1/**********************************************************
2 * Copyright 2007-2009 VMware, Inc.  All rights reserved.
3 *
4 * Permission is hereby granted, free of charge, to any person
5 * obtaining a copy of this software and associated documentation
6 * files (the "Software"), to deal in the Software without
7 * restriction, including without limitation the rights to use, copy,
8 * modify, merge, publish, distribute, sublicense, and/or sell copies
9 * of the Software, and to permit persons to whom the Software is
10 * furnished to do so, subject to the following conditions:
11 *
12 * The above copyright notice and this permission notice shall be
13 * included in all copies or substantial portions of the Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
16 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
17 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
18 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
19 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
20 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
21 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22 * SOFTWARE.
23 *
24 **********************************************************/
25
26/*
27 * svga3d_shaderdefs.h --
28 *
29 * SVGA3D byte code format and limit definitions.
30 *
31 * The format of the byte code directly corresponds to that defined
32 * by Microsoft DirectX SDK 9.0c (file d3d9types.h). The format can
33 * also be extended so that different shader formats can be supported
34 * for example GLSL, ARB vp/fp, NV/ATI shader formats, etc.
35 *
36 */
37
38#ifndef __SVGA3D_SHADER_DEFS__
39#define __SVGA3D_SHADER_DEFS__
40
41/* SVGA3D shader hardware limits. */
42
43#define SVGA3D_INPUTREG_MAX            16
44#define SVGA3D_OUTPUTREG_MAX           12
45#define SVGA3D_VERTEX_SAMPLERREG_MAX   4
46#define SVGA3D_PIXEL_SAMPLERREG_MAX    16
47#define SVGA3D_SAMPLERREG_MAX          (SVGA3D_PIXEL_SAMPLERREG_MAX+\
48                                        SVGA3D_VERTEX_SAMPLERREG_MAX)
49#define SVGA3D_TEMPREG_MAX             32
50#define SVGA3D_CONSTREG_MAX            256
51#define SVGA3D_CONSTINTREG_MAX         16
52#define SVGA3D_CONSTBOOLREG_MAX        16
53#define SVGA3D_ADDRREG_MAX             1
54#define SVGA3D_PREDREG_MAX             1
55
56/* SVGA3D byte code specific limits */
57
58#define SVGA3D_MAX_SRC_REGS      4
59#define SVGA3D_MAX_NESTING_LEVEL 32
60
61/* SVGA3D version information. */
62
63#define SVGA3D_VS_TYPE  0xFFFE
64#define SVGA3D_PS_TYPE  0xFFFF
65
66typedef struct {
67   union {
68      struct {
69         uint32 minor : 8;
70         uint32 major : 8;
71         uint32 type : 16;
72      };
73
74      uint32 value;
75   };
76} SVGA3dShaderVersion;
77
78#define SVGA3D_VS_10 ((SVGA3D_VS_TYPE << 16) | 1 << 8)
79#define SVGA3D_VS_11 (SVGA3D_VS_10 | 1)
80#define SVGA3D_VS_20 ((SVGA3D_VS_TYPE << 16) | 2 << 8)
81#define SVGA3D_VS_30 ((SVGA3D_VS_TYPE << 16) | 3 << 8)
82
83#define SVGA3D_PS_10 ((SVGA3D_PS_TYPE << 16) | 1 << 8)
84#define SVGA3D_PS_11 (SVGA3D_PS_10 | 1)
85#define SVGA3D_PS_12 (SVGA3D_PS_10 | 2)
86#define SVGA3D_PS_13 (SVGA3D_PS_10 | 3)
87#define SVGA3D_PS_14 (SVGA3D_PS_10 | 4)
88#define SVGA3D_PS_20 ((SVGA3D_PS_TYPE << 16) | 2 << 8)
89#define SVGA3D_PS_30 ((SVGA3D_PS_TYPE << 16) | 3 << 8)
90
91/* The *_ENABLED are for backwards compatibility with old drivers */
