-
Notifications
You must be signed in to change notification settings - Fork 10.7k
/
BufferizableOpInterface.cpp
666 lines (574 loc) · 25.8 KB
/
BufferizableOpInterface.cpp
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
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
//===- BufferizableOpInterface.cpp - Bufferizable Ops ---=----------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
#include "mlir/Dialect/Bufferization/IR/BufferizableOpInterface.h"
#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
#include "mlir/Dialect/Func/IR/FuncOps.h"
#include "mlir/Dialect/MemRef/IR/MemRef.h"
#include "mlir/Dialect/Tensor/IR/Tensor.h"
#include "mlir/IR/AsmState.h"
#include "mlir/IR/BlockAndValueMapping.h"
#include "mlir/IR/BuiltinOps.h"
#include "mlir/IR/Operation.h"
#include "mlir/IR/TypeUtilities.h"
#include "mlir/IR/Value.h"
#include "llvm/Support/Debug.h"
//===----------------------------------------------------------------------===//
// BufferizableOpInterface
//===----------------------------------------------------------------------===//
namespace mlir {
namespace bufferization {
#include "mlir/Dialect/Bufferization/IR/BufferizableOpInterface.cpp.inc"
} // namespace bufferization
} // namespace mlir
#define DEBUG_TYPE "bufferizable-op-interface"
#define DBGS() (llvm::dbgs() << '[' << DEBUG_TYPE << "] ")
#define LDBG(X) LLVM_DEBUG(DBGS() << (X))
using namespace mlir;
using namespace bufferization;
/// Attribute name used to mark region arguments that can be bufferized
/// in-place during linalg comprehensive bufferization.
constexpr const ::llvm::StringLiteral
bufferization::BufferizableOpInterface::kInplaceableAttrName;
/// Create an AllocTensorOp for the given shaped value. If `copy` is set, the
/// shaped value is copied. Otherwise, a tensor with undefined contents is
/// allocated.
Value bufferization::allocateTensorForShapedValue(OpBuilder &b, Location loc,
Value shapedValue,
bool escape, bool copy) {
Value tensor;
if (shapedValue.getType().isa<RankedTensorType>()) {
tensor = shapedValue;
} else if (shapedValue.getType().isa<MemRefType>()) {
tensor = b.create<ToTensorOp>(loc, shapedValue);
} else {
llvm_unreachable("expected RankedTensorType or MemRefType");
}
RankedTensorType tensorType = tensor.getType().cast<RankedTensorType>();
SmallVector<Value> dynamicSizes;
if (!copy) {
// Compute the dynamic part of the shape.
// First try to query the shape via ReifyRankedShapedTypeOpInterface.
bool reifiedShapes = false;
if (shapedValue.getType().isa<RankedTensorType>() &&
shapedValue.isa<OpResult>()) {
if (auto rankedOp = dyn_cast_or_null<ReifyRankedShapedTypeOpInterface>(
shapedValue.getDefiningOp())) {
ReifiedRankedShapedTypeDims resultDims;
if (succeeded(rankedOp.reifyResultShapes(b, resultDims))) {
reifiedShapes = true;
auto &shape =
resultDims[shapedValue.cast<OpResult>().getResultNumber()];
for (const auto &dim : enumerate(tensorType.getShape()))
if (ShapedType::isDynamic(dim.value()))
dynamicSizes.push_back(shape[dim.index()]);
}
}
}
// If the shape could not be reified, create DimOps.
if (!reifiedShapes)
populateDynamicDimSizes(b, loc, tensor, dynamicSizes);
}
auto allocTensorOp = b.create<AllocTensorOp>(loc, tensorType, dynamicSizes,
copy ? tensor : Value());
allocTensorOp->setAttr(BufferizationDialect::kEscapeAttrName,
b.getBoolArrayAttr({escape}));
return allocTensorOp;
}
LogicalResult BufferizableOpInterface::resolveTensorOpOperandConflicts(
RewriterBase &rewriter, const AnalysisState &state) {
OpBuilder::InsertionGuard g(rewriter);
Operation *op = getOperation();
SmallVector<OpOperand *> outOfPlaceOpOperands;
DenseSet<OpOperand *> copiedOpOperands;
DenseSet<OpOperand *> escapingOpOperandCopies;
SmallVector<OpResult> outOfPlaceOpResults;
DenseSet<OpResult> copiedOpResults;
DenseSet<OpResult> escapingOpResultCopies;
// Find all out-of-place OpOperands.
