Xmin;
      sfv.scissor.xmax = ctx->DrawBuffer->_Xmax - 1;
      sfv.scissor.ymin = ctx->DrawBuffer->_Ymin;
      sfv.scissor.ymax = ctx->DrawBuffer->_Ymax - 1;
   }
   else {
      /* memory: Y=0=top */
      sfv.scissor.xmin = ctx->DrawBuffer->_Xmin;
      sfv.scissor.xmax = ctx->DrawBuffer->_Xmax - 1;
      sfv.scissor.ymin = ctx->DrawBuffer->Height - ctx->DrawBuffer->_Ymax;
      sfv.scissor.ymax = ctx->DrawBuffer->Height - ctx->DrawBuffer->_Ymin - 1;
   }

   dri_bo_unreference(brw->sf.vp_bo);
   brw->sf.vp_bo = brw_cache_data( &brw->cache, BRW_SF_VP, &sfv, NULL, 0 );
}

const struct brw_tracked_state brw_sf_vp = {
   .dirty = {
      .mesa  = (_NEW_VIEWPORT | 
		_NEW_SCISSOR |
		_NEW_BUFFERS),
      .brw   = 0,
      .cache = 0
   },
   .prepare = upload_sf_vp
};

struct brw_sf_unit_key {
   unsigned int total_grf;
   unsigned int urb_entry_read_length;

   unsigned int nr_urb_entries, urb_size, sfsize;

   GLenum front_face, cull_face;
   unsigned scissor:1;
   unsigned line_smooth:1;
   unsigned point_sprite:1;
   unsigned point_attenuated:1;
   unsigned render_to_fbo:1;
   float line_width;
   float point_size;
};

static void
sf_unit_populate_key(struct brw_context *brw, struct brw_sf_unit_key *key)
{
   GLcontext *ctx = &brw->intel.ctx;
   memset(key, 0, sizeof(*key));

   /* CACHE_NEW_SF_PROG */
   key->total_grf = brw->sf.prog_data->total_grf;
   key->urb_entry_read_length = brw->sf.prog_data->urb_read_length;

   /* BRW_NEW_URB_FENCE */
   key->nr_urb_entries = brw->urb.nr_sf_entries;
   key->urb_size = brw->urb.vsize;
   key->sfsize = brw->urb.sfsize;

   key->scissor = ctx->Scissor.Enabled;
   key->front_face = ctx->Polygon.FrontFace;

   if (ctx->Polygon.CullFlag)
      key->cull_face = ctx->Polygon.CullFaceMode;
   else
      key->cull_face = GL_NONE;

   key->line_width = ctx->Line.Width;
   key->line_smooth = ctx->Line.SmoothFlag;

   key->point_sprite = ctx->Point.PointSprite;
   key->point_size = CLAMP(ctx->Point.Size, ctx->Point.MinSize, ctx->Point.MaxSize);
   key->point_attenuated = ctx->Point._Attenuated;

   key->render_to_fbo = brw->intel.ctx.DrawBuffer->Name != 0;
}

static dri_bo *
sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
			dri_bo **reloc_bufs)
{
   struct brw_sf_unit_state sf;
   dri_bo *bo;

   memset(&sf, 0, sizeof(sf));

   sf.thread0.grf_reg_count = ALIGN(key->total_grf, 16) / 16 - 1;
   sf.thread0.kernel_start_pointer = brw->sf.prog_bo->offset >> 6; /* reloc */

   sf.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;

   sf.thread3.dispatch_grf_start_reg = 3;
   sf.thread3.urb_entry_read_offset = 1;
   sf.thread3.urb_entry_read_length = key->urb_entry_read_length;

   sf.thread4.nr_urb_entries = key->nr_urb_entries;
   sf.thread4.urb_entry_allocation_size = key->sfsize - 1;
   /* Each SF thread produces 1 PUE, and there can be up to 24 threads */
   sf.thread4.max_threads = MIN2(24, key->nr_urb_entries) - 1;

   if (INTEL_DEBUG & DEBUG_SINGLE_THREAD)
      sf.thread4.max_threads = 0;

   if (INTEL_DEBUG & DEBUG_STATS)
      sf.thread4.stats_enable = 1;

   /* CACHE_NEW_SF_VP */
   sf.sf5.sf_viewport_state_offset = brw->sf.vp_bo->offset >> 5; /* reloc */

   sf.sf5.viewport_transform = 1;

   /* _NEW_SCISSOR */
   if (key->scissor)
      sf.sf6.scissor = 1;

   /* _NEW_POLYGON */
   if (key->front_face == GL_CCW)
      sf.sf5.front_winding = BRW_FRONTWINDING_CCW;
   else
      sf.sf5.front_winding = BRW_FRONTWINDING_CW;

