1 /*
2 * Copyright 2009 Henri Verbeet for CodeWeavers
3 *
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.
8 *
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.
13 *
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
17 */
18
19 #include "config.h"
20 #include "wine/port.h"
21
22 #include "wined3d_private.h"
23
24 WINE_DEFAULT_DEBUG_CHANNEL(d3d_shader);
25 WINE_DECLARE_DEBUG_CHANNEL(d3d_bytecode);
26
27 #define WINED3D_SM4_INSTRUCTION_MODIFIER (0x1u << 31)
28
29 #define WINED3D_SM4_MODIFIER_AOFFIMMI 0x1
30 #define WINED3D_SM4_AOFFIMMI_U_SHIFT 9
31 #define WINED3D_SM4_AOFFIMMI_U_MASK (0xfu << WINED3D_SM4_AOFFIMMI_U_SHIFT)
32 #define WINED3D_SM4_AOFFIMMI_V_SHIFT 13
33 #define WINED3D_SM4_AOFFIMMI_V_MASK (0xfu << WINED3D_SM4_AOFFIMMI_V_SHIFT)
34 #define WINED3D_SM4_AOFFIMMI_W_SHIFT 17
35 #define WINED3D_SM4_AOFFIMMI_W_MASK (0xfu << WINED3D_SM4_AOFFIMMI_W_SHIFT)
36
37 #define WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT 24
38 #define WINED3D_SM4_INSTRUCTION_LENGTH_MASK (0x1fu << WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT)
39
40 #define WINED3D_SM4_INSTRUCTION_FLAGS_SHIFT 11
41 #define WINED3D_SM4_INSTRUCTION_FLAGS_MASK (0x7u << WINED3D_SM4_INSTRUCTION_FLAGS_SHIFT)
42
43 #define WINED3D_SM4_RESOURCE_TYPE_SHIFT 11
44 #define WINED3D_SM4_RESOURCE_TYPE_MASK (0xfu << WINED3D_SM4_RESOURCE_TYPE_SHIFT)
45
46 #define WINED3D_SM4_PRIMITIVE_TYPE_SHIFT 11
47 #define WINED3D_SM4_PRIMITIVE_TYPE_MASK (0x3fu << WINED3D_SM4_PRIMITIVE_TYPE_SHIFT)
48
49 #define WINED3D_SM4_INDEX_TYPE_SHIFT 11
50 #define WINED3D_SM4_INDEX_TYPE_MASK (0x1u << WINED3D_SM4_INDEX_TYPE_SHIFT)
51
52 #define WINED3D_SM4_SAMPLER_MODE_SHIFT 11
53 #define WINED3D_SM4_SAMPLER_MODE_MASK (0xfu << WINED3D_SM4_SAMPLER_MODE_SHIFT)
54
55 #define WINED3D_SM4_SHADER_DATA_TYPE_SHIFT 11
56 #define WINED3D_SM4_SHADER_DATA_TYPE_MASK (0xfu << WINED3D_SM4_SHADER_DATA_TYPE_SHIFT)
57
58 #define WINED3D_SM4_INTERPOLATION_MODE_SHIFT 11
59 #define WINED3D_SM4_INTERPOLATION_MODE_MASK (0xfu << WINED3D_SM4_INTERPOLATION_MODE_SHIFT)
60
61 #define WINED3D_SM4_GLOBAL_FLAGS_SHIFT 11
62 #define WINED3D_SM4_GLOBAL_FLAGS_MASK (0xffu << WINED3D_SM4_GLOBAL_FLAGS_SHIFT)
63
64 #define WINED3D_SM5_CONTROL_POINT_COUNT_SHIFT 11
65 #define WINED3D_SM5_CONTROL_POINT_COUNT_MASK (0xffu << WINED3D_SM5_CONTROL_POINT_COUNT_SHIFT)
66
67 #define WINED3D_SM5_FP_ARRAY_SIZE_SHIFT 16
68 #define WINED3D_SM5_FP_TABLE_COUNT_MASK 0xffffu
69
70 #define WINED3D_SM5_UAV_FLAGS_SHIFT 15
71 #define WINED3D_SM5_UAV_FLAGS_MASK (0x1ffu << WINED3D_SM5_UAV_FLAGS_SHIFT)
72
73 #define WINED3D_SM5_SYNC_FLAGS_SHIFT 11
74 #define WINED3D_SM5_SYNC_FLAGS_MASK (0xffu << WINED3D_SM5_SYNC_FLAGS_SHIFT)
75
76 #define WINED3D_SM5_TESSELLATOR_SHIFT 11
77 #define WINED3D_SM5_TESSELLATOR_MASK (0xfu << WINED3D_SM5_TESSELLATOR_SHIFT)
78
79 #define WINED3D_SM4_OPCODE_MASK 0xff
80
81 #define WINED3D_SM4_REGISTER_MODIFIER (0x1u << 31)
82
83 #define WINED3D_SM4_ADDRESSING_SHIFT1 25
84 #define WINED3D_SM4_ADDRESSING_MASK1 (0x3u << WINED3D_SM4_ADDRESSING_SHIFT1)
85
86 #define WINED3D_SM4_ADDRESSING_SHIFT0 22
87 #define WINED3D_SM4_ADDRESSING_MASK0 (0x3u << WINED3D_SM4_ADDRESSING_SHIFT0)
88
89 #define WINED3D_SM4_REGISTER_ORDER_SHIFT 20
90 #define WINED3D_SM4_REGISTER_ORDER_MASK (0x3u << WINED3D_SM4_REGISTER_ORDER_SHIFT)
91
92 #define WINED3D_SM4_REGISTER_TYPE_SHIFT 12
93 #define WINED3D_SM4_REGISTER_TYPE_MASK (0xffu << WINED3D_SM4_REGISTER_TYPE_SHIFT)
94
95 #define WINED3D_SM4_SWIZZLE_TYPE_SHIFT 2
96 #define WINED3D_SM4_SWIZZLE_TYPE_MASK (0x3u << WINED3D_SM4_SWIZZLE_TYPE_SHIFT)
97
98 #define WINED3D_SM4_IMMCONST_TYPE_SHIFT 0
99 #define WINED3D_SM4_IMMCONST_TYPE_MASK (0x3u << WINED3D_SM4_IMMCONST_TYPE_SHIFT)
100
101 #define WINED3D_SM4_WRITEMASK_SHIFT 4
102 #define WINED3D_SM4_WRITEMASK_MASK (0xfu << WINED3D_SM4_WRITEMASK_SHIFT)
103
104 #define WINED3D_SM4_SWIZZLE_SHIFT 4
105 #define WINED3D_SM4_SWIZZLE_MASK (0xffu << WINED3D_SM4_SWIZZLE_SHIFT)
106
107 #define WINED3D_SM4_VERSION_MAJOR(version) (((version) >> 4) & 0xf)
108 #define WINED3D_SM4_VERSION_MINOR(version) (((version) >> 0) & 0xf)
109
110 #define WINED3D_SM4_ADDRESSING_RELATIVE 0x2
111 #define WINED3D_SM4_ADDRESSING_OFFSET 0x1
112
113 #define WINED3D_SM4_INSTRUCTION_FLAG_SATURATE 0x4
114
115 #define WINED3D_SM4_CONDITIONAL_NZ (0x1u << 18)
116
117 enum wined3d_sm4_opcode
118 {
119 WINED3D_SM4_OP_ADD = 0x00,
120 WINED3D_SM4_OP_AND = 0x01,
121 WINED3D_SM4_OP_BREAK = 0x02,
122 WINED3D_SM4_OP_BREAKC = 0x03,
123 WINED3D_SM4_OP_CASE = 0x06,
124 WINED3D_SM4_OP_CONTINUE = 0x07,
125 WINED3D_SM4_OP_CONTINUEC = 0x08,
126 WINED3D_SM4_OP_CUT = 0x09,
127 WINED3D_SM4_OP_DEFAULT = 0x0a,
128 WINED3D_SM4_OP_DERIV_RTX = 0x0b,
129 WINED3D_SM4_OP_DERIV_RTY = 0x0c,
130 WINED3D_SM4_OP_DISCARD = 0x0d,
131 WINED3D_SM4_OP_DIV = 0x0e,
132 WINED3D_SM4_OP_DP2 = 0x0f,
133 WINED3D_SM4_OP_DP3 = 0x10,
134 WINED3D_SM4_OP_DP4 = 0x11,
135 WINED3D_SM4_OP_ELSE = 0x12,
136 WINED3D_SM4_OP_EMIT = 0x13,
137 WINED3D_SM4_OP_ENDIF = 0x15,
138 WINED3D_SM4_OP_ENDLOOP = 0x16,
139 WINED3D_SM4_OP_ENDSWITCH = 0x17,
140 WINED3D_SM4_OP_EQ = 0x18,
141 WINED3D_SM4_OP_EXP = 0x19,
142 WINED3D_SM4_OP_FRC = 0x1a,
143 WINED3D_SM4_OP_FTOI = 0x1b,
144 WINED3D_SM4_OP_FTOU = 0x1c,
145 WINED3D_SM4_OP_GE = 0x1d,
146 WINED3D_SM4_OP_IADD = 0x1e,
147 WINED3D_SM4_OP_IF = 0x1f,
148 WINED3D_SM4_OP_IEQ = 0x20,
149 WINED3D_SM4_OP_IGE = 0x21,
150 WINED3D_SM4_OP_ILT = 0x22,
151 WINED3D_SM4_OP_IMAD = 0x23,
152 WINED3D_SM4_OP_IMAX = 0x24,
153 WINED3D_SM4_OP_IMIN = 0x25,
154 WINED3D_SM4_OP_IMUL = 0x26,
155 WINED3D_SM4_OP_INE = 0x27,
156 WINED3D_SM4_OP_INEG = 0x28,
157 WINED3D_SM4_OP_ISHL = 0x29,
158 WINED3D_SM4_OP_ISHR = 0x2a,
159 WINED3D_SM4_OP_ITOF = 0x2b,
160 WINED3D_SM4_OP_LABEL = 0x2c,
161 WINED3D_SM4_OP_LD = 0x2d,
162 WINED3D_SM4_OP_LD2DMS = 0x2e,
163 WINED3D_SM4_OP_LOG = 0x2f,
164 WINED3D_SM4_OP_LOOP = 0x30,
165 WINED3D_SM4_OP_LT = 0x31,
166 WINED3D_SM4_OP_MAD = 0x32,
167 WINED3D_SM4_OP_MIN = 0x33,
168 WINED3D_SM4_OP_MAX = 0x34,
169 WINED3D_SM4_OP_SHADER_DATA = 0x35,
170 WINED3D_SM4_OP_MOV = 0x36,
171 WINED3D_SM4_OP_MOVC = 0x37,
172 WINED3D_SM4_OP_MUL = 0x38,
173 WINED3D_SM4_OP_NE = 0x39,
174 WINED3D_SM4_OP_NOP = 0x3a,
175 WINED3D_SM4_OP_NOT = 0x3b,
176 WINED3D_SM4_OP_OR = 0x3c,
177 WINED3D_SM4_OP_RESINFO = 0x3d,
178 WINED3D_SM4_OP_RET = 0x3e,
179 WINED3D_SM4_OP_RETC = 0x3f,
180 WINED3D_SM4_OP_ROUND_NE = 0x40,
181 WINED3D_SM4_OP_ROUND_NI = 0x41,
182 WINED3D_SM4_OP_ROUND_PI = 0x42,
183 WINED3D_SM4_OP_ROUND_Z = 0x43,
184 WINED3D_SM4_OP_RSQ = 0x44,
185 WINED3D_SM4_OP_SAMPLE = 0x45,
186 WINED3D_SM4_OP_SAMPLE_C = 0x46,
187 WINED3D_SM4_OP_SAMPLE_C_LZ = 0x47,
188 WINED3D_SM4_OP_SAMPLE_LOD = 0x48,
189 WINED3D_SM4_OP_SAMPLE_GRAD = 0x49,
190 WINED3D_SM4_OP_SAMPLE_B = 0x4a,
191 WINED3D_SM4_OP_SQRT = 0x4b,
192 WINED3D_SM4_OP_SWITCH = 0x4c,
193 WINED3D_SM4_OP_SINCOS = 0x4d,
194 WINED3D_SM4_OP_UDIV = 0x4e,
195 WINED3D_SM4_OP_ULT = 0x4f,
196 WINED3D_SM4_OP_UGE = 0x50,
197 WINED3D_SM4_OP_UMUL = 0x51,
198 WINED3D_SM4_OP_UMAX = 0x53,
199 WINED3D_SM4_OP_UMIN = 0x54,
200 WINED3D_SM4_OP_USHR = 0x55,
201 WINED3D_SM4_OP_UTOF = 0x56,
202 WINED3D_SM4_OP_XOR = 0x57,
203 WINED3D_SM4_OP_DCL_RESOURCE = 0x58,
204 WINED3D_SM4_OP_DCL_CONSTANT_BUFFER = 0x59,
205 WINED3D_SM4_OP_DCL_SAMPLER = 0x5a,
206 WINED3D_SM4_OP_DCL_INDEX_RANGE = 0x5b,
207 WINED3D_SM4_OP_DCL_OUTPUT_TOPOLOGY = 0x5c,
208 WINED3D_SM4_OP_DCL_INPUT_PRIMITIVE = 0x5d,
209 WINED3D_SM4_OP_DCL_VERTICES_OUT = 0x5e,
210 WINED3D_SM4_OP_DCL_INPUT = 0x5f,
211 WINED3D_SM4_OP_DCL_INPUT_SGV = 0x60,
212 WINED3D_SM4_OP_DCL_INPUT_SIV = 0x61,
213 WINED3D_SM4_OP_DCL_INPUT_PS = 0x62,
214 WINED3D_SM4_OP_DCL_INPUT_PS_SGV = 0x63,
215 WINED3D_SM4_OP_DCL_INPUT_PS_SIV = 0x64,
216 WINED3D_SM4_OP_DCL_OUTPUT = 0x65,
217 WINED3D_SM4_OP_DCL_OUTPUT_SIV = 0x67,
218 WINED3D_SM4_OP_DCL_TEMPS = 0x68,
219 WINED3D_SM4_OP_DCL_INDEXABLE_TEMP = 0x69,
220 WINED3D_SM4_OP_DCL_GLOBAL_FLAGS = 0x6a,
221 WINED3D_SM4_OP_LOD = 0x6c,
222 WINED3D_SM4_OP_GATHER4 = 0x6d,
223 WINED3D_SM4_OP_SAMPLE_POS = 0x6e,
224 WINED3D_SM4_OP_SAMPLE_INFO = 0x6f,
225 WINED3D_SM5_OP_HS_DECLS = 0x71,
226 WINED3D_SM5_OP_HS_CONTROL_POINT_PHASE = 0x72,
227 WINED3D_SM5_OP_HS_FORK_PHASE = 0x73,
228 WINED3D_SM5_OP_HS_JOIN_PHASE = 0x74,
229 WINED3D_SM5_OP_EMIT_STREAM = 0x75,
230 WINED3D_SM5_OP_CUT_STREAM = 0x76,
231 WINED3D_SM5_OP_FCALL = 0x78,
232 WINED3D_SM5_OP_BUFINFO = 0x79,
233 WINED3D_SM5_OP_DERIV_RTX_COARSE = 0x7a,
234 WINED3D_SM5_OP_DERIV_RTX_FINE = 0x7b,
235 WINED3D_SM5_OP_DERIV_RTY_COARSE = 0x7c,
236 WINED3D_SM5_OP_DERIV_RTY_FINE = 0x7d,
237 WINED3D_SM5_OP_GATHER4_C = 0x7e,
238 WINED3D_SM5_OP_GATHER4_PO = 0x7f,
239 WINED3D_SM5_OP_GATHER4_PO_C = 0x80,
240 WINED3D_SM5_OP_RCP = 0x81,
241 WINED3D_SM5_OP_F32TOF16 = 0x82,
242 WINED3D_SM5_OP_F16TOF32 = 0x83,
243 WINED3D_SM5_OP_COUNTBITS = 0x86,
244 WINED3D_SM5_OP_FIRSTBIT_HI = 0x87,
245 WINED3D_SM5_OP_FIRSTBIT_LO = 0x88,
246 WINED3D_SM5_OP_FIRSTBIT_SHI = 0x89,
247 WINED3D_SM5_OP_UBFE = 0x8a,
248 WINED3D_SM5_OP_IBFE = 0x8b,
249 WINED3D_SM5_OP_BFI = 0x8c,
