1 | //===- ScheduleDAG.cpp - Implement the ScheduleDAG class ------------------===// |
2 | // |
3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
4 | // See https://llvm.org/LICENSE.txt for license information. |
5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
6 | // |
7 | //===----------------------------------------------------------------------===// |
8 | // |
9 | /// \file Implements the ScheduleDAG class, which is a base class used by |
10 | /// scheduling implementation classes. |
11 | // |
12 | //===----------------------------------------------------------------------===// |
13 | |
14 | #include "llvm/CodeGen/ScheduleDAG.h" |
15 | #include "llvm/ADT/STLExtras.h" |
16 | #include "llvm/ADT/SmallVector.h" |
17 | #include "llvm/ADT/Statistic.h" |
18 | #include "llvm/CodeGen/MachineFunction.h" |
19 | #include "llvm/CodeGen/ScheduleHazardRecognizer.h" |
20 | #include "llvm/CodeGen/SelectionDAGNodes.h" |
21 | #include "llvm/CodeGen/TargetInstrInfo.h" |
22 | #include "llvm/CodeGen/TargetRegisterInfo.h" |
23 | #include "llvm/CodeGen/TargetSubtargetInfo.h" |
24 | #include "llvm/Config/llvm-config.h" |
25 | #include "llvm/Support/CommandLine.h" |
26 | #include "llvm/Support/Compiler.h" |
27 | #include "llvm/Support/Debug.h" |
28 | #include "llvm/Support/raw_ostream.h" |
29 | #include <algorithm> |
30 | #include <cassert> |
31 | #include <iterator> |
32 | #include <limits> |
33 | #include <utility> |
34 | #include <vector> |
35 | |
36 | using namespace llvm; |
37 | |
38 | #define DEBUG_TYPE "pre-RA-sched" |
39 | |
40 | STATISTIC(NumNewPredsAdded, "Number of times a single predecessor was added" ); |
41 | STATISTIC(NumTopoInits, |
42 | "Number of times the topological order has been recomputed" ); |
43 | |
44 | #ifndef NDEBUG |
45 | static cl::opt<bool> StressSchedOpt( |
46 | "stress-sched" , cl::Hidden, cl::init(false), |
47 | cl::desc("Stress test instruction scheduling" )); |
48 | #endif |
49 | |
50 | void SchedulingPriorityQueue::anchor() {} |
51 | |
52 | ScheduleDAG::ScheduleDAG(MachineFunction &mf) |
53 | : TM(mf.getTarget()), TII(mf.getSubtarget().getInstrInfo()), |
54 | TRI(mf.getSubtarget().getRegisterInfo()), MF(mf), |
55 | MRI(mf.getRegInfo()) { |
56 | #ifndef NDEBUG |
57 | StressSched = StressSchedOpt; |
58 | #endif |
59 | } |
60 | |
61 | ScheduleDAG::~ScheduleDAG() = default; |
62 | |
63 | void ScheduleDAG::clearDAG() { |
64 | SUnits.clear(); |
65 | EntrySU = SUnit(); |
66 | ExitSU = SUnit(); |
67 | } |
68 | |
69 | const MCInstrDesc *ScheduleDAG::getNodeDesc(const SDNode *Node) const { |
70 | if (!Node || !Node->isMachineOpcode()) return nullptr; |
71 | return &TII->get(Opcode: Node->getMachineOpcode()); |
72 | } |
73 | |
74 | LLVM_DUMP_METHOD void SDep::dump(const TargetRegisterInfo *TRI) const { |
75 | switch (getKind()) { |
76 | case Data: dbgs() << "Data" ; break; |
77 | case Anti: dbgs() << "Anti" ; break; |
78 | case Output: dbgs() << "Out " ; break; |
79 | case Order: dbgs() << "Ord " ; break; |
80 | } |
81 | |
82 | switch (getKind()) { |
83 | case Data: |
84 | dbgs() << " Latency=" << getLatency(); |
85 | if (TRI && isAssignedRegDep()) |
86 | dbgs() << " Reg=" << printReg(Reg: getReg(), TRI); |
87 | break; |
