Merge branch 'llvm-cliptest-viewport'
[mesa.git] / src / mesa / drivers / dri / r300 / r300_context.h
1 /*
2 Copyright (C) The Weather Channel, Inc. 2002. All Rights Reserved.
3
4 The Weather Channel (TM) funded Tungsten Graphics to develop the
5 initial release of the Radeon 8500 driver under the XFree86 license.
6 This notice must be preserved.
7
8 Permission is hereby granted, free of charge, to any person obtaining
9 a copy of this software and associated documentation files (the
10 "Software"), to deal in the Software without restriction, including
11 without limitation the rights to use, copy, modify, merge, publish,
12 distribute, sublicense, and/or sell copies of the Software, and to
13 permit persons to whom the Software is furnished to do so, subject to
14 the following conditions:
15
16 The above copyright notice and this permission notice (including the
17 next paragraph) shall be included in all copies or substantial
18 portions of the Software.
19
20 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
21 EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
22 MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
23 IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
24 LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
25 OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
26 WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
27
28 **************************************************************************/
29
30 /**
31 * \file
32 *
33 * \author Keith Whitwell <keith@tungstengraphics.com>
34 * \author Nicolai Haehnle <prefect_@gmx.net>
35 */
36
37 #ifndef __R300_CONTEXT_H__
38 #define __R300_CONTEXT_H__
39
40 #include "drm.h"
41 #include "radeon_drm.h"
42 #include "dri_util.h"
43 #include "radeon_common.h"
44
45 #include "main/mtypes.h"
46 #include "program/prog_instruction.h"
47 #include "compiler/radeon_code.h"
48
49 struct r300_context;
50 typedef struct r300_context r300ContextRec;
51 typedef struct r300_context *r300ContextPtr;
52
53
54 #include "r300_vertprog.h"
55
56
57 /* The blit width for texture uploads
58 */
59 #define R300_BLIT_WIDTH_BYTES 1024
60 #define R300_MAX_TEXTURE_UNITS 8
61
62
63
64 #define R300_VPT_CMD_0 0
65 #define R300_VPT_XSCALE 1
66 #define R300_VPT_XOFFSET 2
67 #define R300_VPT_YSCALE 3
68 #define R300_VPT_YOFFSET 4
69 #define R300_VPT_ZSCALE 5
70 #define R300_VPT_ZOFFSET 6
71 #define R300_VPT_CMDSIZE 7
72
73 #define R300_VIR_CMD_0 0 /* vir is variable size (at least 1) */
74 #define R300_VIR_CNTL_0 1
75 #define R300_VIR_CNTL_1 2
76 #define R300_VIR_CNTL_2 3
77 #define R300_VIR_CNTL_3 4
78 #define R300_VIR_CNTL_4 5
79 #define R300_VIR_CNTL_5 6
80 #define R300_VIR_CNTL_6 7
81 #define R300_VIR_CNTL_7 8
82 #define R300_VIR_CMDSIZE 9
83
84 #define R300_VIC_CMD_0 0