92typedef enum {
93   SVGA3DPSVERSION_NONE = 0,
94   SVGA3DPSVERSION_ENABLED = 1,
95   SVGA3DPSVERSION_11 = 3,
96   SVGA3DPSVERSION_12 = 5,
97   SVGA3DPSVERSION_13 = 7,
98   SVGA3DPSVERSION_14 = 9,
99   SVGA3DPSVERSION_20 = 11,
100   SVGA3DPSVERSION_30 = 13,
101   SVGA3DPSVERSION_40 = 15,
102   SVGA3DPSVERSION_MAX
103} SVGA3dPixelShaderVersion;
104
105typedef enum {
106   SVGA3DVSVERSION_NONE = 0,
107   SVGA3DVSVERSION_ENABLED = 1,
108   SVGA3DVSVERSION_11 = 3,
109   SVGA3DVSVERSION_20 = 5,
110   SVGA3DVSVERSION_30 = 7,
111   SVGA3DVSVERSION_40 = 9,
112   SVGA3DVSVERSION_MAX
113} SVGA3dVertexShaderVersion;
114
115/* SVGA3D instruction op codes. */
116
117typedef enum {
118   SVGA3DOP_NOP = 0,
119   SVGA3DOP_MOV,
120   SVGA3DOP_ADD,
121   SVGA3DOP_SUB,
122   SVGA3DOP_MAD,
123   SVGA3DOP_MUL,
124   SVGA3DOP_RCP,
125   SVGA3DOP_RSQ,
126   SVGA3DOP_DP3,
127   SVGA3DOP_DP4,
128   SVGA3DOP_MIN,
129   SVGA3DOP_MAX,
130   SVGA3DOP_SLT,
131   SVGA3DOP_SGE,
132   SVGA3DOP_EXP,
133   SVGA3DOP_LOG,
134   SVGA3DOP_LIT,
135   SVGA3DOP_DST,
136   SVGA3DOP_LRP,
137   SVGA3DOP_FRC,
138   SVGA3DOP_M4x4,
139   SVGA3DOP_M4x3,
140   SVGA3DOP_M3x4,
141   SVGA3DOP_M3x3,
142   SVGA3DOP_M3x2,
143   SVGA3DOP_CALL,
144   SVGA3DOP_CALLNZ,
145   SVGA3DOP_LOOP,
146   SVGA3DOP_RET,
147   SVGA3DOP_ENDLOOP,
148   SVGA3DOP_LABEL,
149   SVGA3DOP_DCL,
150   SVGA3DOP_POW,
151   SVGA3DOP_CRS,
152   SVGA3DOP_SGN,
153   SVGA3DOP_ABS,
154   SVGA3DOP_NRM,
155   SVGA3DOP_SINCOS,
156   SVGA3DOP_REP,
157   SVGA3DOP_ENDREP,
158   SVGA3DOP_IF,
159   SVGA3DOP_IFC,
160   SVGA3DOP_ELSE,
161   SVGA3DOP_ENDIF,
162   SVGA3DOP_BREAK,
163   SVGA3DOP_BREAKC,
164   SVGA3DOP_MOVA,
165   SVGA3DOP_DEFB,
166   SVGA3DOP_DEFI,
167   SVGA3DOP_TEXCOORD = 64,
168   SVGA3DOP_TEXKILL,
169   SVGA3DOP_TEX,
170   SVGA3DOP_TEXBEM,
171   SVGA3DOP_TEXBEML,
172   SVGA3DOP_TEXREG2AR,
173   SVGA3DOP_TEXREG2GB = 70,
174   SVGA3DOP_TEXM3x2PAD,
175   SVGA3DOP_TEXM3x2TEX,
176   SVGA3DOP_TEXM3x3PAD,
177   SVGA3DOP_TEXM3x3TEX,
178   SVGA3DOP_RESERVED0,
179   SVGA3DOP_TEXM3x3SPEC,
180   SVGA3DOP_TEXM3x3VSPEC,
181   SVGA3DOP_EXPP,
182   SVGA3DOP_LOGP,
183   SVGA3DOP_CND = 80,
184   SVGA3DOP_DEF,
185   SVGA3DOP_TEXREG2RGB,
186   SVGA3DOP_TEXDP3TEX,
187   SVGA3DOP_TEXM3x2DEPTH,
188   SVGA3DOP_TEXDP3,
189   SVGA3DOP_TEXM3x3,
190   SVGA3DOP_TEXDEPTH,
191   SVGA3DOP_CMP,
192   SVGA3DOP_BEM,
193   SVGA3DOP_DP2ADD = 90,
194   SVGA3DOP_DSX,
195   SVGA3DOP_DSY,
196   SVGA3DOP_TEXLDD,
197   SVGA3DOP_SETP,
198   SVGA3DOP_TEXLDL,
199   SVGA3DOP_BREAKP = 96,
200   SVGA3DOP_LAST_INST,
201   SVGA3DOP_PHASE = 0xFFFD,
202   SVGA3DOP_COMMENT = 0xFFFE,
203   SVGA3DOP_END = 0xFFFF,
204} SVGA3dShaderOpCodeType;
205
206/* SVGA3D operation control/comparison function types */
207
208typedef enum {
209   SVGA3DOPCONT_NONE,
210   SVGA3DOPCONT_PROJECT,   /* Projective texturing */
211   SVGA3DOPCONT_BIAS,      /* Texturing with a LOD bias */
212} SVGA3dShaderOpCodeControlFnType;
213
214typedef enum {
215   SVGA3DOPCOMP_RESERVED0 = 0,
216   SVGA3DOPCOMP_GT,
217   SVGA3DOPCOMP_EQ,
218   SVGA3DOPCOMP_GE,
219   SVGA3DOPCOMP_LT,
220   SVGA3DOPCOMPC_NE,
221   SVGA3DOPCOMP_LE,
222   SVGA3DOPCOMP_RESERVED1
223} SVGA3dShaderOpCodeCompFnType;
224
225/* SVGA3D register types */
226
227typedef enum {
228    SVGA3DREG_TEMP = 0,       /* Temporary register file */
