summaryrefslogtreecommitdiffstats
path: root/src/video_core/shader/node.h
blob: 0397f4c6e045f8af37827d7a8b84f5953c313a85 (plain) (blame)
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
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
// Copyright 2019 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.

#pragma once

#include <array>
#include <cstddef>
#include <memory>
#include <string>
#include <tuple>
#include <utility>
#include <variant>
#include <vector>

#include "common/common_types.h"
#include "video_core/engines/shader_bytecode.h"

namespace VideoCommon::Shader {

enum class OperationCode {
    Assign, /// (float& dest, float src) -> void

    Select, /// (MetaArithmetic, bool pred, float a, float b) -> float

    FAdd,          /// (MetaArithmetic, float a, float b) -> float
    FMul,          /// (MetaArithmetic, float a, float b) -> float
    FDiv,          /// (MetaArithmetic, float a, float b) -> float
    FFma,          /// (MetaArithmetic, float a, float b, float c) -> float
    FNegate,       /// (MetaArithmetic, float a) -> float
    FAbsolute,     /// (MetaArithmetic, float a) -> float
    FClamp,        /// (MetaArithmetic, float value, float min, float max) -> float
    FCastHalf0,    /// (MetaArithmetic, f16vec2 a) -> float
    FCastHalf1,    /// (MetaArithmetic, f16vec2 a) -> float
    FMin,          /// (MetaArithmetic, float a, float b) -> float
    FMax,          /// (MetaArithmetic, float a, float b) -> float
    FCos,          /// (MetaArithmetic, float a) -> float
    FSin,          /// (MetaArithmetic, float a) -> float
    FExp2,         /// (MetaArithmetic, float a) -> float
    FLog2,         /// (MetaArithmetic, float a) -> float
    FInverseSqrt,  /// (MetaArithmetic, float a) -> float
    FSqrt,         /// (MetaArithmetic, float a) -> float
    FRoundEven,    /// (MetaArithmetic, float a) -> float
    FFloor,        /// (MetaArithmetic, float a) -> float
    FCeil,         /// (MetaArithmetic, float a) -> float
    FTrunc,        /// (MetaArithmetic, float a) -> float
    FCastInteger,  /// (MetaArithmetic, int a) -> float
    FCastUInteger, /// (MetaArithmetic, uint a) -> float

    IAdd,                  /// (MetaArithmetic, int a, int b) -> int
    IMul,                  /// (MetaArithmetic, int a, int b) -> int
    IDiv,                  /// (MetaArithmetic, int a, int b) -> int
    INegate,               /// (MetaArithmetic, int a) -> int
    IAbsolute,             /// (MetaArithmetic, int a) -> int
    IMin,                  /// (MetaArithmetic, int a, int b) -> int
    IMax,                  /// (MetaArithmetic, int a, int b) -> int
    ICastFloat,            /// (MetaArithmetic, float a) -> int
    ICastUnsigned,         /// (MetaArithmetic, uint a) -> int
    ILogicalShiftLeft,     /// (MetaArithmetic, int a, uint b) -> int
    ILogicalShiftRight,    /// (MetaArithmetic, int a, uint b) -> int
    IArithmeticShiftRight, /// (MetaArithmetic, int a, uint b) -> int
    IBitwiseAnd,           /// (MetaArithmetic, int a, int b) -> int
    IBitwiseOr,            /// (MetaArithmetic, int a, int b) -> int
    IBitwiseXor,           /// (MetaArithmetic, int a, int b) -> int
    IBitwiseNot,           /// (MetaArithmetic, int a) -> int
    IBitfieldInsert,       /// (MetaArithmetic, int base, int insert, int offset, int bits) -> int
    IBitfieldExtract,      /// (MetaArithmetic, int value, int offset, int offset) -> int
    IBitCount,             /// (MetaArithmetic, int) -> int

