rasapi32: Update spec file.
[wine/zf.git] / dlls / wined3d / shader_sm4.c
blob4635910269fbb98b4eb835106e651beb0f7c1ac1
1 /*
2 * Copyright 2009 Henri Verbeet for CodeWeavers
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
9 * This library is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with this library; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
19 #include "config.h"
20 #include "wine/port.h"
22 #include "wined3d_private.h"
24 WINE_DEFAULT_DEBUG_CHANNEL(d3d_shader);
25 WINE_DECLARE_DEBUG_CHANNEL(d3d_bytecode);
27 #define WINED3D_SM4_INSTRUCTION_MODIFIER (0x1u << 31)
29 #define WINED3D_SM4_MODIFIER_MASK 0x3fu
31 #define WINED3D_SM5_MODIFIER_DATA_TYPE_SHIFT 6
32 #define WINED3D_SM5_MODIFIER_DATA_TYPE_MASK (0xffffu << WINED3D_SM5_MODIFIER_DATA_TYPE_SHIFT)
34 #define WINED3D_SM5_MODIFIER_RESOURCE_TYPE_SHIFT 6
35 #define WINED3D_SM5_MODIFIER_RESOURCE_TYPE_MASK (0xfu << WINED3D_SM5_MODIFIER_RESOURCE_TYPE_SHIFT)
37 #define WINED3D_SM4_AOFFIMMI_U_SHIFT 9
38 #define WINED3D_SM4_AOFFIMMI_U_MASK (0xfu << WINED3D_SM4_AOFFIMMI_U_SHIFT)
39 #define WINED3D_SM4_AOFFIMMI_V_SHIFT 13
40 #define WINED3D_SM4_AOFFIMMI_V_MASK (0xfu << WINED3D_SM4_AOFFIMMI_V_SHIFT)
41 #define WINED3D_SM4_AOFFIMMI_W_SHIFT 17
42 #define WINED3D_SM4_AOFFIMMI_W_MASK (0xfu << WINED3D_SM4_AOFFIMMI_W_SHIFT)
44 #define WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT 24
45 #define WINED3D_SM4_INSTRUCTION_LENGTH_MASK (0x1fu << WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT)
47 #define WINED3D_SM4_INSTRUCTION_FLAGS_SHIFT 11
48 #define WINED3D_SM4_INSTRUCTION_FLAGS_MASK (0x7u << WINED3D_SM4_INSTRUCTION_FLAGS_SHIFT)
50 #define WINED3D_SM4_RESOURCE_TYPE_SHIFT 11
51 #define WINED3D_SM4_RESOURCE_TYPE_MASK (0xfu << WINED3D_SM4_RESOURCE_TYPE_SHIFT)
53 #define WINED3D_SM4_PRIMITIVE_TYPE_SHIFT 11
54 #define WINED3D_SM4_PRIMITIVE_TYPE_MASK (0x3fu << WINED3D_SM4_PRIMITIVE_TYPE_SHIFT)
56 #define WINED3D_SM4_INDEX_TYPE_SHIFT 11
57 #define WINED3D_SM4_INDEX_TYPE_MASK (0x1u << WINED3D_SM4_INDEX_TYPE_SHIFT)
59 #define WINED3D_SM4_SAMPLER_MODE_SHIFT 11
60 #define WINED3D_SM4_SAMPLER_MODE_MASK (0xfu << WINED3D_SM4_SAMPLER_MODE_SHIFT)
62 #define WINED3D_SM4_SHADER_DATA_TYPE_SHIFT 11
63 #define WINED3D_SM4_SHADER_DATA_TYPE_MASK (0xfu << WINED3D_SM4_SHADER_DATA_TYPE_SHIFT)
65 #define WINED3D_SM4_INTERPOLATION_MODE_SHIFT 11
66 #define WINED3D_SM4_INTERPOLATION_MODE_MASK (0xfu << WINED3D_SM4_INTERPOLATION_MODE_SHIFT)
68 #define WINED3D_SM4_GLOBAL_FLAGS_SHIFT 11
69 #define WINED3D_SM4_GLOBAL_FLAGS_MASK (0xffu << WINED3D_SM4_GLOBAL_FLAGS_SHIFT)
71 #define WINED3D_SM5_PRECISE_SHIFT 19
72 #define WINED3D_SM5_PRECISE_MASK (0xfu << WINED3D_SM5_PRECISE_SHIFT)
74 #define WINED3D_SM5_CONTROL_POINT_COUNT_SHIFT 11
75 #define WINED3D_SM5_CONTROL_POINT_COUNT_MASK (0xffu << WINED3D_SM5_CONTROL_POINT_COUNT_SHIFT)
77 #define WINED3D_SM5_FP_ARRAY_SIZE_SHIFT 16
78 #define WINED3D_SM5_FP_TABLE_COUNT_MASK 0xffffu
80 #define WINED3D_SM5_UAV_FLAGS_SHIFT 15
81 #define WINED3D_SM5_UAV_FLAGS_MASK (0x1ffu << WINED3D_SM5_UAV_FLAGS_SHIFT)
83 #define WINED3D_SM5_SYNC_FLAGS_SHIFT 11
84 #define WINED3D_SM5_SYNC_FLAGS_MASK (0xffu << WINED3D_SM5_SYNC_FLAGS_SHIFT)
86 #define WINED3D_SM5_TESSELLATOR_SHIFT 11
87 #define WINED3D_SM5_TESSELLATOR_MASK (0xfu << WINED3D_SM5_TESSELLATOR_SHIFT)
89 #define WINED3D_SM4_OPCODE_MASK 0xff
91 #define WINED3D_SM4_REGISTER_MODIFIER (0x1u << 31)
93 #define WINED3D_SM4_ADDRESSING_SHIFT1 25
94 #define WINED3D_SM4_ADDRESSING_MASK1 (0x3u << WINED3D_SM4_ADDRESSING_SHIFT1)
96 #define WINED3D_SM4_ADDRESSING_SHIFT0 22
97 #define WINED3D_SM4_ADDRESSING_MASK0 (0x3u << WINED3D_SM4_ADDRESSING_SHIFT0)
99 #define WINED3D_SM4_REGISTER_ORDER_SHIFT 20
100 #define WINED3D_SM4_REGISTER_ORDER_MASK (0x3u << WINED3D_SM4_REGISTER_ORDER_SHIFT)
102 #define WINED3D_SM4_REGISTER_TYPE_SHIFT 12
103 #define WINED3D_SM4_REGISTER_TYPE_MASK (0xffu << WINED3D_SM4_REGISTER_TYPE_SHIFT)
105 #define WINED3D_SM4_SWIZZLE_TYPE_SHIFT 2
106 #define WINED3D_SM4_SWIZZLE_TYPE_MASK (0x3u << WINED3D_SM4_SWIZZLE_TYPE_SHIFT)
108 #define WINED3D_SM4_DIMENSION_SHIFT 0
109 #define WINED3D_SM4_DIMENSION_MASK (0x3u << WINED3D_SM4_DIMENSION_SHIFT)
111 #define WINED3D_SM4_WRITEMASK_SHIFT 4
112 #define WINED3D_SM4_WRITEMASK_MASK (0xfu << WINED3D_SM4_WRITEMASK_SHIFT)
114 #define WINED3D_SM4_SWIZZLE_SHIFT 4
115 #define WINED3D_SM4_SWIZZLE_MASK (0xffu << WINED3D_SM4_SWIZZLE_SHIFT)
117 #define WINED3D_SM4_VERSION_MAJOR(version) (((version) >> 4) & 0xf)
118 #define WINED3D_SM4_VERSION_MINOR(version) (((version) >> 0) & 0xf)
120 #define WINED3D_SM4_ADDRESSING_RELATIVE 0x2
121 #define WINED3D_SM4_ADDRESSING_OFFSET 0x1
123 #define WINED3D_SM4_INSTRUCTION_FLAG_SATURATE 0x4
125 #define WINED3D_SM4_CONDITIONAL_NZ (0x1u << 18)
127 enum wined3d_sm4_opcode
129 WINED3D_SM4_OP_ADD = 0x00,
130 WINED3D_SM4_OP_AND = 0x01,
131 WINED3D_SM4_OP_BREAK = 0x02,
132 WINED3D_SM4_OP_BREAKC = 0x03,
133 WINED3D_SM4_OP_CASE = 0x06,
134 WINED3D_SM4_OP_CONTINUE = 0x07,
135 WINED3D_SM4_OP_CONTINUEC = 0x08,
136 WINED3D_SM4_OP_CUT = 0x09,
137 WINED3D_SM4_OP_DEFAULT = 0x0a,
138 WINED3D_SM4_OP_DERIV_RTX = 0x0b,
139 WINED3D_SM4_OP_DERIV_RTY = 0x0c,
140 WINED3D_SM4_OP_DISCARD = 0x0d,
141 WINED3D_SM4_OP_DIV = 0x0e,
142 WINED3D_SM4_OP_DP2 = 0x0f,
143 WINED3D_SM4_OP_DP3 = 0x10,
144 WINED3D_SM4_OP_DP4 = 0x11,
145 WINED3D_SM4_OP_ELSE = 0x12,
146 WINED3D_SM4_OP_EMIT = 0x13,
147 WINED3D_SM4_OP_ENDIF = 0x15,
148 WINED3D_SM4_OP_ENDLOOP = 0x16,
149 WINED3D_SM4_OP_ENDSWITCH = 0x17,
150 WINED3D_SM4_OP_EQ = 0x18,
151 WINED3D_SM4_OP_EXP = 0x19,
152 WINED3D_SM4_OP_FRC = 0x1a,
153 WINED3D_SM4_OP_FTOI = 0x1b,
154 WINED3D_SM4_OP_FTOU = 0x1c,
155 WINED3D_SM4_OP_GE = 0x1d,
156 WINED3D_SM4_OP_IADD = 0x1e,
157 WINED3D_SM4_OP_IF = 0x1f,
158 WINED3D_SM4_OP_IEQ = 0x20,
159 WINED3D_SM4_OP_IGE = 0x21,
160 WINED3D_SM4_OP_ILT = 0x22,
161 WINED3D_SM4_OP_IMAD = 0x23,
162 WINED3D_SM4_OP_IMAX = 0x24,
163 WINED3D_SM4_OP_IMIN = 0x25,
164 WINED3D_SM4_OP_IMUL = 0x26,
165 WINED3D_SM4_OP_INE = 0x27,
166 WINED3D_SM4_OP_INEG = 0x28,
167 WINED3D_SM4_OP_ISHL = 0x29,
168 WINED3D_SM4_OP_ISHR = 0x2a,
169 WINED3D_SM4_OP_ITOF = 0x2b,
170 WINED3D_SM4_OP_LABEL = 0x2c,
171 WINED3D_SM4_OP_LD = 0x2d,
172 WINED3D_SM4_OP_LD2DMS = 0x2e,
173 WINED3D_SM4_OP_LOG = 0x2f,
174 WINED3D_SM4_OP_LOOP = 0x30,
175 WINED3D_SM4_OP_LT = 0x31,
176 WINED3D_SM4_OP_MAD = 0x32,
177 WINED3D_SM4_OP_MIN = 0x33,
178 WINED3D_SM4_OP_MAX = 0x34,
179 WINED3D_SM4_OP_SHADER_DATA = 0x35,
180 WINED3D_SM4_OP_MOV = 0x36,
181 WINED3D_SM4_OP_MOVC = 0x37,
182 WINED3D_SM4_OP_MUL = 0x38,
183 WINED3D_SM4_OP_NE = 0x39,
184 WINED3D_SM4_OP_NOP = 0x3a,
185 WINED3D_SM4_OP_NOT = 0x3b,
186 WINED3D_SM4_OP_OR = 0x3c,
187 WINED3D_SM4_OP_RESINFO = 0x3d,
188 WINED3D_SM4_OP_RET = 0x3e,
189 WINED3D_SM4_OP_RETC = 0x3f,
190 WINED3D_SM4_OP_ROUND_NE = 0x40,
191 WINED3D_SM4_OP_ROUND_NI = 0x41,
192 WINED3D_SM4_OP_ROUND_PI = 0x42,
193 WINED3D_SM4_OP_ROUND_Z = 0x43,
194 WINED3D_SM4_OP_RSQ = 0x44,
195 WINED3D_SM4_OP_SAMPLE = 0x45,
196 WINED3D_SM4_OP_SAMPLE_C = 0x46,
197 WINED3D_SM4_OP_SAMPLE_C_LZ = 0x47,
198 WINED3D_SM4_OP_SAMPLE_LOD = 0x48,
199 WINED3D_SM4_OP_SAMPLE_GRAD = 0x49,
200 WINED3D_SM4_OP_SAMPLE_B = 0x4a,
201 WINED3D_SM4_OP_SQRT = 0x4b,
202 WINED3D_SM4_OP_SWITCH = 0x4c,
203 WINED3D_SM4_OP_SINCOS = 0x4d,
204 WINED3D_SM4_OP_UDIV = 0x4e,
205 WINED3D_SM4_OP_ULT = 0x4f,
206 WINED3D_SM4_OP_UGE = 0x50,
207 WINED3D_SM4_OP_UMUL = 0x51,
208 WINED3D_SM4_OP_UMAX = 0x53,
209 WINED3D_SM4_OP_UMIN = 0x54,
210 WINED3D_SM4_OP_USHR = 0x55,
211 WINED3D_SM4_OP_UTOF = 0x56,
212 WINED3D_SM4_OP_XOR = 0x57,
213 WINED3D_SM4_OP_DCL_RESOURCE = 0x58,
214 WINED3D_SM4_OP_DCL_CONSTANT_BUFFER = 0x59,
215 WINED3D_SM4_OP_DCL_SAMPLER = 0x5a,
216 WINED3D_SM4_OP_DCL_INDEX_RANGE = 0x5b,
217 WINED3D_SM4_OP_DCL_OUTPUT_TOPOLOGY = 0x5c,
218 WINED3D_SM4_OP_DCL_INPUT_PRIMITIVE = 0x5d,
219 WINED3D_SM4_OP_DCL_VERTICES_OUT = 0x5e,
220 WINED3D_SM4_OP_DCL_INPUT = 0x5f,
221 WINED3D_SM4_OP_DCL_INPUT_SGV = 0x60,
222 WINED3D_SM4_OP_DCL_INPUT_SIV = 0x61,
223 WINED3D_SM4_OP_DCL_INPUT_PS = 0x62,
224 WINED3D_SM4_OP_DCL_INPUT_PS_SGV = 0x63,
225 WINED3D_SM4_OP_DCL_INPUT_PS_SIV = 0x64,
226 WINED3D_SM4_OP_DCL_OUTPUT = 0x65,
227 WINED3D_SM4_OP_DCL_OUTPUT_SIV = 0x67,
228 WINED3D_SM4_OP_DCL_TEMPS = 0x68,
229 WINED3D_SM4_OP_DCL_INDEXABLE_TEMP = 0x69,
230 WINED3D_SM4_OP_DCL_GLOBAL_FLAGS = 0x6a,
231 WINED3D_SM4_OP_LOD = 0x6c,
232 WINED3D_SM4_OP_GATHER4 = 0x6d,
233 WINED3D_SM4_OP_SAMPLE_POS = 0x6e,
234 WINED3D_SM4_OP_SAMPLE_INFO = 0x6f,
235 WINED3D_SM5_OP_HS_DECLS = 0x71,
236 WINED3D_SM5_OP_HS_CONTROL_POINT_PHASE = 0x72,
237 WINED3D_SM5_OP_HS_FORK_PHASE = 0x73,
238 WINED3D_SM5_OP_HS_JOIN_PHASE = 0x74,
239 WINED3D_SM5_OP_EMIT_STREAM = 0x75,
240 WINED3D_SM5_OP_CUT_STREAM = 0x76,
241 WINED3D_SM5_OP_FCALL = 0x78,
242 WINED3D_SM5_OP_BUFINFO = 0x79,
243 WINED3D_SM5_OP_DERIV_RTX_COARSE = 0x7a,
244 WINED3D_SM5_OP_DERIV_RTX_FINE = 0x7b,
245 WINED3D_SM5_OP_DERIV_RTY_COARSE = 0x7c,
246 WINED3D_SM5_OP_DERIV_RTY_FINE = 0x7d,
247 WINED3D_SM5_OP_GATHER4_C = 0x7e,
248 WINED3D_SM5_OP_GATHER4_PO = 0x7f,
249 WINED3D_SM5_OP_GATHER4_PO_C = 0x80,
250 WINED3D_SM5_OP_RCP = 0x81,
251 WINED3D_SM5_OP_F32TOF16 = 0x82,
252 WINED3D_SM5_OP_F16TOF32 = 0x83,
253 WINED3D_SM5_OP_COUNTBITS = 0x86,
254 WINED3D_SM5_OP_FIRSTBIT_HI = 0x87,
255 WINED3D_SM5_OP_FIRSTBIT_LO = 0x88,
256 WINED3D_SM5_OP_FIRSTBIT_SHI = 0x89,
257 WINED3D_SM5_OP_UBFE = 0x8a,
258 WINED3D_SM5_OP_IBFE = 0x8b,
259 WINED3D_SM5_OP_BFI = 0x8c,
260 WINED3D_SM5_OP_BFREV = 0x8d,
261 WINED3D_SM5_OP_SWAPC = 0x8e,
262 WINED3D_SM5_OP_DCL_STREAM = 0x8f,
263 WINED3D_SM5_OP_DCL_FUNCTION_BODY = 0x90,