for (OpOperand &opOperand : op->getOpOperands()) {
Type operandType = opOperand.get().getType();
if (!operandType.isa<TensorType>())
continue;
if (state.isInPlace(opOperand))
continue;
if (operandType.isa<UnrankedTensorType>())
return op->emitError("copies of unranked tensors are not supported");
SmallVector<OpResult> aliasingOpResults =
state.getAliasingOpResult(opOperand);
// Is the result yielded from a block? Or are deallocations turned off
// entirely? In either case, mark the allocation as "escaping", so that it
// will not be deallocated.
bool escape = !state.getOptions().createDeallocs ||
llvm::any_of(aliasingOpResults, [&](Value v) {
return state.isTensorYielded(v);
});
if (aliasingOpResults.size() == 1 &&
!state.bufferizesToMemoryWrite(opOperand) &&
state.getAliasingOpOperand(aliasingOpResults.front()).size() == 1) {
// The op itself does not write but may create exactly one alias. Instead
// of copying the OpOperand, copy the OpResult. The OpResult can sometimes
// be smaller than the OpOperand (e.g., in the case of an extract_slice,
// where the result is usually a smaller part of the source).
outOfPlaceOpResults.push_back(aliasingOpResults.front());
if (!state.canOmitTensorCopy(opOperand))
copiedOpResults.insert(aliasingOpResults.front());
if (escape)
escapingOpResultCopies.insert(aliasingOpResults.front());
} else {
// In all other cases, make a copy of the OpOperand.
outOfPlaceOpOperands.push_back(&opOperand);
if (!state.canOmitTensorCopy(opOperand))
copiedOpOperands.insert(&opOperand);
if (escape)
escapingOpOperandCopies.insert(&opOperand);
}
}
// Insert copies of OpOperands.
rewriter.setInsertionPoint(op);
for (OpOperand *opOperand : outOfPlaceOpOperands) {
Value copy = allocateTensorForShapedValue(
rewriter, op->getLoc(), opOperand->get(),
escapingOpOperandCopies.contains(opOperand),
copiedOpOperands.contains(opOperand));
rewriter.updateRootInPlace(op, [&]() { opOperand->set(copy); });
}
// Insert copies of OpResults.
rewriter.setInsertionPointAfter(op);
for (OpResult opResult : outOfPlaceOpResults) {
Value copy =
allocateTensorForShapedValue(rewriter, op->getLoc(), opResult,
escapingOpResultCopies.contains(opResult),
copiedOpResults.count(opResult));
SmallVector<OpOperand *> uses = llvm::to_vector(llvm::map_range(
opResult.getUses(), [](OpOperand &use) { return &use; }));
for (OpOperand *use : uses) {
// Do not update the alloc_tensor op that we just created.
if (use->getOwner() != copy.getDefiningOp())
rewriter.updateRootInPlace(use->getOwner(), [&]() { use->set(copy); });
}
}
return success();
}
//===----------------------------------------------------------------------===//
// OpFilter
//===----------------------------------------------------------------------===//
bool OpFilter::isOpAllowed(Operation *op) const {
// All other ops: Allow/disallow according to filter.
bool isAllowed = !hasAllowRule();
for (const Entry &entry : entries) {
bool filterResult = entry.fn(op);
switch (entry.type) {
case Entry::ALLOW:
isAllowed |= filterResult;
break;
case Entry::DENY:
if (filterResult)
// DENY filter matches. This op is no allowed. (Even if other ALLOW
// filters may match.)
return false;
};
}
return isAllowed;
}
//===----------------------------------------------------------------------===//
// BufferizationOptions
//===----------------------------------------------------------------------===//
// Default constructor for BufferizationOptions.