250 WINED3D_SM5_OP_BFREV = 0x8d,
251 WINED3D_SM5_OP_SWAPC = 0x8e,
252 WINED3D_SM5_OP_DCL_STREAM = 0x8f,
253 WINED3D_SM5_OP_DCL_FUNCTION_BODY = 0x90,
254 WINED3D_SM5_OP_DCL_FUNCTION_TABLE = 0x91,
255 WINED3D_SM5_OP_DCL_INTERFACE = 0x92,
256 WINED3D_SM5_OP_DCL_INPUT_CONTROL_POINT_COUNT = 0x93,
257 WINED3D_SM5_OP_DCL_OUTPUT_CONTROL_POINT_COUNT = 0x94,
258 WINED3D_SM5_OP_DCL_TESSELLATOR_DOMAIN = 0x95,
259 WINED3D_SM5_OP_DCL_TESSELLATOR_PARTITIONING = 0x96,
260 WINED3D_SM5_OP_DCL_TESSELLATOR_OUTPUT_PRIMITIVE = 0x97,
261 WINED3D_SM5_OP_DCL_HS_MAX_TESSFACTOR = 0x98,
262 WINED3D_SM5_OP_DCL_HS_FORK_PHASE_INSTANCE_COUNT = 0x99,
263 WINED3D_SM5_OP_DCL_HS_JOIN_PHASE_INSTANCE_COUNT = 0x9a,
264 WINED3D_SM5_OP_DCL_THREAD_GROUP = 0x9b,
265 WINED3D_SM5_OP_DCL_UAV_TYPED = 0x9c,
266 WINED3D_SM5_OP_DCL_UAV_RAW = 0x9d,
267 WINED3D_SM5_OP_DCL_UAV_STRUCTURED = 0x9e,
268 WINED3D_SM5_OP_DCL_TGSM_RAW = 0x9f,
269 WINED3D_SM5_OP_DCL_TGSM_STRUCTURED = 0xa0,
270 WINED3D_SM5_OP_DCL_RESOURCE_RAW = 0xa1,
271 WINED3D_SM5_OP_DCL_RESOURCE_STRUCTURED = 0xa2,
272 WINED3D_SM5_OP_LD_UAV_TYPED = 0xa3,
273 WINED3D_SM5_OP_STORE_UAV_TYPED = 0xa4,
274 WINED3D_SM5_OP_LD_RAW = 0xa5,
275 WINED3D_SM5_OP_STORE_RAW = 0xa6,
276 WINED3D_SM5_OP_LD_STRUCTURED = 0xa7,
277 WINED3D_SM5_OP_STORE_STRUCTURED = 0xa8,
278 WINED3D_SM5_OP_ATOMIC_AND = 0xa9,
279 WINED3D_SM5_OP_ATOMIC_OR = 0xaa,
280 WINED3D_SM5_OP_ATOMIC_XOR = 0xab,
281 WINED3D_SM5_OP_ATOMIC_CMP_STORE = 0xac,
282 WINED3D_SM5_OP_ATOMIC_IADD = 0xad,
283 WINED3D_SM5_OP_ATOMIC_IMAX = 0xae,
284 WINED3D_SM5_OP_ATOMIC_IMIN = 0xaf,
285 WINED3D_SM5_OP_ATOMIC_UMAX = 0xb0,
286 WINED3D_SM5_OP_ATOMIC_UMIN = 0xb1,
287 WINED3D_SM5_OP_IMM_ATOMIC_ALLOC = 0xb2,
288 WINED3D_SM5_OP_IMM_ATOMIC_CONSUME = 0xb3,
289 WINED3D_SM5_OP_IMM_ATOMIC_IADD = 0xb4,
290 WINED3D_SM5_OP_IMM_ATOMIC_AND = 0xb5,
291 WINED3D_SM5_OP_IMM_ATOMIC_OR = 0xb6,
292 WINED3D_SM5_OP_IMM_ATOMIC_XOR = 0xb7,
293 WINED3D_SM5_OP_IMM_ATOMIC_EXCH = 0xb8,
294 WINED3D_SM5_OP_IMM_ATOMIC_CMP_EXCH = 0xb9,
295 WINED3D_SM5_OP_IMM_ATOMIC_IMAX = 0xba,
296 WINED3D_SM5_OP_IMM_ATOMIC_IMIN = 0xbb,
297 WINED3D_SM5_OP_IMM_ATOMIC_UMAX = 0xbc,
298 WINED3D_SM5_OP_IMM_ATOMIC_UMIN = 0xbd,
299 WINED3D_SM5_OP_SYNC = 0xbe,
300 WINED3D_SM5_OP_EVAL_SAMPLE_INDEX = 0xcc,
301 WINED3D_SM5_OP_DCL_GS_INSTANCES = 0xce,
302 };
303
304 enum wined3d_sm4_register_type
305 {
306 WINED3D_SM4_RT_TEMP = 0x00,
307 WINED3D_SM4_RT_INPUT = 0x01,
308 WINED3D_SM4_RT_OUTPUT = 0x02,
309 WINED3D_SM4_RT_INDEXABLE_TEMP = 0x03,
310 WINED3D_SM4_RT_IMMCONST = 0x04,
311 WINED3D_SM4_RT_SAMPLER = 0x06,
312 WINED3D_SM4_RT_RESOURCE = 0x07,
313 WINED3D_SM4_RT_CONSTBUFFER = 0x08,
314 WINED3D_SM4_RT_IMMCONSTBUFFER = 0x09,
315 WINED3D_SM4_RT_PRIMID = 0x0b,
316 WINED3D_SM4_RT_DEPTHOUT = 0x0c,
317 WINED3D_SM4_RT_NULL = 0x0d,
318 WINED3D_SM4_RT_OMASK = 0x0f,
319 WINED3D_SM5_RT_STREAM = 0x10,
320 WINED3D_SM5_RT_FUNCTION_BODY = 0x11,
321 WINED3D_SM5_RT_FUNCTION_POINTER = 0x13,
322 WINED3D_SM5_RT_OUTPUT_CONTROL_POINT_ID = 0x16,
323 WINED3D_SM5_RT_FORK_INSTANCE_ID = 0x17,
324 WINED3D_SM5_RT_JOIN_INSTANCE_ID = 0x18,
325 WINED3D_SM5_RT_INPUT_CONTROL_POINT = 0x19,
326 WINED3D_SM5_RT_OUTPUT_CONTROL_POINT = 0x1a,
327 WINED3D_SM5_RT_PATCH_CONSTANT_DATA = 0x1b,
328 WINED3D_SM5_RT_DOMAIN_LOCATION = 0x1c,
329 WINED3D_SM5_RT_UAV = 0x1e,
330 WINED3D_SM5_RT_SHARED_MEMORY = 0x1f,
331 WINED3D_SM5_RT_THREAD_ID = 0x20,
332 WINED3D_SM5_RT_THREAD_GROUP_ID = 0x21,
333 WINED3D_SM5_RT_LOCAL_THREAD_ID = 0x22,
334 WINED3D_SM5_RT_COVERAGE = 0x23,
335 WINED3D_SM5_RT_LOCAL_THREAD_INDEX = 0x24,
336 WINED3D_SM5_RT_GS_INSTANCE_ID = 0x25,
337 WINED3D_SM5_RT_DEPTHOUT_GREATER_EQUAL = 0x26,
338 WINED3D_SM5_RT_DEPTHOUT_LESS_EQUAL = 0x27,
339 };
340
341 enum wined3d_sm4_output_primitive_type
342 {
343 WINED3D_SM4_OUTPUT_PT_POINTLIST = 0x1,
344 WINED3D_SM4_OUTPUT_PT_LINESTRIP = 0x3,
345 WINED3D_SM4_OUTPUT_PT_TRIANGLESTRIP = 0x5,
346 };
347
348 enum wined3d_sm4_input_primitive_type
349 {
350 WINED3D_SM4_INPUT_PT_POINT = 0x01,
351 WINED3D_SM4_INPUT_PT_LINE = 0x02,
352 WINED3D_SM4_INPUT_PT_TRIANGLE = 0x03,
353 WINED3D_SM4_INPUT_PT_LINEADJ = 0x06,
354 WINED3D_SM4_INPUT_PT_TRIANGLEADJ = 0x07,
355 WINED3D_SM5_INPUT_PT_PATCH1 = 0x08,
356 WINED3D_SM5_INPUT_PT_PATCH2 = 0x09,
357 WINED3D_SM5_INPUT_PT_PATCH3 = 0x0a,
358 WINED3D_SM5_INPUT_PT_PATCH4 = 0x0b,
359 WINED3D_SM5_INPUT_PT_PATCH5 = 0x0c,
360 WINED3D_SM5_INPUT_PT_PATCH6 = 0x0d,
361 WINED3D_SM5_INPUT_PT_PATCH7 = 0x0e,
362 WINED3D_SM5_INPUT_PT_PATCH8 = 0x0f,
363 WINED3D_SM5_INPUT_PT_PATCH9 = 0x10,
364 WINED3D_SM5_INPUT_PT_PATCH10 = 0x11,
365 WINED3D_SM5_INPUT_PT_PATCH11 = 0x12,
366 WINED3D_SM5_INPUT_PT_PATCH12 = 0x13,
367 WINED3D_SM5_INPUT_PT_PATCH13 = 0x14,
368 WINED3D_SM5_INPUT_PT_PATCH14 = 0x15,
369 WINED3D_SM5_INPUT_PT_PATCH15 = 0x16,
370 WINED3D_SM5_INPUT_PT_PATCH16 = 0x17,
371 WINED3D_SM5_INPUT_PT_PATCH17 = 0x18,
372 WINED3D_SM5_INPUT_PT_PATCH18 = 0x19,
373 WINED3D_SM5_INPUT_PT_PATCH19 = 0x1a,
374 WINED3D_SM5_INPUT_PT_PATCH20 = 0x1b,
375 WINED3D_SM5_INPUT_PT_PATCH21 = 0x1c,
376 WINED3D_SM5_INPUT_PT_PATCH22 = 0x1d,
377 WINED3D_SM5_INPUT_PT_PATCH23 = 0x1e,
378 WINED3D_SM5_INPUT_PT_PATCH24 = 0x1f,
379 WINED3D_SM5_INPUT_PT_PATCH25 = 0x20,
380 WINED3D_SM5_INPUT_PT_PATCH26 = 0x21,
381 WINED3D_SM5_INPUT_PT_PATCH27 = 0x22,
382 WINED3D_SM5_INPUT_PT_PATCH28 = 0x23,
383 WINED3D_SM5_INPUT_PT_PATCH29 = 0x24,
384 WINED3D_SM5_INPUT_PT_PATCH30 = 0x25,
385 WINED3D_SM5_INPUT_PT_PATCH31 = 0x26,
386 WINED3D_SM5_INPUT_PT_PATCH32 = 0x27,
387 };
388
389 enum wined3d_sm4_swizzle_type
390 {
391 WINED3D_SM4_SWIZZLE_NONE = 0x0,
392 WINED3D_SM4_SWIZZLE_VEC4 = 0x1,
393 WINED3D_SM4_SWIZZLE_SCALAR = 0x2,
394 };
395
396 enum wined3d_sm4_immconst_type
397 {
398 WINED3D_SM4_IMMCONST_SCALAR = 0x1,
399 WINED3D_SM4_IMMCONST_VEC4 = 0x2,
400 };
401
402 enum wined3d_sm4_resource_type
403 {
404 WINED3D_SM4_RESOURCE_BUFFER = 0x1,
405 WINED3D_SM4_RESOURCE_TEXTURE_1D = 0x2,
406 WINED3D_SM4_RESOURCE_TEXTURE_2D = 0x3,
407 WINED3D_SM4_RESOURCE_TEXTURE_2DMS = 0x4,
408 WINED3D_SM4_RESOURCE_TEXTURE_3D = 0x5,
409 WINED3D_SM4_RESOURCE_TEXTURE_CUBE = 0x6,
410 WINED3D_SM4_RESOURCE_TEXTURE_1DARRAY = 0x7,
411 WINED3D_SM4_RESOURCE_TEXTURE_2DARRAY = 0x8,
412 WINED3D_SM4_RESOURCE_TEXTURE_2DMSARRAY = 0x9,
413 WINED3D_SM4_RESOURCE_TEXTURE_CUBEARRAY = 0xa,
414 };
415
416 enum wined3d_sm4_data_type
417 {
418 WINED3D_SM4_DATA_UNORM = 0x1,
419 WINED3D_SM4_DATA_SNORM = 0x2,
420 WINED3D_SM4_DATA_INT = 0x3,
421 WINED3D_SM4_DATA_UINT = 0x4,
422 WINED3D_SM4_DATA_FLOAT = 0x5,
423 };
424
425 enum wined3d_sm4_sampler_mode
426 {
427 WINED3D_SM4_SAMPLER_DEFAULT = 0x0,
428 WINED3D_SM4_SAMPLER_COMPARISON = 0x1,
429 };
430
431 enum wined3d_sm4_shader_data_type
432 {
433 WINED3D_SM4_SHADER_DATA_IMMEDIATE_CONSTANT_BUFFER = 0x3,
434 WINED3D_SM4_SHADER_DATA_MESSAGE = 0x4,
435 };
436
437 struct wined3d_shader_src_param_entry
438 {
439 struct list entry;
440 struct wined3d_shader_src_param param;
441 };
442
443 struct wined3d_sm4_data
444 {
445 struct wined3d_shader_version shader_version;
446 const DWORD *start, *end;
447
448 unsigned int output_map[MAX_REG_OUTPUT];
449
450 struct wined3d_shader_src_param src_param[5];
451 struct wined3d_shader_dst_param dst_param[2];
452 struct list src_free;
453 struct list src;
454 struct wined3d_shader_immediate_constant_buffer icb;
455 };
456
457 struct wined3d_sm4_opcode_info
458 {
459 enum wined3d_sm4_opcode opcode;
460 enum WINED3D_SHADER_INSTRUCTION_HANDLER handler_idx;
461 const char *dst_info;
462 const char *src_info;
463 void (*read_opcode_func)(struct wined3d_shader_instruction *ins,
464 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
465 struct wined3d_sm4_data *priv);
466 };
467
468 static const enum wined3d_primitive_type output_primitive_type_table[] =
469 {
470 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
471 /* WINED3D_SM4_OUTPUT_PT_POINTLIST */ WINED3D_PT_POINTLIST,
472 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
473 /* WINED3D_SM4_OUTPUT_PT_LINESTRIP */ WINED3D_PT_LINESTRIP,
474 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
475 /* WINED3D_SM4_OUTPUT_PT_TRIANGLESTRIP */ WINED3D_PT_TRIANGLESTRIP,
476 };
477
478 static const enum wined3d_primitive_type input_primitive_type_table[] =
479 {
480 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
481 /* WINED3D_SM4_INPUT_PT_POINT */ WINED3D_PT_POINTLIST,
482 /* WINED3D_SM4_INPUT_PT_LINE */ WINED3D_PT_LINELIST,
483 /* WINED3D_SM4_INPUT_PT_TRIANGLE */ WINED3D_PT_TRIANGLELIST,
484 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
485 /* UNKNOWN */ WINED3D_PT_UNDEFINED,
486 /* WINED3D_SM4_INPUT_PT_LINEADJ */ WINED3D_PT_LINELIST_ADJ,
487 /* WINED3D_SM4_INPUT_PT_TRIANGLEADJ */ WINED3D_PT_TRIANGLELIST_ADJ,
488 };
489
490 static const enum wined3d_shader_resource_type resource_type_table[] =
491 {
492 /* 0 */ WINED3D_SHADER_RESOURCE_NONE,
493 /* WINED3D_SM4_RESOURCE_BUFFER */ WINED3D_SHADER_RESOURCE_BUFFER,
494 /* WINED3D_SM4_RESOURCE_TEXTURE_1D */ WINED3D_SHADER_RESOURCE_TEXTURE_1D,