264 WINED3D_SM5_OP_DCL_FUNCTION_TABLE = 0x91,
265 WINED3D_SM5_OP_DCL_INTERFACE = 0x92,
266 WINED3D_SM5_OP_DCL_INPUT_CONTROL_POINT_COUNT = 0x93,
267 WINED3D_SM5_OP_DCL_OUTPUT_CONTROL_POINT_COUNT = 0x94,
268 WINED3D_SM5_OP_DCL_TESSELLATOR_DOMAIN = 0x95,
269 WINED3D_SM5_OP_DCL_TESSELLATOR_PARTITIONING = 0x96,
270 WINED3D_SM5_OP_DCL_TESSELLATOR_OUTPUT_PRIMITIVE = 0x97,
271 WINED3D_SM5_OP_DCL_HS_MAX_TESSFACTOR = 0x98,
272 WINED3D_SM5_OP_DCL_HS_FORK_PHASE_INSTANCE_COUNT = 0x99,
273 WINED3D_SM5_OP_DCL_HS_JOIN_PHASE_INSTANCE_COUNT = 0x9a,
274 WINED3D_SM5_OP_DCL_THREAD_GROUP = 0x9b,
275 WINED3D_SM5_OP_DCL_UAV_TYPED = 0x9c,
276 WINED3D_SM5_OP_DCL_UAV_RAW = 0x9d,
277 WINED3D_SM5_OP_DCL_UAV_STRUCTURED = 0x9e,
278 WINED3D_SM5_OP_DCL_TGSM_RAW = 0x9f,
279 WINED3D_SM5_OP_DCL_TGSM_STRUCTURED = 0xa0,
280 WINED3D_SM5_OP_DCL_RESOURCE_RAW = 0xa1,
281 WINED3D_SM5_OP_DCL_RESOURCE_STRUCTURED = 0xa2,
282 WINED3D_SM5_OP_LD_UAV_TYPED = 0xa3,
283 WINED3D_SM5_OP_STORE_UAV_TYPED = 0xa4,
284 WINED3D_SM5_OP_LD_RAW = 0xa5,
285 WINED3D_SM5_OP_STORE_RAW = 0xa6,
286 WINED3D_SM5_OP_LD_STRUCTURED = 0xa7,
287 WINED3D_SM5_OP_STORE_STRUCTURED = 0xa8,
288 WINED3D_SM5_OP_ATOMIC_AND = 0xa9,
289 WINED3D_SM5_OP_ATOMIC_OR = 0xaa,
290 WINED3D_SM5_OP_ATOMIC_XOR = 0xab,
291 WINED3D_SM5_OP_ATOMIC_CMP_STORE = 0xac,
292 WINED3D_SM5_OP_ATOMIC_IADD = 0xad,
293 WINED3D_SM5_OP_ATOMIC_IMAX = 0xae,
294 WINED3D_SM5_OP_ATOMIC_IMIN = 0xaf,
295 WINED3D_SM5_OP_ATOMIC_UMAX = 0xb0,
296 WINED3D_SM5_OP_ATOMIC_UMIN = 0xb1,
297 WINED3D_SM5_OP_IMM_ATOMIC_ALLOC = 0xb2,
298 WINED3D_SM5_OP_IMM_ATOMIC_CONSUME = 0xb3,
299 WINED3D_SM5_OP_IMM_ATOMIC_IADD = 0xb4,
300 WINED3D_SM5_OP_IMM_ATOMIC_AND = 0xb5,
301 WINED3D_SM5_OP_IMM_ATOMIC_OR = 0xb6,
302 WINED3D_SM5_OP_IMM_ATOMIC_XOR = 0xb7,
303 WINED3D_SM5_OP_IMM_ATOMIC_EXCH = 0xb8,
304 WINED3D_SM5_OP_IMM_ATOMIC_CMP_EXCH = 0xb9,
305 WINED3D_SM5_OP_IMM_ATOMIC_IMAX = 0xba,
306 WINED3D_SM5_OP_IMM_ATOMIC_IMIN = 0xbb,
307 WINED3D_SM5_OP_IMM_ATOMIC_UMAX = 0xbc,
308 WINED3D_SM5_OP_IMM_ATOMIC_UMIN = 0xbd,
309 WINED3D_SM5_OP_SYNC = 0xbe,
310 WINED3D_SM5_OP_EVAL_SAMPLE_INDEX = 0xcc,
311 WINED3D_SM5_OP_DCL_GS_INSTANCES = 0xce,
314 enum wined3d_sm4_instruction_modifier
316 WINED3D_SM4_MODIFIER_AOFFIMMI = 0x1,
317 WINED3D_SM5_MODIFIER_RESOURCE_TYPE = 0x2,
318 WINED3D_SM5_MODIFIER_DATA_TYPE = 0x3,
321 enum wined3d_sm4_register_type
323 WINED3D_SM4_RT_TEMP = 0x00,
324 WINED3D_SM4_RT_INPUT = 0x01,
325 WINED3D_SM4_RT_OUTPUT = 0x02,
326 WINED3D_SM4_RT_INDEXABLE_TEMP = 0x03,
327 WINED3D_SM4_RT_IMMCONST = 0x04,
328 WINED3D_SM4_RT_SAMPLER = 0x06,
329 WINED3D_SM4_RT_RESOURCE = 0x07,
330 WINED3D_SM4_RT_CONSTBUFFER = 0x08,
331 WINED3D_SM4_RT_IMMCONSTBUFFER = 0x09,
332 WINED3D_SM4_RT_PRIMID = 0x0b,
333 WINED3D_SM4_RT_DEPTHOUT = 0x0c,
334 WINED3D_SM4_RT_NULL = 0x0d,
335 WINED3D_SM4_RT_RASTERIZER = 0x0e,
336 WINED3D_SM4_RT_OMASK = 0x0f,
337 WINED3D_SM5_RT_STREAM = 0x10,
338 WINED3D_SM5_RT_FUNCTION_BODY = 0x11,
339 WINED3D_SM5_RT_FUNCTION_POINTER = 0x13,
340 WINED3D_SM5_RT_OUTPUT_CONTROL_POINT_ID = 0x16,
341 WINED3D_SM5_RT_FORK_INSTANCE_ID = 0x17,
342 WINED3D_SM5_RT_JOIN_INSTANCE_ID = 0x18,
343 WINED3D_SM5_RT_INPUT_CONTROL_POINT = 0x19,
344 WINED3D_SM5_RT_OUTPUT_CONTROL_POINT = 0x1a,
345 WINED3D_SM5_RT_PATCH_CONSTANT_DATA = 0x1b,
346 WINED3D_SM5_RT_DOMAIN_LOCATION = 0x1c,
347 WINED3D_SM5_RT_UAV = 0x1e,
348 WINED3D_SM5_RT_SHARED_MEMORY = 0x1f,
349 WINED3D_SM5_RT_THREAD_ID = 0x20,
350 WINED3D_SM5_RT_THREAD_GROUP_ID = 0x21,
351 WINED3D_SM5_RT_LOCAL_THREAD_ID = 0x22,
352 WINED3D_SM5_RT_COVERAGE = 0x23,
353 WINED3D_SM5_RT_LOCAL_THREAD_INDEX = 0x24,
354 WINED3D_SM5_RT_GS_INSTANCE_ID = 0x25,
355 WINED3D_SM5_RT_DEPTHOUT_GREATER_EQUAL = 0x26,
356 WINED3D_SM5_RT_DEPTHOUT_LESS_EQUAL = 0x27,
359 enum wined3d_sm4_output_primitive_type
361 WINED3D_SM4_OUTPUT_PT_POINTLIST = 0x1,
362 WINED3D_SM4_OUTPUT_PT_LINESTRIP = 0x3,
363 WINED3D_SM4_OUTPUT_PT_TRIANGLESTRIP = 0x5,
366 enum wined3d_sm4_input_primitive_type
368 WINED3D_SM4_INPUT_PT_POINT = 0x01,
369 WINED3D_SM4_INPUT_PT_LINE = 0x02,
370 WINED3D_SM4_INPUT_PT_TRIANGLE = 0x03,
371 WINED3D_SM4_INPUT_PT_LINEADJ = 0x06,
372 WINED3D_SM4_INPUT_PT_TRIANGLEADJ = 0x07,
373 WINED3D_SM5_INPUT_PT_PATCH1 = 0x08,
374 WINED3D_SM5_INPUT_PT_PATCH2 = 0x09,
375 WINED3D_SM5_INPUT_PT_PATCH3 = 0x0a,
376 WINED3D_SM5_INPUT_PT_PATCH4 = 0x0b,
377 WINED3D_SM5_INPUT_PT_PATCH5 = 0x0c,
378 WINED3D_SM5_INPUT_PT_PATCH6 = 0x0d,
379 WINED3D_SM5_INPUT_PT_PATCH7 = 0x0e,
380 WINED3D_SM5_INPUT_PT_PATCH8 = 0x0f,
381 WINED3D_SM5_INPUT_PT_PATCH9 = 0x10,
382 WINED3D_SM5_INPUT_PT_PATCH10 = 0x11,
383 WINED3D_SM5_INPUT_PT_PATCH11 = 0x12,
384 WINED3D_SM5_INPUT_PT_PATCH12 = 0x13,
385 WINED3D_SM5_INPUT_PT_PATCH13 = 0x14,
386 WINED3D_SM5_INPUT_PT_PATCH14 = 0x15,
387 WINED3D_SM5_INPUT_PT_PATCH15 = 0x16,
388 WINED3D_SM5_INPUT_PT_PATCH16 = 0x17,
389 WINED3D_SM5_INPUT_PT_PATCH17 = 0x18,
390 WINED3D_SM5_INPUT_PT_PATCH18 = 0x19,
391 WINED3D_SM5_INPUT_PT_PATCH19 = 0x1a,
392 WINED3D_SM5_INPUT_PT_PATCH20 = 0x1b,
393 WINED3D_SM5_INPUT_PT_PATCH21 = 0x1c,
394 WINED3D_SM5_INPUT_PT_PATCH22 = 0x1d,
395 WINED3D_SM5_INPUT_PT_PATCH23 = 0x1e,
396 WINED3D_SM5_INPUT_PT_PATCH24 = 0x1f,
397 WINED3D_SM5_INPUT_PT_PATCH25 = 0x20,
398 WINED3D_SM5_INPUT_PT_PATCH26 = 0x21,
399 WINED3D_SM5_INPUT_PT_PATCH27 = 0x22,
400 WINED3D_SM5_INPUT_PT_PATCH28 = 0x23,
401 WINED3D_SM5_INPUT_PT_PATCH29 = 0x24,
402 WINED3D_SM5_INPUT_PT_PATCH30 = 0x25,
403 WINED3D_SM5_INPUT_PT_PATCH31 = 0x26,
404 WINED3D_SM5_INPUT_PT_PATCH32 = 0x27,
407 enum wined3d_sm4_swizzle_type
409 WINED3D_SM4_SWIZZLE_NONE = 0x0,
410 WINED3D_SM4_SWIZZLE_VEC4 = 0x1,
411 WINED3D_SM4_SWIZZLE_SCALAR = 0x2,
414 enum wined3d_sm4_dimension
416 WINED3D_SM4_DIMENSION_SCALAR = 0x1,
417 WINED3D_SM4_DIMENSION_VEC4 = 0x2,
420 enum wined3d_sm4_resource_type
422 WINED3D_SM4_RESOURCE_BUFFER = 0x1,
423 WINED3D_SM4_RESOURCE_TEXTURE_1D = 0x2,
424 WINED3D_SM4_RESOURCE_TEXTURE_2D = 0x3,
425 WINED3D_SM4_RESOURCE_TEXTURE_2DMS = 0x4,
426 WINED3D_SM4_RESOURCE_TEXTURE_3D = 0x5,
427 WINED3D_SM4_RESOURCE_TEXTURE_CUBE = 0x6,
428 WINED3D_SM4_RESOURCE_TEXTURE_1DARRAY = 0x7,
429 WINED3D_SM4_RESOURCE_TEXTURE_2DARRAY = 0x8,
430 WINED3D_SM4_RESOURCE_TEXTURE_2DMSARRAY = 0x9,
431 WINED3D_SM4_RESOURCE_TEXTURE_CUBEARRAY = 0xa,
434 enum wined3d_sm4_data_type
436 WINED3D_SM4_DATA_UNORM = 0x1,
437 WINED3D_SM4_DATA_SNORM = 0x2,
438 WINED3D_SM4_DATA_INT = 0x3,
439 WINED3D_SM4_DATA_UINT = 0x4,
440 WINED3D_SM4_DATA_FLOAT = 0x5,
443 enum wined3d_sm4_sampler_mode
445 WINED3D_SM4_SAMPLER_DEFAULT = 0x0,
446 WINED3D_SM4_SAMPLER_COMPARISON = 0x1,
449 enum wined3d_sm4_shader_data_type
451 WINED3D_SM4_SHADER_DATA_IMMEDIATE_CONSTANT_BUFFER = 0x3,
452 WINED3D_SM4_SHADER_DATA_MESSAGE = 0x4,
455 struct wined3d_shader_src_param_entry
457 struct list entry;
458 struct wined3d_shader_src_param param;
461 struct wined3d_sm4_data
463 struct wined3d_shader_version shader_version;
464 const DWORD *start, *end;
466 unsigned int output_map[MAX_REG_OUTPUT];
468 struct wined3d_shader_src_param src_param[5];
469 struct wined3d_shader_dst_param dst_param[2];
470 struct list src_free;
471 struct list src;
472 struct wined3d_shader_immediate_constant_buffer icb;
475 struct wined3d_sm4_opcode_info
477 enum wined3d_sm4_opcode opcode;
478 enum WINED3D_SHADER_INSTRUCTION_HANDLER handler_idx;
479 const char *dst_info;
480 const char *src_info;
481 void (*read_opcode_func)(struct wined3d_shader_instruction *ins,
482 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
483 struct wined3d_sm4_data *priv);
486 static const enum wined3d_primitive_type output_primitive_type_table[] =
488 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
489 /* WINED3D_SM4_OUTPUT_PT_POINTLIST */ WINED3D_PT_POINTLIST,
490 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
491 /* WINED3D_SM4_OUTPUT_PT_LINESTRIP */ WINED3D_PT_LINESTRIP,
492 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
493 /* WINED3D_SM4_OUTPUT_PT_TRIANGLESTRIP */ WINED3D_PT_TRIANGLESTRIP,
496 static const enum wined3d_primitive_type input_primitive_type_table[] =
498 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
499 /* WINED3D_SM4_INPUT_PT_POINT */ WINED3D_PT_POINTLIST,
500 /* WINED3D_SM4_INPUT_PT_LINE */ WINED3D_PT_LINELIST,
501 /* WINED3D_SM4_INPUT_PT_TRIANGLE */ WINED3D_PT_TRIANGLELIST,
502 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
503 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
504 /* WINED3D_SM4_INPUT_PT_LINEADJ */ WINED3D_PT_LINELIST_ADJ,
505 /* WINED3D_SM4_INPUT_PT_TRIANGLEADJ */ WINED3D_PT_TRIANGLELIST_ADJ,
508 static const enum wined3d_shader_resource_type resource_type_table[] =
510 /* 0 */ WINED3D_SHADER_RESOURCE_NONE,
511 /* WINED3D_SM4_RESOURCE_BUFFER */ WINED3D_SHADER_RESOURCE_BUFFER,
512 /* WINED3D_SM4_RESOURCE_TEXTURE_1D */ WINED3D_SHADER_RESOURCE_TEXTURE_1D,
513 /* WINED3D_SM4_RESOURCE_TEXTURE_2D */ WINED3D_SHADER_RESOURCE_TEXTURE_2D,
514 /* WINED3D_SM4_RESOURCE_TEXTURE_2DMS */ WINED3D_SHADER_RESOURCE_TEXTURE_2DMS,
515 /* WINED3D_SM4_RESOURCE_TEXTURE_3D */ WINED3D_SHADER_RESOURCE_TEXTURE_3D,
516 /* WINED3D_SM4_RESOURCE_TEXTURE_CUBE */ WINED3D_SHADER_RESOURCE_TEXTURE_CUBE,
517 /* WINED3D_SM4_RESOURCE_TEXTURE_1DARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_1DARRAY,
518 /* WINED3D_SM4_RESOURCE_TEXTURE_2DARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_2DARRAY,
519 /* WINED3D_SM4_RESOURCE_TEXTURE_2DMSARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_2DMSARRAY,
520 /* WINED3D_SM4_RESOURCE_TEXTURE_CUBEARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_CUBEARRAY,
523 static const enum wined3d_data_type data_type_table[] =
