RISCVISelDAGToDAG.cpp
22 KB
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
//===-- RISCVISelDAGToDAG.cpp - A dag to dag inst selector for RISCV ------===//
//
// 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
//
//===----------------------------------------------------------------------===//
//
// This file defines an instruction selector for the RISCV target.
//
//===----------------------------------------------------------------------===//
#include "RISCVISelDAGToDAG.h"
#include "MCTargetDesc/RISCVMCTargetDesc.h"
#include "Utils/RISCVMatInt.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/Support/Alignment.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
using namespace llvm;
#define DEBUG_TYPE "riscv-isel"
void RISCVDAGToDAGISel::PostprocessISelDAG() {
doPeepholeLoadStoreADDI();
}
static SDNode *selectImm(SelectionDAG *CurDAG, const SDLoc &DL, int64_t Imm,
MVT XLenVT) {
RISCVMatInt::InstSeq Seq;
RISCVMatInt::generateInstSeq(Imm, XLenVT == MVT::i64, Seq);
SDNode *Result = nullptr;
SDValue SrcReg = CurDAG->getRegister(RISCV::X0, XLenVT);
for (RISCVMatInt::Inst &Inst : Seq) {
SDValue SDImm = CurDAG->getTargetConstant(Inst.Imm, DL, XLenVT);
if (Inst.Opc == RISCV::LUI)
Result = CurDAG->getMachineNode(RISCV::LUI, DL, XLenVT, SDImm);
else
Result = CurDAG->getMachineNode(Inst.Opc, DL, XLenVT, SrcReg, SDImm);
// Only the first instruction has X0 as its source.
SrcReg = SDValue(Result, 0);
}
return Result;
}
// Returns true if the Node is an ISD::AND with a constant argument. If so,
// set Mask to that constant value.
static bool isConstantMask(SDNode *Node, uint64_t &Mask) {
if (Node->getOpcode() == ISD::AND &&
Node->getOperand(1).getOpcode() == ISD::Constant) {
Mask = cast<ConstantSDNode>(Node->getOperand(1))->getZExtValue();
return true;
}
return false;
}
void RISCVDAGToDAGISel::Select(SDNode *Node) {
// If we have a custom node, we have already selected.
if (Node->isMachineOpcode()) {
LLVM_DEBUG(dbgs() << "== "; Node->dump(CurDAG); dbgs() << "\n");
Node->setNodeId(-1);
return;
}
// Instruction Selection not handled by the auto-generated tablegen selection
// should be handled here.
unsigned Opcode = Node->getOpcode();
MVT XLenVT = Subtarget->getXLenVT();
SDLoc DL(Node);
EVT VT = Node->getValueType(0);
switch (Opcode) {
case ISD::ADD: {
// Optimize (add r, imm) to (addi (addi r, imm0) imm1) if applicable. The
// immediate must be in specific ranges and have a single use.
if (auto *ConstOp = dyn_cast<ConstantSDNode>(Node->getOperand(1))) {
if (!(ConstOp->hasOneUse()))
break;
// The imm must be in range [-4096,-2049] or [2048,4094].
int64_t Imm = ConstOp->getSExtValue();
if (!(-4096 <= Imm && Imm <= -2049) && !(2048 <= Imm && Imm <= 4094))
break;
// Break the imm to imm0+imm1.
