1//===----- ScheduleDAGRRList.cpp - Reg pressure reduction list scheduler --===// 2// 3// The LLVM Compiler Infrastructure 4// 5// This file is distributed under the University of Illinois Open Source 6// License. See LICENSE.TXT for details. 7// 8//===----------------------------------------------------------------------===// 9// 10// This implements bottom-up and top-down register pressure reduction list 11// schedulers, using standard algorithms. The basic approach uses a priority 12// queue of available nodes to schedule. One at a time, nodes are taken from 13// the priority queue (thus in priority order), checked for legality to 14// schedule, and emitted if legal. 15// 16//===----------------------------------------------------------------------===// 17 18#define DEBUG_TYPE "pre-RA-sched" 19#include "llvm/CodeGen/SchedulerRegistry.h" 20#include "ScheduleDAGSDNodes.h" 21#include "llvm/ADT/STLExtras.h" 22#include "llvm/ADT/SmallSet.h" 23#include "llvm/ADT/Statistic.h" 24#include "llvm/CodeGen/MachineRegisterInfo.h" 25#include "llvm/CodeGen/ScheduleHazardRecognizer.h" 26#include "llvm/CodeGen/SelectionDAGISel.h" 27#include "llvm/IR/DataLayout.h" 28#include "llvm/IR/InlineAsm.h" 29#include "llvm/Support/Debug.h" 30#include "llvm/Support/ErrorHandling.h" 31#include "llvm/Support/raw_ostream.h" 32#include "llvm/Target/TargetInstrInfo.h" 33#include "llvm/Target/TargetLowering.h" 34#include "llvm/Target/TargetMachine.h" 35#include "llvm/Target/TargetRegisterInfo.h" 36#include <climits> 37using namespace llvm; 38 39STATISTIC(NumBacktracks, "Number of times scheduler backtracked"); 40STATISTIC(NumUnfolds, "Number of nodes unfolded"); 41STATISTIC(NumDups, "Number of duplicated nodes"); 42STATISTIC(NumPRCopies, "Number of physical register copies"); 43 44static RegisterScheduler 45 burrListDAGScheduler("list-burr", 46 "Bottom-up register reduction list scheduling", 47 createBURRListDAGScheduler); 48static RegisterScheduler 49 sourceListDAGScheduler("source", 50 "Similar to list-burr but schedules in source " 51 "order when possible", 52 createSourceListDAGScheduler); 53 54static RegisterScheduler 55 hybridListDAGScheduler("list-hybrid", 56 "Bottom-up register pressure aware list scheduling " 57 "which tries to balance latency and register pressure", 58 createHybridListDAGScheduler); 59 60static RegisterScheduler 61 ILPListDAGScheduler("list-ilp", 62 "Bottom-up register pressure aware list scheduling " 63 "which tries to balance ILP and register pressure", 64 createILPListDAGScheduler); 65 66static cl::opt<bool> DisableSchedCycles( 67 "disable-sched-cycles", cl::Hidden, cl::init(false), 68 cl::desc("Disable cycle-level precision during preRA scheduling")); 69 70// Temporary sched=list-ilp flags until the heuristics are robust. 71// Some options are also available under sched=list-hybrid. 72static cl::opt<bool> DisableSchedRegPressure( 73 "disable-sched-reg-pressure", cl::Hidden, cl::init(false), 74 cl::desc("Disable regpressure priority in sched=list-ilp")); 75static cl::opt<bool> DisableSchedLiveUses( 76 "disable-sched-live-uses", cl::Hidden, cl::init(true), 77 cl::desc("Disable live use priority in sched=list-ilp")); 78static cl::opt<bool> DisableSchedVRegCycle( 79 "disable-sched-vrcycle", cl::Hidden, cl::init(false), 80 cl::desc("Disable virtual register cycle interference checks")); 81static cl::opt<bool> DisableSchedPhysRegJoin( 82 "disable-sched-physreg-join", cl::Hidden, cl::init(false), 83 cl::desc("Disable physreg def-use affinity")); 84static cl::opt<bool> DisableSchedStalls( 85 "disable-sched-stalls", cl::Hidden, cl::init(true), 86 cl::desc("Disable no-stall priority in sched=list-ilp")); 87static cl::opt<bool> DisableSchedCriticalPath( 88 "disable-sched-critical-path", cl::Hidden, cl::init(false), 89 cl::desc("Disable critical path priority in sched=list-ilp")); 90static cl::opt<bool> DisableSchedHeight( 91 "disable-sched-height", cl::Hidden, cl::init(false), 92 cl::desc("Disable scheduled-height priority in sched=list-ilp")); 93static cl::opt<bool> Disable2AddrHack( 94 "disable-2addr-hack", cl::Hidden, cl::init(true), 95 cl::desc("Disable scheduler's two-address hack")); 96 97static cl::opt<int> MaxReorderWindow( 98 "max-sched-reorder", cl::Hidden, cl::init(6), 99 cl::desc("Number of instructions to allow ahead of the critical path " 100 "in sched=list-ilp")); 101 102static cl::opt<unsigned> AvgIPC( 103 "sched-avg-ipc", cl::Hidden, cl::init(1), 104 cl::desc("Average inst/cycle whan no target itinerary exists.")); 105 106namespace { 107//===----------------------------------------------------------------------===// 108/// ScheduleDAGRRList - The actual register reduction list scheduler 109/// implementation. This supports both top-down and bottom-up scheduling. 110/// 111class ScheduleDAGRRList : public ScheduleDAGSDNodes { 112private: 113 /// NeedLatency - True if the scheduler will make use of latency information. 114 /// 115 bool NeedLatency; 116 117 /// AvailableQueue - The priority queue to use for the available SUnits. 118 SchedulingPriorityQueue *AvailableQueue; 119 120 /// PendingQueue - This contains all of the instructions whose operands have 121 /// been issued, but their results are not ready yet (due to the latency of 122 /// the operation). Once the operands becomes available, the instruction is 123 /// added to the AvailableQueue. 124 std::vector<SUnit*> PendingQueue; 125 126 /// HazardRec - The hazard recognizer to use. 127 ScheduleHazardRecognizer *HazardRec; 128 129 /// CurCycle - The current scheduler state corresponds to this cycle. 130 unsigned CurCycle; 131 132 /// MinAvailableCycle - Cycle of the soonest available instruction. 133 unsigned MinAvailableCycle; 134 135 /// IssueCount - Count instructions issued in this cycle 136 /// Currently valid only for bottom-up scheduling. 137 unsigned IssueCount; 138 139 /// LiveRegDefs - A set of physical registers and their definition 140 /// that are "live". These nodes must be scheduled before any other nodes that 141 /// modifies the registers can be scheduled. 142 unsigned NumLiveRegs; 143 std::vector<SUnit*> LiveRegDefs; 144 std::vector<SUnit*> LiveRegGens; 145 146 // Collect interferences between physical register use/defs. 147 // Each interference is an SUnit and set of physical registers. 148 SmallVector<SUnit*, 4> Interferences; 149 typedef DenseMap<SUnit*, SmallVector<unsigned, 4> > LRegsMapT; 150 LRegsMapT LRegsMap; 151 152 /// Topo - A topological ordering for SUnits which permits fast IsReachable 153 /// and similar queries. 154 ScheduleDAGTopologicalSort Topo; 155 156 // Hack to keep track of the inverse of FindCallSeqStart without more crazy 157 // DAG crawling. 158 DenseMap<SUnit*, SUnit*> CallSeqEndForStart; 159 160public: 161 ScheduleDAGRRList(MachineFunction &mf, bool needlatency, 162 SchedulingPriorityQueue *availqueue, 163 CodeGenOpt::Level OptLevel) 164 : ScheduleDAGSDNodes(mf), 165 NeedLatency(needlatency), AvailableQueue(availqueue), CurCycle(0), 166 Topo(SUnits, NULL) { 167 168 const TargetMachine &tm = mf.getTarget(); 169 if (DisableSchedCycles || !NeedLatency) 170 HazardRec = new ScheduleHazardRecognizer(); 171 else 172 HazardRec = tm.getInstrInfo()->CreateTargetHazardRecognizer(&tm, this); 173 } 174 175 ~ScheduleDAGRRList() { 176 delete HazardRec; 177 delete AvailableQueue; 178 } 179 180 void Schedule(); 181 182 ScheduleHazardRecognizer *getHazardRec() { return HazardRec; } 183 184 /// IsReachable - Checks if SU is reachable from TargetSU. 185 bool IsReachable(const SUnit *SU, const SUnit *TargetSU) { 186 return Topo.IsReachable(SU, TargetSU); 187 } 188 189 /// WillCreateCycle - Returns true if adding an edge from SU to TargetSU will 190 /// create a cycle. 