525 /* 0 */ WINED3D_DATA_FLOAT,
526 /* WINED3D_SM4_DATA_UNORM */ WINED3D_DATA_UNORM,
527 /* WINED3D_SM4_DATA_SNORM */ WINED3D_DATA_SNORM,
528 /* WINED3D_SM4_DATA_INT */ WINED3D_DATA_INT,
529 /* WINED3D_SM4_DATA_UINT */ WINED3D_DATA_UINT,
530 /* WINED3D_SM4_DATA_FLOAT */ WINED3D_DATA_FLOAT,
533 static BOOL shader_sm4_read_src_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
534 enum wined3d_data_type data_type, struct wined3d_shader_src_param *src_param);
535 static BOOL shader_sm4_read_dst_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
536 enum wined3d_data_type data_type, struct wined3d_shader_dst_param *dst_param);
538 static void shader_sm4_read_conditional_op(struct wined3d_shader_instruction *ins,
539 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
540 struct wined3d_sm4_data *priv)
542 shader_sm4_read_src_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_UINT, &priv->src_param[0]);
543 ins->flags = (opcode_token & WINED3D_SM4_CONDITIONAL_NZ) ?
544 WINED3D_SHADER_CONDITIONAL_OP_NZ : WINED3D_SHADER_CONDITIONAL_OP_Z;
547 static void shader_sm4_read_shader_data(struct wined3d_shader_instruction *ins,
548 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
549 struct wined3d_sm4_data *priv)
551 enum wined3d_sm4_shader_data_type type;
552 unsigned int icb_size;
554 type = (opcode_token & WINED3D_SM4_SHADER_DATA_TYPE_MASK) >> WINED3D_SM4_SHADER_DATA_TYPE_SHIFT;
555 if (type != WINED3D_SM4_SHADER_DATA_IMMEDIATE_CONSTANT_BUFFER)
557 FIXME("Ignoring shader data type %#x.\n", type);
558 ins->handler_idx = WINED3DSIH_NOP;
559 return;
562 ++tokens;
563 icb_size = token_count - 1;
564 if (icb_size % 4 || icb_size > MAX_IMMEDIATE_CONSTANT_BUFFER_SIZE)
566 FIXME("Unexpected immediate constant buffer size %u.\n", icb_size);
567 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
568 return;
571 priv->icb.vec4_count = icb_size / 4;
572 memcpy(priv->icb.data, tokens, sizeof(*tokens) * icb_size);
573 ins->declaration.icb = &priv->icb;
576 static void shader_sm4_read_dcl_resource(struct wined3d_shader_instruction *ins,
577 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
578 struct wined3d_sm4_data *priv)
580 enum wined3d_sm4_resource_type resource_type;
581 enum wined3d_sm4_data_type data_type;
582 enum wined3d_data_type reg_data_type;
583 DWORD components;
585 resource_type = (opcode_token & WINED3D_SM4_RESOURCE_TYPE_MASK) >> WINED3D_SM4_RESOURCE_TYPE_SHIFT;
586 if (!resource_type || (resource_type >= ARRAY_SIZE(resource_type_table)))
588 FIXME("Unhandled resource type %#x.\n", resource_type);
589 ins->declaration.semantic.resource_type = WINED3D_SHADER_RESOURCE_NONE;
591 else
593 ins->declaration.semantic.resource_type = resource_type_table[resource_type];
595 reg_data_type = opcode == WINED3D_SM4_OP_DCL_RESOURCE ? WINED3D_DATA_RESOURCE : WINED3D_DATA_UAV;
596 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], reg_data_type, &ins->declaration.semantic.reg);
598 components = *tokens++;
599 if ((components & 0xfff0) != (components & 0xf) * 0x1110)
600 FIXME("Components (%#x) have different data types.\n", components);
601 data_type = components & 0xf;
603 if (!data_type || (data_type >= ARRAY_SIZE(data_type_table)))
605 FIXME("Unhandled data type %#x.\n", data_type);
606 ins->declaration.semantic.resource_data_type = WINED3D_DATA_FLOAT;
608 else
610 ins->declaration.semantic.resource_data_type = data_type_table[data_type];
613 if (reg_data_type == WINED3D_DATA_UAV)
614 ins->flags = (opcode_token & WINED3D_SM5_UAV_FLAGS_MASK) >> WINED3D_SM5_UAV_FLAGS_SHIFT;
617 static void shader_sm4_read_dcl_constant_buffer(struct wined3d_shader_instruction *ins,
618 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
619 struct wined3d_sm4_data *priv)
621 shader_sm4_read_src_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.src);
622 if (opcode_token & WINED3D_SM4_INDEX_TYPE_MASK)
623 ins->flags |= WINED3DSI_INDEXED_DYNAMIC;
626 static void shader_sm4_read_dcl_sampler(struct wined3d_shader_instruction *ins,
627 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
628 struct wined3d_sm4_data *priv)
630 ins->flags = (opcode_token & WINED3D_SM4_SAMPLER_MODE_MASK) >> WINED3D_SM4_SAMPLER_MODE_SHIFT;
631 if (ins->flags & ~WINED3D_SM4_SAMPLER_COMPARISON)
632 FIXME("Unhandled sampler mode %#x.\n", ins->flags);
633 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_SAMPLER, &ins->declaration.dst);
636 static void shader_sm4_read_dcl_index_range(struct wined3d_shader_instruction *ins,
637 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
638 struct wined3d_sm4_data *priv)
640 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_OPAQUE,
641 &ins->declaration.index_range.first_register);
642 ins->declaration.index_range.last_register = *tokens;
645 static void shader_sm4_read_dcl_output_topology(struct wined3d_shader_instruction *ins,
646 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
647 struct wined3d_sm4_data *priv)
649 enum wined3d_sm4_output_primitive_type primitive_type;
651 primitive_type = (opcode_token & WINED3D_SM4_PRIMITIVE_TYPE_MASK) >> WINED3D_SM4_PRIMITIVE_TYPE_SHIFT;
652 if (primitive_type >= ARRAY_SIZE(output_primitive_type_table))
653 ins->declaration.primitive_type.type = WINED3D_PT_UNDEFINED;
654 else
655 ins->declaration.primitive_type.type = output_primitive_type_table[primitive_type];
657 if (ins->declaration.primitive_type.type == WINED3D_PT_UNDEFINED)
658 FIXME("Unhandled output primitive type %#x.\n", primitive_type);
661 static void shader_sm4_read_dcl_input_primitive(struct wined3d_shader_instruction *ins,
662 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
663 struct wined3d_sm4_data *priv)
665 enum wined3d_sm4_input_primitive_type primitive_type;
667 primitive_type = (opcode_token & WINED3D_SM4_PRIMITIVE_TYPE_MASK) >> WINED3D_SM4_PRIMITIVE_TYPE_SHIFT;
668 if (WINED3D_SM5_INPUT_PT_PATCH1 <= primitive_type && primitive_type <= WINED3D_SM5_INPUT_PT_PATCH32)
670 ins->declaration.primitive_type.type = WINED3D_PT_PATCH;
671 ins->declaration.primitive_type.patch_vertex_count = primitive_type - WINED3D_SM5_INPUT_PT_PATCH1 + 1;
673 else if (primitive_type >= ARRAY_SIZE(input_primitive_type_table))
675 ins->declaration.primitive_type.type = WINED3D_PT_UNDEFINED;
677 else
679 ins->declaration.primitive_type.type = input_primitive_type_table[primitive_type];
682 if (ins->declaration.primitive_type.type == WINED3D_PT_UNDEFINED)
683 FIXME("Unhandled input primitive type %#x.\n", primitive_type);
686 static void shader_sm4_read_declaration_count(struct wined3d_shader_instruction *ins,
687 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
688 struct wined3d_sm4_data *priv)
690 ins->declaration.count = *tokens;
693 static void shader_sm4_read_declaration_dst(struct wined3d_shader_instruction *ins,
694 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
695 struct wined3d_sm4_data *priv)
697 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.dst);
700 static void shader_sm4_read_declaration_register_semantic(struct wined3d_shader_instruction *ins,
701 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
702 struct wined3d_sm4_data *priv)
704 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT,
705 &ins->declaration.register_semantic.reg);
706 ins->declaration.register_semantic.sysval_semantic = *tokens;
709 static void shader_sm4_read_dcl_input_ps(struct wined3d_shader_instruction *ins,
710 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
711 struct wined3d_sm4_data *priv)
713 ins->flags = (opcode_token & WINED3D_SM4_INTERPOLATION_MODE_MASK) >> WINED3D_SM4_INTERPOLATION_MODE_SHIFT;
714 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.dst);
717 static void shader_sm4_read_dcl_input_ps_siv(struct wined3d_shader_instruction *ins,
718 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
719 struct wined3d_sm4_data *priv)
721 ins->flags = (opcode_token & WINED3D_SM4_INTERPOLATION_MODE_MASK) >> WINED3D_SM4_INTERPOLATION_MODE_SHIFT;
722 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT,
723 &ins->declaration.register_semantic.reg);
724 ins->declaration.register_semantic.sysval_semantic = *tokens;
727 static void shader_sm4_read_dcl_indexable_temp(struct wined3d_shader_instruction *ins,
728 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
729 struct wined3d_sm4_data *priv)
731 ins->declaration.indexable_temp.register_idx = *tokens++;
732 ins->declaration.indexable_temp.register_size = *tokens++;
733 ins->declaration.indexable_temp.component_count = *tokens;
736 static void shader_sm4_read_dcl_global_flags(struct wined3d_shader_instruction *ins,
737 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
738 struct wined3d_sm4_data *priv)
740 ins->flags = (opcode_token & WINED3D_SM4_GLOBAL_FLAGS_MASK) >> WINED3D_SM4_GLOBAL_FLAGS_SHIFT;
743 static void shader_sm5_read_fcall(struct wined3d_shader_instruction *ins,
744 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
745 struct wined3d_sm4_data *priv)
747 priv->src_param[0].reg.u.fp_body_idx = *tokens++;
748 shader_sm4_read_src_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_OPAQUE, &priv->src_param[0]);
751 static void shader_sm5_read_dcl_function_body(struct wined3d_shader_instruction *ins,
752 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
753 struct wined3d_sm4_data *priv)
755 ins->declaration.index = *tokens;
758 static void shader_sm5_read_dcl_function_table(struct wined3d_shader_instruction *ins,
759 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
760 struct wined3d_sm4_data *priv)