88 | case Anti: |
89 | case Output: |
90 | dbgs() << " Latency=" << getLatency(); |
91 | break; |
92 | case Order: |
93 | dbgs() << " Latency=" << getLatency(); |
94 | switch(Contents.OrdKind) { |
95 | case Barrier: dbgs() << " Barrier" ; break; |
96 | case MayAliasMem: |
97 | case MustAliasMem: dbgs() << " Memory" ; break; |
98 | case Artificial: dbgs() << " Artificial" ; break; |
99 | case Weak: dbgs() << " Weak" ; break; |
100 | case Cluster: dbgs() << " Cluster" ; break; |
101 | } |
102 | break; |
103 | } |
104 | } |
105 | |
106 | bool SUnit::addPred(const SDep &D, bool Required) { |
107 | // If this node already has this dependence, don't add a redundant one. |
108 | for (SDep &PredDep : Preds) { |
109 | // Zero-latency weak edges may be added purely for heuristic ordering. Don't |
110 | // add them if another kind of edge already exists. |
111 | if (!Required && PredDep.getSUnit() == D.getSUnit()) |
112 | return false; |
113 | if (PredDep.overlaps(Other: D)) { |
114 | // Extend the latency if needed. Equivalent to |
115 | // removePred(PredDep) + addPred(D). |
116 | if (PredDep.getLatency() < D.getLatency()) { |
117 | SUnit *PredSU = PredDep.getSUnit(); |
118 | // Find the corresponding successor in N. |
119 | SDep ForwardD = PredDep; |
120 | ForwardD.setSUnit(this); |
121 | for (SDep &SuccDep : PredSU->Succs) { |
122 | if (SuccDep == ForwardD) { |
123 | SuccDep.setLatency(D.getLatency()); |
124 | break; |
125 | } |
126 | } |
127 | PredDep.setLatency(D.getLatency()); |
128 | } |
129 | return false; |
130 | } |
131 | } |
132 | // Now add a corresponding succ to N. |
133 | SDep P = D; |
134 | P.setSUnit(this); |
135 | SUnit *N = D.getSUnit(); |
136 | // Update the bookkeeping. |
137 | if (D.getKind() == SDep::Data) { |
138 | assert(NumPreds < std::numeric_limits<unsigned>::max() && |
139 | "NumPreds will overflow!" ); |
140 | assert(N->NumSuccs < std::numeric_limits<unsigned>::max() && |
141 | "NumSuccs will overflow!" ); |
142 | ++NumPreds; |
143 | ++N->NumSuccs; |
144 | } |
145 | if (!N->isScheduled) { |
146 | if (D.isWeak()) { |
147 | ++WeakPredsLeft; |
148 | } |
149 | else { |
150 | assert(NumPredsLeft < std::numeric_limits<unsigned>::max() && |
151 | "NumPredsLeft will overflow!" ); |
152 | ++NumPredsLeft; |
153 | } |
154 | } |
155 | if (!isScheduled) { |
156 | if (D.isWeak()) { |
157 | ++N->WeakSuccsLeft; |
158 | } |
159 | else { |
160 | assert(N->NumSuccsLeft < std::numeric_limits<unsigned>::max() && |
161 | "NumSuccsLeft will overflow!" ); |
162 | ++N->NumSuccsLeft; |
163 | } |
164 | } |
165 | Preds.push_back(Elt: D); |
166 | N->Succs.push_back(Elt: P); |
167 | if (P.getLatency() != 0) { |
168 | this->setDepthDirty(); |
169 | N->setHeightDirty(); |
170 | } |
171 | return true; |
172 | } |
173 | |
174 | void SUnit::removePred(const SDep &D) { |
175 | // Find the matching predecessor. |
176 | SmallVectorImpl<SDep>::iterator I = llvm::find(Range&: Preds, Val: D); |
177 | if (I == Preds.end()) |
178 | return; |
179 | // Find the corresponding successor in N. |
180 | SDep P = D; |
181 | P.setSUnit(this); |
182 | SUnit *N = D.getSUnit(); |
183 | SmallVectorImpl<SDep>::iterator Succ = llvm::find(Range&: N->Succs, Val: P); |
