summaryrefslogtreecommitdiffstats
path: root/src/mesa/drivers/dri/i965/brw_shader.h
blob: 80b41bca755be5fee30b3cd2f30961f7d0331f01 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
/*
 * Copyright © 2010 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 */

#include <stdint.h>
#include "brw_reg.h"
#include "brw_defines.h"
#include "brw_context.h"
#include "main/compiler.h"
#include "glsl/ir.h"
#include "program/prog_parameter.h"

#ifdef __cplusplus
#include "brw_ir_allocator.h"
#endif

#pragma once

#define MAX_SAMPLER_MESSAGE_SIZE 11
#define MAX_VGRF_SIZE 16

struct brw_compiler {
   const struct brw_device_info *devinfo;

   struct {
      struct ra_regs *regs;

      /**
       * Array of the ra classes for the unaligned contiguous register
       * block sizes used.
       */
      int *classes;

      /**
       * Mapping for register-allocated objects in *regs to the first
       * GRF for that object.
       */
      uint8_t *ra_reg_to_grf;
   } vec4_reg_set;

   struct {
      struct ra_regs *regs;

      /**
       * Array of the ra classes for the unaligned contiguous register
       * block sizes used, indexed by register size.
       */
      int classes[16];

      /**
       * Mapping from classes to ra_reg ranges.  Each of the per-size
       * classes corresponds to a range of ra_reg nodes.  This array stores
       * those ranges in the form of first ra_reg in each class and the
       * total number of ra_reg elements in the last array element.  This
       * way the range of the i'th class is given by:
       * [ class_to_ra_reg_range[i], class_to_ra_reg_range[i+1] )
       */
      int class_to_ra_reg_range[17];

      /**
       * Mapping for register-allocated objects in *regs to the first
       * GRF for that object.
       */
      uint8_t *ra_reg_to_grf;

      /**
       * ra class for the aligned pairs we use for PLN, which doesn't
       * appear in *classes.
       */
      int aligned_pairs_class;
   } fs_reg_sets[2];

   void (*shader_debug_log)(void *, const char *str, ...) PRINTFLIKE(2, 3);
   void (*shader_perf_log)(void *, const char *str, ...) PRINTFLIKE(2, 3);

   bool scalar_vs;
   struct gl_shader_compiler_options glsl_compiler_options[MESA_SHADER_STAGES];
};

enum PACKED register_file {
   BAD_FILE,
   GRF,
   MRF,
   IMM,
   HW_REG, /* a struct brw_reg */
   ATTR,
   UNIFORM, /* prog_data->params[reg] */
};

struct backend_reg
{
#ifdef __cplusplus
   bool is_zero() const;
   bool is_one() const;
   bool is_negative_one() const;
   bool is_null() const;
   bool is_accumulator() const;
   bool in_range(const backend_reg &r, unsigned n) const;
#endif

   enum register_file file; /**< Register file: GRF, MRF, IMM. */
   enum brw_reg_type type;  /**< Register type: BRW_REGISTER_TYPE_* */

   /**
    * Register number.
    *
    * For GRF, it's a virtual register number until register allocation.
    *
    * For MRF, it's the hardware register.
    */
   uint16_t reg;

   /**
    * Offset within the virtual register.
    *
    * In the scalar backend, this is in units of a float per pixel for pre-
    * register allocation registers (i.e., one register in SIMD8 mode and two
    * registers in SIMD16 mode).
    *
    * For uniforms, this is in units of 1 float.
    */
   uint16_t reg_offset;

   struct brw_reg fixed_hw_reg;

   bool negate;
   bool abs;
};

struct cfg_t;
struct bblock_t;

#ifdef __cplusplus
struct backend_instruction : public exec_node {
   bool is_3src() const;
   bool is_tex() const;
   bool is_math() const;
   bool is_control_flow() const;
   bool is_commutative() const;
   bool can_do_source_mods() const;
   bool can_do_saturate() const;
   bool can_do_cmod() const;
   bool reads_accumulator_implicitly() const;
   bool writes_accumulator_implicitly(const struct brw_device_info *devinfo) const;

   void remove(bblock_t *block);
   void insert_after(bblock_t *block, backend_instruction *inst);
   void insert_before(bblock_t *block, backend_instruction *inst);
   void insert_before(bblock_t *block, exec_list *list);