BufferizationOptions::BufferizationOptions() = default;
bool BufferizationOptions::isOpAllowed(Operation *op) const {
// Special case: If function boundary bufferization is deactivated, do not
// allow ops that belong to the `func` dialect.
bool isFuncBoundaryOp = isa_and_nonnull<func::FuncDialect>(op->getDialect());
if (!bufferizeFunctionBoundaries && isFuncBoundaryOp)
return false;
return opFilter.isOpAllowed(op);
}
BufferizableOpInterface
BufferizationOptions::dynCastBufferizableOp(Operation *op) const {
auto bufferizableOp = dyn_cast<BufferizableOpInterface>(op);
if (!bufferizableOp)
return nullptr;
if (!isOpAllowed(op))
return nullptr;
return bufferizableOp;
}
BufferizableOpInterface
BufferizationOptions::dynCastBufferizableOp(Value value) const {
if (auto bufferizableOp = value.getDefiningOp<BufferizableOpInterface>())
if (isOpAllowed(bufferizableOp.getOperation()))
return bufferizableOp;
return nullptr;
}
void BufferizationOptions::addDialectStateInitializer(
StringRef name, const DialectStateInitFn &fn) {
stateInitializers.push_back(
[=](AnalysisState &state) { state.insertDialectState(name, fn()); });
}
//===----------------------------------------------------------------------===//
// Helper functions for BufferizableOpInterface
//===----------------------------------------------------------------------===//
static void setInsertionPointAfter(OpBuilder &b, Value value) {
if (auto bbArg = value.dyn_cast<BlockArgument>()) {
b.setInsertionPointToStart(bbArg.getOwner());
} else {
b.setInsertionPointAfter(value.getDefiningOp());
}
}
/// Determine which OpOperand* will alias with `result` if the op is bufferized
/// in place. Return an empty vector if the op is not bufferizable.
SmallVector<OpOperand *>
AnalysisState::getAliasingOpOperand(OpResult result) const {
if (Operation *op = result.getDefiningOp())
if (auto bufferizableOp = getOptions().dynCastBufferizableOp(op))
return bufferizableOp.getAliasingOpOperand(result, *this);
return {};
}
/// Determine which OpResult will alias with `opOperand` if the op is bufferized
/// in place. Return an empty vector if the op is not bufferizable.
SmallVector<OpResult>
AnalysisState::getAliasingOpResult(OpOperand &opOperand) const {
if (auto bufferizableOp =
getOptions().dynCastBufferizableOp(opOperand.getOwner()))
return bufferizableOp.getAliasingOpResult(opOperand, *this);
return {};
}
/// Return true if `opOperand` bufferizes to a memory read. Return `true` if the
/// op is not bufferizable.
bool AnalysisState::bufferizesToMemoryRead(OpOperand &opOperand) const {
if (auto bufferizableOp =
getOptions().dynCastBufferizableOp(opOperand.getOwner()))
return bufferizableOp.bufferizesToMemoryRead(opOperand, *this);
// Unknown op that returns a tensor. The inplace analysis does not support it.
// Conservatively return true.
return true;
}
/// Return true if `opOperand` bufferizes to a memory write. Return
/// `true` if the op is not bufferizable.
bool AnalysisState::bufferizesToMemoryWrite(OpOperand &opOperand) const {
if (auto bufferizableOp =
getOptions().dynCastBufferizableOp(opOperand.getOwner()))
return bufferizableOp.bufferizesToMemoryWrite(opOperand, *this);
// Unknown op that returns a tensor. The inplace analysis does not support it.
// Conservatively return true.
return true;
}
/// Return true if `opOperand` does neither read nor write but bufferizes to an
/// alias. Return false if the op is not bufferizable.
bool AnalysisState::bufferizesToAliasOnly(OpOperand &opOperand) const {
if (auto bufferizableOp =
getOptions().dynCastBufferizableOp(opOperand.getOwner()))
return bufferizableOp.bufferizesToAliasOnly(opOperand, *this);
// Unknown op that returns a tensor. The inplace analysis does not support it.
// Conservatively return false.
return false;
}
/// Return true if the given value is read by an op that bufferizes to a memory
/// read. Also takes into account ops that create an alias but do not read by
/// themselves (e.g., ExtractSliceOp).
bool AnalysisState::isValueRead(Value value) const {
assert(value.getType().isa<TensorType>() && "expected TensorType");
SmallVector<OpOperand *> workingSet;
for (OpOperand &use : value.getUses())
workingSet.push_back(&use);
while (!workingSet.empty()) {
OpOperand *uMaybeReading = workingSet.pop_back_val();
// Skip over all ops that neither read nor write (but create an alias).