495 /* WINED3D_SM4_RESOURCE_TEXTURE_2D */ WINED3D_SHADER_RESOURCE_TEXTURE_2D,
496 /* WINED3D_SM4_RESOURCE_TEXTURE_2DMS */ WINED3D_SHADER_RESOURCE_TEXTURE_2DMS,
497 /* WINED3D_SM4_RESOURCE_TEXTURE_3D */ WINED3D_SHADER_RESOURCE_TEXTURE_3D,
498 /* WINED3D_SM4_RESOURCE_TEXTURE_CUBE */ WINED3D_SHADER_RESOURCE_TEXTURE_CUBE,
499 /* WINED3D_SM4_RESOURCE_TEXTURE_1DARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_1DARRAY,
500 /* WINED3D_SM4_RESOURCE_TEXTURE_2DARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_2DARRAY,
501 /* WINED3D_SM4_RESOURCE_TEXTURE_2DMSARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_2DMSARRAY,
502 /* WINED3D_SM4_RESOURCE_TEXTURE_CUBEARRAY */ WINED3D_SHADER_RESOURCE_TEXTURE_CUBEARRAY,
503 };
504
505 static const enum wined3d_data_type data_type_table[] =
506 {
507 /* 0 */ WINED3D_DATA_FLOAT,
508 /* WINED3D_SM4_DATA_UNORM */ WINED3D_DATA_UNORM,
509 /* WINED3D_SM4_DATA_SNORM */ WINED3D_DATA_SNORM,
510 /* WINED3D_SM4_DATA_INT */ WINED3D_DATA_INT,
511 /* WINED3D_SM4_DATA_UINT */ WINED3D_DATA_UINT,
512 /* WINED3D_SM4_DATA_FLOAT */ WINED3D_DATA_FLOAT,
513 };
514
515 static BOOL shader_sm4_read_src_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
516 enum wined3d_data_type data_type, struct wined3d_shader_src_param *src_param);
517 static BOOL shader_sm4_read_dst_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
518 enum wined3d_data_type data_type, struct wined3d_shader_dst_param *dst_param);
519
shader_sm4_read_conditional_op(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)520 static void shader_sm4_read_conditional_op(struct wined3d_shader_instruction *ins,
521 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
522 struct wined3d_sm4_data *priv)
523 {
524 shader_sm4_read_src_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_UINT, &priv->src_param[0]);
525 ins->flags = (opcode_token & WINED3D_SM4_CONDITIONAL_NZ) ?
526 WINED3D_SHADER_CONDITIONAL_OP_NZ : WINED3D_SHADER_CONDITIONAL_OP_Z;
527 }
528
shader_sm4_read_shader_data(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)529 static void shader_sm4_read_shader_data(struct wined3d_shader_instruction *ins,
530 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
531 struct wined3d_sm4_data *priv)
532 {
533 enum wined3d_sm4_shader_data_type type;
534 unsigned int icb_size;
535
536 type = (opcode_token & WINED3D_SM4_SHADER_DATA_TYPE_MASK) >> WINED3D_SM4_SHADER_DATA_TYPE_SHIFT;
537 if (type != WINED3D_SM4_SHADER_DATA_IMMEDIATE_CONSTANT_BUFFER)
538 {
539 FIXME("Ignoring shader data type %#x.\n", type);
540 ins->handler_idx = WINED3DSIH_NOP;
541 return;
542 }
543
544 ++tokens;
545 icb_size = token_count - 1;
546 if (icb_size % 4 || icb_size > MAX_IMMEDIATE_CONSTANT_BUFFER_SIZE)
547 {
548 FIXME("Unexpected immediate constant buffer size %u.\n", icb_size);
549 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
550 return;
551 }
552
553 priv->icb.vec4_count = icb_size / 4;
554 memcpy(priv->icb.data, tokens, sizeof(*tokens) * icb_size);
555 ins->declaration.icb = &priv->icb;
556 }
557
shader_sm4_read_dcl_resource(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)558 static void shader_sm4_read_dcl_resource(struct wined3d_shader_instruction *ins,
559 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
560 struct wined3d_sm4_data *priv)
561 {
562 enum wined3d_sm4_resource_type resource_type;
563 enum wined3d_sm4_data_type data_type;
564 enum wined3d_data_type reg_data_type;
565 DWORD components;
566
567 resource_type = (opcode_token & WINED3D_SM4_RESOURCE_TYPE_MASK) >> WINED3D_SM4_RESOURCE_TYPE_SHIFT;
568 if (!resource_type || (resource_type >= ARRAY_SIZE(resource_type_table)))
569 {
570 FIXME("Unhandled resource type %#x.\n", resource_type);
571 ins->declaration.semantic.resource_type = WINED3D_SHADER_RESOURCE_NONE;
572 }
573 else
574 {
575 ins->declaration.semantic.resource_type = resource_type_table[resource_type];
576 }
577 reg_data_type = opcode == WINED3D_SM4_OP_DCL_RESOURCE ? WINED3D_DATA_RESOURCE : WINED3D_DATA_UAV;
578 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], reg_data_type, &ins->declaration.semantic.reg);
579
580 components = *tokens++;
581 if ((components & 0xfff0) != (components & 0xf) * 0x1110)
582 FIXME("Components (%#x) have different data types.\n", components);
583 data_type = components & 0xf;
584
585 if (!data_type || (data_type >= ARRAY_SIZE(data_type_table)))
586 {
587 FIXME("Unhandled data type %#x.\n", data_type);
588 ins->declaration.semantic.resource_data_type = WINED3D_DATA_FLOAT;
589 }
590 else
591 {
592 ins->declaration.semantic.resource_data_type = data_type_table[data_type];
593 }
594
595 if (reg_data_type == WINED3D_DATA_UAV)
596 ins->flags = (opcode_token & WINED3D_SM5_UAV_FLAGS_MASK) >> WINED3D_SM5_UAV_FLAGS_SHIFT;
597 }
598
shader_sm4_read_dcl_constant_buffer(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)599 static void shader_sm4_read_dcl_constant_buffer(struct wined3d_shader_instruction *ins,
600 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
601 struct wined3d_sm4_data *priv)
602 {
603 shader_sm4_read_src_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.src);
604 if (opcode_token & WINED3D_SM4_INDEX_TYPE_MASK)
605 ins->flags |= WINED3DSI_INDEXED_DYNAMIC;
606 }
607
shader_sm4_read_dcl_sampler(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)608 static void shader_sm4_read_dcl_sampler(struct wined3d_shader_instruction *ins,
609 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
610 struct wined3d_sm4_data *priv)
611 {
612 ins->flags = (opcode_token & WINED3D_SM4_SAMPLER_MODE_MASK) >> WINED3D_SM4_SAMPLER_MODE_SHIFT;
613 if (ins->flags & ~WINED3D_SM4_SAMPLER_COMPARISON)
614 FIXME("Unhandled sampler mode %#x.\n", ins->flags);
615 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_SAMPLER, &ins->declaration.dst);
616 }
617
shader_sm4_read_dcl_index_range(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)618 static void shader_sm4_read_dcl_index_range(struct wined3d_shader_instruction *ins,
619 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
620 struct wined3d_sm4_data *priv)
621 {
622 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_OPAQUE,
623 &ins->declaration.index_range.first_register);
624 ins->declaration.index_range.last_register = *tokens;
625 }
626
shader_sm4_read_dcl_output_topology(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)627 static void shader_sm4_read_dcl_output_topology(struct wined3d_shader_instruction *ins,
628 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
629 struct wined3d_sm4_data *priv)
630 {
631 enum wined3d_sm4_output_primitive_type primitive_type;
632
633 primitive_type = (opcode_token & WINED3D_SM4_PRIMITIVE_TYPE_MASK) >> WINED3D_SM4_PRIMITIVE_TYPE_SHIFT;
634 if (primitive_type >= ARRAY_SIZE(output_primitive_type_table))
635 ins->declaration.primitive_type.type = WINED3D_PT_UNDEFINED;
636 else
637 ins->declaration.primitive_type.type = output_primitive_type_table[primitive_type];
638
639 if (ins->declaration.primitive_type.type == WINED3D_PT_UNDEFINED)
640 FIXME("Unhandled output primitive type %#x.\n", primitive_type);
641 }
642
shader_sm4_read_dcl_input_primitive(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)643 static void shader_sm4_read_dcl_input_primitive(struct wined3d_shader_instruction *ins,
644 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
645 struct wined3d_sm4_data *priv)
646 {
647 enum wined3d_sm4_input_primitive_type primitive_type;
648
649 primitive_type = (opcode_token & WINED3D_SM4_PRIMITIVE_TYPE_MASK) >> WINED3D_SM4_PRIMITIVE_TYPE_SHIFT;
650 if (WINED3D_SM5_INPUT_PT_PATCH1 <= primitive_type && primitive_type <= WINED3D_SM5_INPUT_PT_PATCH32)
651 {
652 ins->declaration.primitive_type.type = WINED3D_PT_PATCH;
653 ins->declaration.primitive_type.patch_vertex_count = primitive_type - WINED3D_SM5_INPUT_PT_PATCH1 + 1;
654 }
655 else if (primitive_type >= ARRAY_SIZE(input_primitive_type_table))
656 {
657 ins->declaration.primitive_type.type = WINED3D_PT_UNDEFINED;
658 }
659 else
660 {
661 ins->declaration.primitive_type.type = input_primitive_type_table[primitive_type];
662 }
663
664 if (ins->declaration.primitive_type.type == WINED3D_PT_UNDEFINED)
665 FIXME("Unhandled input primitive type %#x.\n", primitive_type);
666 }
667
shader_sm4_read_declaration_count(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)668 static void shader_sm4_read_declaration_count(struct wined3d_shader_instruction *ins,
669 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
670 struct wined3d_sm4_data *priv)
671 {
672 ins->declaration.count = *tokens;
673 }
674
shader_sm4_read_declaration_dst(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)675 static void shader_sm4_read_declaration_dst(struct wined3d_shader_instruction *ins,
676 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
677 struct wined3d_sm4_data *priv)
678 {
679 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.dst);
680 }
681