191 bool WillCreateCycle(SUnit *SU, SUnit *TargetSU) { 192 return Topo.WillCreateCycle(SU, TargetSU); 193 } 194 195 /// AddPred - adds a predecessor edge to SUnit SU. 196 /// This returns true if this is a new predecessor. 197 /// Updates the topological ordering if required. 198 void AddPred(SUnit *SU, const SDep &D) { 199 Topo.AddPred(SU, D.getSUnit()); 200 SU->addPred(D); 201 } 202 203 /// RemovePred - removes a predecessor edge from SUnit SU. 204 /// This returns true if an edge was removed. 205 /// Updates the topological ordering if required. 206 void RemovePred(SUnit *SU, const SDep &D) { 207 Topo.RemovePred(SU, D.getSUnit()); 208 SU->removePred(D); 209 } 210 211private: 212 bool isReady(SUnit *SU) { 213 return DisableSchedCycles || !AvailableQueue->hasReadyFilter() || 214 AvailableQueue->isReady(SU); 215 } 216 217 void ReleasePred(SUnit *SU, const SDep *PredEdge); 218 void ReleasePredecessors(SUnit *SU); 219 void ReleasePending(); 220 void AdvanceToCycle(unsigned NextCycle); 221 void AdvancePastStalls(SUnit *SU); 222 void EmitNode(SUnit *SU); 223 void ScheduleNodeBottomUp(SUnit*); 224 void CapturePred(SDep *PredEdge); 225 void UnscheduleNodeBottomUp(SUnit*); 226 void RestoreHazardCheckerBottomUp(); 227 void BacktrackBottomUp(SUnit*, SUnit*); 228 SUnit *CopyAndMoveSuccessors(SUnit*); 229 void InsertCopiesAndMoveSuccs(SUnit*, unsigned, 230 const TargetRegisterClass*, 231 const TargetRegisterClass*, 232 SmallVectorImpl<SUnit*>&); 233 bool DelayForLiveRegsBottomUp(SUnit*, SmallVectorImpl<unsigned>&); 234 235 void releaseInterferences(unsigned Reg = 0); 236 237 SUnit *PickNodeToScheduleBottomUp(); 238 void ListScheduleBottomUp(); 239 240 /// CreateNewSUnit - Creates a new SUnit and returns a pointer to it. 241 /// Updates the topological ordering if required. 242 SUnit *CreateNewSUnit(SDNode *N) { 243 unsigned NumSUnits = SUnits.size(); 244 SUnit *NewNode = newSUnit(N); 245 // Update the topological ordering. 246 if (NewNode->NodeNum >= NumSUnits) 247 Topo.InitDAGTopologicalSorting(); 248 return NewNode; 249 } 250 251 /// CreateClone - Creates a new SUnit from an existing one. 252 /// Updates the topological ordering if required. 253 SUnit *CreateClone(SUnit *N) { 254 unsigned NumSUnits = SUnits.size(); 255 SUnit *NewNode = Clone(N); 256 // Update the topological ordering. 257 if (NewNode->NodeNum >= NumSUnits) 258 Topo.InitDAGTopologicalSorting(); 259 return NewNode; 260 } 261 262 /// forceUnitLatencies - Register-pressure-reducing scheduling doesn't 263 /// need actual latency information but the hybrid scheduler does. 264 bool forceUnitLatencies() const { 265 return !NeedLatency; 266 } 267}; 268} // end anonymous namespace 269 270/// GetCostForDef - Looks up the register class and cost for a given definition. 271/// Typically this just means looking up the representative register class, 272/// but for untyped values (MVT::Untyped) it means inspecting the node's 273/// opcode to determine what register class is being generated. 274static void GetCostForDef(const ScheduleDAGSDNodes::RegDefIter &RegDefPos, 275 const TargetLowering *TLI, 276 const TargetInstrInfo *TII, 277 const TargetRegisterInfo *TRI, 278 unsigned &RegClass, unsigned &Cost, 279 const MachineFunction &MF) { 280 MVT VT = RegDefPos.GetValue(); 281 282 // Special handling for untyped values. These values can only come from 283 // the expansion of custom DAG-to-DAG patterns. 284 if (VT == MVT::Untyped) { 285 const SDNode *Node = RegDefPos.GetNode(); 286 287 // Special handling for CopyFromReg of untyped values. 288 if (!Node->isMachineOpcode() && Node->getOpcode() == ISD::CopyFromReg) { 289 unsigned Reg = cast<RegisterSDNode>(Node->getOperand(1))->getReg(); 290 const TargetRegisterClass *RC = MF.getRegInfo().getRegClass(Reg); 291 RegClass = RC->getID(); 292 Cost = 1; 293 return; 294 } 295 296 unsigned Opcode = Node->getMachineOpcode(); 297 if (Opcode == TargetOpcode::REG_SEQUENCE) { 298 unsigned DstRCIdx = cast<ConstantSDNode>(Node->getOperand(0))->getZExtValue(); 299 const TargetRegisterClass *RC = TRI->getRegClass(DstRCIdx); 300 RegClass = RC->getID(); 301 Cost = 1; 302 return; 303 } 304 305 unsigned Idx = RegDefPos.GetIdx(); 306 const MCInstrDesc Desc = TII->get(Opcode); 307 const TargetRegisterClass *RC = TII->getRegClass(Desc, Idx, TRI, MF); 308 RegClass = RC->getID(); 309 // FIXME: Cost arbitrarily set to 1 because there doesn't seem to be a 310 // better way to determine it. 311 Cost = 1; 312 } else { 313 RegClass = TLI->getRepRegClassFor(VT)->getID(); 314 Cost = TLI->getRepRegClassCostFor(VT); 315 } 316} 317 318/// Schedule - Schedule the DAG using list scheduling. 319void ScheduleDAGRRList::Schedule() { 320 DEBUG(dbgs() 321 << "********** List Scheduling BB#" << BB->getNumber() 322 << " '" << BB->getName() << "' **********\n"); 323 324 CurCycle = 0; 325 IssueCount = 0; 326 MinAvailableCycle = DisableSchedCycles ? 0 : UINT_MAX; 327 NumLiveRegs = 0; 328 // Allocate slots for each physical register, plus one for a special register 329 // to track the virtual resource of a calling sequence. 330 LiveRegDefs.resize(TRI->getNumRegs() + 1, NULL); 331 LiveRegGens.resize(TRI->getNumRegs() + 1, NULL); 332 CallSeqEndForStart.clear(); 333 assert(Interferences.empty() && LRegsMap.empty() && "stale Interferences"); 334 335 // Build the scheduling graph. 336 BuildSchedGraph(NULL); 337 338 DEBUG(for (unsigned su = 0, e = SUnits.size(); su != e; ++su) 339 SUnits[su].dumpAll(this)); 340 Topo.InitDAGTopologicalSorting(); 341 342 AvailableQueue->initNodes(SUnits); 343 344 HazardRec->Reset(); 345 346 // Execute the actual scheduling loop. 347 ListScheduleBottomUp(); 348 349 AvailableQueue->releaseState(); 350 351 DEBUG({ 352 dbgs() << "*** Final schedule ***\n"; 353 dumpSchedule(); 354 dbgs() << '\n'; 355 }); 356} 357 358//===----------------------------------------------------------------------===// 359// Bottom-Up Scheduling 360//===----------------------------------------------------------------------===// 361 362/// ReleasePred - Decrement the NumSuccsLeft count of a predecessor. Add it to 363/// the AvailableQueue if the count reaches zero. Also update its cycle bound. 364void ScheduleDAGRRList::ReleasePred(SUnit *SU, const SDep *PredEdge) { 365 SUnit *PredSU = PredEdge->getSUnit(); 366 367#ifndef NDEBUG 368 if (PredSU->NumSuccsLeft == 0) { 369 dbgs() << "*** Scheduling failed! ***\n"; 370 PredSU->dump(this); 371 dbgs() << " has been released too many times!\n"; 372 llvm_unreachable(0); 373 } 374#endif 375 --PredSU->NumSuccsLeft; 376 377 if (!forceUnitLatencies()) { 378 // Updating predecessor's height. This is now the cycle when the 379 // predecessor can be scheduled without causing a pipeline stall. 380 PredSU->setHeightToAtLeast(SU->getHeight() + PredEdge->getLatency()); 381 } 382 383 // If all the node's successors are scheduled, this node is ready 384 // to be scheduled. Ignore the special EntrySU node. 385 if (PredSU->NumSuccsLeft == 0 && PredSU != &EntrySU) { 386 PredSU->isAvailable = true; 387 388 unsigned Height = PredSU->getHeight(); 389 if (Height < MinAvailableCycle) 390 MinAvailableCycle = Height; 391 392 if (isReady(PredSU)) { 393 AvailableQueue->push(PredSU); 394 } 395 // CapturePred and others may have left the node in the pending queue, avoid 396 // adding it twice. 397 else if (!PredSU->isPending) { 398 PredSU->isPending = true; 399 PendingQueue.push_back(PredSU); 400 } 401 } 402} 403 404/// IsChainDependent - Test if Outer is reachable from Inner through 405/// chain dependencies. 