SDLoc DL(Node);
EVT VT = Node->getValueType(0);
const SDValue ImmOp0 = CurDAG->getTargetConstant(Imm - Imm / 2, DL, VT);
const SDValue ImmOp1 = CurDAG->getTargetConstant(Imm / 2, DL, VT);
auto *NodeAddi0 = CurDAG->getMachineNode(RISCV::ADDI, DL, VT,
Node->getOperand(0), ImmOp0);
auto *NodeAddi1 = CurDAG->getMachineNode(RISCV::ADDI, DL, VT,
SDValue(NodeAddi0, 0), ImmOp1);
ReplaceNode(Node, NodeAddi1);
return;
}
break;
}
case ISD::Constant: {
auto ConstNode = cast<ConstantSDNode>(Node);
if (VT == XLenVT && ConstNode->isNullValue()) {
SDValue New = CurDAG->getCopyFromReg(CurDAG->getEntryNode(), SDLoc(Node),
RISCV::X0, XLenVT);
ReplaceNode(Node, New.getNode());
return;
}
int64_t Imm = ConstNode->getSExtValue();
if (XLenVT == MVT::i64) {
ReplaceNode(Node, selectImm(CurDAG, SDLoc(Node), Imm, XLenVT));
return;
}
break;
}
case ISD::FrameIndex: {
SDValue Imm = CurDAG->getTargetConstant(0, DL, XLenVT);
int FI = cast<FrameIndexSDNode>(Node)->getIndex();
SDValue TFI = CurDAG->getTargetFrameIndex(FI, VT);
ReplaceNode(Node, CurDAG->getMachineNode(RISCV::ADDI, DL, VT, TFI, Imm));
return;
}
case ISD::SRL: {
if (!Subtarget->is64Bit())
break;
SDValue Op0 = Node->getOperand(0);
SDValue Op1 = Node->getOperand(1);
uint64_t Mask;
// Match (srl (and val, mask), imm) where the result would be a
// zero-extended 32-bit integer. i.e. the mask is 0xffffffff or the result
// is equivalent to this (SimplifyDemandedBits may have removed lower bits
// from the mask that aren't necessary due to the right-shifting).
if (Op1.getOpcode() == ISD::Constant &&
isConstantMask(Op0.getNode(), Mask)) {
uint64_t ShAmt = cast<ConstantSDNode>(Op1.getNode())->getZExtValue();
if ((Mask | maskTrailingOnes<uint64_t>(ShAmt)) == 0xffffffff) {
SDValue ShAmtVal =
CurDAG->getTargetConstant(ShAmt, SDLoc(Node), XLenVT);
CurDAG->SelectNodeTo(Node, RISCV::SRLIW, XLenVT, Op0.getOperand(0),
ShAmtVal);
return;
}
}
break;
}
case RISCVISD::READ_CYCLE_WIDE:
assert(!Subtarget->is64Bit() && "READ_CYCLE_WIDE is only used on riscv32");
ReplaceNode(Node, CurDAG->getMachineNode(RISCV::ReadCycleWide, DL, MVT::i32,
MVT::i32, MVT::Other,
Node->getOperand(0)));
return;
}
// Select the default instruction.
SelectCode(Node);
}
bool RISCVDAGToDAGISel::SelectInlineAsmMemoryOperand(
const SDValue &Op, unsigned ConstraintID, std::vector<SDValue> &OutOps) {
switch (ConstraintID) {
case InlineAsm::Constraint_m:
// We just support simple memory operands that have a single address
// operand and need no special handling.
OutOps.push_back(Op);
return false;
case InlineAsm::Constraint_A:
OutOps.push_back(Op);
return false;
default:
break;
}
return true;
}
bool RISCVDAGToDAGISel::SelectAddrFI(SDValue Addr, SDValue &Base) {
if (auto FIN = dyn_cast<FrameIndexSDNode>(Addr)) {
Base = CurDAG->getTargetFrameIndex(FIN->getIndex(), Subtarget->getXLenVT());
return true;
}
return false;
}
// Check that it is a SLOI (Shift Left Ones Immediate). We first check that
// it is the right node tree:
//
// (OR (SHL RS1, VC2), VC1)
//
// and then we check that VC1, the mask used to fill with ones, is compatible
// with VC2, the shamt:
//
// VC1 == maskTrailingOnes<uint64_t>(VC2)
bool RISCVDAGToDAGISel::SelectSLOI(SDValue N, SDValue &RS1, SDValue &Shamt) {
MVT XLenVT = Subtarget->getXLenVT();
if (N.getOpcode() == ISD::OR) {
SDValue Or = N;
if (Or.getOperand(0).getOpcode() == ISD::SHL) {
SDValue Shl = Or.getOperand(0);