184 | assert(Succ != N->Succs.end() && "Mismatching preds / succs lists!" ); |
185 | // Update the bookkeeping. |
186 | if (P.getKind() == SDep::Data) { |
187 | assert(NumPreds > 0 && "NumPreds will underflow!" ); |
188 | assert(N->NumSuccs > 0 && "NumSuccs will underflow!" ); |
189 | --NumPreds; |
190 | --N->NumSuccs; |
191 | } |
192 | if (!N->isScheduled) { |
193 | if (D.isWeak()) { |
194 | assert(WeakPredsLeft > 0 && "WeakPredsLeft will underflow!" ); |
195 | --WeakPredsLeft; |
196 | } else { |
197 | assert(NumPredsLeft > 0 && "NumPredsLeft will underflow!" ); |
198 | --NumPredsLeft; |
199 | } |
200 | } |
201 | if (!isScheduled) { |
202 | if (D.isWeak()) { |
203 | assert(N->WeakSuccsLeft > 0 && "WeakSuccsLeft will underflow!" ); |
204 | --N->WeakSuccsLeft; |
205 | } else { |
206 | assert(N->NumSuccsLeft > 0 && "NumSuccsLeft will underflow!" ); |
207 | --N->NumSuccsLeft; |
208 | } |
209 | } |
210 | N->Succs.erase(CI: Succ); |
211 | Preds.erase(CI: I); |
212 | if (P.getLatency() != 0) { |
213 | this->setDepthDirty(); |
214 | N->setHeightDirty(); |
215 | } |
216 | } |
217 | |
218 | void SUnit::setDepthDirty() { |
219 | if (!isDepthCurrent) return; |
220 | SmallVector<SUnit*, 8> WorkList; |
221 | WorkList.push_back(Elt: this); |
222 | do { |
223 | SUnit *SU = WorkList.pop_back_val(); |
224 | SU->isDepthCurrent = false; |
225 | for (SDep &SuccDep : SU->Succs) { |
226 | SUnit *SuccSU = SuccDep.getSUnit(); |
227 | if (SuccSU->isDepthCurrent) |
228 | WorkList.push_back(Elt: SuccSU); |
229 | } |
230 | } while (!WorkList.empty()); |
231 | } |
232 | |
233 | void SUnit::setHeightDirty() { |
234 | if (!isHeightCurrent) return; |
235 | SmallVector<SUnit*, 8> WorkList; |
236 | WorkList.push_back(Elt: this); |
237 | do { |
238 | SUnit *SU = WorkList.pop_back_val(); |
239 | SU->isHeightCurrent = false; |
240 | for (SDep &PredDep : SU->Preds) { |
241 | SUnit *PredSU = PredDep.getSUnit(); |
242 | if (PredSU->isHeightCurrent) |
243 | WorkList.push_back(Elt: PredSU); |
244 | } |
245 | } while (!WorkList.empty()); |
246 | } |
247 | |
248 | void SUnit::setDepthToAtLeast(unsigned NewDepth) { |
249 | if (NewDepth <= getDepth()) |
250 | return; |
251 | setDepthDirty(); |
252 | Depth = NewDepth; |
253 | isDepthCurrent = true; |
254 | } |
255 | |
256 | void SUnit::setHeightToAtLeast(unsigned NewHeight) { |
257 | if (NewHeight <= getHeight()) |
258 | return; |
259 | setHeightDirty(); |
260 | Height = NewHeight; |
261 | isHeightCurrent = true; |
262 | } |
263 | |
264 | /// Calculates the maximal path from the node to the exit. |
265 | void SUnit::ComputeDepth() { |
266 | SmallVector<SUnit*, 8> WorkList; |
267 | WorkList.push_back(Elt: this); |
268 | do { |
269 | SUnit *Cur = WorkList.back(); |
270 | |
271 | bool Done = true; |
272 | unsigned MaxPredDepth = 0; |
273 | for (const SDep &PredDep : Cur->Preds) { |
274 | SUnit *PredSU = PredDep.getSUnit(); |
275 | if (PredSU->isDepthCurrent) |
276 | MaxPredDepth = std::max(a: MaxPredDepth, |
277 | b: PredSU->Depth + PredDep.getLatency()); |
278 | else { |
279 | Done = false; |
280 | WorkList.push_back(Elt: PredSU); |
281 | } |
282 | } |
283 | |
284 | if (Done) { |
285 | WorkList.pop_back(); |
286 | if (MaxPredDepth != Cur->Depth) { |