406static bool IsChainDependent(SDNode *Outer, SDNode *Inner, 407 unsigned NestLevel, 408 const TargetInstrInfo *TII) { 409 SDNode *N = Outer; 410 for (;;) { 411 if (N == Inner) 412 return true; 413 // For a TokenFactor, examine each operand. There may be multiple ways 414 // to get to the CALLSEQ_BEGIN, but we need to find the path with the 415 // most nesting in order to ensure that we find the corresponding match. 416 if (N->getOpcode() == ISD::TokenFactor) { 417 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) 418 if (IsChainDependent(N->getOperand(i).getNode(), Inner, NestLevel, TII)) 419 return true; 420 return false; 421 } 422 // Check for a lowered CALLSEQ_BEGIN or CALLSEQ_END. 423 if (N->isMachineOpcode()) { 424 if (N->getMachineOpcode() == 425 (unsigned)TII->getCallFrameDestroyOpcode()) { 426 ++NestLevel; 427 } else if (N->getMachineOpcode() == 428 (unsigned)TII->getCallFrameSetupOpcode()) { 429 if (NestLevel == 0) 430 return false; 431 --NestLevel; 432 } 433 } 434 // Otherwise, find the chain and continue climbing. 435 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) 436 if (N->getOperand(i).getValueType() == MVT::Other) { 437 N = N->getOperand(i).getNode(); 438 goto found_chain_operand; 439 } 440 return false; 441 found_chain_operand:; 442 if (N->getOpcode() == ISD::EntryToken) 443 return false; 444 } 445} 446 447/// FindCallSeqStart - Starting from the (lowered) CALLSEQ_END node, locate 448/// the corresponding (lowered) CALLSEQ_BEGIN node. 449/// 450/// NestLevel and MaxNested are used in recursion to indcate the current level 451/// of nesting of CALLSEQ_BEGIN and CALLSEQ_END pairs, as well as the maximum 452/// level seen so far. 453/// 454/// TODO: It would be better to give CALLSEQ_END an explicit operand to point 455/// to the corresponding CALLSEQ_BEGIN to avoid needing to search for it. 456static SDNode * 457FindCallSeqStart(SDNode *N, unsigned &NestLevel, unsigned &MaxNest, 458 const TargetInstrInfo *TII) { 459 for (;;) { 460 // For a TokenFactor, examine each operand. There may be multiple ways 461 // to get to the CALLSEQ_BEGIN, but we need to find the path with the 462 // most nesting in order to ensure that we find the corresponding match. 463 if (N->getOpcode() == ISD::TokenFactor) { 464 SDNode *Best = 0; 465 unsigned BestMaxNest = MaxNest; 466 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) { 467 unsigned MyNestLevel = NestLevel; 468 unsigned MyMaxNest = MaxNest; 469 if (SDNode *New = FindCallSeqStart(N->getOperand(i).getNode(), 470 MyNestLevel, MyMaxNest, TII)) 471 if (!Best || (MyMaxNest > BestMaxNest)) { 472 Best = New; 473 BestMaxNest = MyMaxNest; 474 } 475 } 476 assert(Best); 477 MaxNest = BestMaxNest; 478 return Best; 479 } 480 // Check for a lowered CALLSEQ_BEGIN or CALLSEQ_END. 481 if (N->isMachineOpcode()) { 482 if (N->getMachineOpcode() == 483 (unsigned)TII->getCallFrameDestroyOpcode()) { 484 ++NestLevel; 485 MaxNest = std::max(MaxNest, NestLevel); 486 } else if (N->getMachineOpcode() == 487 (unsigned)TII->getCallFrameSetupOpcode()) { 488 assert(NestLevel != 0); 489 --NestLevel; 490 if (NestLevel == 0) 491 return N; 492 } 493 } 494 // Otherwise, find the chain and continue climbing. 495 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) 496 if (N->getOperand(i).getValueType() == MVT::Other) { 497 N = N->getOperand(i).getNode(); 498 goto found_chain_operand; 499 } 500 return 0; 501 found_chain_operand:; 502 if (N->getOpcode() == ISD::EntryToken) 503 return 0; 504 } 505} 506 507/// Call ReleasePred for each predecessor, then update register live def/gen. 508/// Always update LiveRegDefs for a register dependence even if the current SU 509/// also defines the register. This effectively create one large live range 510/// across a sequence of two-address node. This is important because the 511/// entire chain must be scheduled together. Example: 512/// 513/// flags = (3) add 514/// flags = (2) addc flags 515/// flags = (1) addc flags 516/// 517/// results in 518/// 519/// LiveRegDefs[flags] = 3 520/// LiveRegGens[flags] = 1 521/// 522/// If (2) addc is unscheduled, then (1) addc must also be unscheduled to avoid 523/// interference on flags. 524void ScheduleDAGRRList::ReleasePredecessors(SUnit *SU) { 525 // Bottom up: release predecessors 526 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 527 I != E; ++I) { 528 ReleasePred(SU, &*I); 529 if (I->isAssignedRegDep()) { 530 // This is a physical register dependency and it's impossible or 531 // expensive to copy the register. Make sure nothing that can 532 // clobber the register is scheduled between the predecessor and 533 // this node. 534 SUnit *RegDef = LiveRegDefs[I->getReg()]; (void)RegDef; 535 assert((!RegDef || RegDef == SU || RegDef == I->getSUnit()) && 536 "interference on register dependence"); 537 LiveRegDefs[I->getReg()] = I->getSUnit(); 538 if (!LiveRegGens[I->getReg()]) { 539 ++NumLiveRegs; 540 LiveRegGens[I->getReg()] = SU; 541 } 542 } 543 } 544 545 // If we're scheduling a lowered CALLSEQ_END, find the corresponding 546 // CALLSEQ_BEGIN. Inject an artificial physical register dependence between 547 // these nodes, to prevent other calls from being interscheduled with them. 548 unsigned CallResource = TRI->getNumRegs(); 549 if (!LiveRegDefs[CallResource]) 550 for (SDNode *Node = SU->getNode(); Node; Node = Node->getGluedNode()) 551 if (Node->isMachineOpcode() && 552 Node->getMachineOpcode() == (unsigned)TII->getCallFrameDestroyOpcode()) { 553 unsigned NestLevel = 0; 554 unsigned MaxNest = 0; 555 SDNode *N = FindCallSeqStart(Node, NestLevel, MaxNest, TII); 556 557 SUnit *Def = &SUnits[N->getNodeId()]; 558 CallSeqEndForStart[Def] = SU; 559 560 ++NumLiveRegs; 561 LiveRegDefs[CallResource] = Def; 562 LiveRegGens[CallResource] = SU; 563 break; 564 } 565} 566 567/// Check to see if any of the pending instructions are ready to issue. If 568/// so, add them to the available queue. 569void ScheduleDAGRRList::ReleasePending() { 570 if (DisableSchedCycles) { 571 assert(PendingQueue.empty() && "pending instrs not allowed in this mode"); 572 return; 573 } 574 575 // If the available queue is empty, it is safe to reset MinAvailableCycle. 576 if (AvailableQueue->empty()) 577 MinAvailableCycle = UINT_MAX; 578 579 // Check to see if any of the pending instructions are ready to issue. If 580 // so, add them to the available queue. 581 for (unsigned i = 0, e = PendingQueue.size(); i != e; ++i) { 582 unsigned ReadyCycle = PendingQueue[i]->getHeight(); 583 if (ReadyCycle < MinAvailableCycle) 584 MinAvailableCycle = ReadyCycle; 585 586 if (PendingQueue[i]->isAvailable) { 587 if (!isReady(PendingQueue[i])) 588 continue; 589 AvailableQueue->push(PendingQueue[i]); 590 } 591 PendingQueue[i]->isPending = false; 592 PendingQueue[i] = PendingQueue.back(); 593 PendingQueue.pop_back(); 594 --i; --e; 595 } 596} 597 598/// Move the scheduler state forward by the specified number of Cycles. 599void ScheduleDAGRRList::AdvanceToCycle(unsigned NextCycle) { 600 if (NextCycle <= CurCycle) 601 return; 602 603 IssueCount = 0; 604 AvailableQueue->setCurCycle(NextCycle); 605 if (!HazardRec->isEnabled()) { 606 // Bypass lots of virtual calls in case of long latency. 607 CurCycle = NextCycle; 608 } 609 else { 610 for (; CurCycle != NextCycle; ++CurCycle) { 611 HazardRec->RecedeCycle(); 612 } 613 } 614 // FIXME: Instead of visiting the pending Q each time, set a dirty flag on the 615 // available Q to release pending nodes at least once before popping. 616 ReleasePending(); 617} 618 619/// Move the scheduler state forward until the specified node's dependents are 620/// ready and can be scheduled with no resource conflicts. 