if (isa<ConstantSDNode>(Shl.getOperand(1)) &&
isa<ConstantSDNode>(Or.getOperand(1))) {
if (XLenVT == MVT::i64) {
uint64_t VC1 = Or.getConstantOperandVal(1);
uint64_t VC2 = Shl.getConstantOperandVal(1);
if (VC1 == maskTrailingOnes<uint64_t>(VC2)) {
RS1 = Shl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Shl.getOperand(1).getValueType());
return true;
}
}
if (XLenVT == MVT::i32) {
uint32_t VC1 = Or.getConstantOperandVal(1);
uint32_t VC2 = Shl.getConstantOperandVal(1);
if (VC1 == maskTrailingOnes<uint32_t>(VC2)) {
RS1 = Shl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Shl.getOperand(1).getValueType());
return true;
}
}
}
}
}
return false;
}
// Check that it is a SROI (Shift Right Ones Immediate). We first check that
// it is the right node tree:
//
// (OR (SRL RS1, VC2), VC1)
//
// and then we check that VC1, the mask used to fill with ones, is compatible
// with VC2, the shamt:
//
// VC1 == maskLeadingOnes<uint64_t>(VC2)
bool RISCVDAGToDAGISel::SelectSROI(SDValue N, SDValue &RS1, SDValue &Shamt) {
MVT XLenVT = Subtarget->getXLenVT();
if (N.getOpcode() == ISD::OR) {
SDValue Or = N;
if (Or.getOperand(0).getOpcode() == ISD::SRL) {
SDValue Srl = Or.getOperand(0);
if (isa<ConstantSDNode>(Srl.getOperand(1)) &&
isa<ConstantSDNode>(Or.getOperand(1))) {
if (XLenVT == MVT::i64) {
uint64_t VC1 = Or.getConstantOperandVal(1);
uint64_t VC2 = Srl.getConstantOperandVal(1);
if (VC1 == maskLeadingOnes<uint64_t>(VC2)) {
RS1 = Srl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Srl.getOperand(1).getValueType());
return true;
}
}
if (XLenVT == MVT::i32) {
uint32_t VC1 = Or.getConstantOperandVal(1);
uint32_t VC2 = Srl.getConstantOperandVal(1);
if (VC1 == maskLeadingOnes<uint32_t>(VC2)) {
RS1 = Srl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Srl.getOperand(1).getValueType());
return true;
}
}
}
}
}
return false;
}
// Check that it is a RORI (Rotate Right Immediate). We first check that
// it is the right node tree:
//
// (ROTL RS1, VC)
//
// The compiler translates immediate rotations to the right given by the call
// to the rotateright32/rotateright64 intrinsics as rotations to the left.
// Since the rotation to the left can be easily emulated as a rotation to the
// right by negating the constant, there is no encoding for ROLI.
// We then select the immediate left rotations as RORI by the complementary
// constant:
//
// Shamt == XLen - VC
bool RISCVDAGToDAGISel::SelectRORI(SDValue N, SDValue &RS1, SDValue &Shamt) {
MVT XLenVT = Subtarget->getXLenVT();
if (N.getOpcode() == ISD::ROTL) {
if (isa<ConstantSDNode>(N.getOperand(1))) {
if (XLenVT == MVT::i64) {
uint64_t VC = N.getConstantOperandVal(1);
Shamt = CurDAG->getTargetConstant((64 - VC), SDLoc(N),
N.getOperand(1).getValueType());
RS1 = N.getOperand(0);
return true;
}
if (XLenVT == MVT::i32) {
uint32_t VC = N.getConstantOperandVal(1);
Shamt = CurDAG->getTargetConstant((32 - VC), SDLoc(N),
N.getOperand(1).getValueType());
RS1 = N.getOperand(0);
return true;
}
}
}
return false;
}
// Check that it is a SLLIUW (Shift Logical Left Immediate Unsigned i32
// on RV64).
// SLLIUW is the same as SLLI except for the fact that it clears the bits
// XLEN-1:32 of the input RS1 before shifting.
// We first check that it is the right node tree:
//
// (AND (SHL RS1, VC2), VC1)
//
// We check that VC2, the shamt is less than 32, otherwise the pattern is
// exactly the same as SLLI and we give priority to that.