287 | Cur->setDepthDirty(); |
288 | Cur->Depth = MaxPredDepth; |
289 | } |
290 | Cur->isDepthCurrent = true; |
291 | } |
292 | } while (!WorkList.empty()); |
293 | } |
294 | |
295 | /// Calculates the maximal path from the node to the entry. |
296 | void SUnit::ComputeHeight() { |
297 | SmallVector<SUnit*, 8> WorkList; |
298 | WorkList.push_back(Elt: this); |
299 | do { |
300 | SUnit *Cur = WorkList.back(); |
301 | |
302 | bool Done = true; |
303 | unsigned MaxSuccHeight = 0; |
304 | for (const SDep &SuccDep : Cur->Succs) { |
305 | SUnit *SuccSU = SuccDep.getSUnit(); |
306 | if (SuccSU->isHeightCurrent) |
307 | MaxSuccHeight = std::max(a: MaxSuccHeight, |
308 | b: SuccSU->Height + SuccDep.getLatency()); |
309 | else { |
310 | Done = false; |
311 | WorkList.push_back(Elt: SuccSU); |
312 | } |
313 | } |
314 | |
315 | if (Done) { |
316 | WorkList.pop_back(); |
317 | if (MaxSuccHeight != Cur->Height) { |
318 | Cur->setHeightDirty(); |
319 | Cur->Height = MaxSuccHeight; |
320 | } |
321 | Cur->isHeightCurrent = true; |
322 | } |
323 | } while (!WorkList.empty()); |
324 | } |
325 | |
326 | void SUnit::biasCriticalPath() { |
327 | if (NumPreds < 2) |
328 | return; |
329 | |
330 | SUnit::pred_iterator BestI = Preds.begin(); |
331 | unsigned MaxDepth = BestI->getSUnit()->getDepth(); |
332 | for (SUnit::pred_iterator I = std::next(x: BestI), E = Preds.end(); I != E; |
333 | ++I) { |
334 | if (I->getKind() == SDep::Data && I->getSUnit()->getDepth() > MaxDepth) { |
335 | MaxDepth = I->getSUnit()->getDepth(); |
336 | BestI = I; |
337 | } |
338 | } |
339 | if (BestI != Preds.begin()) |
340 | std::swap(a&: *Preds.begin(), b&: *BestI); |
341 | } |
342 | |
343 | #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) |
344 | LLVM_DUMP_METHOD void SUnit::dumpAttributes() const { |
345 | dbgs() << " # preds left : " << NumPredsLeft << "\n" ; |
346 | dbgs() << " # succs left : " << NumSuccsLeft << "\n" ; |
347 | if (WeakPredsLeft) |
348 | dbgs() << " # weak preds left : " << WeakPredsLeft << "\n" ; |
349 | if (WeakSuccsLeft) |
350 | dbgs() << " # weak succs left : " << WeakSuccsLeft << "\n" ; |
351 | dbgs() << " # rdefs left : " << NumRegDefsLeft << "\n" ; |
352 | dbgs() << " Latency : " << Latency << "\n" ; |
353 | dbgs() << " Depth : " << getDepth() << "\n" ; |
354 | dbgs() << " Height : " << getHeight() << "\n" ; |
355 | } |
356 | |
357 | LLVM_DUMP_METHOD void ScheduleDAG::dumpNodeName(const SUnit &SU) const { |
358 | if (&SU == &EntrySU) |
359 | dbgs() << "EntrySU" ; |
360 | else if (&SU == &ExitSU) |
361 | dbgs() << "ExitSU" ; |
362 | else |
363 | dbgs() << "SU(" << SU.NodeNum << ")" ; |
364 | } |
365 | |
366 | LLVM_DUMP_METHOD void ScheduleDAG::dumpNodeAll(const SUnit &SU) const { |
367 | dumpNode(SU); |
368 | SU.dumpAttributes(); |
369 | if (SU.Preds.size() > 0) { |
370 | dbgs() << " Predecessors:\n" ; |
371 | for (const SDep &Dep : SU.Preds) { |
372 | dbgs() << " " ; |
373 | dumpNodeName(*Dep.getSUnit()); |
374 | dbgs() << ": " ; |
375 | Dep.dump(TRI); |
376 | dbgs() << '\n'; |
377 | } |
378 | } |
379 | if (SU.Succs.size() > 0) { |
380 | dbgs() << " Successors:\n" ; |
381 | for (const SDep &Dep : SU.Succs) { |
382 | dbgs() << " " ; |