if (bufferizesToAliasOnly(*uMaybeReading))
for (OpResult opResult : getAliasingOpResult(*uMaybeReading))
for (OpOperand &use : opResult.getUses())
workingSet.push_back(&use);
if (bufferizesToMemoryRead(*uMaybeReading))
return true;
}
return false;
}
// Starting from `value`, follow the use-def chain in reverse, always selecting
// the aliasing OpOperands. Find and return Values for which `condition`
// evaluates to true. OpOperands of such matching Values are not traversed any
// further.
llvm::SetVector<Value> AnalysisState::findValueInReverseUseDefChain(
Value value, llvm::function_ref<bool(Value)> condition) const {
llvm::SetVector<Value> result, workingSet;
workingSet.insert(value);
while (!workingSet.empty()) {
Value value = workingSet.pop_back_val();
if (condition(value) || value.isa<BlockArgument>()) {
result.insert(value);
continue;
}
OpResult opResult = value.cast<OpResult>();
SmallVector<OpOperand *> opOperands = getAliasingOpOperand(opResult);
if (opOperands.empty() || !options.isOpAllowed(value.getDefiningOp())) {
result.insert(value);
continue;
}
for (OpOperand *o : opOperands)
workingSet.insert(o->get());
}
return result;
}
// Find the Values of the last preceding write of a given Value.
llvm::SetVector<Value>
AnalysisState::findLastPrecedingWrite(Value value) const {
return findValueInReverseUseDefChain(value, [&](Value value) {
Operation *op = value.getDefiningOp();
if (!op)
return true;
auto bufferizableOp = options.dynCastBufferizableOp(op);
if (!bufferizableOp)
return true;
return bufferizableOp.isMemoryWrite(value.cast<OpResult>(), *this);
});
}
AnalysisState::AnalysisState(const BufferizationOptions &options)
: options(options) {
for (const BufferizationOptions::AnalysisStateInitFn &fn :
options.stateInitializers)
fn(*this);
}
bool AnalysisState::canOmitTensorCopy(OpOperand &opOperand) const {
// Do not copy if the tensor has undefined contents.
if (hasUndefinedContents(&opOperand))
return true;
// Do not copy if the buffer of the tensor is entirely overwritten (with
// values that do not depend on the old tensor).
if (bufferizesToMemoryWrite(opOperand) && !bufferizesToMemoryRead(opOperand))
return true;
// Do not copy if the tensor is never read.
SmallVector<OpResult> aliasingOpResults = getAliasingOpResult(opOperand);
if (!bufferizesToMemoryRead(opOperand) &&
llvm::none_of(aliasingOpResults,
[&](OpResult opResult) { return isValueRead(opResult); }))
return true;
// Default: Cannot omit the copy.
return false;
}
bool AnalysisState::isInPlace(OpOperand &opOperand) const {
// ToMemrefOps are always in-place.
if (isa<ToMemrefOp>(opOperand.getOwner()))
return true;
// In the absence of analysis information, OpOperands that bufferize to a
// memory write are out-of-place, i.e., an alloc and copy is inserted.
return !bufferizesToMemoryWrite(opOperand);
}
bool AnalysisState::areEquivalentBufferizedValues(Value v1, Value v2) const {
// In the absence of analysis information, we do not know if the values are
// equivalent. The conservative answer is "false".
return false;
}
bool AnalysisState::areAliasingBufferizedValues(Value v1, Value v2) const {
// In the absence of analysis information, we do not know if the values may be
// aliasing. The conservative answer is "true".
return true;
}
bool AnalysisState::hasUndefinedContents(OpOperand *opOperand) const {
// In the absence of analysis information, the conservative answer is "false".
return false;
}
bool AnalysisState::isTensorYielded(Value tensor) const {
// In the absence of analysis information, the conservative answer is "true".
if (!tensor.getDefiningOp<AllocTensorOp>())
return true;
// For AllocTensorOp results, we can do better: They do not alias with any
// preceding value, so we can follow SSA use-def chains and do a simple
// analysis.
SmallVector<OpOperand *> worklist;
for (OpOperand &use : tensor.getUses())
worklist.push_back(&use);
while (!worklist.empty()) {
OpOperand *operand = worklist.pop_back_val();
Operation *op = operand->getOwner();
// If the op is not bufferizable, we can safely assume that the value is not
// yielded. (When bufferizing that op, it must handle such cases.)
if (!options.dynCastBufferizableOp(op))
continue;
// We cannot analyze through ToMemrefOps, so we have to conservatively
// assume that the value is yielded.
if (isa<ToMemrefOp>(op))
return true;
// Check if the op is returning/yielding.
if (isRegionReturnLike(op))
return true;
// Add all aliasing OpResults to the worklist.