85 #define R300_VIC_CNTL_0 1
86 #define R300_VIC_CNTL_1 2
87 #define R300_VIC_CMDSIZE 3
88
89 #define R300_VOF_CMD_0 0
90 #define R300_VOF_CNTL_0 1
91 #define R300_VOF_CNTL_1 2
92 #define R300_VOF_CMDSIZE 3
93
94 #define R300_PVS_CMD_0 0
95 #define R300_PVS_CNTL_1 1
96 #define R300_PVS_CNTL_2 2
97 #define R300_PVS_CNTL_3 3
98 #define R300_PVS_CMDSIZE 4
99
100 #define R300_GB_MISC_CMD_0 0
101 #define R300_GB_MISC_MSPOS_0 1
102 #define R300_GB_MISC_MSPOS_1 2
103 #define R300_GB_MISC_TILE_CONFIG 3
104 #define R300_GB_MISC_CMDSIZE 4
105 #define R300_GB_MISC2_CMD_0 0
106 #define R300_GB_MISC2_SELECT 1
107 #define R300_GB_MISC2_AA_CONFIG 2
108 #define R300_GB_MISC2_CMDSIZE 3
109
110 #define R300_TXE_CMD_0 0
111 #define R300_TXE_ENABLE 1
112 #define R300_TXE_CMDSIZE 2
113
114 #define R300_PS_CMD_0 0
115 #define R300_PS_POINTSIZE 1
116 #define R300_PS_CMDSIZE 2
117
118 #define R300_ZBS_CMD_0 0
119 #define R300_ZBS_T_FACTOR 1
120 #define R300_ZBS_T_CONSTANT 2
121 #define R300_ZBS_W_FACTOR 3
122 #define R300_ZBS_W_CONSTANT 4
123 #define R300_ZBS_CMDSIZE 5
124
125 #define R300_CUL_CMD_0 0
126 #define R300_CUL_CULL 1
127 #define R300_CUL_CMDSIZE 2
128
129 #define R300_RC_CMD_0 0
130 #define R300_RC_CNTL_0 1
131 #define R300_RC_CNTL_1 2
132 #define R300_RC_CMDSIZE 3
133
134 #define R300_RI_CMD_0 0
135 #define R300_RI_INTERP_0 1
136 #define R300_RI_INTERP_1 2
137 #define R300_RI_INTERP_2 3
138 #define R300_RI_INTERP_3 4
139 #define R300_RI_INTERP_4 5
140 #define R300_RI_INTERP_5 6
141 #define R300_RI_INTERP_6 7
142 #define R300_RI_INTERP_7 8
143 #define R300_RI_CMDSIZE 9
144
145 #define R500_RI_CMDSIZE 17
146
147 #define R300_RR_CMD_0 0 /* rr is variable size (at least 1) */
148 #define R300_RR_INST_0 1
149 #define R300_RR_INST_1 2
150 #define R300_RR_INST_2 3
151 #define R300_RR_INST_3 4
152 #define R300_RR_INST_4 5
153 #define R300_RR_INST_5 6
154 #define R300_RR_INST_6 7
155 #define R300_RR_INST_7 8
156 #define R300_RR_CMDSIZE 9
157
158 #define R300_FP_CMD_0 0
159 #define R300_FP_CNTL0 1
160 #define R300_FP_CNTL1 2
161 #define R300_FP_CNTL2 3
162 #define R300_FP_CMD_1 4
163 #define R300_FP_NODE0 5
164 #define R300_FP_NODE1 6
165 #define R300_FP_NODE2 7
166 #define R300_FP_NODE3 8
167 #define R300_FP_CMDSIZE 9
168
169 #define R500_FP_CMD_0 0
170 #define R500_FP_CNTL 1
171 #define R500_FP_PIXSIZE 2
172 #define R500_FP_CMD_1 3
173 #define R500_FP_CODE_ADDR 4
174 #define R500_FP_CODE_RANGE 5
175 #define R500_FP_CODE_OFFSET 6
176 #define R500_FP_CMD_2 7
177 #define R500_FP_FC_CNTL 8
178 #define R500_FP_CMDSIZE 9
179
180 #define R300_FPT_CMD_0 0
181 #define R300_FPT_INSTR_0 1
182 #define R300_FPT_CMDSIZE 65
183
184 #define R300_FPI_CMD_0 0
185 #define R300_FPI_INSTR_0 1
186 #define R300_FPI_CMDSIZE 65
187 /* R500 has space for 512 instructions - 6 dwords per instruction */