229    SVGA3DREG_INPUT,          /* Input register file */
230    SVGA3DREG_CONST,          /* Constant register file */
231    SVGA3DREG_ADDR,           /* Address register for VS */
232    SVGA3DREG_TEXTURE = 3,    /* Texture register file for PS */
233    SVGA3DREG_RASTOUT,        /* Rasterizer register file */
234    SVGA3DREG_ATTROUT,        /* Attribute output register file */
235    SVGA3DREG_TEXCRDOUT,      /* Texture coordinate output register file */
236    SVGA3DREG_OUTPUT = 6,     /* Output register file for VS 3.0+ */
237    SVGA3DREG_CONSTINT,       /* Constant integer vector register file */
238    SVGA3DREG_COLOROUT,       /* Color output register file */
239    SVGA3DREG_DEPTHOUT,       /* Depth output register file */
240    SVGA3DREG_SAMPLER,        /* Sampler state register file */
241    SVGA3DREG_CONST2,         /* Constant register file 2048 - 4095 */
242    SVGA3DREG_CONST3,         /* Constant register file 4096 - 6143 */
243    SVGA3DREG_CONST4,         /* Constant register file 6144 - 8191 */
244    SVGA3DREG_CONSTBOOL,      /* Constant boolean register file */
245    SVGA3DREG_LOOP,           /* Loop counter register file */
246    SVGA3DREG_TEMPFLOAT16,    /* 16-bit float temp register file */
247    SVGA3DREG_MISCTYPE,       /* Miscellaneous (single) registers */
248    SVGA3DREG_LABEL,          /* Label */
249    SVGA3DREG_PREDICATE,      /* Predicate register */
250} SVGA3dShaderRegType;
251
252/* SVGA3D rasterizer output register types */
253
254typedef enum {
255   SVGA3DRASTOUT_POSITION = 0,
256   SVGA3DRASTOUT_FOG,
257   SVGA3DRASTOUT_PSIZE
258} SVGA3dShaderRastOutRegType;
259
260/* SVGA3D miscellaneous register types */
261
262typedef enum {
263   SVGA3DMISCREG_POSITION = 0,   /* Input position x,y,z,rhw (PS) */
264   SVGA3DMISCREG_FACE            /* Floating point primitive area (PS) */
265} SVGA3DShaderMiscRegType;
266
267/* SVGA3D sampler types */
268
269typedef enum {
270   SVGA3DSAMP_UNKNOWN = 0, /* Uninitialized value */
271   SVGA3DSAMP_2D = 2,      /* dcl_2d s# (for declaring a 2-D texture) */
272   SVGA3DSAMP_CUBE,        /* dcl_cube s# (for declaring a cube texture) */
273   SVGA3DSAMP_VOLUME,      /* dcl_volume s# (for declaring a volume texture) */
274} SVGA3dShaderSamplerType;
275
276/* SVGA3D sampler format classes */
277
278typedef enum {
279   SVGA3DSAMPFORMAT_ARGB,        /* ARGB formats */
280   SVGA3DSAMPFORMAT_V8U8,        /* Sign and normalize (SNORM) V & U */
281   SVGA3DSAMPFORMAT_Q8W8V8U8,    /* SNORM all */
282   SVGA3DSAMPFORMAT_CxV8U8,      /* SNORM V & U, C=SQRT(1-U^2-V^2) */
283   SVGA3DSAMPFORMAT_X8L8V8U8,    /* SNORM V & U */
284   SVGA3DSAMPFORMAT_A2W10V10U10, /* SNORM W, V & U */
285   SVGA3DSAMPFORMAT_DXT_PMA,     /* DXT pre-multiplied alpha */
286   SVGA3DSAMPFORMAT_YUV,         /* YUV video format */
287   SVGA3DSAMPFORMAT_UYVY,        /* UYVY video format */
288   SVGA3DSAMPFORMAT_Rx,          /* R16F/32F */
289   SVGA3DSAMPFORMAT_RxGx,        /* R16FG16F, R32FG32F */
290   SVGA3DSAMPFORMAT_V16U16,      /* SNORM all */
291} SVGA3DShaderSamplerFormatClass;
292
293/* SVGA3D write mask */
294
295#define SVGA3DWRITEMASK_0    1 /* Component 0 (X;Red) */
296#define SVGA3DWRITEMASK_1    2 /* Component 1 (Y;Green) */
297#define SVGA3DWRITEMASK_2    4 /* Component 2 (Z;Blue) */