762 ins->declaration.index = *tokens++;
763 FIXME("Ignoring set of function bodies (count %u).\n", *tokens);
766 static void shader_sm5_read_dcl_interface(struct wined3d_shader_instruction *ins,
767 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
768 struct wined3d_sm4_data *priv)
770 ins->declaration.fp.index = *tokens++;
771 ins->declaration.fp.body_count = *tokens++;
772 ins->declaration.fp.array_size = *tokens >> WINED3D_SM5_FP_ARRAY_SIZE_SHIFT;
773 ins->declaration.fp.table_count = *tokens++ & WINED3D_SM5_FP_TABLE_COUNT_MASK;
774 FIXME("Ignoring set of function tables (count %u).\n", ins->declaration.fp.table_count);
777 static void shader_sm5_read_control_point_count(struct wined3d_shader_instruction *ins,
778 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
779 struct wined3d_sm4_data *priv)
781 ins->declaration.count = (opcode_token & WINED3D_SM5_CONTROL_POINT_COUNT_MASK)
782 >> WINED3D_SM5_CONTROL_POINT_COUNT_SHIFT;
785 static void shader_sm5_read_dcl_tessellator_domain(struct wined3d_shader_instruction *ins,
786 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
787 struct wined3d_sm4_data *priv)
789 ins->declaration.tessellator_domain = (opcode_token & WINED3D_SM5_TESSELLATOR_MASK)
790 >> WINED3D_SM5_TESSELLATOR_SHIFT;
793 static void shader_sm5_read_dcl_tessellator_partitioning(struct wined3d_shader_instruction *ins,
794 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
795 struct wined3d_sm4_data *priv)
797 ins->declaration.tessellator_partitioning = (opcode_token & WINED3D_SM5_TESSELLATOR_MASK)
798 >> WINED3D_SM5_TESSELLATOR_SHIFT;
801 static void shader_sm5_read_dcl_tessellator_output_primitive(struct wined3d_shader_instruction *ins,
802 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
803 struct wined3d_sm4_data *priv)
805 ins->declaration.tessellator_output_primitive = (opcode_token & WINED3D_SM5_TESSELLATOR_MASK)
806 >> WINED3D_SM5_TESSELLATOR_SHIFT;
809 static void shader_sm5_read_dcl_hs_max_tessfactor(struct wined3d_shader_instruction *ins,
810 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
811 struct wined3d_sm4_data *priv)
813 ins->declaration.max_tessellation_factor = *(float *)tokens;
816 static void shader_sm5_read_dcl_thread_group(struct wined3d_shader_instruction *ins,
817 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
818 struct wined3d_sm4_data *priv)
820 ins->declaration.thread_group_size.x = *tokens++;
821 ins->declaration.thread_group_size.y = *tokens++;
822 ins->declaration.thread_group_size.z = *tokens++;
825 static void shader_sm5_read_dcl_uav_raw(struct wined3d_shader_instruction *ins,
826 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
827 struct wined3d_sm4_data *priv)
829 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_UAV, &ins->declaration.dst);
830 ins->flags = (opcode_token & WINED3D_SM5_UAV_FLAGS_MASK) >> WINED3D_SM5_UAV_FLAGS_SHIFT;
833 static void shader_sm5_read_dcl_uav_structured(struct wined3d_shader_instruction *ins,
834 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
835 struct wined3d_sm4_data *priv)
837 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_UAV,
838 &ins->declaration.structured_resource.reg);
839 ins->flags = (opcode_token & WINED3D_SM5_UAV_FLAGS_MASK) >> WINED3D_SM5_UAV_FLAGS_SHIFT;
840 ins->declaration.structured_resource.byte_stride = *tokens;
841 if (ins->declaration.structured_resource.byte_stride % 4)
842 FIXME("Byte stride %u is not multiple of 4.\n", ins->declaration.structured_resource.byte_stride);
845 static void shader_sm5_read_dcl_tgsm_raw(struct wined3d_shader_instruction *ins,
846 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
847 struct wined3d_sm4_data *priv)
849 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.tgsm_raw.reg);
850 ins->declaration.tgsm_raw.byte_count = *tokens;
851 if (ins->declaration.tgsm_raw.byte_count % 4)
852 FIXME("Byte count %u is not multiple of 4.\n", ins->declaration.tgsm_raw.byte_count);
855 static void shader_sm5_read_dcl_tgsm_structured(struct wined3d_shader_instruction *ins,
856 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
857 struct wined3d_sm4_data *priv)
859 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT,
860 &ins->declaration.tgsm_structured.reg);
861 ins->declaration.tgsm_structured.byte_stride = *tokens++;
862 ins->declaration.tgsm_structured.structure_count = *tokens;
863 if (ins->declaration.tgsm_structured.byte_stride % 4)
864 FIXME("Byte stride %u is not multiple of 4.\n", ins->declaration.tgsm_structured.byte_stride);
867 static void shader_sm5_read_dcl_resource_structured(struct wined3d_shader_instruction *ins,
868 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
869 struct wined3d_sm4_data *priv)
871 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_RESOURCE,
872 &ins->declaration.structured_resource.reg);
873 ins->declaration.structured_resource.byte_stride = *tokens;
874 if (ins->declaration.structured_resource.byte_stride % 4)
875 FIXME("Byte stride %u is not multiple of 4.\n", ins->declaration.structured_resource.byte_stride);
878 static void shader_sm5_read_dcl_resource_raw(struct wined3d_shader_instruction *ins,
879 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
880 struct wined3d_sm4_data *priv)
882 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_RESOURCE, &ins->declaration.dst);
885 static void shader_sm5_read_sync(struct wined3d_shader_instruction *ins,
886 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
887 struct wined3d_sm4_data *priv)
889 ins->flags = (opcode_token & WINED3D_SM5_SYNC_FLAGS_MASK) >> WINED3D_SM5_SYNC_FLAGS_SHIFT;
893 * f -> WINED3D_DATA_FLOAT
894 * i -> WINED3D_DATA_INT
895 * u -> WINED3D_DATA_UINT
896 * O -> WINED3D_DATA_OPAQUE
897 * R -> WINED3D_DATA_RESOURCE
898 * S -> WINED3D_DATA_SAMPLER
899 * U -> WINED3D_DATA_UAV
901 static const struct wined3d_sm4_opcode_info opcode_table[] =
903 {WINED3D_SM4_OP_ADD, WINED3DSIH_ADD, "f", "ff"},
904 {WINED3D_SM4_OP_AND, WINED3DSIH_AND, "u", "uu"},
905 {WINED3D_SM4_OP_BREAK, WINED3DSIH_BREAK, "", ""},
906 {WINED3D_SM4_OP_BREAKC, WINED3DSIH_BREAKP, "", "u",
907 shader_sm4_read_conditional_op},
908 {WINED3D_SM4_OP_CASE, WINED3DSIH_CASE, "", "u"},
909 {WINED3D_SM4_OP_CONTINUE, WINED3DSIH_CONTINUE, "", ""},
910 {WINED3D_SM4_OP_CONTINUEC, WINED3DSIH_CONTINUEP, "", "u",
911 shader_sm4_read_conditional_op},
912 {WINED3D_SM4_OP_CUT, WINED3DSIH_CUT, "", ""},
913 {WINED3D_SM4_OP_DEFAULT, WINED3DSIH_DEFAULT, "", ""},
914 {WINED3D_SM4_OP_DERIV_RTX, WINED3DSIH_DSX, "f", "f"},
915 {WINED3D_SM4_OP_DERIV_RTY, WINED3DSIH_DSY, "f", "f"},
916 {WINED3D_SM4_OP_DISCARD, WINED3DSIH_TEXKILL, "", "u",
917 shader_sm4_read_conditional_op},
918 {WINED3D_SM4_OP_DIV, WINED3DSIH_DIV, "f", "ff"},
919 {WINED3D_SM4_OP_DP2, WINED3DSIH_DP2, "f", "ff"},
920 {WINED3D_SM4_OP_DP3, WINED3DSIH_DP3, "f", "ff"},
921 {WINED3D_SM4_OP_DP4, WINED3DSIH_DP4, "f", "ff"},
922 {WINED3D_SM4_OP_ELSE, WINED3DSIH_ELSE, "", ""},
923 {WINED3D_SM4_OP_EMIT, WINED3DSIH_EMIT, "", ""},
924 {WINED3D_SM4_OP_ENDIF, WINED3DSIH_ENDIF, "", ""},
925 {WINED3D_SM4_OP_ENDLOOP, WINED3DSIH_ENDLOOP, "", ""},
926 {WINED3D_SM4_OP_ENDSWITCH, WINED3DSIH_ENDSWITCH, "", ""},
927 {WINED3D_SM4_OP_EQ, WINED3DSIH_EQ, "u", "ff"},
928 {WINED3D_SM4_OP_EXP, WINED3DSIH_EXP, "f", "f"},
929 {WINED3D_SM4_OP_FRC, WINED3DSIH_FRC, "f", "f"},
930 {WINED3D_SM4_OP_FTOI, WINED3DSIH_FTOI, "i", "f"},
931 {WINED3D_SM4_OP_FTOU, WINED3DSIH_FTOU, "u", "f"},
932 {WINED3D_SM4_OP_GE, WINED3DSIH_GE, "u", "ff"},
933 {WINED3D_SM4_OP_IADD, WINED3DSIH_IADD, "i", "ii"},
934 {WINED3D_SM4_OP_IF, WINED3DSIH_IF, "", "u",
935 shader_sm4_read_conditional_op},
936 {WINED3D_SM4_OP_IEQ, WINED3DSIH_IEQ, "u", "ii"},
937 {WINED3D_SM4_OP_IGE, WINED3DSIH_IGE, "u", "ii"},
938 {WINED3D_SM4_OP_ILT, WINED3DSIH_ILT, "u", "ii"},
939 {WINED3D_SM4_OP_IMAD, WINED3DSIH_IMAD, "i", "iii"},
940 {WINED3D_SM4_OP_IMAX, WINED3DSIH_IMAX, "i", "ii"},
941 {WINED3D_SM4_OP_IMIN, WINED3DSIH_IMIN, "i", "ii"},
942 {WINED3D_SM4_OP_IMUL, WINED3DSIH_IMUL, "ii", "ii"},
943 {WINED3D_SM4_OP_INE, WINED3DSIH_INE, "u", "ii"},
944 {WINED3D_SM4_OP_INEG, WINED3DSIH_INEG, "i", "i"},
945 {WINED3D_SM4_OP_ISHL, WINED3DSIH_ISHL, "i", "ii"},
946 {WINED3D_SM4_OP_ISHR, WINED3DSIH_ISHR, "i", "ii"},
947 {WINED3D_SM4_OP_ITOF, WINED3DSIH_ITOF, "f", "i"},
948 {WINED3D_SM4_OP_LABEL, WINED3DSIH_LABEL, "", "O"},
949 {WINED3D_SM4_OP_LD, WINED3DSIH_LD, "u", "iR"},
950 {WINED3D_SM4_OP_LD2DMS, WINED3DSIH_LD2DMS, "u", "iRi"},
951 {WINED3D_SM4_OP_LOG, WINED3DSIH_LOG, "f", "f"},
952 {WINED3D_SM4_OP_LOOP, WINED3DSIH_LOOP, "", ""},
953 {WINED3D_SM4_OP_LT, WINED3DSIH_LT, "u", "ff"},
954 {WINED3D_SM4_OP_MAD, WINED3DSIH_MAD, "f", "fff"},
955 {WINED3D_SM4_OP_MIN, WINED3DSIH_MIN, "f", "ff"},
956 {WINED3D_SM4_OP_MAX, WINED3DSIH_MAX, "f", "ff"},
957 {WINED3D_SM4_OP_SHADER_DATA, WINED3DSIH_DCL_IMMEDIATE_CONSTANT_BUFFER, "", "",
958 shader_sm4_read_shader_data},
959 {WINED3D_SM4_OP_MOV, WINED3DSIH_MOV, "f", "f"},
960 {WINED3D_SM4_OP_MOVC, WINED3DSIH_MOVC, "f", "uff"},
961 {WINED3D_SM4_OP_MUL, WINED3DSIH_MUL, "f", "ff"},
962 {WINED3D_SM4_OP_NE, WINED3DSIH_NE, "u", "ff"},
963 {WINED3D_SM4_OP_NOP, WINED3DSIH_NOP, "", ""},
964 {WINED3D_SM4_OP_NOT, WINED3DSIH_NOT, "u", "u"},
965 {WINED3D_SM4_OP_OR, WINED3DSIH_OR, "u", "uu"},
966 {WINED3D_SM4_OP_RESINFO, WINED3DSIH_RESINFO, "f", "iR"},
967 {WINED3D_SM4_OP_RET, WINED3DSIH_RET, "", ""},
968 {WINED3D_SM4_OP_RETC, WINED3DSIH_RETP, "", "u",
969 shader_sm4_read_conditional_op},
970 {WINED3D_SM4_OP_ROUND_NE, WINED3DSIH_ROUND_NE, "f", "f"},
971 {WINED3D_SM4_OP_ROUND_NI, WINED3DSIH_ROUND_NI, "f", "f"},