621void ScheduleDAGRRList::AdvancePastStalls(SUnit *SU) { 622 if (DisableSchedCycles) 623 return; 624 625 // FIXME: Nodes such as CopyFromReg probably should not advance the current 626 // cycle. Otherwise, we can wrongly mask real stalls. If the non-machine node 627 // has predecessors the cycle will be advanced when they are scheduled. 628 // But given the crude nature of modeling latency though such nodes, we 629 // currently need to treat these nodes like real instructions. 630 // if (!SU->getNode() || !SU->getNode()->isMachineOpcode()) return; 631 632 unsigned ReadyCycle = SU->getHeight(); 633 634 // Bump CurCycle to account for latency. We assume the latency of other 635 // available instructions may be hidden by the stall (not a full pipe stall). 636 // This updates the hazard recognizer's cycle before reserving resources for 637 // this instruction. 638 AdvanceToCycle(ReadyCycle); 639 640 // Calls are scheduled in their preceding cycle, so don't conflict with 641 // hazards from instructions after the call. EmitNode will reset the 642 // scoreboard state before emitting the call. 643 if (SU->isCall) 644 return; 645 646 // FIXME: For resource conflicts in very long non-pipelined stages, we 647 // should probably skip ahead here to avoid useless scoreboard checks. 648 int Stalls = 0; 649 while (true) { 650 ScheduleHazardRecognizer::HazardType HT = 651 HazardRec->getHazardType(SU, -Stalls); 652 653 if (HT == ScheduleHazardRecognizer::NoHazard) 654 break; 655 656 ++Stalls; 657 } 658 AdvanceToCycle(CurCycle + Stalls); 659} 660 661/// Record this SUnit in the HazardRecognizer. 662/// Does not update CurCycle. 663void ScheduleDAGRRList::EmitNode(SUnit *SU) { 664 if (!HazardRec->isEnabled()) 665 return; 666 667 // Check for phys reg copy. 668 if (!SU->getNode()) 669 return; 670 671 switch (SU->getNode()->getOpcode()) { 672 default: 673 assert(SU->getNode()->isMachineOpcode() && 674 "This target-independent node should not be scheduled."); 675 break; 676 case ISD::MERGE_VALUES: 677 case ISD::TokenFactor: 678 case ISD::LIFETIME_START: 679 case ISD::LIFETIME_END: 680 case ISD::CopyToReg: 681 case ISD::CopyFromReg: 682 case ISD::EH_LABEL: 683 // Noops don't affect the scoreboard state. Copies are likely to be 684 // removed. 685 return; 686 case ISD::INLINEASM: 687 // For inline asm, clear the pipeline state. 688 HazardRec->Reset(); 689 return; 690 } 691 if (SU->isCall) { 692 // Calls are scheduled with their preceding instructions. For bottom-up 693 // scheduling, clear the pipeline state before emitting. 694 HazardRec->Reset(); 695 } 696 697 HazardRec->EmitInstruction(SU); 698} 699 700static void resetVRegCycle(SUnit *SU); 701 702/// ScheduleNodeBottomUp - Add the node to the schedule. Decrement the pending 703/// count of its predecessors. If a predecessor pending count is zero, add it to 704/// the Available queue. 705void ScheduleDAGRRList::ScheduleNodeBottomUp(SUnit *SU) { 706 DEBUG(dbgs() << "\n*** Scheduling [" << CurCycle << "]: "); 707 DEBUG(SU->dump(this)); 708 709#ifndef NDEBUG 710 if (CurCycle < SU->getHeight()) 711 DEBUG(dbgs() << " Height [" << SU->getHeight() 712 << "] pipeline stall!\n"); 713#endif 714 715 // FIXME: Do not modify node height. It may interfere with 716 // backtracking. Instead add a "ready cycle" to SUnit. Before scheduling the 717 // node its ready cycle can aid heuristics, and after scheduling it can 718 // indicate the scheduled cycle. 719 SU->setHeightToAtLeast(CurCycle); 720 721 // Reserve resources for the scheduled intruction. 722 EmitNode(SU); 723 724 Sequence.push_back(SU); 725 726 AvailableQueue->scheduledNode(SU); 727 728 // If HazardRec is disabled, and each inst counts as one cycle, then 729 // advance CurCycle before ReleasePredecessors to avoid useless pushes to 730 // PendingQueue for schedulers that implement HasReadyFilter. 731 if (!HazardRec->isEnabled() && AvgIPC < 2) 732 AdvanceToCycle(CurCycle + 1); 733 734 // Update liveness of predecessors before successors to avoid treating a 735 // two-address node as a live range def. 736 ReleasePredecessors(SU); 737 738 // Release all the implicit physical register defs that are live. 739 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 740 I != E; ++I) { 741 // LiveRegDegs[I->getReg()] != SU when SU is a two-address node. 742 if (I->isAssignedRegDep() && LiveRegDefs[I->getReg()] == SU) { 743 assert(NumLiveRegs > 0 && "NumLiveRegs is already zero!"); 744 --NumLiveRegs; 745 LiveRegDefs[I->getReg()] = NULL; 746 LiveRegGens[I->getReg()] = NULL; 747 releaseInterferences(I->getReg()); 748 } 749 } 750 // Release the special call resource dependence, if this is the beginning 751 // of a call. 752 unsigned CallResource = TRI->getNumRegs(); 753 if (LiveRegDefs[CallResource] == SU) 754 for (const SDNode *SUNode = SU->getNode(); SUNode; 755 SUNode = SUNode->getGluedNode()) { 756 if (SUNode->isMachineOpcode() && 757 SUNode->getMachineOpcode() == (unsigned)TII->getCallFrameSetupOpcode()) { 758 assert(NumLiveRegs > 0 && "NumLiveRegs is already zero!"); 759 --NumLiveRegs; 760 LiveRegDefs[CallResource] = NULL; 761 LiveRegGens[CallResource] = NULL; 762 releaseInterferences(CallResource); 763 } 764 } 765 766 resetVRegCycle(SU); 767 768 SU->isScheduled = true; 769 770 // Conditions under which the scheduler should eagerly advance the cycle: 771 // (1) No available instructions 772 // (2) All pipelines full, so available instructions must have hazards. 773 // 774 // If HazardRec is disabled, the cycle was pre-advanced before calling 775 // ReleasePredecessors. In that case, IssueCount should remain 0. 776 // 777 // Check AvailableQueue after ReleasePredecessors in case of zero latency. 778 if (HazardRec->isEnabled() || AvgIPC > 1) { 779 if (SU->getNode() && SU->getNode()->isMachineOpcode()) 780 ++IssueCount; 781 if ((HazardRec->isEnabled() && HazardRec->atIssueLimit()) 782 || (!HazardRec->isEnabled() && IssueCount == AvgIPC)) 783 AdvanceToCycle(CurCycle + 1); 784 } 785} 786 787/// CapturePred - This does the opposite of ReleasePred. Since SU is being 788/// unscheduled, incrcease the succ left count of its predecessors. Remove 789/// them from AvailableQueue if necessary. 790void ScheduleDAGRRList::CapturePred(SDep *PredEdge) { 791 SUnit *PredSU = PredEdge->getSUnit(); 792 if (PredSU->isAvailable) { 793 PredSU->isAvailable = false; 794 if (!PredSU->isPending) 795 AvailableQueue->remove(PredSU); 796 } 797 798 assert(PredSU->NumSuccsLeft < UINT_MAX && "NumSuccsLeft will overflow!"); 799 ++PredSU->NumSuccsLeft; 800} 801 802/// UnscheduleNodeBottomUp - Remove the node from the schedule, update its and 803/// its predecessor states to reflect the change. 804void ScheduleDAGRRList::UnscheduleNodeBottomUp(SUnit *SU) { 805 DEBUG(dbgs() << "*** Unscheduling [" << SU->getHeight() << "]: "); 806 DEBUG(SU->dump(this)); 807 808 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 809 I != E; ++I) { 810 CapturePred(&*I); 811 if (I->isAssignedRegDep() && SU == LiveRegGens[I->getReg()]){ 812 assert(NumLiveRegs > 0 && "NumLiveRegs is already zero!"); 813 assert(LiveRegDefs[I->getReg()] == I->getSUnit() && 814 "Physical register dependency violated?"); 815 --NumLiveRegs; 816 LiveRegDefs[I->getReg()] = NULL; 817 LiveRegGens[I->getReg()] = NULL; 818 releaseInterferences(I->getReg()); 819 } 820 } 821 822 // Reclaim the special call resource dependence, if this is the beginning 823 // of a call. 824 unsigned CallResource = TRI->getNumRegs(); 825 for (const SDNode *SUNode = SU->getNode(); SUNode; 826 SUNode = SUNode->getGluedNode()) { 827 if (SUNode->isMachineOpcode() && 828 SUNode->getMachineOpcode() == (unsigned)TII->getCallFrameSetupOpcode()) { 829 ++NumLiveRegs; 830 LiveRegDefs[CallResource] = SU; 831 LiveRegGens[CallResource] = CallSeqEndForStart[SU]; 832 } 833 } 834 835 // Release the special