188 #define R500_FPI_CMDSIZE (512*6+1)
189
190 #define R300_FPP_CMD_0 0
191 #define R300_FPP_PARAM_0 1
192 #define R300_FPP_CMDSIZE (32*4+1)
193 /* R500 has spcae for 256 constants - 4 dwords per constant */
194 #define R500_FPP_CMDSIZE (256*4+1)
195
196 #define R300_FOGS_CMD_0 0
197 #define R300_FOGS_STATE 1
198 #define R300_FOGS_CMDSIZE 2
199
200 #define R300_FOGC_CMD_0 0
201 #define R300_FOGC_R 1
202 #define R300_FOGC_G 2
203 #define R300_FOGC_B 3
204 #define R300_FOGC_CMDSIZE 4
205
206 #define R300_FOGP_CMD_0 0
207 #define R300_FOGP_SCALE 1
208 #define R300_FOGP_START 2
209 #define R300_FOGP_CMDSIZE 3
210
211 #define R300_AT_CMD_0 0
212 #define R300_AT_ALPHA_TEST 1
213 #define R300_AT_UNKNOWN 2
214 #define R300_AT_CMDSIZE 3
215
216 #define R300_BLD_CMD_0 0
217 #define R300_BLD_CBLEND 1
218 #define R300_BLD_ABLEND 2
219 #define R300_BLD_CMDSIZE 3
220
221 #define R300_CMK_CMD_0 0
222 #define R300_CMK_COLORMASK 1
223 #define R300_CMK_CMDSIZE 2
224
225 #define R300_CB_CMD_0 0
226 #define R300_CB_OFFSET 1
227 #define R300_CB_CMD_1 2
228 #define R300_CB_PITCH 3
229 #define R300_CB_CMDSIZE 4
230
231 #define R300_ZS_CMD_0 0
232 #define R300_ZS_CNTL_0 1
233 #define R300_ZS_CNTL_1 2
234 #define R300_ZS_CNTL_2 3
235 #define R300_ZS_CMDSIZE 4
236
237 #define R300_ZSB_CMD_0 0
238 #define R300_ZSB_CNTL_0 1
239 #define R300_ZSB_CMDSIZE 2
240
241 #define R300_ZB_CMD_0 0
242 #define R300_ZB_OFFSET 1
243 #define R300_ZB_PITCH 2
244 #define R300_ZB_CMDSIZE 3
245
246 #define R300_VAP_CNTL_FLUSH 0
247 #define R300_VAP_CNTL_FLUSH_1 1
248 #define R300_VAP_CNTL_CMD 2
249 #define R300_VAP_CNTL_INSTR 3
250 #define R300_VAP_CNTL_SIZE 4
251
252 #define R300_VPI_CMD_0 0
253 #define R300_VPI_INSTR_0 1
254 #define R300_VPI_CMDSIZE 1025 /* 256 16 byte instructions */
255
256 #define R300_VPP_CMD_0 0
257 #define R300_VPP_PARAM_0 1
258 #define R300_VPP_CMDSIZE 1025 /* 256 4-component parameters */
259
260 #define R300_VPUCP_CMD_0 0
261 #define R300_VPUCP_X 1
262 #define R300_VPUCP_Y 2
263 #define R300_VPUCP_Z 3
264 #define R300_VPUCP_W 4
265 #define R300_VPUCP_CMDSIZE 5 /* 256 4-component parameters */
266
267 #define R300_VPS_CMD_0 0
268 #define R300_VPS_ZERO_0 1
269 #define R300_VPS_ZERO_1 2
270 #define R300_VPS_POINTSIZE 3
271 #define R300_VPS_ZERO_3 4
272 #define R300_VPS_CMDSIZE 5
273
274 /* the layout is common for all fields inside tex */
275 #define R300_TEX_CMD_0 0
276 #define R300_TEX_VALUE_0 1
277 /* We don't really use this, instead specify mtu+1 dynamically
278 #define R300_TEX_CMDSIZE (MAX_TEXTURE_UNITS+1)
279 */
280
281 #define R300_QUERYOBJ_CMD_0 0
282 #define R300_QUERYOBJ_DATA_0 1
283 #define R300_QUERYOBJ_CMD_1 2
284 #define R300_QUERYOBJ_DATA_1 3
285 #define R300_QUERYOBJ_CMDSIZE 4
286
287 /**
288 * Cache for hardware register state.