shader_sm4_read_declaration_register_semantic(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)682 static void shader_sm4_read_declaration_register_semantic(struct wined3d_shader_instruction *ins,
683 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
684 struct wined3d_sm4_data *priv)
685 {
686 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT,
687 &ins->declaration.register_semantic.reg);
688 ins->declaration.register_semantic.sysval_semantic = *tokens;
689 }
690
shader_sm4_read_dcl_input_ps(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)691 static void shader_sm4_read_dcl_input_ps(struct wined3d_shader_instruction *ins,
692 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
693 struct wined3d_sm4_data *priv)
694 {
695 ins->flags = (opcode_token & WINED3D_SM4_INTERPOLATION_MODE_MASK) >> WINED3D_SM4_INTERPOLATION_MODE_SHIFT;
696 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.dst);
697 }
698
shader_sm4_read_dcl_input_ps_siv(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)699 static void shader_sm4_read_dcl_input_ps_siv(struct wined3d_shader_instruction *ins,
700 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
701 struct wined3d_sm4_data *priv)
702 {
703 ins->flags = (opcode_token & WINED3D_SM4_INTERPOLATION_MODE_MASK) >> WINED3D_SM4_INTERPOLATION_MODE_SHIFT;
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;
707 }
708
shader_sm4_read_dcl_indexable_temp(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)709 static void shader_sm4_read_dcl_indexable_temp(struct wined3d_shader_instruction *ins,
710 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
711 struct wined3d_sm4_data *priv)
712 {
713 ins->declaration.indexable_temp.register_idx = *tokens++;
714 ins->declaration.indexable_temp.register_size = *tokens++;
715 ins->declaration.indexable_temp.component_count = *tokens;
716 }
717
shader_sm4_read_dcl_global_flags(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)718 static void shader_sm4_read_dcl_global_flags(struct wined3d_shader_instruction *ins,
719 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
720 struct wined3d_sm4_data *priv)
721 {
722 ins->flags = (opcode_token & WINED3D_SM4_GLOBAL_FLAGS_MASK) >> WINED3D_SM4_GLOBAL_FLAGS_SHIFT;
723 }
724
shader_sm5_read_fcall(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)725 static void shader_sm5_read_fcall(struct wined3d_shader_instruction *ins,
726 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
727 struct wined3d_sm4_data *priv)
728 {
729 priv->src_param[0].reg.u.fp_body_idx = *tokens++;
730 shader_sm4_read_src_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_OPAQUE, &priv->src_param[0]);
731 }
732
shader_sm5_read_dcl_function_body(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)733 static void shader_sm5_read_dcl_function_body(struct wined3d_shader_instruction *ins,
734 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
735 struct wined3d_sm4_data *priv)
736 {
737 ins->declaration.index = *tokens;
738 }
739
shader_sm5_read_dcl_function_table(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)740 static void shader_sm5_read_dcl_function_table(struct wined3d_shader_instruction *ins,
741 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
742 struct wined3d_sm4_data *priv)
743 {
744 ins->declaration.index = *tokens++;
745 FIXME("Ignoring set of function bodies (count %u).\n", *tokens);
746 }
747
shader_sm5_read_dcl_interface(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)748 static void shader_sm5_read_dcl_interface(struct wined3d_shader_instruction *ins,
749 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
750 struct wined3d_sm4_data *priv)
751 {
752 ins->declaration.fp.index = *tokens++;
753 ins->declaration.fp.body_count = *tokens++;
754 ins->declaration.fp.array_size = *tokens >> WINED3D_SM5_FP_ARRAY_SIZE_SHIFT;
755 ins->declaration.fp.table_count = *tokens++ & WINED3D_SM5_FP_TABLE_COUNT_MASK;
756 FIXME("Ignoring set of function tables (count %u).\n", ins->declaration.fp.table_count);
757 }
758
shader_sm5_read_control_point_count(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)759 static void shader_sm5_read_control_point_count(struct wined3d_shader_instruction *ins,
760 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
761 struct wined3d_sm4_data *priv)
762 {
763 ins->declaration.count = (opcode_token & WINED3D_SM5_CONTROL_POINT_COUNT_MASK)
764 >> WINED3D_SM5_CONTROL_POINT_COUNT_SHIFT;
765 }
766
shader_sm5_read_dcl_tessellator_domain(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)767 static void shader_sm5_read_dcl_tessellator_domain(struct wined3d_shader_instruction *ins,
768 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
769 struct wined3d_sm4_data *priv)
770 {
771 ins->declaration.tessellator_domain = (opcode_token & WINED3D_SM5_TESSELLATOR_MASK)
772 >> WINED3D_SM5_TESSELLATOR_SHIFT;
773 }
774
shader_sm5_read_dcl_tessellator_partitioning(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)775 static void shader_sm5_read_dcl_tessellator_partitioning(struct wined3d_shader_instruction *ins,
776 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
777 struct wined3d_sm4_data *priv)
778 {
779 ins->declaration.tessellator_partitioning = (opcode_token & WINED3D_SM5_TESSELLATOR_MASK)
780 >> WINED3D_SM5_TESSELLATOR_SHIFT;
781 }
782
shader_sm5_read_dcl_tessellator_output_primitive(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)783 static void shader_sm5_read_dcl_tessellator_output_primitive(struct wined3d_shader_instruction *ins,
784 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
785 struct wined3d_sm4_data *priv)
786 {
787 ins->declaration.tessellator_output_primitive = (opcode_token & WINED3D_SM5_TESSELLATOR_MASK)
788 >> WINED3D_SM5_TESSELLATOR_SHIFT;
789 }
790
shader_sm5_read_dcl_hs_max_tessfactor(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)791 static void shader_sm5_read_dcl_hs_max_tessfactor(struct wined3d_shader_instruction *ins,
792 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
793 struct wined3d_sm4_data *priv)
794 {
795 ins->declaration.max_tessellation_factor = *(float *)tokens;
796 }
797
shader_sm5_read_dcl_thread_group(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)798 static void shader_sm5_read_dcl_thread_group(struct wined3d_shader_instruction *ins,
799 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
800 struct wined3d_sm4_data *priv)
801 {
802 ins->declaration.thread_group_size.x = *tokens++;
803 ins->declaration.thread_group_size.y = *tokens++;
804 ins->declaration.thread_group_size.z = *tokens++;
805 }
806
shader_sm5_read_dcl_uav_raw(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)807 static void shader_sm5_read_dcl_uav_raw(struct wined3d_shader_instruction *ins,
808 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
809 struct wined3d_sm4_data *priv)
810 {
811 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_UAV, &ins->declaration.dst);
812 ins->flags = (opcode_token & WINED3D_SM5_UAV_FLAGS_MASK) >> WINED3D_SM5_UAV_FLAGS_SHIFT;
813 }
814
shader_sm5_read_dcl_uav_structured(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)815 static void shader_sm5_read_dcl_uav_structured(struct wined3d_shader_instruction *ins,
816 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
817 struct wined3d_sm4_data *priv)
818 {
819 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_UAV,
820 &ins->declaration.structured_resource.reg);
821 ins->flags = (opcode_token & WINED3D_SM5_UAV_FLAGS_MASK) >> WINED3D_SM5_UAV_FLAGS_SHIFT;
822 ins->declaration.structured_resource.byte_stride = *tokens;
823 if (ins->declaration.structured_resource.byte_stride % 4)
824 FIXME("Byte stride %u is not multiple of 4.\n", ins->declaration.structured_resource.byte_stride);
825 }
826
shader_sm5_read_dcl_tgsm_raw(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)827 static void shader_sm5_read_dcl_tgsm_raw(struct wined3d_shader_instruction *ins,
828 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
829 struct wined3d_sm4_data *priv)
830 {
831 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT, &ins->declaration.tgsm_raw.reg);
832 ins->declaration.tgsm_raw.byte_count = *tokens;
833 if (ins->declaration.tgsm_raw.byte_count % 4)
834 FIXME("Byte count %u is not multiple of 4.\n", ins->declaration.tgsm_raw.byte_count);
835 }
836
shader_sm5_read_dcl_tgsm_structured(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)837 static void shader_sm5_read_dcl_tgsm_structured(struct wined3d_shader_instruction *ins,
838 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
839 struct wined3d_sm4_data *priv)
840 {
841 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_FLOAT,
842 &ins->declaration.tgsm_structured.reg);
843 ins->declaration.tgsm_structured.byte_stride = *tokens++;
844 ins->declaration.tgsm_structured.structure_count = *tokens;
845 if (ins->declaration.tgsm_structured.byte_stride % 4)