383 | dumpNodeName(*Dep.getSUnit()); |
384 | dbgs() << ": " ; |
385 | Dep.dump(TRI); |
386 | dbgs() << '\n'; |
387 | } |
388 | } |
389 | } |
390 | #endif |
391 | |
392 | #ifndef NDEBUG |
393 | unsigned ScheduleDAG::VerifyScheduledDAG(bool isBottomUp) { |
394 | bool AnyNotSched = false; |
395 | unsigned DeadNodes = 0; |
396 | for (const SUnit &SUnit : SUnits) { |
397 | if (!SUnit.isScheduled) { |
398 | if (SUnit.NumPreds == 0 && SUnit.NumSuccs == 0) { |
399 | ++DeadNodes; |
400 | continue; |
401 | } |
402 | if (!AnyNotSched) |
403 | dbgs() << "*** Scheduling failed! ***\n" ; |
404 | dumpNode(SUnit); |
405 | dbgs() << "has not been scheduled!\n" ; |
406 | AnyNotSched = true; |
407 | } |
408 | if (SUnit.isScheduled && |
409 | (isBottomUp ? SUnit.getHeight() : SUnit.getDepth()) > |
410 | unsigned(std::numeric_limits<int>::max())) { |
411 | if (!AnyNotSched) |
412 | dbgs() << "*** Scheduling failed! ***\n" ; |
413 | dumpNode(SUnit); |
414 | dbgs() << "has an unexpected " |
415 | << (isBottomUp ? "Height" : "Depth" ) << " value!\n" ; |
416 | AnyNotSched = true; |
417 | } |
418 | if (isBottomUp) { |
419 | if (SUnit.NumSuccsLeft != 0) { |
420 | if (!AnyNotSched) |
421 | dbgs() << "*** Scheduling failed! ***\n" ; |
422 | dumpNode(SUnit); |
423 | dbgs() << "has successors left!\n" ; |
424 | AnyNotSched = true; |
425 | } |
426 | } else { |
427 | if (SUnit.NumPredsLeft != 0) { |
428 | if (!AnyNotSched) |
429 | dbgs() << "*** Scheduling failed! ***\n" ; |
430 | dumpNode(SUnit); |
431 | dbgs() << "has predecessors left!\n" ; |
432 | AnyNotSched = true; |
433 | } |
434 | } |
435 | } |
436 | assert(!AnyNotSched); |
437 | return SUnits.size() - DeadNodes; |
438 | } |
439 | #endif |
440 | |
441 | void ScheduleDAGTopologicalSort::InitDAGTopologicalSorting() { |
442 | // The idea of the algorithm is taken from |
443 | // "Online algorithms for managing the topological order of |
444 | // a directed acyclic graph" by David J. Pearce and Paul H.J. Kelly |
445 | // This is the MNR algorithm, which was first introduced by |
446 | // A. Marchetti-Spaccamela, U. Nanni and H. Rohnert in |
447 | // "Maintaining a topological order under edge insertions". |
448 | // |
449 | // Short description of the algorithm: |
450 | // |
451 | // Topological ordering, ord, of a DAG maps each node to a topological |
452 | // index so that for all edges X->Y it is the case that ord(X) < ord(Y). |
453 | // |
454 | // This means that if there is a path from the node X to the node Z, |
455 | // then ord(X) < ord(Z). |
456 | // |
457 | // This property can be used to check for reachability of nodes: |
458 | // if Z is reachable from X, then an insertion of the edge Z->X would |
459 | // create a cycle. |
460 | // |
461 | // The algorithm first computes a topological ordering for the DAG by |
462 | // initializing the Index2Node and Node2Index arrays and then tries to keep |
463 | // the ordering up-to-date after edge insertions by reordering the DAG. |
464 | // |
465 | // On insertion of the edge X->Y, the algorithm first marks by calling DFS |
466 | // the nodes reachable from Y, and then shifts them using Shift to lie |
467 | // immediately after X in Index2Node. |
468 | |
469 | // Cancel pending updates, mark as valid. |