// Eventually we check that that VC1, the mask used to clear the upper 32 bits
// of RS1, is correct:
//
// VC1 == (0xFFFFFFFF << VC2)
bool RISCVDAGToDAGISel::SelectSLLIUW(SDValue N, SDValue &RS1, SDValue &Shamt) {
if (N.getOpcode() == ISD::AND && Subtarget->getXLenVT() == MVT::i64) {
SDValue And = N;
if (And.getOperand(0).getOpcode() == ISD::SHL) {
SDValue Shl = And.getOperand(0);
if (isa<ConstantSDNode>(Shl.getOperand(1)) &&
isa<ConstantSDNode>(And.getOperand(1))) {
uint64_t VC1 = And.getConstantOperandVal(1);
uint64_t VC2 = Shl.getConstantOperandVal(1);
if (VC2 < 32 && VC1 == ((uint64_t)0xFFFFFFFF << VC2)) {
RS1 = Shl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Shl.getOperand(1).getValueType());
return true;
}
}
}
}
return false;
}
// Check that it is a SLOIW (Shift Left Ones Immediate i32 on RV64).
// We first check that it is the right node tree:
//
// (SIGN_EXTEND_INREG (OR (SHL RS1, VC2), VC1))
//
// and then we check that VC1, the mask used to fill with ones, is compatible
// with VC2, the shamt:
//
// VC1 == maskTrailingOnes<uint32_t>(VC2)
bool RISCVDAGToDAGISel::SelectSLOIW(SDValue N, SDValue &RS1, SDValue &Shamt) {
if (Subtarget->getXLenVT() == MVT::i64 &&
N.getOpcode() == ISD::SIGN_EXTEND_INREG &&
cast<VTSDNode>(N.getOperand(1))->getVT() == MVT::i32) {
if (N.getOperand(0).getOpcode() == ISD::OR) {
SDValue Or = N.getOperand(0);
if (Or.getOperand(0).getOpcode() == ISD::SHL) {
SDValue Shl = Or.getOperand(0);
if (isa<ConstantSDNode>(Shl.getOperand(1)) &&
isa<ConstantSDNode>(Or.getOperand(1))) {
uint32_t VC1 = Or.getConstantOperandVal(1);
uint32_t VC2 = Shl.getConstantOperandVal(1);
if (VC1 == maskTrailingOnes<uint32_t>(VC2)) {
RS1 = Shl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Shl.getOperand(1).getValueType());
return true;
}
}
}
}
}
return false;
}
// Check that it is a SROIW (Shift Right Ones Immediate i32 on RV64).
// We first check that it is the right node tree:
//
// (OR (SHL RS1, VC2), VC1)
//
// and then we check that VC1, the mask used to fill with ones, is compatible
// with VC2, the shamt:
//
// VC1 == maskLeadingOnes<uint32_t>(VC2)
bool RISCVDAGToDAGISel::SelectSROIW(SDValue N, SDValue &RS1, SDValue &Shamt) {
if (N.getOpcode() == ISD::OR && Subtarget->getXLenVT() == MVT::i64) {
SDValue Or = N;
if (Or.getOperand(0).getOpcode() == ISD::SRL) {
SDValue Srl = Or.getOperand(0);
if (isa<ConstantSDNode>(Srl.getOperand(1)) &&
isa<ConstantSDNode>(Or.getOperand(1))) {
uint32_t VC1 = Or.getConstantOperandVal(1);
uint32_t VC2 = Srl.getConstantOperandVal(1);
if (VC1 == maskLeadingOnes<uint32_t>(VC2)) {
RS1 = Srl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC2, SDLoc(N),
Srl.getOperand(1).getValueType());
return true;
}
}
}
}
return false;
}
// Check that it is a RORIW (i32 Right Rotate Immediate on RV64).
// We first check that it is the right node tree:
//
// (SIGN_EXTEND_INREG (OR (SHL (AsserSext RS1, i32), VC2),
// (SRL (AND (AssertSext RS2, i32), VC3), VC1)))
//
// Then we check that the constant operands respect these constraints:
//
// VC2 == 32 - VC1
// VC3 == maskLeadingOnes<uint32_t>(VC2)
//
// being VC1 the Shamt we need, VC2 the complementary of Shamt over 32
// and VC3 a 32 bit mask of (32 - VC1) leading ones.