972 {WINED3D_SM4_OP_ROUND_PI, WINED3DSIH_ROUND_PI, "f", "f"},
973 {WINED3D_SM4_OP_ROUND_Z, WINED3DSIH_ROUND_Z, "f", "f"},
974 {WINED3D_SM4_OP_RSQ, WINED3DSIH_RSQ, "f", "f"},
975 {WINED3D_SM4_OP_SAMPLE, WINED3DSIH_SAMPLE, "u", "fRS"},
976 {WINED3D_SM4_OP_SAMPLE_C, WINED3DSIH_SAMPLE_C, "f", "fRSf"},
977 {WINED3D_SM4_OP_SAMPLE_C_LZ, WINED3DSIH_SAMPLE_C_LZ, "f", "fRSf"},
978 {WINED3D_SM4_OP_SAMPLE_LOD, WINED3DSIH_SAMPLE_LOD, "u", "fRSf"},
979 {WINED3D_SM4_OP_SAMPLE_GRAD, WINED3DSIH_SAMPLE_GRAD, "u", "fRSff"},
980 {WINED3D_SM4_OP_SAMPLE_B, WINED3DSIH_SAMPLE_B, "u", "fRSf"},
981 {WINED3D_SM4_OP_SQRT, WINED3DSIH_SQRT, "f", "f"},
982 {WINED3D_SM4_OP_SWITCH, WINED3DSIH_SWITCH, "", "u"},
983 {WINED3D_SM4_OP_SINCOS, WINED3DSIH_SINCOS, "ff", "f"},
984 {WINED3D_SM4_OP_UDIV, WINED3DSIH_UDIV, "uu", "uu"},
985 {WINED3D_SM4_OP_ULT, WINED3DSIH_ULT, "u", "uu"},
986 {WINED3D_SM4_OP_UGE, WINED3DSIH_UGE, "u", "uu"},
987 {WINED3D_SM4_OP_UMUL, WINED3DSIH_UMUL, "uu", "uu"},
988 {WINED3D_SM4_OP_UMAX, WINED3DSIH_UMAX, "u", "uu"},
989 {WINED3D_SM4_OP_UMIN, WINED3DSIH_UMIN, "u", "uu"},
990 {WINED3D_SM4_OP_USHR, WINED3DSIH_USHR, "u", "uu"},
991 {WINED3D_SM4_OP_UTOF, WINED3DSIH_UTOF, "f", "u"},
992 {WINED3D_SM4_OP_XOR, WINED3DSIH_XOR, "u", "uu"},
993 {WINED3D_SM4_OP_DCL_RESOURCE, WINED3DSIH_DCL, "R", "",
994 shader_sm4_read_dcl_resource},
995 {WINED3D_SM4_OP_DCL_CONSTANT_BUFFER, WINED3DSIH_DCL_CONSTANT_BUFFER, "", "",
996 shader_sm4_read_dcl_constant_buffer},
997 {WINED3D_SM4_OP_DCL_SAMPLER, WINED3DSIH_DCL_SAMPLER, "", "",
998 shader_sm4_read_dcl_sampler},
999 {WINED3D_SM4_OP_DCL_INDEX_RANGE, WINED3DSIH_DCL_INDEX_RANGE, "", "",
1000 shader_sm4_read_dcl_index_range},
1001 {WINED3D_SM4_OP_DCL_OUTPUT_TOPOLOGY, WINED3DSIH_DCL_OUTPUT_TOPOLOGY, "", "",
1002 shader_sm4_read_dcl_output_topology},
1003 {WINED3D_SM4_OP_DCL_INPUT_PRIMITIVE, WINED3DSIH_DCL_INPUT_PRIMITIVE, "", "",
1004 shader_sm4_read_dcl_input_primitive},
1005 {WINED3D_SM4_OP_DCL_VERTICES_OUT, WINED3DSIH_DCL_VERTICES_OUT, "", "",
1006 shader_sm4_read_declaration_count},
1007 {WINED3D_SM4_OP_DCL_INPUT, WINED3DSIH_DCL_INPUT, "", "",
1008 shader_sm4_read_declaration_dst},
1009 {WINED3D_SM4_OP_DCL_INPUT_SGV, WINED3DSIH_DCL_INPUT_SGV, "", "",
1010 shader_sm4_read_declaration_register_semantic},
1011 {WINED3D_SM4_OP_DCL_INPUT_SIV, WINED3DSIH_DCL_INPUT_SIV, "", "",
1012 shader_sm4_read_declaration_register_semantic},
1013 {WINED3D_SM4_OP_DCL_INPUT_PS, WINED3DSIH_DCL_INPUT_PS, "", "",
1014 shader_sm4_read_dcl_input_ps},
1015 {WINED3D_SM4_OP_DCL_INPUT_PS_SGV, WINED3DSIH_DCL_INPUT_PS_SGV, "", "",
1016 shader_sm4_read_declaration_register_semantic},
1017 {WINED3D_SM4_OP_DCL_INPUT_PS_SIV, WINED3DSIH_DCL_INPUT_PS_SIV, "", "",
1018 shader_sm4_read_dcl_input_ps_siv},
1019 {WINED3D_SM4_OP_DCL_OUTPUT, WINED3DSIH_DCL_OUTPUT, "", "",
1020 shader_sm4_read_declaration_dst},
1021 {WINED3D_SM4_OP_DCL_OUTPUT_SIV, WINED3DSIH_DCL_OUTPUT_SIV, "", "",
1022 shader_sm4_read_declaration_register_semantic},
1023 {WINED3D_SM4_OP_DCL_TEMPS, WINED3DSIH_DCL_TEMPS, "", "",
1024 shader_sm4_read_declaration_count},
1025 {WINED3D_SM4_OP_DCL_INDEXABLE_TEMP, WINED3DSIH_DCL_INDEXABLE_TEMP, "", "",
1026 shader_sm4_read_dcl_indexable_temp},
1027 {WINED3D_SM4_OP_DCL_GLOBAL_FLAGS, WINED3DSIH_DCL_GLOBAL_FLAGS, "", "",
1028 shader_sm4_read_dcl_global_flags},
1029 {WINED3D_SM4_OP_LOD, WINED3DSIH_LOD, "f", "fRS"},
1030 {WINED3D_SM4_OP_GATHER4, WINED3DSIH_GATHER4, "u", "fRS"},
1031 {WINED3D_SM4_OP_SAMPLE_POS, WINED3DSIH_SAMPLE_POS, "f", "Ru"},
1032 {WINED3D_SM4_OP_SAMPLE_INFO, WINED3DSIH_SAMPLE_INFO, "f", "R"},
1033 {WINED3D_SM5_OP_HS_DECLS, WINED3DSIH_HS_DECLS, "", ""},
1034 {WINED3D_SM5_OP_HS_CONTROL_POINT_PHASE, WINED3DSIH_HS_CONTROL_POINT_PHASE, "", ""},
1035 {WINED3D_SM5_OP_HS_FORK_PHASE, WINED3DSIH_HS_FORK_PHASE, "", ""},
1036 {WINED3D_SM5_OP_HS_JOIN_PHASE, WINED3DSIH_HS_JOIN_PHASE, "", ""},
1037 {WINED3D_SM5_OP_EMIT_STREAM, WINED3DSIH_EMIT_STREAM, "", "f"},
1038 {WINED3D_SM5_OP_CUT_STREAM, WINED3DSIH_CUT_STREAM, "", "f"},
1039 {WINED3D_SM5_OP_FCALL, WINED3DSIH_FCALL, "", "O",
1040 shader_sm5_read_fcall},
1041 {WINED3D_SM5_OP_BUFINFO, WINED3DSIH_BUFINFO, "i", "U"},
1042 {WINED3D_SM5_OP_DERIV_RTX_COARSE, WINED3DSIH_DSX_COARSE, "f", "f"},
1043 {WINED3D_SM5_OP_DERIV_RTX_FINE, WINED3DSIH_DSX_FINE, "f", "f"},
1044 {WINED3D_SM5_OP_DERIV_RTY_COARSE, WINED3DSIH_DSY_COARSE, "f", "f"},
1045 {WINED3D_SM5_OP_DERIV_RTY_FINE, WINED3DSIH_DSY_FINE, "f", "f"},
1046 {WINED3D_SM5_OP_GATHER4_C, WINED3DSIH_GATHER4_C, "f", "fRSf"},
1047 {WINED3D_SM5_OP_GATHER4_PO, WINED3DSIH_GATHER4_PO, "f", "fiRS"},
1048 {WINED3D_SM5_OP_GATHER4_PO_C, WINED3DSIH_GATHER4_PO_C, "f", "fiRSf"},
1049 {WINED3D_SM5_OP_RCP, WINED3DSIH_RCP, "f", "f"},
1050 {WINED3D_SM5_OP_F32TOF16, WINED3DSIH_F32TOF16, "u", "f"},
1051 {WINED3D_SM5_OP_F16TOF32, WINED3DSIH_F16TOF32, "f", "u"},
1052 {WINED3D_SM5_OP_COUNTBITS, WINED3DSIH_COUNTBITS, "u", "u"},
1053 {WINED3D_SM5_OP_FIRSTBIT_HI, WINED3DSIH_FIRSTBIT_HI, "u", "u"},
1054 {WINED3D_SM5_OP_FIRSTBIT_LO, WINED3DSIH_FIRSTBIT_LO, "u", "u"},
1055 {WINED3D_SM5_OP_FIRSTBIT_SHI, WINED3DSIH_FIRSTBIT_SHI, "u", "i"},
1056 {WINED3D_SM5_OP_UBFE, WINED3DSIH_UBFE, "u", "iiu"},
1057 {WINED3D_SM5_OP_IBFE, WINED3DSIH_IBFE, "i", "iii"},
1058 {WINED3D_SM5_OP_BFI, WINED3DSIH_BFI, "u", "iiuu"},
1059 {WINED3D_SM5_OP_BFREV, WINED3DSIH_BFREV, "u", "u"},
1060 {WINED3D_SM5_OP_SWAPC, WINED3DSIH_SWAPC, "ff", "uff"},
1061 {WINED3D_SM5_OP_DCL_STREAM, WINED3DSIH_DCL_STREAM, "", "O"},
1062 {WINED3D_SM5_OP_DCL_FUNCTION_BODY, WINED3DSIH_DCL_FUNCTION_BODY, "", "",
1063 shader_sm5_read_dcl_function_body},
1064 {WINED3D_SM5_OP_DCL_FUNCTION_TABLE, WINED3DSIH_DCL_FUNCTION_TABLE, "", "",
1065 shader_sm5_read_dcl_function_table},
1066 {WINED3D_SM5_OP_DCL_INTERFACE, WINED3DSIH_DCL_INTERFACE, "", "",
1067 shader_sm5_read_dcl_interface},
1068 {WINED3D_SM5_OP_DCL_INPUT_CONTROL_POINT_COUNT, WINED3DSIH_DCL_INPUT_CONTROL_POINT_COUNT, "", "",
1069 shader_sm5_read_control_point_count},
1070 {WINED3D_SM5_OP_DCL_OUTPUT_CONTROL_POINT_COUNT, WINED3DSIH_DCL_OUTPUT_CONTROL_POINT_COUNT, "", "",
1071 shader_sm5_read_control_point_count},
1072 {WINED3D_SM5_OP_DCL_TESSELLATOR_DOMAIN, WINED3DSIH_DCL_TESSELLATOR_DOMAIN, "", "",
1073 shader_sm5_read_dcl_tessellator_domain},
1074 {WINED3D_SM5_OP_DCL_TESSELLATOR_PARTITIONING, WINED3DSIH_DCL_TESSELLATOR_PARTITIONING, "", "",
1075 shader_sm5_read_dcl_tessellator_partitioning},
1076 {WINED3D_SM5_OP_DCL_TESSELLATOR_OUTPUT_PRIMITIVE, WINED3DSIH_DCL_TESSELLATOR_OUTPUT_PRIMITIVE, "", "",
1077 shader_sm5_read_dcl_tessellator_output_primitive},
1078 {WINED3D_SM5_OP_DCL_HS_MAX_TESSFACTOR, WINED3DSIH_DCL_HS_MAX_TESSFACTOR, "", "",
1079 shader_sm5_read_dcl_hs_max_tessfactor},
1080 {WINED3D_SM5_OP_DCL_HS_FORK_PHASE_INSTANCE_COUNT, WINED3DSIH_DCL_HS_FORK_PHASE_INSTANCE_COUNT, "", "",
1081 shader_sm4_read_declaration_count},
1082 {WINED3D_SM5_OP_DCL_HS_JOIN_PHASE_INSTANCE_COUNT, WINED3DSIH_DCL_HS_JOIN_PHASE_INSTANCE_COUNT, "", "",
1083 shader_sm4_read_declaration_count},
1084 {WINED3D_SM5_OP_DCL_THREAD_GROUP, WINED3DSIH_DCL_THREAD_GROUP, "", "",
1085 shader_sm5_read_dcl_thread_group},
1086 {WINED3D_SM5_OP_DCL_UAV_TYPED, WINED3DSIH_DCL_UAV_TYPED, "", "",
1087 shader_sm4_read_dcl_resource},
1088 {WINED3D_SM5_OP_DCL_UAV_RAW, WINED3DSIH_DCL_UAV_RAW, "", "",
1089 shader_sm5_read_dcl_uav_raw},
1090 {WINED3D_SM5_OP_DCL_UAV_STRUCTURED, WINED3DSIH_DCL_UAV_STRUCTURED, "", "",
1091 shader_sm5_read_dcl_uav_structured},
1092 {WINED3D_SM5_OP_DCL_TGSM_RAW, WINED3DSIH_DCL_TGSM_RAW, "", "",
1093 shader_sm5_read_dcl_tgsm_raw},
1094 {WINED3D_SM5_OP_DCL_TGSM_STRUCTURED, WINED3DSIH_DCL_TGSM_STRUCTURED, "", "",
1095 shader_sm5_read_dcl_tgsm_structured},
1096 {WINED3D_SM5_OP_DCL_RESOURCE_RAW, WINED3DSIH_DCL_RESOURCE_RAW, "", "",
1097 shader_sm5_read_dcl_resource_raw},
1098 {WINED3D_SM5_OP_DCL_RESOURCE_STRUCTURED, WINED3DSIH_DCL_RESOURCE_STRUCTURED, "", "",
1099 shader_sm5_read_dcl_resource_structured},
1100 {WINED3D_SM5_OP_LD_UAV_TYPED, WINED3DSIH_LD_UAV_TYPED, "u", "iU"},
1101 {WINED3D_SM5_OP_STORE_UAV_TYPED, WINED3DSIH_STORE_UAV_TYPED, "U", "iu"},
1102 {WINED3D_SM5_OP_LD_RAW, WINED3DSIH_LD_RAW, "u", "iU"},
1103 {WINED3D_SM5_OP_STORE_RAW, WINED3DSIH_STORE_RAW, "U", "iu"},
1104 {WINED3D_SM5_OP_LD_STRUCTURED, WINED3DSIH_LD_STRUCTURED, "u", "iiR"},
1105 {WINED3D_SM5_OP_STORE_STRUCTURED, WINED3DSIH_STORE_STRUCTURED, "U", "iiu"},
1106 {WINED3D_SM5_OP_ATOMIC_AND, WINED3DSIH_ATOMIC_AND, "U", "iu"},
1107 {WINED3D_SM5_OP_ATOMIC_OR, WINED3DSIH_ATOMIC_OR, "U", "iu"},
1108 {WINED3D_SM5_OP_ATOMIC_XOR, WINED3DSIH_ATOMIC_XOR, "U", "iu"},
1109 {WINED3D_SM5_OP_ATOMIC_CMP_STORE, WINED3DSIH_ATOMIC_CMP_STORE, "U", "iuu"},
1110 {WINED3D_SM5_OP_ATOMIC_IADD, WINED3DSIH_ATOMIC_IADD, "U", "ii"},
1111 {WINED3D_SM5_OP_ATOMIC_IMAX, WINED3DSIH_ATOMIC_IMAX, "U", "ii"},
1112 {WINED3D_SM5_OP_ATOMIC_IMIN, WINED3DSIH_ATOMIC_IMIN, "U", "ii"},
1113 {WINED3D_SM5_OP_ATOMIC_UMAX, WINED3DSIH_ATOMIC_UMAX, "U", "iu"},
1114 {WINED3D_SM5_OP_ATOMIC_UMIN, WINED3DSIH_ATOMIC_UMIN, "U", "iu"},
1115 {WINED3D_SM5_OP_IMM_ATOMIC_ALLOC, WINED3DSIH_IMM_ATOMIC_ALLOC, "u", "U"},
1116 {WINED3D_SM5_OP_IMM_ATOMIC_CONSUME, WINED3DSIH_IMM_ATOMIC_CONSUME, "u", "U"},
1117 {WINED3D_SM5_OP_IMM_ATOMIC_IADD, WINED3DSIH_IMM_ATOMIC_IADD, "uU", "ii"},
1118 {WINED3D_SM5_OP_IMM_ATOMIC_AND, WINED3DSIH_IMM_ATOMIC_AND, "uU", "iu"},
1119 {WINED3D_SM5_OP_IMM_ATOMIC_OR, WINED3DSIH_IMM_ATOMIC_OR, "uU", "iu"},
1120 {WINED3D_SM5_OP_IMM_ATOMIC_XOR, WINED3DSIH_IMM_ATOMIC_XOR, "uU", "iu"},
1121 {WINED3D_SM5_OP_IMM_ATOMIC_EXCH, WINED3DSIH_IMM_ATOMIC_EXCH, "uU", "iu"},