    UAdd,                  /// (MetaArithmetic, uint a, uint b) -> uint
    UMul,                  /// (MetaArithmetic, uint a, uint b) -> uint
    UDiv,                  /// (MetaArithmetic, uint a, uint b) -> uint
    UMin,                  /// (MetaArithmetic, uint a, uint b) -> uint
    UMax,                  /// (MetaArithmetic, uint a, uint b) -> uint
    UCastFloat,            /// (MetaArithmetic, float a) -> uint
    UCastSigned,           /// (MetaArithmetic, int a) -> uint
    ULogicalShiftLeft,     /// (MetaArithmetic, uint a, uint b) -> uint
    ULogicalShiftRight,    /// (MetaArithmetic, uint a, uint b) -> uint
    UArithmeticShiftRight, /// (MetaArithmetic, uint a, uint b) -> uint
    UBitwiseAnd,           /// (MetaArithmetic, uint a, uint b) -> uint
    UBitwiseOr,            /// (MetaArithmetic, uint a, uint b) -> uint
    UBitwiseXor,           /// (MetaArithmetic, uint a, uint b) -> uint
    UBitwiseNot,           /// (MetaArithmetic, uint a) -> uint
    UBitfieldInsert,  /// (MetaArithmetic, uint base, uint insert, int offset, int bits) -> uint
    UBitfieldExtract, /// (MetaArithmetic, uint value, int offset, int offset) -> uint
    UBitCount,        /// (MetaArithmetic, uint) -> uint

    HAdd,       /// (MetaArithmetic, f16vec2 a, f16vec2 b) -> f16vec2
    HMul,       /// (MetaArithmetic, f16vec2 a, f16vec2 b) -> f16vec2
    HFma,       /// (MetaArithmetic, f16vec2 a, f16vec2 b, f16vec2 c) -> f16vec2
    HAbsolute,  /// (f16vec2 a) -> f16vec2
    HNegate,    /// (f16vec2 a, bool first, bool second) -> f16vec2
    HClamp,     /// (f16vec2 src, float min, float max) -> f16vec2
    HCastFloat, /// (MetaArithmetic, float a) -> f16vec2
    HUnpack,    /// (Tegra::Shader::HalfType, T value) -> f16vec2
    HMergeF32,  /// (f16vec2 src) -> float
    HMergeH0,   /// (f16vec2 dest, f16vec2 src) -> f16vec2
    HMergeH1,   /// (f16vec2 dest, f16vec2 src) -> f16vec2
    HPack2,     /// (float a, float b) -> f16vec2

    LogicalAssign, /// (bool& dst, bool src) -> void
    LogicalAnd,    /// (bool a, bool b) -> bool
    LogicalOr,     /// (bool a, bool b) -> bool
    LogicalXor,    /// (bool a, bool b) -> bool
    LogicalNegate, /// (bool a) -> bool
    LogicalPick2,  /// (bool2 pair, uint index) -> bool
    LogicalAnd2,   /// (bool2 a) -> bool

    LogicalFLessThan,     /// (float a, float b) -> bool
    LogicalFEqual,        /// (float a, float b) -> bool
    LogicalFLessEqual,    /// (float a, float b) -> bool
    LogicalFGreaterThan,  /// (float a, float b) -> bool
    LogicalFNotEqual,     /// (float a, float b) -> bool
    LogicalFGreaterEqual, /// (float a, float b) -> bool
    LogicalFIsNan,        /// (float a) -> bool

    LogicalILessThan,     /// (int a, int b) -> bool
    LogicalIEqual,        /// (int a, int b) -> bool
    LogicalILessEqual,    /// (int a, int b) -> bool
    LogicalIGreaterThan,  /// (int a, int b) -> bool
    LogicalINotEqual,     /// (int a, int b) -> bool
    LogicalIGreaterEqual, /// (int a, int b) -> bool

    LogicalULessThan,     /// (uint a, uint b) -> bool
    LogicalUEqual,        /// (uint a, uint b) -> bool
    LogicalULessEqual,    /// (uint a, uint b) -> bool
    LogicalUGreaterThan,  /// (uint a, uint b) -> bool
    LogicalUNotEqual,     /// (uint a, uint b) -> bool
    LogicalUGreaterEqual, /// (uint a, uint b) -> bool

    Logical2HLessThan,            /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HEqual,               /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HLessEqual,           /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HGreaterThan,         /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HNotEqual,            /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HGreaterEqual,        /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HLessThanWithNan,     /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HEqualWithNan,        /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HLessEqualWithNan,    /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HGreaterThanWithNan,  /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HNotEqualWithNan,     /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2
    Logical2HGreaterEqualWithNan, /// (MetaHalfArithmetic, f16vec2 a, f16vec2) -> bool2