bool RISCVDAGToDAGISel::SelectRORIW(SDValue N, SDValue &RS1, SDValue &Shamt) {
if (N.getOpcode() == ISD::SIGN_EXTEND_INREG &&
Subtarget->getXLenVT() == MVT::i64 &&
cast<VTSDNode>(N.getOperand(1))->getVT() == MVT::i32) {
if (N.getOperand(0).getOpcode() == ISD::OR) {
SDValue Or = N.getOperand(0);
if (Or.getOperand(0).getOpcode() == ISD::SHL &&
Or.getOperand(1).getOpcode() == ISD::SRL) {
SDValue Shl = Or.getOperand(0);
SDValue Srl = Or.getOperand(1);
if (Srl.getOperand(0).getOpcode() == ISD::AND) {
SDValue And = Srl.getOperand(0);
if (isa<ConstantSDNode>(Srl.getOperand(1)) &&
isa<ConstantSDNode>(Shl.getOperand(1)) &&
isa<ConstantSDNode>(And.getOperand(1))) {
uint32_t VC1 = Srl.getConstantOperandVal(1);
uint32_t VC2 = Shl.getConstantOperandVal(1);
uint32_t VC3 = And.getConstantOperandVal(1);
if (VC2 == (32 - VC1) &&
VC3 == maskLeadingOnes<uint32_t>(VC2)) {
RS1 = Shl.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC1, SDLoc(N),
Srl.getOperand(1).getValueType());
return true;
}
}
}
}
}
}
return false;
}
// Check that it is a FSRIW (i32 Funnel Shift Right Immediate on RV64).
// We first check that it is the right node tree:
//
// (SIGN_EXTEND_INREG (OR (SHL (AsserSext RS1, i32), VC2),
// (SRL (AND (AssertSext RS2, i32), VC3), VC1)))
//
// Then we check that the constant operands respect these constraints:
//
// VC2 == 32 - VC1
// VC3 == maskLeadingOnes<uint32_t>(VC2)
//
// being VC1 the Shamt we need, VC2 the complementary of Shamt over 32
// and VC3 a 32 bit mask of (32 - VC1) leading ones.
bool RISCVDAGToDAGISel::SelectFSRIW(SDValue N, SDValue &RS1, SDValue &RS2,
SDValue &Shamt) {
if (N.getOpcode() == ISD::SIGN_EXTEND_INREG &&
Subtarget->getXLenVT() == MVT::i64 &&
cast<VTSDNode>(N.getOperand(1))->getVT() == MVT::i32) {
if (N.getOperand(0).getOpcode() == ISD::OR) {
SDValue Or = N.getOperand(0);
if (Or.getOperand(0).getOpcode() == ISD::SHL &&
Or.getOperand(1).getOpcode() == ISD::SRL) {
SDValue Shl = Or.getOperand(0);
SDValue Srl = Or.getOperand(1);
if (Srl.getOperand(0).getOpcode() == ISD::AND) {
SDValue And = Srl.getOperand(0);
if (isa<ConstantSDNode>(Srl.getOperand(1)) &&
isa<ConstantSDNode>(Shl.getOperand(1)) &&
isa<ConstantSDNode>(And.getOperand(1))) {
uint32_t VC1 = Srl.getConstantOperandVal(1);
uint32_t VC2 = Shl.getConstantOperandVal(1);
uint32_t VC3 = And.getConstantOperandVal(1);
if (VC2 == (32 - VC1) &&
VC3 == maskLeadingOnes<uint32_t>(VC2)) {
RS1 = Shl.getOperand(0);
RS2 = And.getOperand(0);
Shamt = CurDAG->getTargetConstant(VC1, SDLoc(N),
Srl.getOperand(1).getValueType());
return true;
}
}
}
}
}
}
return false;
}
// Merge an ADDI into the offset of a load/store instruction where possible.