1122 {WINED3D_SM5_OP_IMM_ATOMIC_CMP_EXCH, WINED3DSIH_IMM_ATOMIC_CMP_EXCH, "uU", "iuu"},
1123 {WINED3D_SM5_OP_IMM_ATOMIC_IMAX, WINED3DSIH_IMM_ATOMIC_IMAX, "iU", "ii"},
1124 {WINED3D_SM5_OP_IMM_ATOMIC_IMIN, WINED3DSIH_IMM_ATOMIC_IMIN, "iU", "ii"},
1125 {WINED3D_SM5_OP_IMM_ATOMIC_UMAX, WINED3DSIH_IMM_ATOMIC_UMAX, "uU", "iu"},
1126 {WINED3D_SM5_OP_IMM_ATOMIC_UMIN, WINED3DSIH_IMM_ATOMIC_UMIN, "uU", "iu"},
1127 {WINED3D_SM5_OP_SYNC, WINED3DSIH_SYNC, "", "",
1128 shader_sm5_read_sync},
1129 {WINED3D_SM5_OP_EVAL_SAMPLE_INDEX, WINED3DSIH_EVAL_SAMPLE_INDEX, "f", "fi"},
1130 {WINED3D_SM5_OP_DCL_GS_INSTANCES, WINED3DSIH_DCL_GS_INSTANCES, "", "",
1131 shader_sm4_read_declaration_count},
1134 static const enum wined3d_shader_register_type register_type_table[] =
1136 /* WINED3D_SM4_RT_TEMP */ WINED3DSPR_TEMP,
1137 /* WINED3D_SM4_RT_INPUT */ WINED3DSPR_INPUT,
1138 /* WINED3D_SM4_RT_OUTPUT */ WINED3DSPR_OUTPUT,
1139 /* WINED3D_SM4_RT_INDEXABLE_TEMP */ WINED3DSPR_IDXTEMP,
1140 /* WINED3D_SM4_RT_IMMCONST */ WINED3DSPR_IMMCONST,
1141 /* UNKNOWN */ ~0u,
1142 /* WINED3D_SM4_RT_SAMPLER */ WINED3DSPR_SAMPLER,
1143 /* WINED3D_SM4_RT_RESOURCE */ WINED3DSPR_RESOURCE,
1144 /* WINED3D_SM4_RT_CONSTBUFFER */ WINED3DSPR_CONSTBUFFER,
1145 /* WINED3D_SM4_RT_IMMCONSTBUFFER */ WINED3DSPR_IMMCONSTBUFFER,
1146 /* UNKNOWN */ ~0u,
1147 /* WINED3D_SM4_RT_PRIMID */ WINED3DSPR_PRIMID,
1148 /* WINED3D_SM4_RT_DEPTHOUT */ WINED3DSPR_DEPTHOUT,
1149 /* WINED3D_SM4_RT_NULL */ WINED3DSPR_NULL,
1150 /* WINED3D_SM4_RT_RASTERIZER */ WINED3DSPR_RASTERIZER,
1151 /* WINED3D_SM4_RT_OMASK */ WINED3DSPR_SAMPLEMASK,
1152 /* WINED3D_SM5_RT_STREAM */ WINED3DSPR_STREAM,
1153 /* WINED3D_SM5_RT_FUNCTION_BODY */ WINED3DSPR_FUNCTIONBODY,
1154 /* UNKNOWN */ ~0u,
1155 /* WINED3D_SM5_RT_FUNCTION_POINTER */ WINED3DSPR_FUNCTIONPOINTER,
1156 /* UNKNOWN */ ~0u,
1157 /* UNKNOWN */ ~0u,
1158 /* WINED3D_SM5_RT_OUTPUT_CONTROL_POINT_ID */ WINED3DSPR_OUTPOINTID,
1159 /* WINED3D_SM5_RT_FORK_INSTANCE_ID */ WINED3DSPR_FORKINSTID,
1160 /* WINED3D_SM5_RT_JOIN_INSTANCE_ID */ WINED3DSPR_JOININSTID,
1161 /* WINED3D_SM5_RT_INPUT_CONTROL_POINT */ WINED3DSPR_INCONTROLPOINT,
1162 /* WINED3D_SM5_RT_OUTPUT_CONTROL_POINT */ WINED3DSPR_OUTCONTROLPOINT,
1163 /* WINED3D_SM5_RT_PATCH_CONSTANT_DATA */ WINED3DSPR_PATCHCONST,
1164 /* WINED3D_SM5_RT_DOMAIN_LOCATION */ WINED3DSPR_TESSCOORD,
1165 /* UNKNOWN */ ~0u,
1166 /* WINED3D_SM5_RT_UAV */ WINED3DSPR_UAV,
1167 /* WINED3D_SM5_RT_SHARED_MEMORY */ WINED3DSPR_GROUPSHAREDMEM,
1168 /* WINED3D_SM5_RT_THREAD_ID */ WINED3DSPR_THREADID,
1169 /* WINED3D_SM5_RT_THREAD_GROUP_ID */ WINED3DSPR_THREADGROUPID,
1170 /* WINED3D_SM5_RT_LOCAL_THREAD_ID */ WINED3DSPR_LOCALTHREADID,
1171 /* WINED3D_SM5_RT_COVERAGE */ WINED3DSPR_COVERAGE,
1172 /* WINED3D_SM5_RT_LOCAL_THREAD_INDEX */ WINED3DSPR_LOCALTHREADINDEX,
1173 /* WINED3D_SM5_RT_GS_INSTANCE_ID */ WINED3DSPR_GSINSTID,
1174 /* WINED3D_SM5_RT_DEPTHOUT_GREATER_EQUAL */ WINED3DSPR_DEPTHOUTGE,
1175 /* WINED3D_SM5_RT_DEPTHOUT_LESS_EQUAL */ WINED3DSPR_DEPTHOUTLE,
1178 static const struct wined3d_sm4_opcode_info *get_opcode_info(enum wined3d_sm4_opcode opcode)
1180 unsigned int i;
1182 for (i = 0; i < ARRAY_SIZE(opcode_table); ++i)
1184 if (opcode == opcode_table[i].opcode) return &opcode_table[i];
1187 return NULL;
1190 static void map_register(const struct wined3d_sm4_data *priv, struct wined3d_shader_register *reg)
1192 switch (priv->shader_version.type)
1194 case WINED3D_SHADER_TYPE_PIXEL:
1195 if (reg->type == WINED3DSPR_OUTPUT)
1197 unsigned int reg_idx = reg->idx[0].offset;
1199 if (reg_idx >= ARRAY_SIZE(priv->output_map))
1201 ERR("Invalid output index %u.\n", reg_idx);
1202 break;
1205 reg->type = WINED3DSPR_COLOROUT;
1206 reg->idx[0].offset = priv->output_map[reg_idx];
1208 break;
1210 default:
1211 break;
1215 static enum wined3d_data_type map_data_type(char t)
1217 switch (t)
1219 case 'f':
1220 return WINED3D_DATA_FLOAT;
1221 case 'i':
1222 return WINED3D_DATA_INT;
1223 case 'u':
1224 return WINED3D_DATA_UINT;
1225 case 'O':
1226 return WINED3D_DATA_OPAQUE;
1227 case 'R':
1228 return WINED3D_DATA_RESOURCE;
1229 case 'S':
1230 return WINED3D_DATA_SAMPLER;
1231 case 'U':
1232 return WINED3D_DATA_UAV;
1233 default:
1234 ERR("Invalid data type '%c'.\n", t);
1235 return WINED3D_DATA_FLOAT;
1239 static enum wined3d_shader_type wined3d_get_sm4_shader_type(const DWORD *byte_code, size_t byte_code_size)
1241 DWORD shader_type;
1243 if (byte_code_size / sizeof(*byte_code) < 1)
1245 WARN("Invalid byte code size %lu.\n", (long)byte_code_size);
1246 return WINED3D_SHADER_TYPE_INVALID;
1249 shader_type = byte_code[0] >> 16;
1250 switch (shader_type)
1252 case WINED3D_SM4_PS:
1253 return WINED3D_SHADER_TYPE_PIXEL;
1254 break;
1255 case WINED3D_SM4_VS:
1256 return WINED3D_SHADER_TYPE_VERTEX;
1257 break;
1258 case WINED3D_SM4_GS:
1259 return WINED3D_SHADER_TYPE_GEOMETRY;
1260 break;
1261 case WINED3D_SM5_HS:
1262 return WINED3D_SHADER_TYPE_HULL;
1263 break;
1264 case WINED3D_SM5_DS:
1265 return WINED3D_SHADER_TYPE_DOMAIN;
1266 break;
1267 case WINED3D_SM5_CS:
1268 return WINED3D_SHADER_TYPE_COMPUTE;
1269 break;
1270 default:
1271 FIXME("Unrecognised shader type %#x.\n", shader_type);
1272 return WINED3D_SHADER_TYPE_INVALID;
1276 static void *shader_sm4_init(const DWORD *byte_code, size_t byte_code_size,
1277 const struct wined3d_shader_signature *output_signature)
1279 DWORD version_token, token_count;
1280 struct wined3d_sm4_data *priv;
1281 unsigned int i;
1283 if (byte_code_size / sizeof(*byte_code) < 2)
1285 WARN("Invalid byte code size %lu.\n", (long)byte_code_size);
1286 return NULL;
1289 version_token = byte_code[0];
1290 TRACE("Version: 0x%08x.\n", version_token);
1291 token_count = byte_code[1];
1292 TRACE("Token count: %u.\n", token_count);
1294 if (token_count < 2 || byte_code_size / sizeof(*byte_code) < token_count)
1296 WARN("Invalid token count %u.\n", token_count);
1297 return NULL;
1300 if (!(priv = heap_alloc(sizeof(*priv))))
1302 ERR("Failed to allocate private data\n");
1303 return NULL;
1306 priv->start = &byte_code[2];
1307 priv->end = &byte_code[token_count];
1309 priv->shader_version.type = wined3d_get_sm4_shader_type(byte_code, byte_code_size);
1310 if (priv->shader_version.type == WINED3D_SHADER_TYPE_INVALID)
1312 heap_free(priv);
1313 return NULL;
1316 priv->shader_version.major = WINED3D_SM4_VERSION_MAJOR(version_token);
1317 priv->shader_version.minor = WINED3D_SM4_VERSION_MINOR(version_token);
1319 memset(priv->output_map, 0xff, sizeof(priv->output_map));
1320 for (i = 0; i < output_signature->element_count; ++i)
1322 struct wined3d_shader_signature_element *e = &output_signature->elements[i];
1324 if (e->register_idx >= ARRAY_SIZE(priv->output_map))
1326 WARN("Invalid output index %u.\n", e->register_idx);
1327 continue;
1330 priv->output_map[e->register_idx] = e->semantic_idx;
1333 list_init(&priv->src_free);
1334 list_init(&priv->src);
1336 return priv;
1339 static void shader_sm4_free(void *data)
1341 struct wined3d_shader_src_param_entry *e1, *e2;
1342 struct wined3d_sm4_data *priv = data;
1344 list_move_head(&priv->src_free, &priv->src);
1345 LIST_FOR_EACH_ENTRY_SAFE(e1, e2, &priv->src_free, struct wined3d_shader_src_param_entry, entry)
1347 heap_free(e1);
1349 heap_free(priv);
1352 static struct wined3d_shader_src_param *get_src_param(struct wined3d_sm4_data *priv)
1354 struct wined3d_shader_src_param_entry *e;
1355 struct list *elem;
1357 if (!list_empty(&priv->src_free))
1359 elem = list_head(&priv->src_free);
1360 list_remove(elem);
1362 else
1364 if (!(e = heap_alloc(sizeof(*e))))
1365 return NULL;
1366 elem = &e->entry;
1369 list_add_tail(&priv->src, elem);
1370 e = LIST_ENTRY(elem, struct wined3d_shader_src_param_entry, entry);
1371 return &e->param;
1374 static void shader_sm4_read_header(void *data, const DWORD **ptr, struct wined3d_shader_version *shader_version)
1376 struct wined3d_sm4_data *priv = data;
1378 *ptr = priv->start;
1379 *shader_version = priv->shader_version;
1382 static BOOL shader_sm4_read_reg_idx(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1383 DWORD addressing, struct wined3d_shader_register_index *reg_idx)
1385 if (addressing & WINED3D_SM4_ADDRESSING_RELATIVE)
1387 struct wined3d_shader_src_param *rel_addr = get_src_param(priv);
1389 if (!(reg_idx->rel_addr = rel_addr))
1391 ERR("Failed to get src param for relative addressing.\n");
1392 return FALSE;
1395 if (addressing & WINED3D_SM4_ADDRESSING_OFFSET)
1396 reg_idx->offset = *(*ptr)++;
1397 else
1398 reg_idx->offset = 0;
1399 shader_sm4_read_src_param(priv, ptr, end, WINED3D_DATA_INT, rel_addr);
1401 else
1403 reg_idx->rel_addr = NULL;
1404 reg_idx->offset = *(*ptr)++;
1407 return TRUE;
1410 static BOOL shader_sm4_read_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1411 enum wined3d_data_type data_type, struct wined3d_shader_register *param,
1412 enum wined3d_shader_src_modifier *modifier)
1414 enum wined3d_sm4_register_type register_type;
1415 DWORD token, order;
1417 if (*ptr >= end)
1419 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1420 return FALSE;
1422 token = *(*ptr)++;
1424 register_type = (token & WINED3D_SM4_REGISTER_TYPE_MASK) >> WINED3D_SM4_REGISTER_TYPE_SHIFT;
1425 if (register_type >= ARRAY_SIZE(register_type_table)
1426 || register_type_table[register_type] == ~0u)
1428 FIXME("Unhandled register type %#x.\n", register_type);
1429 param->type = WINED3DSPR_TEMP;
1431 else
1433 param->type = register_type_table[register_type];
1435 param->data_type = data_type;
1437 if (token & WINED3D_SM4_REGISTER_MODIFIER)