call resource dependence, if this is the end 836 // of a call. 837 if (LiveRegGens[CallResource] == SU) 838 for (const SDNode *SUNode = SU->getNode(); SUNode; 839 SUNode = SUNode->getGluedNode()) { 840 if (SUNode->isMachineOpcode() && 841 SUNode->getMachineOpcode() == (unsigned)TII->getCallFrameDestroyOpcode()) { 842 assert(NumLiveRegs > 0 && "NumLiveRegs is already zero!"); 843 --NumLiveRegs; 844 LiveRegDefs[CallResource] = NULL; 845 LiveRegGens[CallResource] = NULL; 846 releaseInterferences(CallResource); 847 } 848 } 849 850 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 851 I != E; ++I) { 852 if (I->isAssignedRegDep()) { 853 if (!LiveRegDefs[I->getReg()]) 854 ++NumLiveRegs; 855 // This becomes the nearest def. Note that an earlier def may still be 856 // pending if this is a two-address node. 857 LiveRegDefs[I->getReg()] = SU; 858 if (LiveRegGens[I->getReg()] == NULL || 859 I->getSUnit()->getHeight() < LiveRegGens[I->getReg()]->getHeight()) 860 LiveRegGens[I->getReg()] = I->getSUnit(); 861 } 862 } 863 if (SU->getHeight() < MinAvailableCycle) 864 MinAvailableCycle = SU->getHeight(); 865 866 SU->setHeightDirty(); 867 SU->isScheduled = false; 868 SU->isAvailable = true; 869 if (!DisableSchedCycles && AvailableQueue->hasReadyFilter()) { 870 // Don't make available until backtracking is complete. 871 SU->isPending = true; 872 PendingQueue.push_back(SU); 873 } 874 else { 875 AvailableQueue->push(SU); 876 } 877 AvailableQueue->unscheduledNode(SU); 878} 879 880/// After backtracking, the hazard checker needs to be restored to a state 881/// corresponding the current cycle. 882void ScheduleDAGRRList::RestoreHazardCheckerBottomUp() { 883 HazardRec->Reset(); 884 885 unsigned LookAhead = std::min((unsigned)Sequence.size(), 886 HazardRec->getMaxLookAhead()); 887 if (LookAhead == 0) 888 return; 889 890 std::vector<SUnit*>::const_iterator I = (Sequence.end() - LookAhead); 891 unsigned HazardCycle = (*I)->getHeight(); 892 for (std::vector<SUnit*>::const_iterator E = Sequence.end(); I != E; ++I) { 893 SUnit *SU = *I; 894 for (; SU->getHeight() > HazardCycle; ++HazardCycle) { 895 HazardRec->RecedeCycle(); 896 } 897 EmitNode(SU); 898 } 899} 900 901/// BacktrackBottomUp - Backtrack scheduling to a previous cycle specified in 902/// BTCycle in order to schedule a specific node. 903void ScheduleDAGRRList::BacktrackBottomUp(SUnit *SU, SUnit *BtSU) { 904 SUnit *OldSU = Sequence.back(); 905 while (true) { 906 Sequence.pop_back(); 907 // FIXME: use ready cycle instead of height 908 CurCycle = OldSU->getHeight(); 909 UnscheduleNodeBottomUp(OldSU); 910 AvailableQueue->setCurCycle(CurCycle); 911 if (OldSU == BtSU) 912 break; 913 OldSU = Sequence.back(); 914 } 915 916 assert(!SU->isSucc(OldSU) && "Something is wrong!"); 917 918 RestoreHazardCheckerBottomUp(); 919 920 ReleasePending(); 921 922 ++NumBacktracks; 923} 924 925static bool isOperandOf(const SUnit *SU, SDNode *N) { 926 for (const SDNode *SUNode = SU->getNode(); SUNode; 927 SUNode = SUNode->getGluedNode()) { 928 if (SUNode->isOperandOf(N)) 929 return true; 930 } 931 return false; 932} 933 934/// CopyAndMoveSuccessors - Clone the specified node and move its scheduled 935/// successors to the newly created node. 936SUnit *ScheduleDAGRRList::CopyAndMoveSuccessors(SUnit *SU) { 937 SDNode *N = SU->getNode(); 938 if (!N) 939 return NULL; 940 941 if (SU->getNode()->getGluedNode()) 942 return NULL; 943 944 SUnit *NewSU; 945 bool TryUnfold = false; 946 for (unsigned i = 0, e = N->getNumValues(); i != e; ++i) { 947 EVT VT = N->getValueType(i); 948 if (VT == MVT::Glue) 949 return NULL; 950 else if (VT == MVT::Other) 951 TryUnfold = true; 952 } 953 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) { 954 const SDValue &Op = N->getOperand(i); 955 EVT VT = Op.getNode()->getValueType(Op.getResNo()); 956 if (VT == MVT::Glue) 957 return NULL; 958 } 959 960 if (TryUnfold) { 961 SmallVector<SDNode*, 2> NewNodes; 962 if (!TII->unfoldMemoryOperand(*DAG, N, NewNodes)) 963 return NULL; 964 965 // unfolding an x86 DEC64m operation results in store, dec, load which 966 // can't be handled here so quit 967 if (NewNodes.size() == 3) 968 return NULL; 969 970 DEBUG(dbgs() << "Unfolding SU #" << SU->NodeNum << "\n"); 971 assert(NewNodes.size() == 2 && "Expected a load folding node!"); 972 973 N = NewNodes[1]; 974 SDNode *LoadNode = NewNodes[0]; 975 unsigned NumVals = N->getNumValues(); 976 unsigned OldNumVals = SU->getNode()->getNumValues(); 977 for (unsigned i = 0; i != NumVals; ++i) 978 DAG->ReplaceAllUsesOfValueWith(SDValue(SU->getNode(), i), SDValue(N, i)); 979 DAG->ReplaceAllUsesOfValueWith(SDValue(SU->getNode(), OldNumVals-1), 980 SDValue(LoadNode, 1)); 981 982 // LoadNode may already exist. This can happen when there is another 983 // load from the same location and producing the same type of value 984 // but it has different alignment or volatileness. 985 bool isNewLoad = true; 986 SUnit *LoadSU; 987 if (LoadNode->getNodeId() != -1) { 988 LoadSU = &SUnits[LoadNode->getNodeId()]; 989 isNewLoad = false; 990 } else { 991 LoadSU = CreateNewSUnit(LoadNode); 992 LoadNode->setNodeId(LoadSU->NodeNum); 993 994 InitNumRegDefsLeft(LoadSU); 995 computeLatency(LoadSU); 996 } 997 998 SUnit *NewSU = CreateNewSUnit(N); 999 assert(N->getNodeId() == -1 && "Node already inserted!"); 1000 N->setNodeId(NewSU->NodeNum); 1001 1002 const MCInstrDesc &MCID = TII->get(N->getMachineOpcode()); 1003 for (unsigned i = 0; i != MCID.getNumOperands(); ++i) { 1004 if (MCID.getOperandConstraint(i, MCOI::TIED_TO) != -1) { 1005 NewSU->isTwoAddress = true; 1006 break; 1007 } 1008 } 1009 if (MCID.isCommutable()) 1010 NewSU->isCommutable = true; 1011 1012 InitNumRegDefsLeft(NewSU); 1013 computeLatency(NewSU); 1014 1015 // Record all the edges to and from the old SU, by category. 1016 SmallVector<SDep, 4> ChainPreds; 1017 SmallVector<SDep, 4> ChainSuccs; 1018 SmallVector<SDep, 4> LoadPreds; 1019 SmallVector<SDep, 4> NodePreds; 1020 SmallVector<SDep, 4> NodeSuccs; 1021 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 1022 I != E; ++I) { 1023 if (I->isCtrl()) 1024 ChainPreds.push_back(*I); 1025 else if (isOperandOf(I->getSUnit(), LoadNode)) 1026 LoadPreds.push_back(*I); 1027 else 1028 NodePreds.push_back(*I); 1029 } 1030 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 1031 I != E; ++I) { 1032 if (I->isCtrl()) 1033 ChainSuccs.push_back(*I); 1034 else 1035 NodeSuccs.push_back(*I); 1036 } 1037 1038 // Now assign edges to the newly-created nodes. 1039 for (unsigned i = 0, e = ChainPreds.size(); i != e; ++i) { 1040 const SDep &Pred = ChainPreds[i]; 1041 RemovePred(SU, Pred); 1042 if (isNewLoad) 1043 AddPred(LoadSU, Pred); 1044 } 1045 for (unsigned i = 0, e = LoadPreds.size(); i != e; ++i) { 1046 const SDep &Pred = LoadPreds[i]; 1047 RemovePred(SU, Pred); 1048 if (isNewLoad) 1049 AddPred(LoadSU, Pred); 1050 } 1051 for (unsigned i = 0, e = NodePreds.size(); i != e; ++i) { 1052 const SDep &Pred = NodePreds[i]; 1053 RemovePred(SU, Pred); 1054 AddPred(NewSU, Pred); 1055 } 1056 for (unsigned i = 0, e = NodeSuccs.size(); i != e; ++i) { 1057 SDep D = NodeSuccs[i]; 1058 SUnit *SuccDep = D.getSUnit(); 1059 D.setSUnit(SU); 1060 RemovePred(SuccDep, D); 1061 D.setSUnit(NewSU); 1062 AddPred(SuccDep, D); 1063 // Balance register pressure. 1064 if (AvailableQueue->tracksRegPressure() && SuccDep->isScheduled 1065 && !D.isCtrl() && NewSU->NumRegDefsLeft > 0) 1066 --NewSU->NumRegDefsLeft; 1067 } 1068 for (unsigned i = 0, e = ChainSuccs.size(); i != e; ++i) { 1069 SDep D = ChainSuccs[i]; 1070 SUnit *SuccDep = D.getSUnit(); 1071 D.setSUnit(SU); 1072 RemovePred(SuccDep, D); 1073 if (isNewLoad) { 1074 D.setSUnit(LoadSU); 1075 AddPred(SuccDep, D); 1076 } 1077 } 1078 1079 // Add a data dependency to reflect that NewSU reads the value defined 1080 // by LoadSU. 