846 FIXME("Byte stride %u is not multiple of 4.\n", ins->declaration.tgsm_structured.byte_stride);
847 }
848
shader_sm5_read_dcl_resource_structured(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)849 static void shader_sm5_read_dcl_resource_structured(struct wined3d_shader_instruction *ins,
850 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
851 struct wined3d_sm4_data *priv)
852 {
853 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_RESOURCE,
854 &ins->declaration.structured_resource.reg);
855 ins->declaration.structured_resource.byte_stride = *tokens;
856 if (ins->declaration.structured_resource.byte_stride % 4)
857 FIXME("Byte stride %u is not multiple of 4.\n", ins->declaration.structured_resource.byte_stride);
858 }
859
shader_sm5_read_dcl_resource_raw(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)860 static void shader_sm5_read_dcl_resource_raw(struct wined3d_shader_instruction *ins,
861 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
862 struct wined3d_sm4_data *priv)
863 {
864 shader_sm4_read_dst_param(priv, &tokens, &tokens[token_count], WINED3D_DATA_RESOURCE, &ins->declaration.dst);
865 }
866
shader_sm5_read_sync(struct wined3d_shader_instruction * ins,DWORD opcode,DWORD opcode_token,const DWORD * tokens,unsigned int token_count,struct wined3d_sm4_data * priv)867 static void shader_sm5_read_sync(struct wined3d_shader_instruction *ins,
868 DWORD opcode, DWORD opcode_token, const DWORD *tokens, unsigned int token_count,
869 struct wined3d_sm4_data *priv)
870 {
871 ins->flags = (opcode_token & WINED3D_SM5_SYNC_FLAGS_MASK) >> WINED3D_SM5_SYNC_FLAGS_SHIFT;
872 }
873
874 /*
875 * f -> WINED3D_DATA_FLOAT
876 * i -> WINED3D_DATA_INT
877 * u -> WINED3D_DATA_UINT
878 * O -> WINED3D_DATA_OPAQUE
879 * R -> WINED3D_DATA_RESOURCE
880 * S -> WINED3D_DATA_SAMPLER
881 * U -> WINED3D_DATA_UAV
882 */
883 static const struct wined3d_sm4_opcode_info opcode_table[] =
884 {
885 {WINED3D_SM4_OP_ADD, WINED3DSIH_ADD, "f", "ff"},
886 {WINED3D_SM4_OP_AND, WINED3DSIH_AND, "u", "uu"},
887 {WINED3D_SM4_OP_BREAK, WINED3DSIH_BREAK, "", ""},
888 {WINED3D_SM4_OP_BREAKC, WINED3DSIH_BREAKP, "", "u",
889 shader_sm4_read_conditional_op},
890 {WINED3D_SM4_OP_CASE, WINED3DSIH_CASE, "", "u"},
891 {WINED3D_SM4_OP_CONTINUE, WINED3DSIH_CONTINUE, "", ""},
892 {WINED3D_SM4_OP_CONTINUEC, WINED3DSIH_CONTINUEP, "", "u",
893 shader_sm4_read_conditional_op},
894 {WINED3D_SM4_OP_CUT, WINED3DSIH_CUT, "", ""},
895 {WINED3D_SM4_OP_DEFAULT, WINED3DSIH_DEFAULT, "", ""},
896 {WINED3D_SM4_OP_DERIV_RTX, WINED3DSIH_DSX, "f", "f"},
897 {WINED3D_SM4_OP_DERIV_RTY, WINED3DSIH_DSY, "f", "f"},
898 {WINED3D_SM4_OP_DISCARD, WINED3DSIH_TEXKILL, "", "u",
899 shader_sm4_read_conditional_op},
900 {WINED3D_SM4_OP_DIV, WINED3DSIH_DIV, "f", "ff"},
901 {WINED3D_SM4_OP_DP2, WINED3DSIH_DP2, "f", "ff"},
902 {WINED3D_SM4_OP_DP3, WINED3DSIH_DP3, "f", "ff"},
903 {WINED3D_SM4_OP_DP4, WINED3DSIH_DP4, "f", "ff"},
904 {WINED3D_SM4_OP_ELSE, WINED3DSIH_ELSE, "", ""},
905 {WINED3D_SM4_OP_EMIT, WINED3DSIH_EMIT, "", ""},
906 {WINED3D_SM4_OP_ENDIF, WINED3DSIH_ENDIF, "", ""},
907 {WINED3D_SM4_OP_ENDLOOP, WINED3DSIH_ENDLOOP, "", ""},
908 {WINED3D_SM4_OP_ENDSWITCH, WINED3DSIH_ENDSWITCH, "", ""},
909 {WINED3D_SM4_OP_EQ, WINED3DSIH_EQ, "u", "ff"},
910 {WINED3D_SM4_OP_EXP, WINED3DSIH_EXP, "f", "f"},
911 {WINED3D_SM4_OP_FRC, WINED3DSIH_FRC, "f", "f"},
912 {WINED3D_SM4_OP_FTOI, WINED3DSIH_FTOI, "i", "f"},
913 {WINED3D_SM4_OP_FTOU, WINED3DSIH_FTOU, "u", "f"},
914 {WINED3D_SM4_OP_GE, WINED3DSIH_GE, "u", "ff"},
915 {WINED3D_SM4_OP_IADD, WINED3DSIH_IADD, "i", "ii"},
916 {WINED3D_SM4_OP_IF, WINED3DSIH_IF, "", "u",
917 shader_sm4_read_conditional_op},
918 {WINED3D_SM4_OP_IEQ, WINED3DSIH_IEQ, "u", "ii"},
919 {WINED3D_SM4_OP_IGE, WINED3DSIH_IGE, "u", "ii"},
920 {WINED3D_SM4_OP_ILT, WINED3DSIH_ILT, "u", "ii"},
921 {WINED3D_SM4_OP_IMAD, WINED3DSIH_IMAD, "i", "iii"},
922 {WINED3D_SM4_OP_IMAX, WINED3DSIH_IMAX, "i", "ii"},
923 {WINED3D_SM4_OP_IMIN, WINED3DSIH_IMIN, "i", "ii"},
924 {WINED3D_SM4_OP_IMUL, WINED3DSIH_IMUL, "ii", "ii"},
925 {WINED3D_SM4_OP_INE, WINED3DSIH_INE, "u", "ii"},
926 {WINED3D_SM4_OP_INEG, WINED3DSIH_INEG, "i", "i"},
927 {WINED3D_SM4_OP_ISHL, WINED3DSIH_ISHL, "i", "ii"},
928 {WINED3D_SM4_OP_ISHR, WINED3DSIH_ISHR, "i", "ii"},
929 {WINED3D_SM4_OP_ITOF, WINED3DSIH_ITOF, "f", "i"},
930 {WINED3D_SM4_OP_LABEL, WINED3DSIH_LABEL, "", "O"},
931 {WINED3D_SM4_OP_LD, WINED3DSIH_LD, "u", "iR"},
932 {WINED3D_SM4_OP_LD2DMS, WINED3DSIH_LD2DMS, "u", "iRi"},
933 {WINED3D_SM4_OP_LOG, WINED3DSIH_LOG, "f", "f"},
934 {WINED3D_SM4_OP_LOOP, WINED3DSIH_LOOP, "", ""},
935 {WINED3D_SM4_OP_LT, WINED3DSIH_LT, "u", "ff"},
936 {WINED3D_SM4_OP_MAD, WINED3DSIH_MAD, "f", "fff"},
937 {WINED3D_SM4_OP_MIN, WINED3DSIH_MIN, "f", "ff"},
938 {WINED3D_SM4_OP_MAX, WINED3DSIH_MAX, "f", "ff"},
939 {WINED3D_SM4_OP_SHADER_DATA, WINED3DSIH_DCL_IMMEDIATE_CONSTANT_BUFFER, "", "",
940 shader_sm4_read_shader_data},
941 {WINED3D_SM4_OP_MOV, WINED3DSIH_MOV, "f", "f"},
942 {WINED3D_SM4_OP_MOVC, WINED3DSIH_MOVC, "f", "uff"},
943 {WINED3D_SM4_OP_MUL, WINED3DSIH_MUL, "f", "ff"},
944 {WINED3D_SM4_OP_NE, WINED3DSIH_NE, "u", "ff"},
945 {WINED3D_SM4_OP_NOP, WINED3DSIH_NOP, "", ""},
946 {WINED3D_SM4_OP_NOT, WINED3DSIH_NOT, "u", "u"},
947 {WINED3D_SM4_OP_OR, WINED3DSIH_OR, "u", "uu"},
948 {WINED3D_SM4_OP_RESINFO, WINED3DSIH_RESINFO, "f", "iR"},
949 {WINED3D_SM4_OP_RET, WINED3DSIH_RET, "", ""},
950 {WINED3D_SM4_OP_RETC, WINED3DSIH_RETP, "", "u",
951 shader_sm4_read_conditional_op},
952 {WINED3D_SM4_OP_ROUND_NE, WINED3DSIH_ROUND_NE, "f", "f"},
953 {WINED3D_SM4_OP_ROUND_NI, WINED3DSIH_ROUND_NI, "f", "f"},
954 {WINED3D_SM4_OP_ROUND_PI, WINED3DSIH_ROUND_PI, "f", "f"},
955 {WINED3D_SM4_OP_ROUND_Z, WINED3DSIH_ROUND_Z, "f", "f"},
956 {WINED3D_SM4_OP_RSQ, WINED3DSIH_RSQ, "f", "f"},
957 {WINED3D_SM4_OP_SAMPLE, WINED3DSIH_SAMPLE, "u", "fRS"},
958 {WINED3D_SM4_OP_SAMPLE_C, WINED3DSIH_SAMPLE_C, "f", "fRSf"},
959 {WINED3D_SM4_OP_SAMPLE_C_LZ, WINED3DSIH_SAMPLE_C_LZ, "f", "fRSf"},
960 {WINED3D_SM4_OP_SAMPLE_LOD, WINED3DSIH_SAMPLE_LOD, "u", "fRSf"},
961 {WINED3D_SM4_OP_SAMPLE_GRAD, WINED3DSIH_SAMPLE_GRAD, "u", "fRSff"},
962 {WINED3D_SM4_OP_SAMPLE_B, WINED3DSIH_SAMPLE_B, "u", "fRSf"},
963 {WINED3D_SM4_OP_SQRT, WINED3DSIH_SQRT, "f", "f"},
964 {WINED3D_SM4_OP_SWITCH, WINED3DSIH_SWITCH, "", "u"},
965 {WINED3D_SM4_OP_SINCOS, WINED3DSIH_SINCOS, "ff", "f"},
966 {WINED3D_SM4_OP_UDIV, WINED3DSIH_UDIV, "uu", "uu"},
967 {WINED3D_SM4_OP_ULT, WINED3DSIH_ULT, "u", "uu"},
968 {WINED3D_SM4_OP_UGE, WINED3DSIH_UGE, "u", "uu"},
969 {WINED3D_SM4_OP_UMUL, WINED3DSIH_UMUL, "uu", "uu"},
970 {WINED3D_SM4_OP_UMAX, WINED3DSIH_UMAX, "u", "uu"},
971 {WINED3D_SM4_OP_UMIN, WINED3DSIH_UMIN, "u", "uu"},
972 {WINED3D_SM4_OP_USHR, WINED3DSIH_USHR, "u", "uu"},
973 {WINED3D_SM4_OP_UTOF, WINED3DSIH_UTOF, "f", "u"},
974 {WINED3D_SM4_OP_XOR, WINED3DSIH_XOR, "u", "uu"},
975 {WINED3D_SM4_OP_DCL_RESOURCE, WINED3DSIH_DCL, "R", "",
976 shader_sm4_read_dcl_resource},
977 {WINED3D_SM4_OP_DCL_CONSTANT_BUFFER, WINED3DSIH_DCL_CONSTANT_BUFFER, "", "",
978 shader_sm4_read_dcl_constant_buffer},
979 {WINED3D_SM4_OP_DCL_SAMPLER, WINED3DSIH_DCL_SAMPLER, "", "",
980 shader_sm4_read_dcl_sampler},
981 {WINED3D_SM4_OP_DCL_INDEX_RANGE, WINED3DSIH_DCL_INDEX_RANGE, "", "",
982 shader_sm4_read_dcl_index_range},
983 {WINED3D_SM4_OP_DCL_OUTPUT_TOPOLOGY, WINED3DSIH_DCL_OUTPUT_TOPOLOGY, "", "",
984 shader_sm4_read_dcl_output_topology},
985 {WINED3D_SM4_OP_DCL_INPUT_PRIMITIVE, WINED3DSIH_DCL_INPUT_PRIMITIVE, "", "",
986 shader_sm4_read_dcl_input_primitive},
987 {WINED3D_SM4_OP_DCL_VERTICES_OUT, WINED3DSIH_DCL_VERTICES_OUT, "", "",
988 shader_sm4_read_declaration_count},
989 {WINED3D_SM4_OP_DCL_INPUT, WINED3DSIH_DCL_INPUT, "", "",
990 shader_sm4_read_declaration_dst},
991 {WINED3D_SM4_OP_DCL_INPUT_SGV, WINED3DSIH_DCL_INPUT_SGV, "", "",
992 shader_sm4_read_declaration_register_semantic},
993 {WINED3D_SM4_OP_DCL_INPUT_SIV, WINED3DSIH_DCL_INPUT_SIV, "", "",
994 shader_sm4_read_declaration_register_semantic},
995 {WINED3D_SM4_OP_DCL_INPUT_PS, WINED3DSIH_DCL_INPUT_PS, "", "",
996 shader_sm4_read_dcl_input_ps},
997 {WINED3D_SM4_OP_DCL_INPUT_PS_SGV, WINED3DSIH_DCL_INPUT_PS_SGV, "", "",
998 shader_sm4_read_declaration_register_semantic},
999 {WINED3D_SM4_OP_DCL_INPUT_PS_SIV, WINED3DSIH_DCL_INPUT_PS_SIV, "", "",
1000 shader_sm4_read_dcl_input_ps_siv},
1001 {WINED3D_SM4_OP_DCL_OUTPUT, WINED3DSIH_DCL_OUTPUT, "", "",
1002 shader_sm4_read_declaration_dst},
1003 {WINED3D_SM4_OP_DCL_OUTPUT_SIV, WINED3DSIH_DCL_OUTPUT_SIV, "", "",
1004 shader_sm4_read_declaration_register_semantic},
1005 {WINED3D_SM4_OP_DCL_TEMPS, WINED3DSIH_DCL_TEMPS, "", "",
1006 shader_sm4_read_declaration_count},
1007 {WINED3D_SM4_OP_DCL_INDEXABLE_TEMP, WINED3DSIH_DCL_INDEXABLE_TEMP, "", "",
1008 shader_sm4_read_dcl_indexable_temp},
1009 {WINED3D_SM4_OP_DCL_GLOBAL_FLAGS, WINED3DSIH_DCL_GLOBAL_FLAGS, "", "",
1010 shader_sm4_read_dcl_global_flags},
1011 {WINED3D_SM4_OP_LOD, WINED3DSIH_LOD, "f", "fRS"},