470 | Dirty = false; |
471 | Updates.clear(); |
472 | |
473 | unsigned DAGSize = SUnits.size(); |
474 | std::vector<SUnit*> WorkList; |
475 | WorkList.reserve(n: DAGSize); |
476 | |
477 | Index2Node.resize(new_size: DAGSize); |
478 | Node2Index.resize(new_size: DAGSize); |
479 | |
480 | // Initialize the data structures. |
481 | if (ExitSU) |
482 | WorkList.push_back(x: ExitSU); |
483 | for (SUnit &SU : SUnits) { |
484 | int NodeNum = SU.NodeNum; |
485 | unsigned Degree = SU.Succs.size(); |
486 | // Temporarily use the Node2Index array as scratch space for degree counts. |
487 | Node2Index[NodeNum] = Degree; |
488 | |
489 | // Is it a node without dependencies? |
490 | if (Degree == 0) { |
491 | assert(SU.Succs.empty() && "SUnit should have no successors" ); |
492 | // Collect leaf nodes. |
493 | WorkList.push_back(x: &SU); |
494 | } |
495 | } |
496 | |
497 | int Id = DAGSize; |
498 | while (!WorkList.empty()) { |
499 | SUnit *SU = WorkList.back(); |
500 | WorkList.pop_back(); |
501 | if (SU->NodeNum < DAGSize) |
502 | Allocate(n: SU->NodeNum, index: --Id); |
503 | for (const SDep &PredDep : SU->Preds) { |
504 | SUnit *SU = PredDep.getSUnit(); |
505 | if (SU->NodeNum < DAGSize && !--Node2Index[SU->NodeNum]) |
506 | // If all dependencies of the node are processed already, |
507 | // then the node can be computed now. |
508 | WorkList.push_back(x: SU); |
509 | } |
510 | } |
511 | |
512 | Visited.resize(N: DAGSize); |
513 | NumTopoInits++; |
514 | |
515 | #ifndef NDEBUG |
516 | // Check correctness of the ordering |
517 | for (SUnit &SU : SUnits) { |
518 | for (const SDep &PD : SU.Preds) { |
519 | assert(Node2Index[SU.NodeNum] > Node2Index[PD.getSUnit()->NodeNum] && |
520 | "Wrong topological sorting" ); |
521 | } |
522 | } |
523 | #endif |
524 | } |
525 | |
526 | void ScheduleDAGTopologicalSort::FixOrder() { |
527 | // Recompute from scratch after new nodes have been added. |
528 | if (Dirty) { |
529 | InitDAGTopologicalSorting(); |
530 | return; |
531 | } |
532 | |
533 | // Otherwise apply updates one-by-one. |
534 | for (auto &U : Updates) |
535 | AddPred(Y: U.first, X: U.second); |
536 | Updates.clear(); |
537 | } |
538 | |
539 | void ScheduleDAGTopologicalSort::AddPredQueued(SUnit *Y, SUnit *X) { |
540 | // Recomputing the order from scratch is likely more efficient than applying |
541 | // updates one-by-one for too many updates. The current cut-off is arbitrarily |
542 | // chosen. |
543 | Dirty = Dirty || Updates.size() > 10; |
544 | |
545 | if (Dirty) |
546 | return; |
547 | |
548 | Updates.emplace_back(Args&: Y, Args&: X); |
549 | } |
550 | |
551 | void ScheduleDAGTopologicalSort::AddPred(SUnit *Y, SUnit *X) { |
552 | int UpperBound, LowerBound; |
553 | LowerBound = Node2Index[Y->NodeNum]; |
554 | UpperBound = Node2Index[X->NodeNum]; |
555 | bool HasLoop = false; |
556 | // Is Ord(X) < Ord(Y) ? |
557 | if (LowerBound < UpperBound) { |
558 | // Update the topological order. |
559 | Visited.reset(); |
560 | DFS(SU: Y, UpperBound, HasLoop); |
561 | assert(!HasLoop && "Inserted edge creates a loop!" ); |
562 | // Recompute topological indexes. |
563 | Shift(Visited, LowerBound, UpperBound); |
564 | } |
565 | |