   /* The viewport is inverted for rendering to a FBO, and that inverts
    * polygon front/back orientation.
    */
   sf.sf5.front_winding ^= key->render_to_fbo;

   switch (key->cull_face) {
   case GL_FRONT:
      sf.sf6.cull_mode = BRW_CULLMODE_FRONT;
      break;
   case GL_BACK:
      sf.sf6.cull_mode = BRW_CULLMODE_BACK;
      break;
   case GL_FRONT_AND_BACK:
      sf.sf6.cull_mode = BRW_CULLMODE_BOTH;
      break;
   case GL_NONE:
      sf.sf6.cull_mode = BRW_CULLMODE_NONE;
      break;
   default:
      assert(0);
      break;
   }

   /* _NEW_LINE */
   /* XXX use ctx->Const.Min/MaxLineWidth here */
   sf.sf6.line_width = CLAMP(key->line_width, 1.0, 5.0) * (1<<1);

   sf.sf6.line_endcap_aa_region_width = 1;
   if (key->line_smooth)
      sf.sf6.aa_enable = 1;
   else if (sf.sf6.line_width <= 0x2)
       sf.sf6.line_width = 0;

   /* _NEW_POINT */
   key->render_to_fbo = brw->intel.ctx.DrawBuffer->Name != 0;
   if (!key->render_to_fbo) {
      /* Rendering to an OpenGL window */
      sf.sf6.point_rast_rule = BRW_RASTRULE_UPPER_RIGHT;
   }
   else {
      /* If rendering to an FBO, the pixel coordinate system is
       * inverted with respect to the normal OpenGL coordinate
       * system, so BRW_RASTRULE_LOWER_RIGHT is correct.
       * But this value is listed as "Reserved, but not seen as useful"
       * in Intel documentation (page 212, "Point Rasterization Rule",
       * section 7.4 "SF Pipeline State Summary", of document
       * "Intel® 965 Express Chipset Family and Intel® G35 Express
       * Chipset Graphics Controller Programmer's Reference Manual,
       * Volume 2: 3D/Media", Revision 1.0b as of January 2008,
       * available at 
       *     http://intellinuxgraphics.org/documentation.html
       * at the time of this writing).
       *
       * It does work on at least some devices, if not all;
       * if devices that don't support it can be identified,
       * the likely failure case is that points are rasterized
       * incorrectly, which is no worse than occurs without
       * the value, so we're using it here.
       */
      sf.sf6.point_rast_rule = BRW_RASTRULE_LOWER_RIGHT;
   }
   /* XXX clamp max depends on AA vs. non-AA */

   sf.sf7.sprite_point = key->point_sprite;
   sf.sf7.point_size = CLAMP(rint(key->point_size), 1, 255) * (1<<3);
   sf.sf7.use_point_size_state = !key->point_attenuated;
   sf.sf7.aa_line_distance_mode = 0;

   /* might be BRW_NEW_PRIMITIVE if we have to adjust pv for polygons:
    */
   sf.sf7.trifan_pv = 2;
   sf.sf7.linestrip_pv = 1;
   sf.sf7.tristrip_pv = 2;
   sf.sf7.line_last_pixel_enable = 0;

   /* Set bias for OpenGL rasterization rules:
    */
   sf.sf6.dest_org_vbias = 0x8;
   sf.sf6.dest_org_hbias = 0x8;

   bo = brw_upload_cache(&brw->cache, BRW_SF_UNIT,
			 key, sizeof(*key),
			 reloc_bufs, 2,
			 &sf, sizeof(sf),
			 NULL, NULL);

   /* Emit SF program relocation */
   dri_bo_emit_reloc(bo,
		     I915_GEM_DOMAIN_INSTRUCTION, 0,
		     sf.thread0.grf_reg_count << 1,
		     offsetof(struct brw_sf_unit_state, thread0),
		     brw->sf.prog_bo);

   /* Emit SF viewport relocation */
   dri_bo_emit_reloc(bo,
		     I915_GEM_DOMAIN_INSTRUCTION, 0,
		     sf.sf5.front_winding | (sf.sf5.viewport_transform << 1),
		     offsetof(struct brw_sf_unit_state, sf5),
		     brw->sf.vp_bo);

   return bo;
}

static void upload_sf_unit( struct brw_context *brw )
{
   struct brw_sf_unit_key key;
   dri_bo *reloc_bufs[2];

   sf_unit_populate_key(brw, &key);

   reloc_bufs[0] = brw->sf.prog_bo;
   reloc_bufs[1] = brw->sf.vp_bo;

   dri_bo_unreference(brw->sf.state_bo);
   brw->sf.state_bo = brw_search_cache(&brw->cache, BRW_SF_UNIT,
				       &key, sizeof(key),
				       reloc_bufs, 2,
				       NULL);
   if (brw->sf.state_bo == NULL) {
      brw->sf.state_bo = sf_unit_create_from_key(brw, &key, reloc_bufs);
   }
}

const struct brw_tracked_state brw_sf_unit = {
   .dirty = {
      .mesa  = (_NEW_POLYGON | 
		_NEW_LINE | 
		_NEW_POINT | 
		_NEW_SCISSOR),
      .brw   = BRW_NEW_URB_FENCE,
      .cache = (CACHE_NEW_SF_VP |
		CACHE_NEW_SF_PROG)
   },
   .prepare = upload_sf_unit,
};