298#define SVGA3DWRITEMASK_3    8 /* Component 3 (W;Alpha) */
299#define SVGA3DWRITEMASK_ALL 15 /* All components */
300
301/* SVGA3D destination modifiers */
302
303#define SVGA3DDSTMOD_NONE              0 /* nop */
304#define SVGA3DDSTMOD_SATURATE          1 /* clamp to [0, 1] */
305#define SVGA3DDSTMOD_PARTIALPRECISION  2 /* Partial precision hint */
306
307/*
308 * Relevant to multisampling only:
309 * When the pixel center is not covered, sample
310 * attribute or compute gradients/LOD
311 * using multisample "centroid" location.
312 * "Centroid" is some location within the covered
313 * region of the pixel.
314 */
315
316#define SVGA3DDSTMOD_MSAMPCENTROID     4
317
318/* SVGA3D source swizzle */
319
320#define SVGA3DSWIZZLE_REPLICATEX 0x00
321#define SVGA3DSWIZZLE_REPLICATEY 0x55
322#define SVGA3DSWIZZLE_REPLICATEZ 0xAA
323#define SVGA3DSWIZZLE_REPLICATEW 0xFF
324#define SVGA3DSWIZZLE_NONE       0xE4
325#define SVGA3DSWIZZLE_YZXW       0xC9
326#define SVGA3DSWIZZLE_ZXYW       0xD2
327#define SVGA3DSWIZZLE_WXYZ       0x1B
328
329/* SVGA3D source modifiers */
330
331typedef enum {
332    SVGA3DSRCMOD_NONE = 0, /* nop */
333    SVGA3DSRCMOD_NEG,      /* negate */
334    SVGA3DSRCMOD_BIAS,     /* bias */
335    SVGA3DSRCMOD_BIASNEG,  /* bias and negate */
336    SVGA3DSRCMOD_SIGN,     /* sign */
337    SVGA3DSRCMOD_SIGNNEG,  /* sign and negate */
338    SVGA3DSRCMOD_COMP,     /* complement */
339    SVGA3DSRCMOD_X2,       /* x2 */
340    SVGA3DSRCMOD_X2NEG,    /* x2 and negate */
341    SVGA3DSRCMOD_DZ,       /* divide through by z component */
342    SVGA3DSRCMOD_DW,       /* divide through by w component */
343    SVGA3DSRCMOD_ABS,      /* abs() */
344    SVGA3DSRCMOD_ABSNEG,   /* -abs() */
345    SVGA3DSRCMOD_NOT,      /* ! (for predicate register) */
346} SVGA3dShaderSrcModType;
347
348/* SVGA3D instruction token */
349
350typedef struct {
351   union {
352      struct {
353         uint32 comment_op : 16;
354         uint32 comment_size : 16;
355      };
356
357      struct {
358         uint32 op : 16;
359         uint32 control : 3;
360         uint32 reserved2 : 5;
361         uint32 size : 4;
362         uint32 predicated : 1;
363         uint32 reserved1 : 1;
364         uint32 coissue : 1;
365         uint32 reserved0 : 1;
366      };
367
368      uint32 value;
369   };
370} SVGA3dShaderInstToken;
371
372/* SVGA3D destination parameter token */
373
374typedef struct {
375   union {
376      struct {
377         uint32 num : 11;
378         uint32 type_upper : 2;
379         uint32 relAddr : 1;
380         uint32 reserved1 : 2;
381         uint32 mask : 4;
382         uint32 dstMod : 4;
383         uint32 shfScale : 4;
384         uint32 type_lower : 3;
385         uint32 reserved0 : 1;
386      };
387
388      uint32 value;
389   };
390} SVGA3dShaderDestToken;
391
392/* SVGA3D source parameter token */
393
394typedef struct {
395   union {
396      struct {
397         uint32 num : 11;
398         uint32 type_upper : 2;
399         uint32 relAddr : 1;
400         uint32 reserved1 : 2;
401         uint32 swizzle : 8;
402         uint32 srcMod : 4;
403         uint32 type_lower : 3;
404         uint32 reserved0 : 1;
405      };
406
407      uint32 value;
408   };
409} SVGA3dShaderSrcToken;
410
411/* SVGA3DOP_DCL parameter tokens */