1439 DWORD m;
1441 if (*ptr >= end)
1443 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1444 return FALSE;
1446 m = *(*ptr)++;
1448 switch (m)
1450 case 0x41:
1451 *modifier = WINED3DSPSM_NEG;
1452 break;
1454 case 0x81:
1455 *modifier = WINED3DSPSM_ABS;
1456 break;
1458 case 0xc1:
1459 *modifier = WINED3DSPSM_ABSNEG;
1460 break;
1462 default:
1463 FIXME("Skipping modifier 0x%08x.\n", m);
1464 *modifier = WINED3DSPSM_NONE;
1465 break;
1468 else
1470 *modifier = WINED3DSPSM_NONE;
1473 order = (token & WINED3D_SM4_REGISTER_ORDER_MASK) >> WINED3D_SM4_REGISTER_ORDER_SHIFT;
1475 if (order < 1)
1476 param->idx[0].offset = ~0U;
1477 else
1479 DWORD addressing = (token & WINED3D_SM4_ADDRESSING_MASK0) >> WINED3D_SM4_ADDRESSING_SHIFT0;
1480 if (!(shader_sm4_read_reg_idx(priv, ptr, end, addressing, &param->idx[0])))
1482 ERR("Failed to read register index.\n");
1483 return FALSE;
1487 if (order < 2)
1488 param->idx[1].offset = ~0U;
1489 else
1491 DWORD addressing = (token & WINED3D_SM4_ADDRESSING_MASK1) >> WINED3D_SM4_ADDRESSING_SHIFT1;
1492 if (!(shader_sm4_read_reg_idx(priv, ptr, end, addressing, &param->idx[1])))
1494 ERR("Failed to read register index.\n");
1495 return FALSE;
1499 if (order > 2)
1500 FIXME("Unhandled order %u.\n", order);
1502 if (register_type == WINED3D_SM4_RT_IMMCONST)
1504 enum wined3d_sm4_dimension dimension = (token & WINED3D_SM4_DIMENSION_MASK) >> WINED3D_SM4_DIMENSION_SHIFT;
1506 switch (dimension)
1508 case WINED3D_SM4_DIMENSION_SCALAR:
1509 param->immconst_type = WINED3D_IMMCONST_SCALAR;
1510 if (end - *ptr < 1)
1512 WARN("Invalid ptr %p, end %p.\n", *ptr, end);
1513 return FALSE;
1515 memcpy(param->u.immconst_data, *ptr, 1 * sizeof(DWORD));
1516 *ptr += 1;
1517 break;
1519 case WINED3D_SM4_DIMENSION_VEC4:
1520 param->immconst_type = WINED3D_IMMCONST_VEC4;
1521 if (end - *ptr < 4)
1523 WARN("Invalid ptr %p, end %p.\n", *ptr, end);
1524 return FALSE;
1526 memcpy(param->u.immconst_data, *ptr, 4 * sizeof(DWORD));
1527 *ptr += 4;
1528 break;
1530 default:
1531 FIXME("Unhandled dimension %#x.\n", dimension);
1532 break;
1536 map_register(priv, param);
1538 return TRUE;
1541 static BOOL shader_sm4_read_src_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1542 enum wined3d_data_type data_type, struct wined3d_shader_src_param *src_param)
1544 DWORD token;
1546 if (*ptr >= end)
1548 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1549 return FALSE;
1551 token = **ptr;
1553 if (!shader_sm4_read_param(priv, ptr, end, data_type, &src_param->reg, &src_param->modifiers))
1555 ERR("Failed to read parameter.\n");
1556 return FALSE;
1559 if (src_param->reg.type == WINED3DSPR_IMMCONST)
1561 src_param->swizzle = WINED3DSP_NOSWIZZLE;
1563 else
1565 enum wined3d_sm4_swizzle_type swizzle_type =
1566 (token & WINED3D_SM4_SWIZZLE_TYPE_MASK) >> WINED3D_SM4_SWIZZLE_TYPE_SHIFT;
1568 switch (swizzle_type)
1570 case WINED3D_SM4_SWIZZLE_NONE:
1571 src_param->swizzle = WINED3DSP_NOSWIZZLE;
1572 break;
1574 case WINED3D_SM4_SWIZZLE_SCALAR:
1575 src_param->swizzle = (token & WINED3D_SM4_SWIZZLE_MASK) >> WINED3D_SM4_SWIZZLE_SHIFT;
1576 src_param->swizzle = (src_param->swizzle & 0x3) * 0x55;
1577 break;
1579 case WINED3D_SM4_SWIZZLE_VEC4:
1580 src_param->swizzle = (token & WINED3D_SM4_SWIZZLE_MASK) >> WINED3D_SM4_SWIZZLE_SHIFT;
1581 break;
1583 default:
1584 FIXME("Unhandled swizzle type %#x.\n", swizzle_type);
1585 break;
1589 return TRUE;
1592 static BOOL shader_sm4_read_dst_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1593 enum wined3d_data_type data_type, struct wined3d_shader_dst_param *dst_param)
1595 enum wined3d_shader_src_modifier modifier;
1596 DWORD token;
1598 if (*ptr >= end)
1600 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1601 return FALSE;
1603 token = **ptr;
1605 if (!shader_sm4_read_param(priv, ptr, end, data_type, &dst_param->reg, &modifier))
1607 ERR("Failed to read parameter.\n");
1608 return FALSE;
1611 if (modifier != WINED3DSPSM_NONE)
1613 ERR("Invalid source modifier %#x on destination register.\n", modifier);
1614 return FALSE;
1617 dst_param->write_mask = (token & WINED3D_SM4_WRITEMASK_MASK) >> WINED3D_SM4_WRITEMASK_SHIFT;
1618 dst_param->modifiers = 0;
1619 dst_param->shift = 0;
1621 return TRUE;
1624 static void shader_sm4_read_instruction_modifier(DWORD modifier, struct wined3d_shader_instruction *ins)
1626 enum wined3d_sm4_instruction_modifier modifier_type = modifier & WINED3D_SM4_MODIFIER_MASK;
1628 switch (modifier_type)
1630 case WINED3D_SM4_MODIFIER_AOFFIMMI:
1632 static const DWORD recognized_bits = WINED3D_SM4_INSTRUCTION_MODIFIER
1633 | WINED3D_SM4_MODIFIER_MASK
1634 | WINED3D_SM4_AOFFIMMI_U_MASK
1635 | WINED3D_SM4_AOFFIMMI_V_MASK
1636 | WINED3D_SM4_AOFFIMMI_W_MASK;
1638 /* Bit fields are used for sign extension. */
1639 struct
1641 int u : 4;
1642 int v : 4;
1643 int w : 4;
1644 } aoffimmi;
1646 if (modifier & ~recognized_bits)
1647 FIXME("Unhandled instruction modifier %#x.\n", modifier);
1649 aoffimmi.u = (modifier & WINED3D_SM4_AOFFIMMI_U_MASK) >> WINED3D_SM4_AOFFIMMI_U_SHIFT;
1650 aoffimmi.v = (modifier & WINED3D_SM4_AOFFIMMI_V_MASK) >> WINED3D_SM4_AOFFIMMI_V_SHIFT;
1651 aoffimmi.w = (modifier & WINED3D_SM4_AOFFIMMI_W_MASK) >> WINED3D_SM4_AOFFIMMI_W_SHIFT;
1652 ins->texel_offset.u = aoffimmi.u;
1653 ins->texel_offset.v = aoffimmi.v;
1654 ins->texel_offset.w = aoffimmi.w;
1655 break;
1658 case WINED3D_SM5_MODIFIER_DATA_TYPE:
1660 DWORD components = (modifier & WINED3D_SM5_MODIFIER_DATA_TYPE_MASK) >> WINED3D_SM5_MODIFIER_DATA_TYPE_SHIFT;
1661 enum wined3d_sm4_data_type data_type = components & 0xf;
1663 if ((components & 0xfff0) != (components & 0xf) * 0x1110)
1664 FIXME("Components (%#x) have different data types.\n", components);
1665 ins->resource_data_type = data_type_table[data_type];
1666 break;
1669 case WINED3D_SM5_MODIFIER_RESOURCE_TYPE:
1671 enum wined3d_sm4_resource_type resource_type
1672 = (modifier & WINED3D_SM5_MODIFIER_RESOURCE_TYPE_MASK) >> WINED3D_SM5_MODIFIER_RESOURCE_TYPE_SHIFT;
1674 ins->resource_type = resource_type_table[resource_type];
1675 break;
1678 default:
1679 FIXME("Unhandled instruction modifier %#x.\n", modifier);
1683 static void shader_sm4_read_instruction(void *data, const DWORD **ptr, struct wined3d_shader_instruction *ins)
1685 const struct wined3d_sm4_opcode_info *opcode_info;
1686 DWORD opcode_token, opcode, previous_token;
1687 struct wined3d_sm4_data *priv = data;
1688 unsigned int i, len;
1689 SIZE_T remaining;
1690 const DWORD *p;
1691 DWORD precise;
1693 list_move_head(&priv->src_free, &priv->src);
1695 if (*ptr >= priv->end)
1697 WARN("End of byte-code, failed to read opcode.\n");
1698 goto fail;
1700 remaining = priv->end - *ptr;
1702 opcode_token = *(*ptr)++;
1703 opcode = opcode_token & WINED3D_SM4_OPCODE_MASK;
1705 len = ((opcode_token & WINED3D_SM4_INSTRUCTION_LENGTH_MASK) >> WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT);
1706 if (!len)
1708 if (remaining < 2)
1710 WARN("End of byte-code, failed to read length token.\n");
1711 goto fail;
1713 len = **ptr;
1715 if (!len || remaining < len)
1717 WARN("Read invalid length %u (remaining %lu).\n", len, remaining);
1718 goto fail;
1720 --len;
1722 if (TRACE_ON(d3d_bytecode))
1724 TRACE_(d3d_bytecode)("[ %08x ", opcode_token);
1725 for (i = 0; i < len; ++i)
1727 TRACE_(d3d_bytecode)("%08x ", (*ptr)[i]);
1729 TRACE_(d3d_bytecode)("]\n");
1732 if (!(opcode_info = get_opcode_info(opcode)))
1734 FIXME("Unrecognized opcode %#x, opcode_token 0x%08x.\n", opcode, opcode_token);
1735 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1736 *ptr += len;
1737 return;
1740 ins->handler_idx = opcode_info->handler_idx;
1741 ins->flags = 0;
1742 ins->coissue = 0;
1743 ins->predicate = NULL;
1744 ins->dst_count = strlen(opcode_info->dst_info);
1745 ins->dst = priv->dst_param;
1746 ins->src_count = strlen(opcode_info->src_info);
1747 ins->src = priv->src_param;
1748 ins->resource_type = WINED3D_SHADER_RESOURCE_NONE;
1749 ins->resource_data_type = WINED3D_DATA_FLOAT;
1750 memset(&ins->texel_offset, 0, sizeof(ins->texel_offset));
1752 p = *ptr;
1753 *ptr += len;
1755 if (opcode_info->read_opcode_func)
1757 opcode_info->read_opcode_func(ins, opcode, opcode_token, p, len, priv);
1759 else
1761 enum wined3d_shader_dst_modifier instruction_dst_modifier = WINED3DSPDM_NONE;
1763 previous_token = opcode_token;
1764 while (previous_token & WINED3D_SM4_INSTRUCTION_MODIFIER && p != *ptr)
1765 shader_sm4_read_instruction_modifier(previous_token = *p++, ins);
1767 ins->flags = (opcode_token & WINED3D_SM4_INSTRUCTION_FLAGS_MASK) >> WINED3D_SM4_INSTRUCTION_FLAGS_SHIFT;
1768 if (ins->flags & WINED3D_SM4_INSTRUCTION_FLAG_SATURATE)
1770 ins->flags &= ~WINED3D_SM4_INSTRUCTION_FLAG_SATURATE;
1771 instruction_dst_modifier = WINED3DSPDM_SATURATE;
1773 precise = (opcode_token & WINED3D_SM5_PRECISE_MASK) >> WINED3D_SM5_PRECISE_SHIFT;
1774 ins->flags |= precise << WINED3DSI_PRECISE_SHIFT;
1776 for (i = 0; i < ins->dst_count; ++i)
1778 if (!(shader_sm4_read_dst_param(priv, &p, *ptr, map_data_type(opcode_info->dst_info[i]),
1779 &priv->dst_param[i])))
1781 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1782 return;
1784 priv->dst_param[i].modifiers |= instruction_dst_modifier;
1787 for (i = 0; i < ins->src_count; ++i)
1789 if (!(shader_sm4_read_src_param(priv, &p, *ptr, map_data_type(opcode_info->src_info[i]),
1790 &priv->src_param[i])))
1792 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1793 return;
1798 return;
1800 fail:
1801 *ptr = priv->end;
1802 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1803 return;