1081 SDep D(LoadSU, SDep::Data, 0); 1082 D.setLatency(LoadSU->Latency); 1083 AddPred(NewSU, D); 1084 1085 if (isNewLoad) 1086 AvailableQueue->addNode(LoadSU); 1087 AvailableQueue->addNode(NewSU); 1088 1089 ++NumUnfolds; 1090 1091 if (NewSU->NumSuccsLeft == 0) { 1092 NewSU->isAvailable = true; 1093 return NewSU; 1094 } 1095 SU = NewSU; 1096 } 1097 1098 DEBUG(dbgs() << " Duplicating SU #" << SU->NodeNum << "\n"); 1099 NewSU = CreateClone(SU); 1100 1101 // New SUnit has the exact same predecessors. 1102 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 1103 I != E; ++I) 1104 if (!I->isArtificial()) 1105 AddPred(NewSU, *I); 1106 1107 // Only copy scheduled successors. Cut them from old node's successor 1108 // list and move them over. 1109 SmallVector<std::pair<SUnit *, SDep>, 4> DelDeps; 1110 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 1111 I != E; ++I) { 1112 if (I->isArtificial()) 1113 continue; 1114 SUnit *SuccSU = I->getSUnit(); 1115 if (SuccSU->isScheduled) { 1116 SDep D = *I; 1117 D.setSUnit(NewSU); 1118 AddPred(SuccSU, D); 1119 D.setSUnit(SU); 1120 DelDeps.push_back(std::make_pair(SuccSU, D)); 1121 } 1122 } 1123 for (unsigned i = 0, e = DelDeps.size(); i != e; ++i) 1124 RemovePred(DelDeps[i].first, DelDeps[i].second); 1125 1126 AvailableQueue->updateNode(SU); 1127 AvailableQueue->addNode(NewSU); 1128 1129 ++NumDups; 1130 return NewSU; 1131} 1132 1133/// InsertCopiesAndMoveSuccs - Insert register copies and move all 1134/// scheduled successors of the given SUnit to the last copy. 1135void ScheduleDAGRRList::InsertCopiesAndMoveSuccs(SUnit *SU, unsigned Reg, 1136 const TargetRegisterClass *DestRC, 1137 const TargetRegisterClass *SrcRC, 1138 SmallVectorImpl<SUnit*> &Copies) { 1139 SUnit *CopyFromSU = CreateNewSUnit(NULL); 1140 CopyFromSU->CopySrcRC = SrcRC; 1141 CopyFromSU->CopyDstRC = DestRC; 1142 1143 SUnit *CopyToSU = CreateNewSUnit(NULL); 1144 CopyToSU->CopySrcRC = DestRC; 1145 CopyToSU->CopyDstRC = SrcRC; 1146 1147 // Only copy scheduled successors. Cut them from old node's successor 1148 // list and move them over. 1149 SmallVector<std::pair<SUnit *, SDep>, 4> DelDeps; 1150 for (SUnit::succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 1151 I != E; ++I) { 1152 if (I->isArtificial()) 1153 continue; 1154 SUnit *SuccSU = I->getSUnit(); 1155 if (SuccSU->isScheduled) { 1156 SDep D = *I; 1157 D.setSUnit(CopyToSU); 1158 AddPred(SuccSU, D); 1159 DelDeps.push_back(std::make_pair(SuccSU, *I)); 1160 } 1161 else { 1162 // Avoid scheduling the def-side copy before other successors. Otherwise 1163 // we could introduce another physreg interference on the copy and 1164 // continue inserting copies indefinitely. 1165 AddPred(SuccSU, SDep(CopyFromSU, SDep::Artificial)); 1166 } 1167 } 1168 for (unsigned i = 0, e = DelDeps.size(); i != e; ++i) 1169 RemovePred(DelDeps[i].first, DelDeps[i].second); 1170 1171 SDep FromDep(SU, SDep::Data, Reg); 1172 FromDep.setLatency(SU->Latency); 1173 AddPred(CopyFromSU, FromDep); 1174 SDep ToDep(CopyFromSU, SDep::Data, 0); 1175 ToDep.setLatency(CopyFromSU->Latency); 1176 AddPred(CopyToSU, ToDep); 1177 1178 AvailableQueue->updateNode(SU); 1179 AvailableQueue->addNode(CopyFromSU); 1180 AvailableQueue->addNode(CopyToSU); 1181 Copies.push_back(CopyFromSU); 1182 Copies.push_back(CopyToSU); 1183 1184 ++NumPRCopies; 1185} 1186 1187/// getPhysicalRegisterVT - Returns the ValueType of the physical register 1188/// definition of the specified node. 1189/// FIXME: Move to SelectionDAG? 1190static EVT getPhysicalRegisterVT(SDNode *N, unsigned Reg, 1191 const TargetInstrInfo *TII) { 1192 const MCInstrDesc &MCID = TII->get(N->getMachineOpcode()); 1193 assert(MCID.ImplicitDefs && "Physical reg def must be in implicit def list!"); 1194 unsigned NumRes = MCID.getNumDefs(); 1195 for (const uint16_t *ImpDef = MCID.getImplicitDefs(); *ImpDef; ++ImpDef) { 1196 if (Reg == *ImpDef) 1197 break; 1198 ++NumRes; 1199 } 1200 return N->getValueType(NumRes); 1201} 1202 1203/// CheckForLiveRegDef - Return true and update live register vector if the 1204/// specified register def of the specified SUnit clobbers any "live" registers. 1205static void CheckForLiveRegDef(SUnit *SU, unsigned Reg, 1206 std::vector<SUnit*> &LiveRegDefs, 1207 SmallSet<unsigned, 4> &RegAdded, 1208 SmallVectorImpl<unsigned> &LRegs, 1209 const TargetRegisterInfo *TRI) { 1210 for (MCRegAliasIterator AliasI(Reg, TRI, true); AliasI.isValid(); ++AliasI) { 1211 1212 // Check if Ref is live. 1213 if (!LiveRegDefs[*AliasI]) continue; 1214 1215 // Allow multiple uses of the same def. 1216 if (LiveRegDefs[*AliasI] == SU) continue; 1217 1218 // Add Reg to the set of interfering live regs. 1219 if (RegAdded.insert(*AliasI)) { 1220 LRegs.push_back(*AliasI); 1221 } 1222 } 1223} 1224 1225/// CheckForLiveRegDefMasked - Check for any live physregs that are clobbered 1226/// by RegMask, and add them to LRegs. 1227static void CheckForLiveRegDefMasked(SUnit *SU, const uint32_t *RegMask, 1228 std::vector<SUnit*> &LiveRegDefs, 1229 SmallSet<unsigned, 4> &RegAdded, 1230 SmallVectorImpl<unsigned> &LRegs) { 1231 // Look at all live registers. Skip Reg0 and the special CallResource. 1232 for (unsigned i = 1, e = LiveRegDefs.size()-1; i != e; ++i) { 1233 if (!LiveRegDefs[i]) continue; 1234 if (LiveRegDefs[i] == SU) continue; 1235 if (!MachineOperand::clobbersPhysReg(RegMask, i)) continue; 1236 if (RegAdded.insert(i)) 1237 LRegs.push_back(i); 1238 } 1239} 1240 1241/// getNodeRegMask - Returns the register mask attached to an SDNode, if any. 1242static const uint32_t *getNodeRegMask(const SDNode *N) { 1243 for (unsigned i = 0, e = N->getNumOperands(); i != e; ++i) 1244 if (const RegisterMaskSDNode *Op = 1245 dyn_cast<RegisterMaskSDNode>(N->getOperand(i).getNode())) 1246 return Op->getRegMask(); 1247 return NULL; 1248} 1249 1250/// DelayForLiveRegsBottomUp - Returns true if it is necessary to delay 1251/// scheduling of the given node to satisfy live physical register dependencies. 1252/// If the specific node is the last one that's available to schedule, do 1253/// whatever is necessary (i.e. backtracking or cloning) to make it possible. 1254bool ScheduleDAGRRList:: 1255DelayForLiveRegsBottomUp(SUnit *SU, SmallVectorImpl<unsigned> &LRegs) { 1256 if (NumLiveRegs == 0) 1257 return false; 1258 1259 SmallSet<unsigned, 4> RegAdded; 1260 // If this node would clobber any "live" register, then it's not ready. 1261 // 1262 // If SU is the currently live definition of the same register that it uses, 1263 // then we are free to schedule it. 1264 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 1265 I != E; ++I) { 1266 if (I->isAssignedRegDep() && LiveRegDefs[I->getReg()] != SU) 1267 CheckForLiveRegDef(I->getSUnit(), I->getReg(), LiveRegDefs, 1268 RegAdded, LRegs, TRI); 1269 } 1270 1271 for (SDNode *Node = SU->getNode(); Node; Node = Node->getGluedNode()) { 1272 if (Node->getOpcode() == ISD::INLINEASM) { 1273 // Inline asm can clobber physical defs. 1274 unsigned NumOps = Node->getNumOperands(); 1275 if (Node->getOperand(NumOps-1).getValueType() == MVT::Glue) 1276 --NumOps; // Ignore the glue operand. 1277 1278 for (unsigned i = InlineAsm::Op_FirstOperand; i != NumOps;) { 1279 unsigned Flags = 1280 cast<ConstantSDNode>(Node->getOperand(i))->getZExtValue(); 1281 unsigned NumVals = InlineAsm::getNumOperandRegisters(Flags); 1282 1283 ++i; // Skip the ID value. 1284 if (InlineAsm::isRegDefKind(Flags) || 1285 InlineAsm::isRegDefEarlyClobberKind(Flags) || 1286 InlineAsm::isClobberKind(Flags)) { 1287 // Check for def of register or earlyclobber register. 