289 */
290 struct r300_hw_state {
291 struct radeon_state_atom vpt; /* viewport (1D98) */
292 struct radeon_state_atom vap_cntl;
293 struct radeon_state_atom vap_index_offset; /* 0x208c r5xx only */
294 struct radeon_state_atom vof; /* VAP output format register 0x2090 */
295 struct radeon_state_atom vte; /* (20B0) */
296 struct radeon_state_atom vap_vf_max_vtx_indx; /* Maximum Vertex Indx Clamp (2134) */
297 struct radeon_state_atom vap_cntl_status;
298 struct radeon_state_atom vir[2]; /* vap input route (2150/21E0) */
299 struct radeon_state_atom vic; /* vap input control (2180) */
300 struct radeon_state_atom vap_psc_sgn_norm_cntl; /* Programmable Stream Control Signed Normalize Control (21DC) */
301 struct radeon_state_atom vap_clip_cntl;
302 struct radeon_state_atom vap_clip;
303 struct radeon_state_atom vap_pvs_vtx_timeout_reg; /* Vertex timeout register (2288) */
304 struct radeon_state_atom pvs; /* pvs_cntl (22D0) */
305 struct radeon_state_atom gb_enable; /* (4008) */
306 struct radeon_state_atom gb_misc; /* Multisampling position shifts ? (4010) */
307 struct radeon_state_atom gb_misc2; /* Multisampling position shifts ? (4010) */
308 struct radeon_state_atom ga_point_s0; /* S Texture Coordinate of Vertex 0 for Point texture stuffing (LLC) (4200) */
309 struct radeon_state_atom ga_triangle_stipple; /* (4214) */
310 struct radeon_state_atom ps; /* pointsize (421C) */
311 struct radeon_state_atom ga_point_minmax; /* (4230) */
312 struct radeon_state_atom lcntl; /* line control */
313 struct radeon_state_atom ga_line_stipple; /* (4260) */
314 struct radeon_state_atom shade;
315 struct radeon_state_atom shade2;
316 struct radeon_state_atom polygon_mode;
317 struct radeon_state_atom fogp; /* fog parameters (4294) */
318 struct radeon_state_atom ga_soft_reset; /* (429C) */
319 struct radeon_state_atom zbias_cntl;
320 struct radeon_state_atom zbs; /* zbias (42A4) */
321 struct radeon_state_atom occlusion_cntl;
322 struct radeon_state_atom cul; /* cull cntl (42B8) */
323 struct radeon_state_atom su_depth_scale; /* (42C0) */
324 struct radeon_state_atom rc; /* rs control (4300) */
325 struct radeon_state_atom ri; /* rs interpolators (4310) */
326 struct radeon_state_atom rr; /* rs route (4330) */
327 struct radeon_state_atom sc_hyperz; /* (43A4) */
328 struct radeon_state_atom sc_screendoor; /* (43E8) */
329 struct radeon_state_atom fp; /* fragment program cntl + nodes (4600) */
330 struct radeon_state_atom fpt; /* texi - (4620) */
331 struct radeon_state_atom us_out_fmt; /* (46A4) */
332 struct radeon_state_atom r500fp; /* r500 fp instructions */
333 struct radeon_state_atom r500fp_const; /* r500 fp constants */
334 struct radeon_state_atom fpi[4]; /* fp instructions (46C0/47C0/48C0/49C0) */
335 struct radeon_state_atom fogs; /* fog state (4BC0) */
336 struct radeon_state_atom fogc; /* fog color (4BC8) */
337 struct radeon_state_atom at; /* alpha test (4BD4) */
338 struct radeon_state_atom fg_depth_src; /* (4BD8) */
339 struct radeon_state_atom fpp; /* 0x4C00 and following */
340 struct radeon_state_atom rb3d_cctl; /* (4E00) */
341 struct radeon_state_atom bld; /* blending (4E04) */
342 struct radeon_state_atom cmk; /* colormask (4E0C) */
343 struct radeon_state_atom blend_color; /* constant blend color */