566 | NumNewPredsAdded++; |
567 | } |
568 | |
569 | void ScheduleDAGTopologicalSort::RemovePred(SUnit *M, SUnit *N) { |
570 | // InitDAGTopologicalSorting(); |
571 | } |
572 | |
573 | void ScheduleDAGTopologicalSort::DFS(const SUnit *SU, int UpperBound, |
574 | bool &HasLoop) { |
575 | std::vector<const SUnit*> WorkList; |
576 | WorkList.reserve(n: SUnits.size()); |
577 | |
578 | WorkList.push_back(x: SU); |
579 | do { |
580 | SU = WorkList.back(); |
581 | WorkList.pop_back(); |
582 | Visited.set(SU->NodeNum); |
583 | for (const SDep &SuccDep : llvm::reverse(C: SU->Succs)) { |
584 | unsigned s = SuccDep.getSUnit()->NodeNum; |
585 | // Edges to non-SUnits are allowed but ignored (e.g. ExitSU). |
586 | if (s >= Node2Index.size()) |
587 | continue; |
588 | if (Node2Index[s] == UpperBound) { |
589 | HasLoop = true; |
590 | return; |
591 | } |
592 | // Visit successors if not already and in affected region. |
593 | if (!Visited.test(Idx: s) && Node2Index[s] < UpperBound) { |
594 | WorkList.push_back(x: SuccDep.getSUnit()); |
595 | } |
596 | } |
597 | } while (!WorkList.empty()); |
598 | } |
599 | |
600 | std::vector<int> ScheduleDAGTopologicalSort::GetSubGraph(const SUnit &StartSU, |
601 | const SUnit &TargetSU, |
602 | bool &Success) { |
603 | std::vector<const SUnit*> WorkList; |
604 | int LowerBound = Node2Index[StartSU.NodeNum]; |
605 | int UpperBound = Node2Index[TargetSU.NodeNum]; |
606 | bool Found = false; |
607 | BitVector VisitedBack; |
608 | std::vector<int> Nodes; |
609 | |
610 | if (LowerBound > UpperBound) { |
611 | Success = false; |
612 | return Nodes; |
613 | } |
614 | |
615 | WorkList.reserve(n: SUnits.size()); |
616 | Visited.reset(); |
617 | |
618 | // Starting from StartSU, visit all successors up |
619 | // to UpperBound. |
620 | WorkList.push_back(x: &StartSU); |
621 | do { |
622 | const SUnit *SU = WorkList.back(); |
623 | WorkList.pop_back(); |
624 | for (const SDep &SD : llvm::reverse(C: SU->Succs)) { |
625 | const SUnit *Succ = SD.getSUnit(); |
626 | unsigned s = Succ->NodeNum; |
627 | // Edges to non-SUnits are allowed but ignored (e.g. ExitSU). |
628 | if (Succ->isBoundaryNode()) |
629 | continue; |
630 | if (Node2Index[s] == UpperBound) { |
631 | Found = true; |
632 | continue; |
633 | } |
634 | // Visit successors if not already and in affected region. |
635 | if (!Visited.test(Idx: s) && Node2Index[s] < UpperBound) { |
636 | Visited.set(s); |
637 | WorkList.push_back(x: Succ); |
638 | } |
639 | } |
640 | } while (!WorkList.empty()); |
641 | |
642 | if (!Found) { |
643 | Success = false; |
644 | return Nodes; |
645 | } |
646 | |
647 | WorkList.clear(); |
648 | VisitedBack.resize(N: SUnits.size()); |
649 | Found = false; |
650 | |
651 | // Starting from TargetSU, visit all predecessors up |
652 | // to LowerBound. SUs that are visited by the two |
653 | // passes are added to Nodes. |
654 | WorkList.push_back(x: &TargetSU); |
655 | do { |
656 | const SUnit *SU = WorkList.back(); |
657 | WorkList.pop_back(); |
658 | for (const SDep &SD : llvm::reverse(C: SU->Preds)) { |
659 | const SUnit *Pred = SD.getSUnit(); |
660 | unsigned s = Pred->NodeNum; |
661 | // Edges to non-SUnits are allowed but ignored (e.g. EntrySU). |