    Texture,                /// (MetaTexture, float[N] coords) -> float4
    TextureLod,             /// (MetaTexture, float[N] coords) -> float4
    TextureGather,          /// (MetaTexture, float[N] coords) -> float4
    TextureQueryDimensions, /// (MetaTexture, float a) -> float4
    TextureQueryLod,        /// (MetaTexture, float[N] coords) -> float4
    TexelFetch,             /// (MetaTexture, int[N], int) -> float4

    ImageStore, /// (MetaImage, float[N] coords) -> void

    Branch,         /// (uint branch_target) -> void
    BranchIndirect, /// (uint branch_target) -> void
    PushFlowStack,  /// (uint branch_target) -> void
    PopFlowStack,   /// () -> void
    Exit,           /// () -> void
    Discard,        /// () -> void

    EmitVertex,   /// () -> void
    EndPrimitive, /// () -> void

    YNegate,            /// () -> float
    LocalInvocationIdX, /// () -> uint
    LocalInvocationIdY, /// () -> uint
    LocalInvocationIdZ, /// () -> uint
    WorkGroupIdX,       /// () -> uint
    WorkGroupIdY,       /// () -> uint
    WorkGroupIdZ,       /// () -> uint

    BallotThread, /// (bool) -> uint
    VoteAll,      /// (bool) -> bool
    VoteAny,      /// (bool) -> bool
    VoteEqual,    /// (bool) -> bool

    Amount,
};

enum class InternalFlag {
    Zero = 0,
    Sign = 1,
    Carry = 2,
    Overflow = 3,
    Amount = 4,
};

enum class MetaStackClass {
    Ssy,
    Pbk,
};

class OperationNode;
class ConditionalNode;
class GprNode;
class ImmediateNode;
class InternalFlagNode;
class PredicateNode;
class AbufNode;
class CbufNode;
class LmemNode;
class GmemNode;
class CommentNode;

using NodeData =
    std::variant<OperationNode, ConditionalNode, GprNode, ImmediateNode, InternalFlagNode,
                 PredicateNode, AbufNode, CbufNode, LmemNode, GmemNode, CommentNode>;
using Node = std::shared_ptr<NodeData>;
using Node4 = std::array<Node, 4>;
using NodeBlock = std::vector<Node>;

class Sampler {
public:
    /// This constructor is for bound samplers
    explicit Sampler(std::size_t offset, std::size_t index, Tegra::Shader::TextureType type,
                     bool is_array, bool is_shadow)
        : offset{offset}, index{index}, type{type}, is_array{is_array}, is_shadow{is_shadow},
          is_bindless{false} {}

    /// This constructor is for bindless samplers
    explicit Sampler(u32 cbuf_index, u32 cbuf_offset, std::size_t index,
                     Tegra::Shader::TextureType type, bool is_array, bool is_shadow)
        : offset{(static_cast<u64>(cbuf_index) << 32) | cbuf_offset}, index{index}, type{type},
          is_array{is_array}, is_shadow{is_shadow}, is_bindless{true} {}

    /// This constructor is for serialization/deserialization
    explicit Sampler(std::size_t offset, std::size_t index, Tegra::Shader::TextureType type,
                     bool is_array, bool is_shadow, bool is_bindless)
        : offset{offset}, index{index}, type{type}, is_array{is_array}, is_shadow{is_shadow},
          is_bindless{is_bindless} {}

    std::size_t GetOffset() const {
        return offset;
    }

    std::size_t GetIndex() const {
        return index;
    }

    Tegra::Shader::TextureType GetType() const {
        return type;
    }

    bool IsArray() const {
        return is_array;
    }

    bool IsShadow() const {
        return is_shadow;
    }

    bool IsBindless() const {
        return is_bindless;
    }

    std::pair<u32, u32> GetBindlessCBuf() const {
        return {static_cast<u32>(offset >> 32), static_cast<u32>(offset)};
    }

    bool operator<(const Sampler& rhs) const {
        return std::tie(index, offset, type, is_array, is_shadow, is_bindless) <
               std::tie(rhs.index, rhs.offset, rhs.type, rhs.is_array, rhs.is_shadow,
                        rhs.is_bindless);
    }