   /**
    * True if the instruction has side effects other than writing to
    * its destination registers.  You are expected not to reorder or
    * optimize these out unless you know what you are doing.
    */
   bool has_side_effects() const;
#else
struct backend_instruction {
   struct exec_node link;
#endif
   /** @{
    * Annotation for the generated IR.  One of the two can be set.
    */
   const void *ir;
   const char *annotation;
   /** @} */

   uint32_t offset; /**< spill/unspill offset or texture offset bitfield */
   uint8_t mlen; /**< SEND message length */
   int8_t base_mrf; /**< First MRF in the SEND message, if mlen is nonzero. */
   uint8_t target; /**< MRT target. */
   uint8_t regs_written; /**< Number of registers written by the instruction. */

   enum opcode opcode; /* BRW_OPCODE_* or FS_OPCODE_* */
   enum brw_conditional_mod conditional_mod; /**< BRW_CONDITIONAL_* */
   enum brw_predicate predicate;
   bool predicate_inverse:1;
   bool writes_accumulator:1; /**< instruction implicitly writes accumulator */
   bool force_writemask_all:1;
   bool no_dd_clear:1;
   bool no_dd_check:1;
   bool saturate:1;
   bool shadow_compare:1;

   /* Chooses which flag subregister (f0.0 or f0.1) is used for conditional
    * mod and predication.
    */
   unsigned flag_subreg:1;

   /** The number of hardware registers used for a message header. */
   uint8_t header_size;
};

#ifdef __cplusplus

enum instruction_scheduler_mode {
   SCHEDULE_PRE,
   SCHEDULE_PRE_NON_LIFO,
   SCHEDULE_PRE_LIFO,
   SCHEDULE_POST,
};

class backend_shader {
protected:

   backend_shader(const struct brw_compiler *compiler,
                  void *log_data,
                  void *mem_ctx,
                  struct gl_shader_program *shader_prog,
                  struct gl_program *prog,
                  struct brw_stage_prog_data *stage_prog_data,
                  gl_shader_stage stage);

public:

   const struct brw_compiler *compiler;
   void *log_data; /* Passed to compiler->*_log functions */

   const struct brw_device_info * const devinfo;
   nir_shader *nir;
   struct brw_shader * const shader;
   struct gl_shader_program * const shader_prog;
   struct gl_program * const prog;
   struct brw_stage_prog_data * const stage_prog_data;

   /** ralloc context for temporary data used during compile */
   void *mem_ctx;

   /**
    * List of either fs_inst or vec4_instruction (inheriting from
    * backend_instruction)
    */
   exec_list instructions;

   cfg_t *cfg;

   gl_shader_stage stage;
   bool debug_enabled;
   const char *stage_name;
   const char *stage_abbrev;

   brw::simple_allocator alloc;

   virtual void dump_instruction(backend_instruction *inst) = 0;
   virtual void dump_instruction(backend_instruction *inst, FILE *file) = 0;
   virtual void dump_instructions();
   virtual void dump_instructions(const char *name);

   void calculate_cfg();
   void invalidate_cfg();

   virtual void invalidate_live_intervals() = 0;
};

uint32_t brw_texture_offset(int *offsets, unsigned num_components);

void brw_setup_image_uniform_values(gl_shader_stage stage,
                                    struct brw_stage_prog_data *stage_prog_data,
                                    unsigned param_start_index,
                                    const gl_uniform_storage *storage);

#endif /* __cplusplus */

enum brw_reg_type brw_type_for_base_type(const struct glsl_type *type);
enum brw_conditional_mod brw_conditional_for_comparison(unsigned int op);
uint32_t brw_math_function(enum opcode op);
const char *brw_instruction_name(enum opcode op);
bool brw_saturate_immediate(enum brw_reg_type type, struct brw_reg *reg);
bool brw_negate_immediate(enum brw_reg_type type, struct brw_reg *reg);
bool brw_abs_immediate(enum brw_reg_type type, struct brw_reg *reg);

#ifdef __cplusplus
extern "C" {
#endif

struct brw_compiler *
brw_compiler_create(void *mem_ctx, const struct brw_device_info *devinfo);

void
brw_assign_common_binding_table_offsets(gl_shader_stage stage,
                                        const struct brw_device_info *devinfo,
                                        const struct gl_shader_program *shader_prog,
                                        const struct gl_program *prog,
                                        struct brw_stage_prog_data *stage_prog_data,
                                        uint32_t next_binding_table_offset);

bool brw_vs_precompile(struct gl_context *ctx,
                       struct gl_shader_program *shader_prog,
                       struct gl_program *prog);
bool brw_gs_precompile(struct gl_context *ctx,
                       struct gl_shader_program *shader_prog,
                       struct gl_program *prog);
bool brw_fs_precompile(struct gl_context *ctx,
                       struct gl_shader_program *shader_prog,
                       struct gl_program *prog);
bool brw_cs_precompile(struct gl_context *ctx,
                       struct gl_shader_program *shader_prog,
                       struct gl_program *prog);

int type_size_scalar(const struct glsl_type *type);
int type_size_vec4(const struct glsl_type *type);

#ifdef __cplusplus
}
#endif