1012 {WINED3D_SM4_OP_GATHER4, WINED3DSIH_GATHER4, "u", "fRS"},
1013 {WINED3D_SM4_OP_SAMPLE_POS, WINED3DSIH_SAMPLE_POS, "f", "Ru"},
1014 {WINED3D_SM4_OP_SAMPLE_INFO, WINED3DSIH_SAMPLE_INFO, "f", "R"},
1015 {WINED3D_SM5_OP_HS_DECLS, WINED3DSIH_HS_DECLS, "", ""},
1016 {WINED3D_SM5_OP_HS_CONTROL_POINT_PHASE, WINED3DSIH_HS_CONTROL_POINT_PHASE, "", ""},
1017 {WINED3D_SM5_OP_HS_FORK_PHASE, WINED3DSIH_HS_FORK_PHASE, "", ""},
1018 {WINED3D_SM5_OP_HS_JOIN_PHASE, WINED3DSIH_HS_JOIN_PHASE, "", ""},
1019 {WINED3D_SM5_OP_EMIT_STREAM, WINED3DSIH_EMIT_STREAM, "", "f"},
1020 {WINED3D_SM5_OP_CUT_STREAM, WINED3DSIH_CUT_STREAM, "", "f"},
1021 {WINED3D_SM5_OP_FCALL, WINED3DSIH_FCALL, "", "O",
1022 shader_sm5_read_fcall},
1023 {WINED3D_SM5_OP_BUFINFO, WINED3DSIH_BUFINFO, "i", "U"},
1024 {WINED3D_SM5_OP_DERIV_RTX_COARSE, WINED3DSIH_DSX_COARSE, "f", "f"},
1025 {WINED3D_SM5_OP_DERIV_RTX_FINE, WINED3DSIH_DSX_FINE, "f", "f"},
1026 {WINED3D_SM5_OP_DERIV_RTY_COARSE, WINED3DSIH_DSY_COARSE, "f", "f"},
1027 {WINED3D_SM5_OP_DERIV_RTY_FINE, WINED3DSIH_DSY_FINE, "f", "f"},
1028 {WINED3D_SM5_OP_GATHER4_C, WINED3DSIH_GATHER4_C, "f", "fRSf"},
1029 {WINED3D_SM5_OP_GATHER4_PO, WINED3DSIH_GATHER4_PO, "f", "fiRS"},
1030 {WINED3D_SM5_OP_GATHER4_PO_C, WINED3DSIH_GATHER4_PO_C, "f", "fiRSf"},
1031 {WINED3D_SM5_OP_RCP, WINED3DSIH_RCP, "f", "f"},
1032 {WINED3D_SM5_OP_F32TOF16, WINED3DSIH_F32TOF16, "u", "f"},
1033 {WINED3D_SM5_OP_F16TOF32, WINED3DSIH_F16TOF32, "f", "u"},
1034 {WINED3D_SM5_OP_COUNTBITS, WINED3DSIH_COUNTBITS, "u", "u"},
1035 {WINED3D_SM5_OP_FIRSTBIT_HI, WINED3DSIH_FIRSTBIT_HI, "u", "u"},
1036 {WINED3D_SM5_OP_FIRSTBIT_LO, WINED3DSIH_FIRSTBIT_LO, "u", "u"},
1037 {WINED3D_SM5_OP_FIRSTBIT_SHI, WINED3DSIH_FIRSTBIT_SHI, "u", "i"},
1038 {WINED3D_SM5_OP_UBFE, WINED3DSIH_UBFE, "u", "iiu"},
1039 {WINED3D_SM5_OP_IBFE, WINED3DSIH_IBFE, "i", "iii"},
1040 {WINED3D_SM5_OP_BFI, WINED3DSIH_BFI, "u", "iiuu"},
1041 {WINED3D_SM5_OP_BFREV, WINED3DSIH_BFREV, "u", "u"},
1042 {WINED3D_SM5_OP_SWAPC, WINED3DSIH_SWAPC, "ff", "uff"},
1043 {WINED3D_SM5_OP_DCL_STREAM, WINED3DSIH_DCL_STREAM, "", "O"},
1044 {WINED3D_SM5_OP_DCL_FUNCTION_BODY, WINED3DSIH_DCL_FUNCTION_BODY, "", "",
1045 shader_sm5_read_dcl_function_body},
1046 {WINED3D_SM5_OP_DCL_FUNCTION_TABLE, WINED3DSIH_DCL_FUNCTION_TABLE, "", "",
1047 shader_sm5_read_dcl_function_table},
1048 {WINED3D_SM5_OP_DCL_INTERFACE, WINED3DSIH_DCL_INTERFACE, "", "",
1049 shader_sm5_read_dcl_interface},
1050 {WINED3D_SM5_OP_DCL_INPUT_CONTROL_POINT_COUNT, WINED3DSIH_DCL_INPUT_CONTROL_POINT_COUNT, "", "",
1051 shader_sm5_read_control_point_count},
1052 {WINED3D_SM5_OP_DCL_OUTPUT_CONTROL_POINT_COUNT, WINED3DSIH_DCL_OUTPUT_CONTROL_POINT_COUNT, "", "",
1053 shader_sm5_read_control_point_count},
1054 {WINED3D_SM5_OP_DCL_TESSELLATOR_DOMAIN, WINED3DSIH_DCL_TESSELLATOR_DOMAIN, "", "",
1055 shader_sm5_read_dcl_tessellator_domain},
1056 {WINED3D_SM5_OP_DCL_TESSELLATOR_PARTITIONING, WINED3DSIH_DCL_TESSELLATOR_PARTITIONING, "", "",
1057 shader_sm5_read_dcl_tessellator_partitioning},
1058 {WINED3D_SM5_OP_DCL_TESSELLATOR_OUTPUT_PRIMITIVE, WINED3DSIH_DCL_TESSELLATOR_OUTPUT_PRIMITIVE, "", "",
1059 shader_sm5_read_dcl_tessellator_output_primitive},
1060 {WINED3D_SM5_OP_DCL_HS_MAX_TESSFACTOR, WINED3DSIH_DCL_HS_MAX_TESSFACTOR, "", "",
1061 shader_sm5_read_dcl_hs_max_tessfactor},
1062 {WINED3D_SM5_OP_DCL_HS_FORK_PHASE_INSTANCE_COUNT, WINED3DSIH_DCL_HS_FORK_PHASE_INSTANCE_COUNT, "", "",
1063 shader_sm4_read_declaration_count},
1064 {WINED3D_SM5_OP_DCL_HS_JOIN_PHASE_INSTANCE_COUNT, WINED3DSIH_DCL_HS_JOIN_PHASE_INSTANCE_COUNT, "", "",
1065 shader_sm4_read_declaration_count},
1066 {WINED3D_SM5_OP_DCL_THREAD_GROUP, WINED3DSIH_DCL_THREAD_GROUP, "", "",
1067 shader_sm5_read_dcl_thread_group},
1068 {WINED3D_SM5_OP_DCL_UAV_TYPED, WINED3DSIH_DCL_UAV_TYPED, "", "",
1069 shader_sm4_read_dcl_resource},
1070 {WINED3D_SM5_OP_DCL_UAV_RAW, WINED3DSIH_DCL_UAV_RAW, "", "",
1071 shader_sm5_read_dcl_uav_raw},
1072 {WINED3D_SM5_OP_DCL_UAV_STRUCTURED, WINED3DSIH_DCL_UAV_STRUCTURED, "", "",
1073 shader_sm5_read_dcl_uav_structured},
1074 {WINED3D_SM5_OP_DCL_TGSM_RAW, WINED3DSIH_DCL_TGSM_RAW, "", "",
1075 shader_sm5_read_dcl_tgsm_raw},
1076 {WINED3D_SM5_OP_DCL_TGSM_STRUCTURED, WINED3DSIH_DCL_TGSM_STRUCTURED, "", "",
1077 shader_sm5_read_dcl_tgsm_structured},
1078 {WINED3D_SM5_OP_DCL_RESOURCE_RAW, WINED3DSIH_DCL_RESOURCE_RAW, "", "",
1079 shader_sm5_read_dcl_resource_raw},
1080 {WINED3D_SM5_OP_DCL_RESOURCE_STRUCTURED, WINED3DSIH_DCL_RESOURCE_STRUCTURED, "", "",
1081 shader_sm5_read_dcl_resource_structured},
1082 {WINED3D_SM5_OP_LD_UAV_TYPED, WINED3DSIH_LD_UAV_TYPED, "u", "iU"},
1083 {WINED3D_SM5_OP_STORE_UAV_TYPED, WINED3DSIH_STORE_UAV_TYPED, "U", "iu"},
1084 {WINED3D_SM5_OP_LD_RAW, WINED3DSIH_LD_RAW, "u", "iU"},
1085 {WINED3D_SM5_OP_STORE_RAW, WINED3DSIH_STORE_RAW, "U", "iu"},
1086 {WINED3D_SM5_OP_LD_STRUCTURED, WINED3DSIH_LD_STRUCTURED, "u", "iiR"},
1087 {WINED3D_SM5_OP_STORE_STRUCTURED, WINED3DSIH_STORE_STRUCTURED, "U", "iiu"},
1088 {WINED3D_SM5_OP_ATOMIC_AND, WINED3DSIH_ATOMIC_AND, "U", "iu"},
1089 {WINED3D_SM5_OP_ATOMIC_OR, WINED3DSIH_ATOMIC_OR, "U", "iu"},
1090 {WINED3D_SM5_OP_ATOMIC_XOR, WINED3DSIH_ATOMIC_XOR, "U", "iu"},
1091 {WINED3D_SM5_OP_ATOMIC_CMP_STORE, WINED3DSIH_ATOMIC_CMP_STORE, "U", "iuu"},
1092 {WINED3D_SM5_OP_ATOMIC_IADD, WINED3DSIH_ATOMIC_IADD, "U", "ii"},
1093 {WINED3D_SM5_OP_ATOMIC_IMAX, WINED3DSIH_ATOMIC_IMAX, "U", "ii"},
1094 {WINED3D_SM5_OP_ATOMIC_IMIN, WINED3DSIH_ATOMIC_IMIN, "U", "ii"},
1095 {WINED3D_SM5_OP_ATOMIC_UMAX, WINED3DSIH_ATOMIC_UMAX, "U", "iu"},
1096 {WINED3D_SM5_OP_ATOMIC_UMIN, WINED3DSIH_ATOMIC_UMIN, "U", "iu"},
1097 {WINED3D_SM5_OP_IMM_ATOMIC_ALLOC, WINED3DSIH_IMM_ATOMIC_ALLOC, "u", "U"},
1098 {WINED3D_SM5_OP_IMM_ATOMIC_CONSUME, WINED3DSIH_IMM_ATOMIC_CONSUME, "u", "U"},
1099 {WINED3D_SM5_OP_IMM_ATOMIC_IADD, WINED3DSIH_IMM_ATOMIC_IADD, "uU", "ii"},
1100 {WINED3D_SM5_OP_IMM_ATOMIC_AND, WINED3DSIH_IMM_ATOMIC_AND, "uU", "iu"},
1101 {WINED3D_SM5_OP_IMM_ATOMIC_OR, WINED3DSIH_IMM_ATOMIC_OR, "uU", "iu"},
1102 {WINED3D_SM5_OP_IMM_ATOMIC_XOR, WINED3DSIH_IMM_ATOMIC_XOR, "uU", "iu"},
1103 {WINED3D_SM5_OP_IMM_ATOMIC_EXCH, WINED3DSIH_IMM_ATOMIC_EXCH, "uU", "iu"},
1104 {WINED3D_SM5_OP_IMM_ATOMIC_CMP_EXCH, WINED3DSIH_IMM_ATOMIC_CMP_EXCH, "uU", "iuu"},
1105 {WINED3D_SM5_OP_IMM_ATOMIC_IMAX, WINED3DSIH_IMM_ATOMIC_IMAX, "iU", "ii"},
1106 {WINED3D_SM5_OP_IMM_ATOMIC_IMIN, WINED3DSIH_IMM_ATOMIC_IMIN, "iU", "ii"},
1107 {WINED3D_SM5_OP_IMM_ATOMIC_UMAX, WINED3DSIH_IMM_ATOMIC_UMAX, "uU", "iu"},
1108 {WINED3D_SM5_OP_IMM_ATOMIC_UMIN, WINED3DSIH_IMM_ATOMIC_UMIN, "uU", "iu"},
1109 {WINED3D_SM5_OP_SYNC, WINED3DSIH_SYNC, "", "",
1110 shader_sm5_read_sync},
1111 {WINED3D_SM5_OP_EVAL_SAMPLE_INDEX, WINED3DSIH_EVAL_SAMPLE_INDEX, "f", "fi"},
1112 {WINED3D_SM5_OP_DCL_GS_INSTANCES, WINED3DSIH_DCL_GS_INSTANCES, "", "",
1113 shader_sm4_read_declaration_count},
1114 };
1115
1116 static const enum wined3d_shader_register_type register_type_table[] =
1117 {
1118 /* WINED3D_SM4_RT_TEMP */ WINED3DSPR_TEMP,
1119 /* WINED3D_SM4_RT_INPUT */ WINED3DSPR_INPUT,
1120 /* WINED3D_SM4_RT_OUTPUT */ WINED3DSPR_OUTPUT,
1121 /* WINED3D_SM4_RT_INDEXABLE_TEMP */ WINED3DSPR_IDXTEMP,
1122 /* WINED3D_SM4_RT_IMMCONST */ WINED3DSPR_IMMCONST,
1123 /* UNKNOWN */ ~0u,
1124 /* WINED3D_SM4_RT_SAMPLER */ WINED3DSPR_SAMPLER,
1125 /* WINED3D_SM4_RT_RESOURCE */ WINED3DSPR_RESOURCE,
1126 /* WINED3D_SM4_RT_CONSTBUFFER */ WINED3DSPR_CONSTBUFFER,
1127 /* WINED3D_SM4_RT_IMMCONSTBUFFER */ WINED3DSPR_IMMCONSTBUFFER,
1128 /* UNKNOWN */ ~0u,
1129 /* WINED3D_SM4_RT_PRIMID */ WINED3DSPR_PRIMID,
1130 /* WINED3D_SM4_RT_DEPTHOUT */ WINED3DSPR_DEPTHOUT,
1131 /* WINED3D_SM4_RT_NULL */ WINED3DSPR_NULL,
1132 /* UNKNOWN */ ~0u,
1133 /* WINED3D_SM4_RT_OMASK */ WINED3DSPR_SAMPLEMASK,
1134 /* WINED3D_SM5_RT_STREAM */ WINED3DSPR_STREAM,
1135 /* WINED3D_SM5_RT_FUNCTION_BODY */ WINED3DSPR_FUNCTIONBODY,
1136 /* UNKNOWN */ ~0u,
1137 /* WINED3D_SM5_RT_FUNCTION_POINTER */ WINED3DSPR_FUNCTIONPOINTER,
1138 /* UNKNOWN */ ~0u,
1139 /* UNKNOWN */ ~0u,
1140 /* WINED3D_SM5_RT_OUTPUT_CONTROL_POINT_ID */ WINED3DSPR_OUTPOINTID,
1141 /* WINED3D_SM5_RT_FORK_INSTANCE_ID */ WINED3DSPR_FORKINSTID,
1142 /* WINED3D_SM5_RT_JOIN_INSTANCE_ID */ WINED3DSPR_JOININSTID,
1143 /* WINED3D_SM5_RT_INPUT_CONTROL_POINT */ WINED3DSPR_INCONTROLPOINT,
1144 /* WINED3D_SM5_RT_OUTPUT_CONTROL_POINT */ WINED3DSPR_OUTCONTROLPOINT,
1145 /* WINED3D_SM5_RT_PATCH_CONSTANT_DATA */ WINED3DSPR_PATCHCONST,
1146 /* WINED3D_SM5_RT_DOMAIN_LOCATION */ WINED3DSPR_TESSCOORD,
1147 /* UNKNOWN */ ~0u,
1148 /* WINED3D_SM5_RT_UAV */ WINED3DSPR_UAV,
1149 /* WINED3D_SM5_RT_SHARED_MEMORY */ WINED3DSPR_GROUPSHAREDMEM,
1150 /* WINED3D_SM5_RT_THREAD_ID */ WINED3DSPR_THREADID,
1151 /* WINED3D_SM5_RT_THREAD_GROUP_ID */ WINED3DSPR_THREADGROUPID,
1152 /* WINED3D_SM5_RT_LOCAL_THREAD_ID */ WINED3DSPR_LOCALTHREADID,
1153 /* WINED3D_SM5_RT_COVERAGE */ WINED3DSPR_COVERAGE,
1154 /* WINED3D_SM5_RT_LOCAL_THREAD_INDEX */ WINED3DSPR_LOCALTHREADINDEX,