private:
    /// Offset in TSC memory from which to read the sampler object, as specified by the sampling
    /// instruction.
    std::size_t offset{};
    std::size_t index{}; ///< Value used to index into the generated GLSL sampler array.
    Tegra::Shader::TextureType type{}; ///< The type used to sample this texture (Texture2D, etc)
    bool is_array{};    ///< Whether the texture is being sampled as an array texture or not.
    bool is_shadow{};   ///< Whether the texture is being sampled as a depth texture or not.
    bool is_bindless{}; ///< Whether this sampler belongs to a bindless texture or not.
};

class Image {
public:
    explicit Image(std::size_t offset, std::size_t index, Tegra::Shader::ImageType type)
        : offset{offset}, index{index}, type{type}, is_bindless{false} {}

    explicit Image(u32 cbuf_index, u32 cbuf_offset, std::size_t index,
                   Tegra::Shader::ImageType type)
        : offset{(static_cast<u64>(cbuf_index) << 32) | cbuf_offset}, index{index}, type{type},
          is_bindless{true} {}

    explicit Image(std::size_t offset, std::size_t index, Tegra::Shader::ImageType type,
                   bool is_bindless)
        : offset{offset}, index{index}, type{type}, is_bindless{is_bindless} {}

    std::size_t GetOffset() const {
        return offset;
    }

    std::size_t GetIndex() const {
        return index;
    }

    Tegra::Shader::ImageType GetType() const {
        return type;
    }

    bool IsBindless() const {
        return is_bindless;
    }

    std::pair<u32, u32> GetBindlessCBuf() const {
        return {static_cast<u32>(offset >> 32), static_cast<u32>(offset)};
    }

    bool operator<(const Image& rhs) const {
        return std::tie(offset, index, type, is_bindless) <
               std::tie(rhs.offset, rhs.index, rhs.type, rhs.is_bindless);
    }

private:
    std::size_t offset{};
    std::size_t index{};
    Tegra::Shader::ImageType type{};
    bool is_bindless{};
};

struct GlobalMemoryBase {
    u32 cbuf_index{};
    u32 cbuf_offset{};

    bool operator<(const GlobalMemoryBase& rhs) const {
        return std::tie(cbuf_index, cbuf_offset) < std::tie(rhs.cbuf_index, rhs.cbuf_offset);
    }
};

/// Parameters describing an arithmetic operation
struct MetaArithmetic {
    bool precise{}; ///< Whether the operation can be constraint or not
};

/// Parameters describing a texture sampler
struct MetaTexture {
    const Sampler& sampler;
    Node array;
    Node depth_compare;
    std::vector<Node> aoffi;
    Node bias;
    Node lod;
    Node component{};
    u32 element{};
};

struct MetaImage {
    const Image& image;
    std::vector<Node> values;
};

/// Parameters that modify an operation but are not part of any particular operand
using Meta =
    std::variant<MetaArithmetic, MetaTexture, MetaImage, MetaStackClass, Tegra::Shader::HalfType>;

/// Holds any kind of operation that can be done in the IR
class OperationNode final {
public:
    explicit OperationNode(OperationCode code) : OperationNode(code, Meta{}) {}

    explicit OperationNode(OperationCode code, Meta meta)
        : OperationNode(code, meta, std::vector<Node>{}) {}

    explicit OperationNode(OperationCode code, std::vector<Node> operands)
        : OperationNode(code, Meta{}, std::move(operands)) {}

    explicit OperationNode(OperationCode code, Meta meta, std::vector<Node> operands)
        : code{code}, meta{std::move(meta)}, operands{std::move(operands)} {}

    template <typename... Args>
    explicit OperationNode(OperationCode code, Meta meta, Args&&... operands)
        : code{code}, meta{std::move(meta)}, operands{operands...} {}

    OperationCode GetCode() const {
        return code;
    }

    const Meta& GetMeta() const {
        return meta;
    }

    std::size_t GetOperandsCount() const {
        return operands.size();
    }

    const Node& operator[](std::size_t operand_index) const {
        return operands.at(operand_index);
    }

private:
    OperationCode code{};
    Meta meta{};
    std::vector<Node> operands;
};