1806 static BOOL shader_sm4_is_end(void *data, const DWORD **ptr)
1808 struct wined3d_sm4_data *priv = data;
1809 return *ptr == priv->end;
1812 const struct wined3d_shader_frontend sm4_shader_frontend =
1814 shader_sm4_init,
1815 shader_sm4_free,
1816 shader_sm4_read_header,
1817 shader_sm4_read_instruction,
1818 shader_sm4_is_end,
1821 #define TAG_AON9 WINEMAKEFOURCC('A', 'o', 'n', '9')
1822 #define TAG_DXBC WINEMAKEFOURCC('D', 'X', 'B', 'C')
1823 #define TAG_ISGN WINEMAKEFOURCC('I', 'S', 'G', 'N')
1824 #define TAG_OSG5 WINEMAKEFOURCC('O', 'S', 'G', '5')
1825 #define TAG_OSGN WINEMAKEFOURCC('O', 'S', 'G', 'N')
1826 #define TAG_PCSG WINEMAKEFOURCC('P', 'C', 'S', 'G')
1827 #define TAG_SHDR WINEMAKEFOURCC('S', 'H', 'D', 'R')
1828 #define TAG_SHEX WINEMAKEFOURCC('S', 'H', 'E', 'X')
1830 struct aon9_header
1832 DWORD chunk_size;
1833 DWORD shader_version;
1834 DWORD unknown;
1835 DWORD byte_code_offset;
1838 struct shader_handler_context
1840 struct wined3d_shader *shader;
1841 enum wined3d_shader_byte_code_format *format;
1842 unsigned int max_version;
1845 static void read_dword(const char **ptr, DWORD *d)
1847 memcpy(d, *ptr, sizeof(*d));
1848 *ptr += sizeof(*d);
1851 static BOOL require_space(size_t offset, size_t count, size_t size, size_t data_size)
1853 return !count || (data_size - offset) / count >= size;
1856 static void skip_dword_unknown(const char **ptr, unsigned int count)
1858 unsigned int i;
1859 DWORD d;
1861 WARN("Skipping %u unknown DWORDs:\n", count);
1862 for (i = 0; i < count; ++i)
1864 read_dword(ptr, &d);
1865 WARN("\t0x%08x\n", d);
1869 static HRESULT parse_dxbc(const char *data, SIZE_T data_size,
1870 HRESULT (*chunk_handler)(const char *data, DWORD data_size, DWORD tag, void *ctx), void *ctx)
1872 const char *ptr = data;
1873 HRESULT hr = S_OK;
1874 DWORD chunk_count;
1875 DWORD total_size;
1876 unsigned int i;
1877 DWORD version;
1878 DWORD tag;
1880 read_dword(&ptr, &tag);
1881 TRACE("tag: %s.\n", debugstr_an((const char *)&tag, 4));
1883 if (tag != TAG_DXBC)
1885 WARN("Wrong tag.\n");
1886 return E_INVALIDARG;
1889 WARN("Ignoring DXBC checksum.\n");
1890 skip_dword_unknown(&ptr, 4);
1892 read_dword(&ptr, &version);
1893 TRACE("version: %#x.\n", version);
1894 if (version != 0x00000001)
1896 WARN("Got unexpected DXBC version %#x.\n", version);
1897 return E_INVALIDARG;
1900 read_dword(&ptr, &total_size);
1901 TRACE("total size: %#x\n", total_size);
1903 read_dword(&ptr, &chunk_count);
1904 TRACE("chunk count: %#x\n", chunk_count);
1906 for (i = 0; i < chunk_count; ++i)
1908 DWORD chunk_tag, chunk_size;
1909 const char *chunk_ptr;
1910 DWORD chunk_offset;
1912 read_dword(&ptr, &chunk_offset);
1913 TRACE("chunk %u at offset %#x\n", i, chunk_offset);
1915 if (chunk_offset >= data_size || !require_space(chunk_offset, 2, sizeof(DWORD), data_size))
1917 WARN("Invalid chunk offset %#x (data size %#lx).\n", chunk_offset, data_size);
1918 return E_FAIL;
1921 chunk_ptr = data + chunk_offset;
1923 read_dword(&chunk_ptr, &chunk_tag);
1924 read_dword(&chunk_ptr, &chunk_size);
1926 if (!require_space(chunk_ptr - data, 1, chunk_size, data_size))
1928 WARN("Invalid chunk size %#x (data size %#lx, chunk offset %#x).\n",
1929 chunk_size, data_size, chunk_offset);
1930 return E_FAIL;
1933 if (FAILED(hr = chunk_handler(chunk_ptr, chunk_size, chunk_tag, ctx)))
1934 break;
1937 return hr;
1940 static const char *shader_get_string(const char *data, size_t data_size, DWORD offset)
1942 if (offset >= data_size)
1944 WARN("Invalid offset %#x (data size %#lx).\n", offset, (long)data_size);
1945 return NULL;
1948 if (!memchr( data + offset, 0, data_size - offset ))
1949 return NULL;
1951 return data + offset;
1954 static HRESULT shader_parse_signature(DWORD tag, const char *data, DWORD data_size,
1955 struct wined3d_shader_signature *s)
1957 struct wined3d_shader_signature_element *e;
1958 const char *ptr = data;
1959 unsigned int i;
1960 DWORD count;
1962 if (!require_space(0, 2, sizeof(DWORD), data_size))
1964 WARN("Invalid data size %#x.\n", data_size);
1965 return E_INVALIDARG;
1968 read_dword(&ptr, &count);
1969 TRACE("%u elements.\n", count);
1971 skip_dword_unknown(&ptr, 1); /* It seems to always be 0x00000008. */
1973 if (!require_space(ptr - data, count, 6 * sizeof(DWORD), data_size))
1975 WARN("Invalid count %#x (data size %#x).\n", count, data_size);
1976 return E_INVALIDARG;
1979 if (!(e = heap_calloc(count, sizeof(*e))))
1981 ERR("Failed to allocate input signature memory.\n");
1982 return E_OUTOFMEMORY;
1985 for (i = 0; i < count; ++i)
1987 DWORD name_offset;
1989 if (tag == TAG_OSG5)
1990 read_dword(&ptr, &e[i].stream_idx);
1991 else
1992 e[i].stream_idx = 0;
1993 read_dword(&ptr, &name_offset);
1994 if (!(e[i].semantic_name = shader_get_string(data, data_size, name_offset)))
1996 WARN("Invalid name offset %#x (data size %#x).\n", name_offset, data_size);
1997 heap_free(e);
1998 return E_INVALIDARG;
2000 read_dword(&ptr, &e[i].semantic_idx);
2001 read_dword(&ptr, &e[i].sysval_semantic);
2002 read_dword(&ptr, &e[i].component_type);
2003 read_dword(&ptr, &e[i].register_idx);
2004 read_dword(&ptr, &e[i].mask);
2006 TRACE("Stream: %u, semantic: %s, semantic idx: %u, sysval_semantic %#x, "
2007 "type %u, register idx: %u, use_mask %#x, input_mask %#x.\n",
2008 e[i].stream_idx, debugstr_a(e[i].semantic_name), e[i].semantic_idx, e[i].sysval_semantic,
2009 e[i].component_type, e[i].register_idx, (e[i].mask >> 8) & 0xff, e[i].mask & 0xff);
2012 s->elements = e;
2013 s->element_count = count;
2015 return S_OK;
2018 static HRESULT shader_dxbc_chunk_handler(const char *data, DWORD data_size, DWORD tag, void *context)
2020 struct shader_handler_context *ctx = context;
2021 struct wined3d_shader *shader = ctx->shader;
2022 HRESULT hr;
2024 switch (tag)
2026 case TAG_ISGN:
2027 if (ctx->max_version < 4)
2029 TRACE("Skipping shader input signature.\n");
2030 break;
2032 if (shader->input_signature.elements)
2034 FIXME("Multiple input signatures.\n");
2035 break;
2037 if (FAILED(hr = shader_parse_signature(tag, data, data_size, &shader->input_signature)))
2038 return hr;
2039 break;
2041 case TAG_OSGN:
2042 case TAG_OSG5:
2043 if (ctx->max_version < 4)
2045 TRACE("Skipping shader output signature.\n");
2046 break;
2048 if (shader->output_signature.elements)
2050 FIXME("Multiple output signatures.\n");
2051 break;
2053 if (FAILED(hr = shader_parse_signature(tag, data, data_size, &shader->output_signature)))
2054 return hr;
2055 break;
2057 case TAG_PCSG:
2058 if (shader->patch_constant_signature.elements)
2060 FIXME("Multiple patch constant signatures.\n");
2061 break;
2063 if (FAILED(hr = shader_parse_signature(tag, data, data_size, &shader->patch_constant_signature)))
2064 return hr;
2065 break;
2067 case TAG_SHDR:
2068 case TAG_SHEX:
2069 if (ctx->max_version < 4)
2071 TRACE("Skipping SM4+ shader.\n");
2072 break;
2074 if (shader->function)
2075 FIXME("Multiple shader code chunks.\n");
2076 shader->function = (const DWORD *)data;
2077 shader->functionLength = data_size;
2078 *ctx->format = WINED3D_SHADER_BYTE_CODE_FORMAT_SM4;
2079 break;
2081 case TAG_AON9:
2082 if (ctx->max_version < 4)
2084 const struct aon9_header *header = (const struct aon9_header *)data;
2085 unsigned int unknown_dword_count;
2086 const char *byte_code;
2088 if (data_size < sizeof(*header))
2090 WARN("Invalid Aon9 data size %#x.\n", data_size);
2091 return E_FAIL;
2093 byte_code = data + header->byte_code_offset;
2094 unknown_dword_count = (header->byte_code_offset - sizeof(*header)) / sizeof(DWORD);
2096 if (data_size - 2 * sizeof(DWORD) < header->byte_code_offset)
2098 WARN("Invalid byte code offset %#x (size %#x).\n", header->byte_code_offset, data_size);
2099 return E_FAIL;
2101 FIXME("Skipping %u unknown DWORDs.\n", unknown_dword_count);
2103 if (shader->function)
2104 FIXME("Multiple shader code chunks.\n");
2105 shader->function = (const DWORD *)byte_code;
2106 shader->functionLength = data_size - header->byte_code_offset;
2107 *ctx->format = WINED3D_SHADER_BYTE_CODE_FORMAT_SM1;
2108 TRACE("Feature level 9 shader version 0%08x, 0%08x.\n",
2109 header->shader_version, *shader->function);
2111 else
2113 TRACE("Skipping feature level 9 shader code.\n");
2115 break;
2117 default:
2118 TRACE("Skipping chunk %s.\n", debugstr_an((const char *)&tag, 4));
2119 break;
2122 return S_OK;
2125 HRESULT shader_extract_from_dxbc(struct wined3d_shader *shader,
2126 unsigned int max_shader_version, enum wined3d_shader_byte_code_format *format)
2128 struct shader_handler_context ctx;
2129 HRESULT hr;
2131 ctx.shader = shader;
2132 ctx.format = format;
2133 ctx.max_version = max_shader_version;
2135 hr = parse_dxbc(shader->byte_code, shader->byte_code_size, shader_dxbc_chunk_handler, &ctx);
2136 if (!shader->function)
2137 hr = E_INVALIDARG;
2139 if (FAILED(hr))
2140 WARN("Failed to parse DXBC, hr %#x.\n", hr);
2142 return hr;
2145 static HRESULT shader_isgn_chunk_handler(const char *data, DWORD data_size, DWORD tag, void *ctx)
2147 struct wined3d_shader_signature *is = ctx;
2149 if (tag != TAG_ISGN)
2150 return S_OK;
2152 if (is->elements)
2154 FIXME("Multiple shader signatures.\n");
2155 return S_OK;
2158 return shader_parse_signature(tag, data, data_size, is);
2161 HRESULT CDECL wined3d_extract_shader_input_signature_from_dxbc(struct wined3d_shader_signature *signature,
2162 const void *code, SIZE_T code_size)
2164 memset(signature, 0, sizeof(*signature));
2165 return parse_dxbc(code, code_size, shader_isgn_chunk_handler, signature);