1155 /* WINED3D_SM5_RT_GS_INSTANCE_ID */ WINED3DSPR_GSINSTID,
1156 /* WINED3D_SM5_RT_DEPTHOUT_GREATER_EQUAL */ WINED3DSPR_DEPTHOUTGE,
1157 /* WINED3D_SM5_RT_DEPTHOUT_LESS_EQUAL */ WINED3DSPR_DEPTHOUTLE,
1158 };
1159
get_opcode_info(enum wined3d_sm4_opcode opcode)1160 static const struct wined3d_sm4_opcode_info *get_opcode_info(enum wined3d_sm4_opcode opcode)
1161 {
1162 unsigned int i;
1163
1164 for (i = 0; i < ARRAY_SIZE(opcode_table); ++i)
1165 {
1166 if (opcode == opcode_table[i].opcode) return &opcode_table[i];
1167 }
1168
1169 return NULL;
1170 }
1171
map_register(const struct wined3d_sm4_data * priv,struct wined3d_shader_register * reg)1172 static void map_register(const struct wined3d_sm4_data *priv, struct wined3d_shader_register *reg)
1173 {
1174 switch (priv->shader_version.type)
1175 {
1176 case WINED3D_SHADER_TYPE_PIXEL:
1177 if (reg->type == WINED3DSPR_OUTPUT)
1178 {
1179 unsigned int reg_idx = reg->idx[0].offset;
1180
1181 if (reg_idx >= ARRAY_SIZE(priv->output_map))
1182 {
1183 ERR("Invalid output index %u.\n", reg_idx);
1184 break;
1185 }
1186
1187 reg->type = WINED3DSPR_COLOROUT;
1188 reg->idx[0].offset = priv->output_map[reg_idx];
1189 }
1190 break;
1191
1192 default:
1193 break;
1194 }
1195 }
1196
map_data_type(char t)1197 static enum wined3d_data_type map_data_type(char t)
1198 {
1199 switch (t)
1200 {
1201 case 'f':
1202 return WINED3D_DATA_FLOAT;
1203 case 'i':
1204 return WINED3D_DATA_INT;
1205 case 'u':
1206 return WINED3D_DATA_UINT;
1207 case 'O':
1208 return WINED3D_DATA_OPAQUE;
1209 case 'R':
1210 return WINED3D_DATA_RESOURCE;
1211 case 'S':
1212 return WINED3D_DATA_SAMPLER;
1213 case 'U':
1214 return WINED3D_DATA_UAV;
1215 default:
1216 ERR("Invalid data type '%c'.\n", t);
1217 return WINED3D_DATA_FLOAT;
1218 }
1219 }
1220
shader_sm4_init(const DWORD * byte_code,size_t byte_code_size,const struct wined3d_shader_signature * output_signature)1221 static void *shader_sm4_init(const DWORD *byte_code, size_t byte_code_size,
1222 const struct wined3d_shader_signature *output_signature)
1223 {
1224 DWORD version_token, token_count;
1225 struct wined3d_sm4_data *priv;
1226 unsigned int i;
1227
1228 if (byte_code_size / sizeof(*byte_code) < 2)
1229 {
1230 WARN("Invalid byte code size %lu.\n", (long)byte_code_size);
1231 return NULL;
1232 }
1233
1234 version_token = byte_code[0];
1235 TRACE("Version: 0x%08x.\n", version_token);
1236 token_count = byte_code[1];
1237 TRACE("Token count: %u.\n", token_count);
1238
1239 if (token_count < 2 || byte_code_size / sizeof(*byte_code) < token_count)
1240 {
1241 WARN("Invalid token count %u.\n", token_count);
1242 return NULL;
1243 }
1244
1245 if (!(priv = heap_alloc(sizeof(*priv))))
1246 {
1247 ERR("Failed to allocate private data\n");
1248 return NULL;
1249 }
1250
1251 priv->start = &byte_code[2];
1252 priv->end = &byte_code[token_count];
1253
1254 switch (version_token >> 16)
1255 {
1256 case WINED3D_SM4_PS:
1257 priv->shader_version.type = WINED3D_SHADER_TYPE_PIXEL;
1258 break;
1259
1260 case WINED3D_SM4_VS:
1261 priv->shader_version.type = WINED3D_SHADER_TYPE_VERTEX;
1262 break;
1263
1264 case WINED3D_SM4_GS:
1265 priv->shader_version.type = WINED3D_SHADER_TYPE_GEOMETRY;
1266 break;
1267
1268 case WINED3D_SM5_HS:
1269 priv->shader_version.type = WINED3D_SHADER_TYPE_HULL;
1270 break;
1271
1272 case WINED3D_SM5_DS:
1273 priv->shader_version.type = WINED3D_SHADER_TYPE_DOMAIN;
1274 break;
1275
1276 case WINED3D_SM5_CS:
1277 priv->shader_version.type = WINED3D_SHADER_TYPE_COMPUTE;
1278 break;
1279
1280 default:
1281 FIXME("Unrecognised shader type %#x.\n", version_token >> 16);
1282 }
1283 priv->shader_version.major = WINED3D_SM4_VERSION_MAJOR(version_token);
1284 priv->shader_version.minor = WINED3D_SM4_VERSION_MINOR(version_token);
1285
1286 memset(priv->output_map, 0xff, sizeof(priv->output_map));
1287 for (i = 0; i < output_signature->element_count; ++i)
1288 {
1289 struct wined3d_shader_signature_element *e = &output_signature->elements[i];
1290
1291 if (e->register_idx >= ARRAY_SIZE(priv->output_map))
1292 {
1293 WARN("Invalid output index %u.\n", e->register_idx);
1294 continue;
1295 }
1296
1297 priv->output_map[e->register_idx] = e->semantic_idx;
1298 }
1299
1300 list_init(&priv->src_free);
1301 list_init(&priv->src);
1302
1303 return priv;
1304 }
1305
shader_sm4_free(void * data)1306 static void shader_sm4_free(void *data)
1307 {
1308 struct wined3d_shader_src_param_entry *e1, *e2;
1309 struct wined3d_sm4_data *priv = data;
1310
1311 list_move_head(&priv->src_free, &priv->src);
1312 LIST_FOR_EACH_ENTRY_SAFE(e1, e2, &priv->src_free, struct wined3d_shader_src_param_entry, entry)
1313 {
1314 heap_free(e1);
1315 }
1316 heap_free(priv);
1317 }
1318
get_src_param(struct wined3d_sm4_data * priv)1319 static struct wined3d_shader_src_param *get_src_param(struct wined3d_sm4_data *priv)
1320 {
1321 struct wined3d_shader_src_param_entry *e;
1322 struct list *elem;
1323
1324 if (!list_empty(&priv->src_free))
1325 {
1326 elem = list_head(&priv->src_free);
1327 list_remove(elem);
1328 }
1329 else
1330 {
1331 if (!(e = heap_alloc(sizeof(*e))))
1332 return NULL;
1333 elem = &e->entry;
1334 }
1335
1336 list_add_tail(&priv->src, elem);
1337 e = LIST_ENTRY(elem, struct wined3d_shader_src_param_entry, entry);
1338 return &e->param;
1339 }
1340
shader_sm4_read_header(void * data,const DWORD ** ptr,struct wined3d_shader_version * shader_version)1341 static void shader_sm4_read_header(void *data, const DWORD **ptr, struct wined3d_shader_version *shader_version)
1342 {
1343 struct wined3d_sm4_data *priv = data;
1344
1345 *ptr = priv->start;
1346 *shader_version = priv->shader_version;
1347 }
1348
shader_sm4_read_reg_idx(struct wined3d_sm4_data * priv,const DWORD ** ptr,const DWORD * end,DWORD addressing,struct wined3d_shader_register_index * reg_idx)1349 static BOOL shader_sm4_read_reg_idx(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1350 DWORD addressing, struct wined3d_shader_register_index *reg_idx)
1351 {
1352 if (addressing & WINED3D_SM4_ADDRESSING_RELATIVE)
1353 {
1354 struct wined3d_shader_src_param *rel_addr = get_src_param(priv);
1355
1356 if (!(reg_idx->rel_addr = rel_addr))
1357 {
1358 ERR("Failed to get src param for relative addressing.\n");
1359 return FALSE;
1360 }
1361
1362 if (addressing & WINED3D_SM4_ADDRESSING_OFFSET)
1363 reg_idx->offset = *(*ptr)++;
1364 else
1365 reg_idx->offset = 0;
1366 shader_sm4_read_src_param(priv, ptr, end, WINED3D_DATA_INT, rel_addr);
1367 }
1368 else
1369 {
1370 reg_idx->rel_addr = NULL;
1371 reg_idx->offset = *(*ptr)++;
1372 }
1373
1374 return TRUE;
1375 }
1376
shader_sm4_read_param(struct wined3d_sm4_data * priv,const DWORD ** ptr,const DWORD * end,enum wined3d_data_type data_type,struct wined3d_shader_register * param,enum wined3d_shader_src_modifier * modifier)1377 static BOOL shader_sm4_read_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1378 enum wined3d_data_type data_type, struct wined3d_shader_register *param,
1379 enum wined3d_shader_src_modifier *modifier)
1380 {
1381 enum wined3d_sm4_register_type register_type;
1382 DWORD token, order;
1383
1384 if (*ptr >= end)
1385 {
1386 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1387 return FALSE;
1388 }
1389 token = *(*ptr)++;
1390
1391 register_type = (token & WINED3D_SM4_REGISTER_TYPE_MASK) >> WINED3D_SM4_REGISTER_TYPE_SHIFT;
1392 if (register_type >= ARRAY_SIZE(register_type_table)
1393 || register_type_table[register_type] == ~0u)
1394 {
1395 FIXME("Unhandled register type %#x.\n", register_type);
1396 param->type = WINED3DSPR_TEMP;
1397 }
1398 else
1399 {
1400 param->type = register_type_table[register_type];
1401 }
1402 param->data_type = data_type;
1403
1404 if (token & WINED3D_SM4_REGISTER_MODIFIER)
1405 {
1406 DWORD m;
1407
1408 if (*ptr >= end)
1409 {
1410 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1411 return FALSE;
1412 }
1413 m = *(*ptr)++;
1414
1415 switch (m)
1416 {
1417 case 0x41:
1418 *modifier = WINED3DSPSM_NEG;
1419 break;
1420
1421 case 0x81:
1422 *modifier = WINED3DSPSM_ABS;
1423 break;
1424
1425 case 0xc1:
1426 *modifier = WINED3DSPSM_ABSNEG;
1427 break;
1428
1429 default:
1430 FIXME("Skipping modifier 0x%08x.\n", m);
1431 *modifier = WINED3DSPSM_NONE;
1432 break;
1433 }
1434 }
1435 else
1436 {
1437 *modifier = WINED3DSPSM_NONE;
1438 }
1439
1440 order = (token & WINED3D_SM4_REGISTER_ORDER_MASK) >> WINED3D_SM4_REGISTER_ORDER_SHIFT;
1441
1442 if (order < 1)
1443 param->idx[0].offset = ~0U;
1444 else
1445 {
1446 DWORD addressing = (token & WINED3D_SM4_ADDRESSING_MASK0) >> WINED3D_SM4_ADDRESSING_SHIFT0;
1447 if (!(shader_sm4_read_reg_idx(priv, ptr, end, addressing, ¶m->idx[0])))
1448 {
1449 ERR("Failed to read register index.\n");
1450 return FALSE;
1451 }
1452 }
1453
1454 if (order < 2)
1455 param->idx[1].offset = ~0U;
1456 else
1457 {
1458 DWORD addressing = (token & WINED3D_SM4_ADDRESSING_MASK1) >> WINED3D_SM4_ADDRESSING_SHIFT1;