662 | if (Pred->isBoundaryNode()) |
663 | continue; |
664 | if (Node2Index[s] == LowerBound) { |
665 | Found = true; |
666 | continue; |
667 | } |
668 | if (!VisitedBack.test(Idx: s) && Visited.test(Idx: s)) { |
669 | VisitedBack.set(s); |
670 | WorkList.push_back(x: Pred); |
671 | Nodes.push_back(x: s); |
672 | } |
673 | } |
674 | } while (!WorkList.empty()); |
675 | |
676 | assert(Found && "Error in SUnit Graph!" ); |
677 | Success = true; |
678 | return Nodes; |
679 | } |
680 | |
681 | void ScheduleDAGTopologicalSort::Shift(BitVector& Visited, int LowerBound, |
682 | int UpperBound) { |
683 | std::vector<int> L; |
684 | int shift = 0; |
685 | int i; |
686 | |
687 | for (i = LowerBound; i <= UpperBound; ++i) { |
688 | // w is node at topological index i. |
689 | int w = Index2Node[i]; |
690 | if (Visited.test(Idx: w)) { |
691 | // Unmark. |
692 | Visited.reset(Idx: w); |
693 | L.push_back(x: w); |
694 | shift = shift + 1; |
695 | } else { |
696 | Allocate(n: w, index: i - shift); |
697 | } |
698 | } |
699 | |
700 | for (unsigned LI : L) { |
701 | Allocate(n: LI, index: i - shift); |
702 | i = i + 1; |
703 | } |
704 | } |
705 | |
706 | bool ScheduleDAGTopologicalSort::WillCreateCycle(SUnit *TargetSU, SUnit *SU) { |
707 | FixOrder(); |
708 | // Is SU reachable from TargetSU via successor edges? |
709 | if (IsReachable(SU, TargetSU)) |
710 | return true; |
711 | for (const SDep &PredDep : TargetSU->Preds) |
712 | if (PredDep.isAssignedRegDep() && |
713 | IsReachable(SU, TargetSU: PredDep.getSUnit())) |
714 | return true; |
715 | return false; |
716 | } |
717 | |
718 | void ScheduleDAGTopologicalSort::AddSUnitWithoutPredecessors(const SUnit *SU) { |
719 | assert(SU->NodeNum == Index2Node.size() && "Node cannot be added at the end" ); |
720 | assert(SU->NumPreds == 0 && "Can only add SU's with no predecessors" ); |
721 | Node2Index.push_back(x: Index2Node.size()); |
722 | Index2Node.push_back(x: SU->NodeNum); |
723 | Visited.resize(N: Node2Index.size()); |
724 | } |
725 | |
726 | bool ScheduleDAGTopologicalSort::IsReachable(const SUnit *SU, |
727 | const SUnit *TargetSU) { |
728 | assert(TargetSU != nullptr && "Invalid target SUnit" ); |
729 | assert(SU != nullptr && "Invalid SUnit" ); |
730 | FixOrder(); |
731 | // If insertion of the edge SU->TargetSU would create a cycle |
732 | // then there is a path from TargetSU to SU. |
733 | int UpperBound, LowerBound; |
734 | LowerBound = Node2Index[TargetSU->NodeNum]; |
735 | UpperBound = Node2Index[SU->NodeNum]; |
736 | bool HasLoop = false; |
737 | // Is Ord(TargetSU) < Ord(SU) ? |
738 | if (LowerBound < UpperBound) { |
739 | Visited.reset(); |
740 | // There may be a path from TargetSU to SU. Check for it. |
741 | DFS(SU: TargetSU, UpperBound, HasLoop); |
742 | } |
743 | return HasLoop; |
744 | } |
745 | |
746 | void ScheduleDAGTopologicalSort::Allocate(int n, int index) { |
747 | Node2Index[n] = index; |
748 | Index2Node[index] = n; |
749 | } |
750 | |
751 | ScheduleDAGTopologicalSort:: |
752 | ScheduleDAGTopologicalSort(std::vector<SUnit> &sunits, SUnit *exitsu) |
753 | : SUnits(sunits), ExitSU(exitsu) {} |
754 | |
755 | ScheduleHazardRecognizer::~ScheduleHazardRecognizer() = default; |
756 | |