// (load (addi base, off1), off2) -> (load base, off1+off2)
// (store val, (addi base, off1), off2) -> (store val, base, off1+off2)
// This is possible when off1+off2 fits a 12-bit immediate.
void RISCVDAGToDAGISel::doPeepholeLoadStoreADDI() {
SelectionDAG::allnodes_iterator Position(CurDAG->getRoot().getNode());
++Position;
while (Position != CurDAG->allnodes_begin()) {
SDNode *N = &*--Position;
// Skip dead nodes and any non-machine opcodes.
if (N->use_empty() || !N->isMachineOpcode())
continue;
int OffsetOpIdx;
int BaseOpIdx;
// Only attempt this optimisation for I-type loads and S-type stores.
switch (N->getMachineOpcode()) {
default:
continue;
case RISCV::LB:
case RISCV::LH:
case RISCV::LW:
case RISCV::LBU:
case RISCV::LHU:
case RISCV::LWU:
case RISCV::LD:
case RISCV::FLW:
case RISCV::FLD:
BaseOpIdx = 0;
OffsetOpIdx = 1;
break;
case RISCV::SB:
case RISCV::SH:
case RISCV::SW:
case RISCV::SD:
case RISCV::FSW:
case RISCV::FSD:
BaseOpIdx = 1;
OffsetOpIdx = 2;
break;
}
if (!isa<ConstantSDNode>(N->getOperand(OffsetOpIdx)))
continue;
SDValue Base = N->getOperand(BaseOpIdx);
// If the base is an ADDI, we can merge it in to the load/store.
if (!Base.isMachineOpcode() || Base.getMachineOpcode() != RISCV::ADDI)
continue;
SDValue ImmOperand = Base.getOperand(1);
uint64_t Offset2 = N->getConstantOperandVal(OffsetOpIdx);
if (auto Const = dyn_cast<ConstantSDNode>(ImmOperand)) {
int64_t Offset1 = Const->getSExtValue();
int64_t CombinedOffset = Offset1 + Offset2;
if (!isInt<12>(CombinedOffset))
continue;
ImmOperand = CurDAG->getTargetConstant(CombinedOffset, SDLoc(ImmOperand),
ImmOperand.getValueType());
} else if (auto GA = dyn_cast<GlobalAddressSDNode>(ImmOperand)) {
// If the off1 in (addi base, off1) is a global variable's address (its
// low part, really), then we can rely on the alignment of that variable
// to provide a margin of safety before off1 can overflow the 12 bits.
// Check if off2 falls within that margin; if so off1+off2 can't overflow.
const DataLayout &DL = CurDAG->getDataLayout();
Align Alignment = GA->getGlobal()->getPointerAlignment(DL);
if (Offset2 != 0 && Alignment <= Offset2)
continue;
int64_t Offset1 = GA->getOffset();
int64_t CombinedOffset = Offset1 + Offset2;
ImmOperand = CurDAG->getTargetGlobalAddress(
GA->getGlobal(), SDLoc(ImmOperand), ImmOperand.getValueType(),
CombinedOffset, GA->getTargetFlags());
} else if (auto CP = dyn_cast<ConstantPoolSDNode>(ImmOperand)) {
// Ditto.
Align Alignment = CP->getAlign();
if (Offset2 != 0 && Alignment <= Offset2)
continue;
int64_t Offset1 = CP->getOffset();
int64_t CombinedOffset = Offset1 + Offset2;
ImmOperand = CurDAG->getTargetConstantPool(
CP->getConstVal(), ImmOperand.getValueType(), CP->getAlign(),
CombinedOffset, CP->getTargetFlags());
} else {
continue;
}
LLVM_DEBUG(dbgs() << "Folding add-immediate into mem-op:\nBase: ");
LLVM_DEBUG(Base->dump(CurDAG));
LLVM_DEBUG(dbgs() << "\nN: ");
LLVM_DEBUG(N->dump(CurDAG));
LLVM_DEBUG(dbgs() << "\n");
// Modify the offset operand of the load/store.
if (BaseOpIdx == 0) // Load
CurDAG->UpdateNodeOperands(N, Base.getOperand(0), ImmOperand,
N->getOperand(2));
else // Store
CurDAG->UpdateNodeOperands(N, N->getOperand(0), Base.getOperand(0),
ImmOperand, N->getOperand(3));
// The add-immediate may now be dead, in which case remove it.
if (Base.getNode()->use_empty())
CurDAG->RemoveDeadNode(Base.getNode());
}
}
// This pass converts a legalized DAG into a RISCV-specific DAG, ready
// for instruction scheduling.
FunctionPass *llvm::createRISCVISelDag(RISCVTargetMachine &TM) {
return new RISCVDAGToDAGISel(TM);
}