// Note: In the absence of detailed analysis information (e.g., there may be
// no function call analysis information), this `getAliasingOpResult` is
// conservative and may report additional OpResults as potentially aliasing.
for (OpResult opResult : getAliasingOpResult(*operand))
for (OpOperand &use : opResult.getUses())
worklist.push_back(&use);
}
// No ReturnLike op found: The value is not yielded.
return false;
}
// bufferization.to_memref is not allowed to change the rank.
static void ensureToMemrefOpIsValid(Value tensor, Type memrefType) {
#ifndef NDEBUG
auto rankedTensorType = tensor.getType().dyn_cast<RankedTensorType>();
assert((!rankedTensorType || memrefType.cast<MemRefType>().getRank() ==
rankedTensorType.getRank()) &&
"to_memref would be invalid: mismatching ranks");
#endif
}
Value bufferization::getBuffer(RewriterBase &rewriter, Value value,
const BufferizationOptions &options) {
auto tensorType = value.getType().dyn_cast<TensorType>();
assert(tensorType && "unexpected non-tensor type");
// Replace "%t = to_tensor %m" with %m.
if (auto toTensorOp = value.getDefiningOp<bufferization::ToTensorOp>())
return toTensorOp.getMemref();
// Insert to_memref op.
OpBuilder::InsertionGuard g(rewriter);
setInsertionPointAfter(rewriter, value);
Type memrefType = getMemRefType(tensorType, options);
ensureToMemrefOpIsValid(value, memrefType);
return rewriter.create<bufferization::ToMemrefOp>(value.getLoc(), memrefType,
value);
}
/// Return the buffer type for a given Value (tensor) after bufferization.
BaseMemRefType
bufferization::getBufferType(Value value, const BufferizationOptions &options) {
auto tensorType = value.getType().dyn_cast<TensorType>();
assert(tensorType && "unexpected non-tensor type");
if (auto toTensorOp = value.getDefiningOp<bufferization::ToTensorOp>())
return toTensorOp.getMemref().getType().cast<BaseMemRefType>();
return getMemRefType(tensorType, options);
}
void bufferization::replaceOpWithBufferizedValues(RewriterBase &rewriter,
Operation *op,
ValueRange values) {
assert(values.size() == op->getNumResults() &&
"expected one value per OpResult");
OpBuilder::InsertionGuard g(rewriter);
// Replace all OpResults with the given values.
SmallVector<Value> replacements;
for (OpResult opResult : op->getOpResults()) {
Value replacement = values[opResult.getResultNumber()];
if (opResult.getType().isa<TensorType>()) {
// The OpResult is a tensor. Such values are replaced with memrefs during
// bufferization.
assert((replacement.getType().isa<MemRefType>() ||
replacement.getType().isa<UnrankedMemRefType>()) &&
"tensor op result should be replaced with a memref value");
// The existing uses of the OpResult still expect a tensor. Insert a
// ToTensorOp. Throughout bufferization, this ToTensorOp will gradually
// loose all of its users and eventually DCE away.
rewriter.setInsertionPointAfter(op);
replacement = rewriter.create<bufferization::ToTensorOp>(
replacement.getLoc(), replacement);
}
replacements.push_back(replacement);
}
rewriter.replaceOp(op, replacements);
}
//===----------------------------------------------------------------------===//
// Bufferization-specific scoped alloc/dealloc insertion support.
//===----------------------------------------------------------------------===//
/// Create a memref allocation with the given type and dynamic extents.
FailureOr<Value> BufferizationOptions::createAlloc(OpBuilder &b, Location loc,
MemRefType type,
ValueRange dynShape) const {
if (allocationFn)
return (*allocationFn)(b, loc, type, dynShape, bufferAlignment);
// Default bufferallocation via AllocOp.
if (bufferAlignment != 0)
return b
.create<memref::AllocOp>(loc, type, dynShape,
b.getI64IntegerAttr(bufferAlignment))
.getResult();
return b.create<memref::AllocOp>(loc, type, dynShape).getResult();
}
/// Creates a memref deallocation. The given memref buffer must have been
/// allocated using `createAlloc`.