1459 if (!(shader_sm4_read_reg_idx(priv, ptr, end, addressing, ¶m->idx[1])))
1460 {
1461 ERR("Failed to read register index.\n");
1462 return FALSE;
1463 }
1464 }
1465
1466 if (order > 2)
1467 FIXME("Unhandled order %u.\n", order);
1468
1469 if (register_type == WINED3D_SM4_RT_IMMCONST)
1470 {
1471 enum wined3d_sm4_immconst_type immconst_type =
1472 (token & WINED3D_SM4_IMMCONST_TYPE_MASK) >> WINED3D_SM4_IMMCONST_TYPE_SHIFT;
1473
1474 switch (immconst_type)
1475 {
1476 case WINED3D_SM4_IMMCONST_SCALAR:
1477 param->immconst_type = WINED3D_IMMCONST_SCALAR;
1478 if (end - *ptr < 1)
1479 {
1480 WARN("Invalid ptr %p, end %p.\n", *ptr, end);
1481 return FALSE;
1482 }
1483 memcpy(param->u.immconst_data, *ptr, 1 * sizeof(DWORD));
1484 *ptr += 1;
1485 break;
1486
1487 case WINED3D_SM4_IMMCONST_VEC4:
1488 param->immconst_type = WINED3D_IMMCONST_VEC4;
1489 if (end - *ptr < 4)
1490 {
1491 WARN("Invalid ptr %p, end %p.\n", *ptr, end);
1492 return FALSE;
1493 }
1494 memcpy(param->u.immconst_data, *ptr, 4 * sizeof(DWORD));
1495 *ptr += 4;
1496 break;
1497
1498 default:
1499 FIXME("Unhandled immediate constant type %#x.\n", immconst_type);
1500 break;
1501 }
1502 }
1503
1504 map_register(priv, param);
1505
1506 return TRUE;
1507 }
1508
shader_sm4_read_src_param(struct wined3d_sm4_data * priv,const DWORD ** ptr,const DWORD * end,enum wined3d_data_type data_type,struct wined3d_shader_src_param * src_param)1509 static BOOL shader_sm4_read_src_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1510 enum wined3d_data_type data_type, struct wined3d_shader_src_param *src_param)
1511 {
1512 DWORD token;
1513
1514 if (*ptr >= end)
1515 {
1516 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1517 return FALSE;
1518 }
1519 token = **ptr;
1520
1521 if (!shader_sm4_read_param(priv, ptr, end, data_type, &src_param->reg, &src_param->modifiers))
1522 {
1523 ERR("Failed to read parameter.\n");
1524 return FALSE;
1525 }
1526
1527 if (src_param->reg.type == WINED3DSPR_IMMCONST)
1528 {
1529 src_param->swizzle = WINED3DSP_NOSWIZZLE;
1530 }
1531 else
1532 {
1533 enum wined3d_sm4_swizzle_type swizzle_type =
1534 (token & WINED3D_SM4_SWIZZLE_TYPE_MASK) >> WINED3D_SM4_SWIZZLE_TYPE_SHIFT;
1535
1536 switch (swizzle_type)
1537 {
1538 case WINED3D_SM4_SWIZZLE_NONE:
1539 src_param->swizzle = WINED3DSP_NOSWIZZLE;
1540 break;
1541
1542 case WINED3D_SM4_SWIZZLE_SCALAR:
1543 src_param->swizzle = (token & WINED3D_SM4_SWIZZLE_MASK) >> WINED3D_SM4_SWIZZLE_SHIFT;
1544 src_param->swizzle = (src_param->swizzle & 0x3) * 0x55;
1545 break;
1546
1547 case WINED3D_SM4_SWIZZLE_VEC4:
1548 src_param->swizzle = (token & WINED3D_SM4_SWIZZLE_MASK) >> WINED3D_SM4_SWIZZLE_SHIFT;
1549 break;
1550
1551 default:
1552 FIXME("Unhandled swizzle type %#x.\n", swizzle_type);
1553 break;
1554 }
1555 }
1556
1557 return TRUE;
1558 }
1559
shader_sm4_read_dst_param(struct wined3d_sm4_data * priv,const DWORD ** ptr,const DWORD * end,enum wined3d_data_type data_type,struct wined3d_shader_dst_param * dst_param)1560 static BOOL shader_sm4_read_dst_param(struct wined3d_sm4_data *priv, const DWORD **ptr, const DWORD *end,
1561 enum wined3d_data_type data_type, struct wined3d_shader_dst_param *dst_param)
1562 {
1563 enum wined3d_shader_src_modifier modifier;
1564 DWORD token;
1565
1566 if (*ptr >= end)
1567 {
1568 WARN("Invalid ptr %p >= end %p.\n", *ptr, end);
1569 return FALSE;
1570 }
1571 token = **ptr;
1572
1573 if (!shader_sm4_read_param(priv, ptr, end, data_type, &dst_param->reg, &modifier))
1574 {
1575 ERR("Failed to read parameter.\n");
1576 return FALSE;
1577 }
1578
1579 if (modifier != WINED3DSPSM_NONE)
1580 {
1581 ERR("Invalid source modifier %#x on destination register.\n", modifier);
1582 return FALSE;
1583 }
1584
1585 dst_param->write_mask = (token & WINED3D_SM4_WRITEMASK_MASK) >> WINED3D_SM4_WRITEMASK_SHIFT;
1586 dst_param->modifiers = 0;
1587 dst_param->shift = 0;
1588
1589 return TRUE;
1590 }
1591
shader_sm4_read_instruction_modifier(DWORD modifier,struct wined3d_shader_instruction * ins)1592 static void shader_sm4_read_instruction_modifier(DWORD modifier, struct wined3d_shader_instruction *ins)
1593 {
1594 static const DWORD recognized_bits = WINED3D_SM4_INSTRUCTION_MODIFIER
1595 | WINED3D_SM4_MODIFIER_AOFFIMMI
1596 | WINED3D_SM4_AOFFIMMI_U_MASK
1597 | WINED3D_SM4_AOFFIMMI_V_MASK
1598 | WINED3D_SM4_AOFFIMMI_W_MASK;
1599
1600 if (modifier & ~recognized_bits)
1601 {
1602 FIXME("Unhandled modifier 0x%08x.\n", modifier);
1603 }
1604 else
1605 {
1606 /* Bit fields are used for sign extension */
1607 struct
1608 {
1609 int u : 4;
1610 int v : 4;
1611 int w : 4;
1612 }
1613 aoffimmi;
1614 aoffimmi.u = (modifier & WINED3D_SM4_AOFFIMMI_U_MASK) >> WINED3D_SM4_AOFFIMMI_U_SHIFT;
1615 aoffimmi.v = (modifier & WINED3D_SM4_AOFFIMMI_V_MASK) >> WINED3D_SM4_AOFFIMMI_V_SHIFT;
1616 aoffimmi.w = (modifier & WINED3D_SM4_AOFFIMMI_W_MASK) >> WINED3D_SM4_AOFFIMMI_W_SHIFT;
1617 ins->texel_offset.u = aoffimmi.u;
1618 ins->texel_offset.v = aoffimmi.v;
1619 ins->texel_offset.w = aoffimmi.w;
1620 }
1621 }
1622
shader_sm4_read_instruction(void * data,const DWORD ** ptr,struct wined3d_shader_instruction * ins)1623 static void shader_sm4_read_instruction(void *data, const DWORD **ptr, struct wined3d_shader_instruction *ins)
1624 {
1625 const struct wined3d_sm4_opcode_info *opcode_info;
1626 DWORD opcode_token, opcode, previous_token;
1627 struct wined3d_sm4_data *priv = data;
1628 unsigned int i, len;
1629 SIZE_T remaining;
1630 const DWORD *p;
1631
1632 list_move_head(&priv->src_free, &priv->src);
1633
1634 if (*ptr >= priv->end)
1635 {
1636 WARN("End of byte-code, failed to read opcode.\n");
1637 goto fail;
1638 }
1639 remaining = priv->end - *ptr;
1640
1641 opcode_token = *(*ptr)++;
1642 opcode = opcode_token & WINED3D_SM4_OPCODE_MASK;
1643
1644 len = ((opcode_token & WINED3D_SM4_INSTRUCTION_LENGTH_MASK) >> WINED3D_SM4_INSTRUCTION_LENGTH_SHIFT);
1645 if (!len)
1646 {
1647 if (remaining < 2)
1648 {
1649 WARN("End of byte-code, failed to read length token.\n");
1650 goto fail;
1651 }
1652 len = **ptr;
1653 }
1654 if (!len || remaining < len)
1655 {
1656 WARN("Read invalid length %u (remaining %lu).\n", len, remaining);
1657 goto fail;
1658 }
1659 --len;
1660
1661 if (TRACE_ON(d3d_bytecode))
1662 {
1663 TRACE_(d3d_bytecode)("[ %08x ", opcode_token);
1664 for (i = 0; i < len; ++i)
1665 {
1666 TRACE_(d3d_bytecode)("%08x ", (*ptr)[i]);
1667 }
1668 TRACE_(d3d_bytecode)("]\n");
1669 }
1670
1671 if (!(opcode_info = get_opcode_info(opcode)))
1672 {
1673 FIXME("Unrecognized opcode %#x, opcode_token 0x%08x.\n", opcode, opcode_token);
1674 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1675 *ptr += len;
1676 return;
1677 }
1678
1679 ins->handler_idx = opcode_info->handler_idx;
1680 ins->flags = 0;
1681 ins->coissue = 0;
1682 ins->predicate = NULL;
1683 ins->dst_count = strlen(opcode_info->dst_info);
1684 ins->dst = priv->dst_param;
1685 ins->src_count = strlen(opcode_info->src_info);
1686 ins->src = priv->src_param;
1687 memset(&ins->texel_offset, 0, sizeof(ins->texel_offset));
1688
1689 p = *ptr;
1690 *ptr += len;
1691
1692 if (opcode_info->read_opcode_func)
1693 {
1694 opcode_info->read_opcode_func(ins, opcode, opcode_token, p, len, priv);
1695 }
1696 else
1697 {
1698 enum wined3d_shader_dst_modifier instruction_dst_modifier = WINED3DSPDM_NONE;
1699
1700 previous_token = opcode_token;
1701 while (previous_token & WINED3D_SM4_INSTRUCTION_MODIFIER && p != *ptr)
1702 shader_sm4_read_instruction_modifier(previous_token = *p++, ins);
1703
1704 ins->flags = (opcode_token & WINED3D_SM4_INSTRUCTION_FLAGS_MASK) >> WINED3D_SM4_INSTRUCTION_FLAGS_SHIFT;
1705
1706 if (ins->flags & WINED3D_SM4_INSTRUCTION_FLAG_SATURATE)
1707 {
1708 ins->flags &= ~WINED3D_SM4_INSTRUCTION_FLAG_SATURATE;
1709 instruction_dst_modifier = WINED3DSPDM_SATURATE;
1710 }
1711
1712 for (i = 0; i < ins->dst_count; ++i)
1713 {
1714 if (!(shader_sm4_read_dst_param(priv, &p, *ptr, map_data_type(opcode_info->dst_info[i]),
1715 &priv->dst_param[i])))
1716 {
1717 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1718 return;
1719 }
1720 priv->dst_param[i].modifiers |= instruction_dst_modifier;
1721 }
1722
1723 for (i = 0; i < ins->src_count; ++i)
1724 {
1725 if (!(shader_sm4_read_src_param(priv, &p, *ptr, map_data_type(opcode_info->src_info[i]),
1726 &priv->src_param[i])))
1727 {
1728 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1729 return;
1730 }
1731 }
1732 }
1733
1734 return;
1735
1736 fail:
1737 *ptr = priv->end;
1738 ins->handler_idx = WINED3DSIH_TABLE_SIZE;
1739 return;
1740 }
1741
shader_sm4_is_end(void * data,const DWORD ** ptr)1742 static BOOL shader_sm4_is_end(void *data, const DWORD **ptr)
1743 {
1744 struct wined3d_sm4_data *priv = data;
1745 return *ptr == priv->end;
1746 }
1747
1748 const struct wined3d_shader_frontend sm4_shader_frontend =
1749 {
1750 shader_sm4_init,
1751 shader_sm4_free,
1752 shader_sm4_read_header,
1753 shader_sm4_read_instruction,
1754 shader_sm4_is_end,
1755 };
1756