344 struct radeon_state_atom rop; /* ropcntl */
345 struct radeon_state_atom cb; /* colorbuffer (4E28) */
346 struct radeon_state_atom rb3d_dither_ctl; /* (4E50) */
347 struct radeon_state_atom rb3d_aaresolve_ctl; /* (4E88) */
348 struct radeon_state_atom rb3d_discard_src_pixel_lte_threshold; /* (4E88) I saw it only written on RV350 hardware.. */
349 struct radeon_state_atom zs; /* zstencil control (4F00) */
350 struct radeon_state_atom zsb; /* zstencil bf */
351 struct radeon_state_atom zstencil_format;
352 struct radeon_state_atom zb; /* z buffer (4F20) */
353 struct radeon_state_atom zb_depthclearvalue; /* (4F28) */
354 struct radeon_state_atom zb_zmask; /* (4F30) */
355 struct radeon_state_atom zb_hiz_offset; /* (4F44) */
356 struct radeon_state_atom zb_hiz_pitch; /* (4F54) */
357
358 struct radeon_state_atom vpi; /* vp instructions */
359 struct radeon_state_atom vpp; /* vp parameters */
360 struct radeon_state_atom vps; /* vertex point size (?) */
361 struct radeon_state_atom vpucp[6]; /* vp user clip plane - 6 */
362 /* 8 texture units */
363 /* the state is grouped by function and not by
364 texture unit. This makes single unit updates
365 really awkward - we are much better off
366 updating the whole thing at once */
367 struct {
368 struct radeon_state_atom filter;
369 struct radeon_state_atom filter_1;
370 struct radeon_state_atom size;
371 struct radeon_state_atom format;
372 struct radeon_state_atom pitch;
373 struct radeon_state_atom offset;
374 struct radeon_state_atom chroma_key;
375 struct radeon_state_atom border_color;
376 } tex;
377 struct radeon_state_atom txe; /* tex enable (4104) */
378 radeonTexObj *textures[R300_MAX_TEXTURE_UNITS];
379 };
380
381 /**
382 * State cache
383 */
384
385 /* Vertex shader state */
386
387 #define COLOR_IS_RGBA
388 #define TAG(x) r300##x
389 #include "tnl_dd/t_dd_vertex.h"
390 #undef TAG
391
392 struct r300_vertex_program_key {
393 GLbitfield FpReads;
394 GLuint FogAttr;
395 GLuint WPosAttr;
396 };
397
398 struct r300_vertex_program {
399 struct gl_vertex_program *Base;
400 struct r300_vertex_program *next;
401
402 struct r300_vertex_program_key key;
403 struct r300_vertex_program_code code;
404
405 GLboolean error;
406 };
407
408 struct r300_vertex_program_cont {
409 /* This is the unmodified vertex program mesa provided us with.
410 * We need to keep it unchanged because we may need to create another
411 * hw specific vertex program based on this.
412 */
413 struct gl_vertex_program mesa_program;
414 /* This is the list of hw specific vertex programs derived from mesa_program */
415 struct r300_vertex_program *progs;
416 };
417
418
419 /**
420 * Store everything about a fragment program that is needed
421 * to render with that program.
422 */
423 struct r300_fragment_program {
424 GLboolean error;
425 struct r300_fragment_program *next;
426 struct r300_fragment_program_external_state state;
427
428 struct rX00_fragment_program_code code;
429 GLbitfield InputsRead;
430
431 /* attribute that we are sending the WPOS in */
432 gl_frag_attrib wpos_attr;
433 /* attribute that we are sending the fog coordinate in */
434 gl_frag_attrib fog_attr;
435 };
436
437 struct r300_fragment_program_cont {
438 /* This is the unmodified fragment program mesa provided us with.