LogicalResult BufferizationOptions::createDealloc(OpBuilder &b, Location loc,
Value allocatedBuffer) const {
if (deallocationFn)
return (*deallocationFn)(b, loc, allocatedBuffer);
// Default buffer deallocation via DeallocOp.
b.create<memref::DeallocOp>(loc, allocatedBuffer);
return success();
}
/// Create a memory copy between two memref buffers.
LogicalResult BufferizationOptions::createMemCpy(OpBuilder &b, Location loc,
Value from, Value to) const {
if (memCpyFn)
return (*memCpyFn)(b, loc, from, to);
b.create<memref::CopyOp>(loc, from, to);
return success();
}
//===----------------------------------------------------------------------===//
// Bufferization-specific BlockAndValueMapping support with debugging.
//===----------------------------------------------------------------------===//
bool bufferization::isFunctionArgument(Value value) {
auto bbArg = value.dyn_cast<BlockArgument>();
if (!bbArg)
return false;
return isa<func::FuncOp>(bbArg.getOwner()->getParentOp());
}
BaseMemRefType bufferization::getMemRefType(TensorType tensorType,
const BufferizationOptions &options,
MemRefLayoutAttrInterface layout,
Attribute memorySpace) {
// Case 1: Unranked memref type.
if (auto unrankedTensorType = tensorType.dyn_cast<UnrankedTensorType>()) {
assert(!layout && "UnrankedTensorType cannot have a layout map");
return UnrankedMemRefType::get(unrankedTensorType.getElementType(),
memorySpace);
}
// Case 2: Ranked memref type with specified layout.
auto rankedTensorType = tensorType.cast<RankedTensorType>();
if (layout) {
return MemRefType::get(rankedTensorType.getShape(),
rankedTensorType.getElementType(), layout,
memorySpace);
}
// Case 3: Configured with "fully dynamic layout maps".
if (options.unknownTypeConversion ==
BufferizationOptions::LayoutMapOption::FullyDynamicLayoutMap)
return getMemRefTypeWithFullyDynamicLayout(tensorType, memorySpace);
// Case 4: Configured with "static identity layout maps".
if (options.unknownTypeConversion ==
BufferizationOptions::LayoutMapOption::IdentityLayoutMap)
return getMemRefTypeWithStaticIdentityLayout(tensorType, memorySpace);
llvm_unreachable("InferLayoutMap is an invalid option");
}
BaseMemRefType
bufferization::getMemRefTypeWithFullyDynamicLayout(TensorType tensorType,
Attribute memorySpace) {
// Case 1: Unranked memref type.
if (auto unrankedTensorType = tensorType.dyn_cast<UnrankedTensorType>()) {
return UnrankedMemRefType::get(unrankedTensorType.getElementType(),
memorySpace);
}
// Case 2: Ranked memref type.
auto rankedTensorType = tensorType.cast<RankedTensorType>();
int64_t dynamicOffset = ShapedType::kDynamicStrideOrOffset;
SmallVector<int64_t> dynamicStrides(rankedTensorType.getRank(),
ShapedType::kDynamicStrideOrOffset);
AffineMap stridedLayout = makeStridedLinearLayoutMap(
dynamicStrides, dynamicOffset, rankedTensorType.getContext());
return MemRefType::get(rankedTensorType.getShape(),
rankedTensorType.getElementType(), stridedLayout,
memorySpace);
}
/// Return a MemRef type with a static identity layout (i.e., no layout map). If
/// the given tensor type is unranked, return an unranked MemRef type.
BaseMemRefType
bufferization::getMemRefTypeWithStaticIdentityLayout(TensorType tensorType,
Attribute memorySpace) {
// Case 1: Unranked memref type.
if (auto unrankedTensorType = tensorType.dyn_cast<UnrankedTensorType>()) {
return UnrankedMemRefType::get(unrankedTensorType.getElementType(),
memorySpace);
}
// Case 2: Ranked memref type.
auto rankedTensorType = tensorType.cast<RankedTensorType>();
MemRefLayoutAttrInterface layout = {};
return MemRefType::get(rankedTensorType.getShape(),
rankedTensorType.getElementType(), layout,
memorySpace);
}