/// Encloses inside any kind of node that returns a boolean conditionally-executed code
class ConditionalNode final {
public:
    explicit ConditionalNode(Node condition, std::vector<Node>&& code)
        : condition{std::move(condition)}, code{std::move(code)} {}

    const Node& GetCondition() const {
        return condition;
    }

    const std::vector<Node>& GetCode() const {
        return code;
    }

private:
    Node condition;         ///< Condition to be satisfied
    std::vector<Node> code; ///< Code to execute
};

/// A general purpose register
class GprNode final {
public:
    explicit constexpr GprNode(Tegra::Shader::Register index) : index{index} {}

    u32 GetIndex() const {
        return static_cast<u32>(index);
    }

private:
    Tegra::Shader::Register index{};
};

/// A 32-bits value that represents an immediate value
class ImmediateNode final {
public:
    explicit constexpr ImmediateNode(u32 value) : value{value} {}

    u32 GetValue() const {
        return value;
    }

private:
    u32 value{};
};

/// One of Maxwell's internal flags
class InternalFlagNode final {
public:
    explicit constexpr InternalFlagNode(InternalFlag flag) : flag{flag} {}

    InternalFlag GetFlag() const {
        return flag;
    }

private:
    InternalFlag flag{};
};

/// A predicate register, it can be negated without additional nodes
class PredicateNode final {
public:
    explicit constexpr PredicateNode(Tegra::Shader::Pred index, bool negated)
        : index{index}, negated{negated} {}

    Tegra::Shader::Pred GetIndex() const {
        return index;
    }

    bool IsNegated() const {
        return negated;
    }

private:
    Tegra::Shader::Pred index{};
    bool negated{};
};

/// Attribute buffer memory (known as attributes or varyings in GLSL terms)
class AbufNode final {
public:
    // Initialize for standard attributes (index is explicit).
    explicit AbufNode(Tegra::Shader::Attribute::Index index, u32 element, Node buffer = {})
        : buffer{std::move(buffer)}, index{index}, element{element} {}

    // Initialize for physical attributes (index is a variable value).
    explicit AbufNode(Node physical_address, Node buffer = {})
        : physical_address{std::move(physical_address)}, buffer{std::move(buffer)} {}

    Tegra::Shader::Attribute::Index GetIndex() const {
        return index;
    }

    u32 GetElement() const {
        return element;
    }

    const Node& GetBuffer() const {
        return buffer;
    }

    bool IsPhysicalBuffer() const {
        return static_cast<bool>(physical_address);
    }

    const Node& GetPhysicalAddress() const {
        return physical_address;
    }

private:
    Node physical_address;
    Node buffer;
    Tegra::Shader::Attribute::Index index{};
    u32 element{};
};

/// Constant buffer node, usually mapped to uniform buffers in GLSL
class CbufNode final {
public:
    explicit CbufNode(u32 index, Node offset) : index{index}, offset{std::move(offset)} {}

    u32 GetIndex() const {
        return index;
    }

    const Node& GetOffset() const {
        return offset;
    }

private:
    u32 index{};
    Node offset;
};

/// Local memory node
class LmemNode final {
public:
    explicit LmemNode(Node address) : address{std::move(address)} {}

    const Node& GetAddress() const {
        return address;
    }

private:
    Node address;
};

/// Global memory node
class GmemNode final {
public:
    explicit GmemNode(Node real_address, Node base_address, const GlobalMemoryBase& descriptor)
        : real_address{std::move(real_address)}, base_address{std::move(base_address)},
          descriptor{descriptor} {}

    const Node& GetRealAddress() const {
        return real_address;
    }

    const Node& GetBaseAddress() const {
        return base_address;
    }

    const GlobalMemoryBase& GetDescriptor() const {
        return descriptor;
    }

private:
    Node real_address;
    Node base_address;
    GlobalMemoryBase descriptor;
};

/// Commentary, can be dropped
class CommentNode final {
public:
    explicit CommentNode(std::string text) : text{std::move(text)} {}

    const std::string& GetText() const {
        return text;
    }

private:
    std::string text;
};

} // namespace VideoCommon::Shader