439 * We need to keep it unchanged because we may need to create another
440 * hw specific fragment program based on this.
441 */
442 struct gl_fragment_program Base;
443 /* This is the list of hw specific fragment programs derived from Base */
444 struct r300_fragment_program *progs;
445 };
446
447
448 #define R300_MAX_AOS_ARRAYS 16
449
450
451 /* r300_swtcl.c
452 */
453 struct r300_swtcl_info {
454 /*
455 * Offset of the 4UB color data within a hardware (swtcl) vertex.
456 */
457 GLuint coloroffset;
458
459 /**
460 * Offset of the 3UB specular color data within a hardware (swtcl) vertex.
461 */
462 GLuint specoffset;
463 };
464
465 struct r300_vtable {
466 void (* SetupRSUnit)(struct gl_context *ctx);
467 void (* SetupFragmentShaderTextures)(struct gl_context *ctx, int *tmu_mappings);
468 void (* SetupPixelShader)(struct gl_context *ctx);
469 };
470
471 struct r300_vertex_buffer {
472 struct vertex_attribute {
473 /* generic */
474 GLubyte element;
475 GLuint stride;
476 GLuint dwords;
477 GLubyte size; /* number of components */
478 GLboolean is_named_bo;
479 struct radeon_bo *bo;
480 GLint bo_offset;
481
482 /* hw specific */
483 uint32_t data_type:4;
484 uint32_t dst_loc:5;
485 uint32_t _signed:1;
486 uint32_t normalize:1;
487 uint32_t swizzle:12;
488 uint32_t write_mask:4;
489 } attribs[VERT_ATTRIB_MAX];
490
491 GLubyte num_attribs;
492 };
493
494 struct r300_index_buffer {
495 struct radeon_bo *bo;
496 int bo_offset;
497
498 GLboolean is_32bit;
499 GLuint count;
500 };
501
502
503 /**
504 * \brief R300 context structure.
505 */
506 struct r300_context {
507 struct radeon_context radeon; /* parent class, must be first */
508
509 struct r300_vtable vtbl;
510
511 struct r300_hw_state hw;
512
513 struct r300_vertex_program *selected_vp;
514 struct r300_fragment_program *selected_fp;
515
516 /* Vertex buffers
517 */
518 GLvector4f dummy_attrib[_TNL_ATTRIB_MAX];
519 GLvector4f *temp_attrib[_TNL_ATTRIB_MAX];
520
521 struct r300_options {
522 uint32_t conformance_mode:1;
523 uint32_t hw_tcl_enabled:1;
524 uint32_t s3tc_force_enabled:1;
525 uint32_t s3tc_force_disabled:1;
526 uint32_t stencil_two_side_disabled:1;
527 } options;
528
529 struct r300_swtcl_info swtcl;
530 struct r300_vertex_buffer vbuf;
531 struct r300_index_buffer ind_buf;
532
533 uint32_t fallback;
534
535 struct {
536 struct r300_vertex_program_code vp_code;
537 struct rX00_fragment_program_code fp_code;
538 } blit;
539
540 DECLARE_RENDERINPUTS(render_inputs_bitset);
541 };
542
543 #define R300_CONTEXT(ctx) ((r300ContextPtr)(ctx->DriverCtx))
544
545 extern void r300DestroyContext(__DRIcontext * driContextPriv);
546 extern GLboolean r300CreateContext(gl_api api,
547 const struct gl_config * glVisual,
548 __DRIcontext * driContextPriv,
549 void *sharedContextPrivate);
550
551 extern void r300InitShaderFuncs(struct dd_function_table *functions);
552
553 extern void r300InitShaderFunctions(r300ContextPtr r300);
554
555 extern void r300InitDraw(struct gl_context *ctx);
556
557 #define r300PackFloat32 radeonPackFloat32
558 #define r300PackFloat24 radeonPackFloat24
559
560 #endif /* __R300_CONTEXT_H__ */