412
413typedef struct {
414   union {
415      struct {
416         union {
417            struct {
418               uint32 usage : 5;
419               uint32 reserved1 : 11;
420               uint32 index : 4;
421               uint32 reserved0 : 12;
422            }; /* input / output declaration */
423
424            struct {
425               uint32 reserved3 : 27;
426               uint32 type : 4;
427               uint32 reserved2 : 1;
428            }; /* sampler declaration */
429         };
430
431         SVGA3dShaderDestToken dst;
432      };
433
434      uint32 values[2];
435   };
436} SVGA3DOpDclArgs;
437
438/* SVGA3DOP_DEF parameter tokens */
439
440typedef struct {
441   union {
442      struct {
443         SVGA3dShaderDestToken dst;
444
445         union {
446            float constValues[4];
447            int constIValues[4];
448            Bool constBValue;
449         };
450      };
451
452      uint32 values[5];
453   };
454} SVGA3DOpDefArgs;
455
456/* SVGA3D shader token */
457
458typedef union {
459   uint32 value;
460   SVGA3dShaderInstToken inst;
461   SVGA3dShaderDestToken dest;
462   SVGA3dShaderSrcToken src;
463} SVGA3dShaderToken;
464
465/* SVGA3D shader program */
466
467typedef struct {
468   SVGA3dShaderVersion version;
469   /* SVGA3dShaderToken stream */
470} SVGA3dShaderProgram;
471
472/* SVGA3D version specific register assignments */
473
474static const uint32 SVGA3D_INPUT_REG_POSITION_VS11 = 0;
475static const uint32 SVGA3D_INPUT_REG_PSIZE_VS11 = 1;
476static const uint32 SVGA3D_INPUT_REG_FOG_VS11 = 3;
477static const uint32 SVGA3D_INPUT_REG_FOG_MASK_VS11 = SVGA3DWRITEMASK_3;
478static const uint32 SVGA3D_INPUT_REG_COLOR_BASE_VS11 = 2;
479static const uint32 SVGA3D_INPUT_REG_TEXCOORD_BASE_VS11 = 4;
480
481static const uint32 SVGA3D_INPUT_REG_COLOR_BASE_PS11 = 0;
482static const uint32 SVGA3D_INPUT_REG_TEXCOORD_BASE_PS11 = 2;
483static const uint32 SVGA3D_OUTPUT_REG_DEPTH_PS11 = 0;
484static const uint32 SVGA3D_OUTPUT_REG_COLOR_PS11 = 1;
485
486static const uint32 SVGA3D_INPUT_REG_COLOR_BASE_PS20 = 0;
487static const uint32 SVGA3D_INPUT_REG_COLOR_NUM_PS20 = 2;
488static const uint32 SVGA3D_INPUT_REG_TEXCOORD_BASE_PS20 = 2;
489static const uint32 SVGA3D_INPUT_REG_TEXCOORD_NUM_PS20 = 8;
490static const uint32 SVGA3D_OUTPUT_REG_COLOR_BASE_PS20 = 1;
491static const uint32 SVGA3D_OUTPUT_REG_COLOR_NUM_PS20 = 4;
492static const uint32 SVGA3D_OUTPUT_REG_DEPTH_BASE_PS20 = 0;
493static const uint32 SVGA3D_OUTPUT_REG_DEPTH_NUM_PS20 = 1;
494
495/*
496 *----------------------------------------------------------------------
497 *
498 * SVGA3dShaderGetRegType --
499 *
500 *      As the register type is split into two non sequential fields,
501 *      this function provides an useful way of accessing the actual
502 *      register type without having to manually concatenate the
503 *      type_upper and type_lower fields.
504 *
505 * Results:
506 *      Returns the register type.
507 *
508 *----------------------------------------------------------------------
509 */
510
511static INLINE SVGA3dShaderRegType
512SVGA3dShaderGetRegType(uint32 token)
513{
514   SVGA3dShaderSrcToken src;
515   src.value = token;
516   return (SVGA3dShaderRegType)(src.type_upper << 3 | src.type_lower);
517}
518
519#endif /* __SVGA3D_SHADER_DEFS__ */
520