1288 for (; NumVals; --NumVals, ++i) { 1289 unsigned Reg = cast<RegisterSDNode>(Node->getOperand(i))->getReg(); 1290 if (TargetRegisterInfo::isPhysicalRegister(Reg)) 1291 CheckForLiveRegDef(SU, Reg, LiveRegDefs, RegAdded, LRegs, TRI); 1292 } 1293 } else 1294 i += NumVals; 1295 } 1296 continue; 1297 } 1298 1299 if (!Node->isMachineOpcode()) 1300 continue; 1301 // If we're in the middle of scheduling a call, don't begin scheduling 1302 // another call. Also, don't allow any physical registers to be live across 1303 // the call. 1304 if (Node->getMachineOpcode() == (unsigned)TII->getCallFrameDestroyOpcode()) { 1305 // Check the special calling-sequence resource. 1306 unsigned CallResource = TRI->getNumRegs(); 1307 if (LiveRegDefs[CallResource]) { 1308 SDNode *Gen = LiveRegGens[CallResource]->getNode(); 1309 while (SDNode *Glued = Gen->getGluedNode()) 1310 Gen = Glued; 1311 if (!IsChainDependent(Gen, Node, 0, TII) && RegAdded.insert(CallResource)) 1312 LRegs.push_back(CallResource); 1313 } 1314 } 1315 if (const uint32_t *RegMask = getNodeRegMask(Node)) 1316 CheckForLiveRegDefMasked(SU, RegMask, LiveRegDefs, RegAdded, LRegs); 1317 1318 const MCInstrDesc &MCID = TII->get(Node->getMachineOpcode()); 1319 if (!MCID.ImplicitDefs) 1320 continue; 1321 for (const uint16_t *Reg = MCID.getImplicitDefs(); *Reg; ++Reg) 1322 CheckForLiveRegDef(SU, *Reg, LiveRegDefs, RegAdded, LRegs, TRI); 1323 } 1324 1325 return !LRegs.empty(); 1326} 1327 1328void ScheduleDAGRRList::releaseInterferences(unsigned Reg) { 1329 // Add the nodes that aren't ready back onto the available list. 1330 for (unsigned i = Interferences.size(); i > 0; --i) { 1331 SUnit *SU = Interferences[i-1]; 1332 LRegsMapT::iterator LRegsPos = LRegsMap.find(SU); 1333 if (Reg) { 1334 SmallVectorImpl<unsigned> &LRegs = LRegsPos->second; 1335 if (std::find(LRegs.begin(), LRegs.end(), Reg) == LRegs.end()) 1336 continue; 1337 } 1338 SU->isPending = false; 1339 // The interfering node may no longer be available due to backtracking. 1340 // Furthermore, it may have been made available again, in which case it is 1341 // now already in the AvailableQueue. 1342 if (SU->isAvailable && !SU->NodeQueueId) { 1343 DEBUG(dbgs() << " Repushing SU #" << SU->NodeNum << '\n'); 1344 AvailableQueue->push(SU); 1345 } 1346 if (i < Interferences.size()) 1347 Interferences[i-1] = Interferences.back(); 1348 Interferences.pop_back(); 1349 LRegsMap.erase(LRegsPos); 1350 } 1351} 1352 1353/// Return a node that can be scheduled in this cycle. Requirements: 1354/// (1) Ready: latency has been satisfied 1355/// (2) No Hazards: resources are available 1356/// (3) No Interferences: may unschedule to break register interferences. 1357SUnit *ScheduleDAGRRList::PickNodeToScheduleBottomUp() { 1358 SUnit *CurSU = AvailableQueue->empty() ? 0 : AvailableQueue->pop(); 1359 while (CurSU) { 1360 SmallVector<unsigned, 4> LRegs; 1361 if (!DelayForLiveRegsBottomUp(CurSU, LRegs)) 1362 break; 1363 DEBUG(dbgs() << " Interfering reg " << 1364 (LRegs[0] == TRI->getNumRegs() ? "CallResource" 1365 : TRI->getName(LRegs[0])) 1366 << " SU #" << CurSU->NodeNum << '\n'); 1367 std::pair<LRegsMapT::iterator, bool> LRegsPair = 1368 LRegsMap.insert(std::make_pair(CurSU, LRegs)); 1369 if (LRegsPair.second) { 1370 CurSU->isPending = true; // This SU is not in AvailableQueue right now. 1371 Interferences.push_back(CurSU); 1372 } 1373 else { 1374 assert(CurSU->isPending && "Intereferences are pending"); 1375 // Update the interference with current live regs. 1376 LRegsPair.first->second = LRegs; 1377 } 1378 CurSU = AvailableQueue->pop(); 1379 } 1380 if (CurSU) 1381 return CurSU; 1382 1383 // All candidates are delayed due to live physical reg dependencies. 1384 // Try backtracking, code duplication, or inserting cross class copies 1385 // to resolve it. 1386 for (unsigned i = 0, e = Interferences.size(); i != e; ++i) { 1387 SUnit *TrySU = Interferences[i]; 1388 SmallVectorImpl<unsigned> &LRegs = LRegsMap[TrySU]; 1389 1390 // Try unscheduling up to the point where it's safe to schedule 1391 // this node. 1392 SUnit *BtSU = NULL; 1393 unsigned LiveCycle = UINT_MAX; 1394 for (unsigned j = 0, ee = LRegs.size(); j != ee; ++j) { 1395 unsigned Reg = LRegs[j]; 1396 if (LiveRegGens[Reg]->getHeight() < LiveCycle) { 1397 BtSU = LiveRegGens[Reg]; 1398 LiveCycle = BtSU->getHeight(); 1399 } 1400 } 1401 if (!WillCreateCycle(TrySU, BtSU)) { 1402 // BacktrackBottomUp mutates Interferences! 1403 BacktrackBottomUp(TrySU, BtSU); 1404 1405 // Force the current node to be scheduled before the node that 1406 // requires the physical reg dep. 1407 if (BtSU->isAvailable) { 1408 BtSU->isAvailable = false; 1409 if (!BtSU->isPending) 1410 AvailableQueue->remove(BtSU); 1411 } 1412 DEBUG(dbgs() << "ARTIFICIAL edge from SU(" << BtSU->NodeNum << ") to SU(" 1413 << TrySU->NodeNum << ")\n"); 1414 AddPred(TrySU, SDep(BtSU, SDep::Artificial)); 1415 1416 // If one or more successors has been unscheduled, then the current 1417 // node is no longer available. 1418 if (!TrySU->isAvailable) 1419 CurSU = AvailableQueue->pop(); 1420 else { 1421 AvailableQueue->remove(TrySU); 1422 CurSU = TrySU; 1423 } 1424 // Interferences has been mutated. We must break. 1425 break; 1426 } 1427 } 1428 1429 if (!CurSU) { 1430 // Can't backtrack. If it's too expensive to copy the value, then try 1431 // duplicate the nodes that produces these "too expensive to copy" 1432 // values to break the dependency. In case even that doesn't work, 1433 // insert cross class copies. 1434 // If it's not too expensive, i.e. cost != -1, issue copies. 1435 SUnit *TrySU = Interferences[0]; 1436 SmallVectorImpl<unsigned> &LRegs = LRegsMap[TrySU]; 1437 assert(LRegs.size() == 1 && "Can't handle this yet!"); 1438 unsigned Reg = LRegs[0]; 1439 SUnit *LRDef = LiveRegDefs[Reg]; 1440 EVT VT = getPhysicalRegisterVT(LRDef->getNode(), Reg, TII); 1441 const TargetRegisterClass *RC = 1442 TRI->getMinimalPhysRegClass(Reg, VT); 1443 const TargetRegisterClass *DestRC = TRI->getCrossCopyRegClass(RC); 1444 1445 // If cross copy register class is the same as RC, then it must be possible 1446 // copy the value directly. Do not try duplicate the def. 1447 // If cross copy register class is not the same as RC, then it's possible to 1448 // copy the value but it require cross register class copies and it is 1449 // expensive. 1450 // If cross copy register class is null, then it's not possible to copy 1451 // the value at all. 1452 SUnit *NewDef = 0; 1453 if (DestRC != RC) { 1454 NewDef = CopyAndMoveSuccessors(LRDef); 1455 if (!DestRC && !NewDef) 1456 report_fatal_error("Can't handle live physical register dependency!"); 1457 } 1458 if (!NewDef) { 1459 // Issue copies, these can be expensive cross register class copies. 1460 SmallVector<SUnit*, 2> Copies; 1461 InsertCopiesAndMoveSuccs(LRDef, Reg, DestRC, RC, Copies); 1462 DEBUG(dbgs() << " Adding an edge from SU #" << TrySU->NodeNum 1463 << " to SU #" << Copies.front()->NodeNum << "\n"); 1464 AddPred(TrySU, SDep(Copies.front(), SDep::Artificial)); 1465 NewDef = Copies.back(); 1466 } 1467 1468 DEBUG(dbgs() << " Adding an edge from SU #" << NewDef->NodeNum 1469 << " to SU #" << TrySU->NodeNum << "\n"); 1470 LiveRegDefs[Reg] = NewDef; 1471 AddPred(NewDef, SDep(TrySU, SDep::Artificial)); 1472 TrySU->isAvailable = false; 1473 CurSU = NewDef; 1474 } 1475 assert(CurSU && "Unable to resolve live physical register dependencies!"); 1476 return CurSU; 1477} 1478 1479/// ListScheduleBottomUp - The main loop of list scheduling for bottom-up 1480/// schedulers. 1481void ScheduleDAGRRList::ListScheduleBottomUp() { 1482 // Release any predecessors of the special Exit node. 1483 ReleasePredecessors(&ExitSU); 1484 1485 // Add root to Available queue. 1486 if (!SUnits.empty()) { 1487 SUnit *RootSU = &SUnits[DAG->getRoot().getNode()->getNodeId()]; 1488 assert(RootSU->Succs.empty() && "Graph root shouldn't have successors!"); 1489 RootSU->isAvailable = true; 1490 AvailableQueue->push(RootSU); 1491 } 1492 1493 // While Available queue is not empty, grab the node with the highest 1494 // priority. If it is not ready put it back. Schedule the node. 1495 Sequence.reserve(SUnits.size()); 1496 while (!AvailableQueue->empty() || !Interferences.empty()) { 1497 DEBUG(dbgs() << "\nExamining Available:\n"; 1498 AvailableQueue->dump(this)); 1499 1500 // Pick the best node to schedule taking all constraints into 1501 // consideration. 1502 SUnit *SU = PickNodeToScheduleBottomUp(); 1503 1504 AdvancePastStalls(SU); 1505 1506 ScheduleNodeBottomUp(SU); 1507 1508 while (AvailableQueue->empty() && !PendingQueue.empty()) { 1509 // Advance the cycle to free resources. Skip ahead to the next ready SU. 1510 assert(MinAvailableCycle < UINT_MAX && "MinAvailableCycle uninitialized"); 1511 AdvanceToCycle(std::max(CurCycle + 1, MinAvailableCycle)); 1512 } 1513 } 1514 1515 // Reverse the order if it is bottom up. 1516 std::reverse(Sequence.begin(), Sequence.end()); 1517 1518#ifndef NDEBUG 1519 VerifyScheduledSequence(/*isBottomUp=*/true); 1520#endif 1521} 1522 1523//===----------------------------------------------------------------------===// 1524// RegReductionPriorityQueue Definition 1525//===----------------------------------------------------------------------===// 1526// 1527// This is a SchedulingPriorityQueue that schedules using Sethi Ullman numbers 1528// to reduce register pressure. 1529// 1530namespace { 1531class RegReductionPQBase; 1532 1533struct queue_sort : public std::binary_function<SUnit*, SUnit*, bool> { 1534 bool isReady(SUnit* SU, unsigned CurCycle) const { return true; } 1535}; 1536 1537#ifndef NDEBUG 1538template<class SF> 1539struct reverse_sort : public queue_sort { 1540 SF &SortFunc; 1541 reverse_sort(SF &sf) : SortFunc(sf) {} 1542 reverse_sort(const reverse_sort &RHS) : SortFunc(RHS.SortFunc) {} 1543 1544 bool operator()(SUnit* left, SUnit* right) const { 1545 // reverse left/right rather than simply !SortFunc(left, right) 1546 // to expose different paths in the comparison logic. 1547 return SortFunc(right, left); 1548 } 1549}; 1550#endif // NDEBUG 1551 1552/// bu_ls_rr_sort - Priority function for bottom up register pressure 1553// reduction scheduler. 1554struct bu_ls_rr_sort : public queue_sort { 1555 enum { 1556 IsBottomUp = true, 1557 HasReadyFilter = false 1558 }; 1559 1560 RegReductionPQBase *SPQ; 1561 bu_ls_rr_sort(RegReductionPQBase *spq) : SPQ(spq) {} 1562 bu_ls_rr_sort(const bu_ls_rr_sort &RHS) : SPQ(RHS.SPQ) {} 1563 1564 bool operator()(SUnit* left, SUnit* right) const; 1565}; 1566 1567// src_ls_rr_sort - Priority function for source order scheduler. 1568struct src_ls_rr_sort : public queue_sort { 1569 enum { 1570 IsBottomUp = true, 1571 HasReadyFilter = false 1572 }; 1573 1574 RegReductionPQBase *SPQ; 1575 src_ls_rr_sort(RegReductionPQBase *spq) 1576 : SPQ(spq) {} 1577 src_ls_rr_sort(const src_ls_rr_sort &RHS) 1578 : SPQ(RHS.SPQ) {} 1579 1580 bool operator()(SUnit* left, SUnit* right) const; 1581}; 1582 1583// hybrid_ls_rr_sort - Priority function for hybrid scheduler. 1584struct hybrid_ls_rr_sort : public queue_sort { 1585 enum { 1586 IsBottomUp = true, 1587 HasReadyFilter = false 1588 }; 1589 1590 RegReductionPQBase *SPQ; 1591 hybrid_ls_rr_sort(RegReductionPQBase *spq) 1592 : SPQ(spq) {} 1593 hybrid_ls_rr_sort(const hybrid_ls_rr_sort &RHS) 1594 : SPQ(RHS.SPQ) {} 1595 1596 bool isReady(SUnit *SU, unsigned CurCycle) const; 1597 1598 bool operator()(SUnit* left, SUnit* right) const; 1599}; 1600 1601// ilp_ls_rr_sort - Priority function for ILP (instruction level parallelism) 1602// scheduler. 1603struct ilp_ls_rr_sort : public queue_sort { 1604 enum { 1605 IsBottomUp = true, 1606 HasReadyFilter = false 1607 }; 1608 1609 RegReductionPQBase *SPQ; 1610 ilp_ls_rr_sort(RegReductionPQBase *spq) 1611 : SPQ(spq) {} 1612 ilp_ls_rr_sort(const ilp_ls_rr_sort &RHS) 1613 : SPQ(RHS.SPQ) {} 1614 1615 bool isReady(SUnit *SU, unsigned CurCycle) const; 1616 1617 bool operator()(SUnit* left, SUnit* right) const; 1618}; 1619 1620class RegReductionPQBase : public SchedulingPriorityQueue { 1621protected: 1622 std::vector<SUnit*> Queue; 1623 unsigned CurQueueId; 1624 bool TracksRegPressure; 1625 bool SrcOrder; 1626 1627 // SUnits - The SUnits for the current graph. 1628 std::vector<SUnit> *SUnits; 1629 1630 MachineFunction &MF; 1631 const TargetInstrInfo *TII; 1632 const TargetRegisterInfo *TRI; 1633 const TargetLowering *TLI; 1634 ScheduleDAGRRList *scheduleDAG; 1635 1636 // SethiUllmanNumbers - The SethiUllman number for each node. 1637 std::vector<unsigned> SethiUllmanNumbers; 1638 1639 /// RegPressure - Tracking current reg pressure per register class. 1640 /// 1641 std::vector<unsigned> RegPressure; 1642 1643 /// RegLimit - Tracking the number of allocatable registers per register 1644 /// class. 1645 std::vector<unsigned> RegLimit; 1646 1647public: 1648 RegReductionPQBase(MachineFunction &mf, 1649 bool hasReadyFilter, 1650 bool tracksrp, 1651 bool srcorder, 1652 const TargetInstrInfo *tii, 1653 const TargetRegisterInfo *tri, 1654 const TargetLowering *tli) 1655 : SchedulingPriorityQueue(hasReadyFilter), 1656 CurQueueId(0), TracksRegPressure(tracksrp), SrcOrder(srcorder), 1657 MF(mf), TII(tii), TRI(tri), TLI(tli), scheduleDAG(NULL) { 1658 if (TracksRegPressure) { 1659 unsigned NumRC = TRI->getNumRegClasses(); 1660 RegLimit.resize(NumRC); 1661 RegPressure.resize(NumRC); 1662 std::fill(RegLimit.begin(), RegLimit.end(), 0); 1663 std::fill(RegPressure.begin(), RegPressure.end(), 0); 1664 for (TargetRegisterInfo::regclass_iterator I = TRI->regclass_begin(), 1665 E = TRI->regclass_end(); I != E; ++I) 1666 RegLimit[(*I)->getID()] = tri->getRegPressureLimit(*I, MF); 1667 } 1668 } 1669 1670 void setScheduleDAG(ScheduleDAGRRList *scheduleDag) { 1671 scheduleDAG = scheduleDag; 1672 } 1673 1674 ScheduleHazardRecognizer* getHazardRec() { 1675 return scheduleDAG->getHazardRec(); 1676 } 1677 1678 void initNodes(std::vector<SUnit> &sunits); 1679 1680 void addNode(const SUnit *SU); 1681 1682 void updateNode(const SUnit *SU); 1683 1684 void releaseState() { 1685 SUnits = 0; 1686 SethiUllmanNumbers.clear(); 1687 std::fill(RegPressure.begin(), RegPressure.end(), 0); 1688 } 1689 1690 unsigned getNodePriority(const SUnit *SU) const; 1691 1692 unsigned getNodeOrdering(const SUnit *SU) const { 1693 if (!SU->getNode()) return 0; 1694 1695 return SU->getNode()->getIROrder(); 1696 } 1697 1698 bool empty() const { return Queue.empty(); } 1699 1700 void push(SUnit *U) { 1701 assert(!U->NodeQueueId && "Node in the queue already"); 1702 U->NodeQueueId = ++CurQueueId; 1703 Queue.push_back(U); 1704 } 1705 1706 void remove(SUnit *SU) { 1707 assert(!Queue.empty() && "Queue is empty!"); 1708 assert(SU->NodeQueueId != 0 && "Not in queue!"); 1709 std::vector<SUnit *>::iterator I = std::find(Queue.begin(), Queue.end(), 1710 SU); 1711 if (I != prior(Queue.end())) 1712 std::swap(*I, Queue.back()); 1713 Queue.pop_back(); 1714 SU->NodeQueueId = 0; 1715 } 1716 1717 bool tracksRegPressure() const { return TracksRegPressure; } 1718 1719 void dumpRegPressure() const; 1720 1721 bool HighRegPressure(const SUnit *SU) const; 1722 1723 bool MayReduceRegPressure(SUnit *SU) const; 1724 1725 int RegPressureDiff(SUnit *SU, unsigned &LiveUses) const; 1726 1727 void scheduledNode(SUnit *SU); 1728 1729 void unscheduledNode(SUnit *SU); 1730 1731protected: 1732 bool canClobber(const SUnit *SU, const SUnit *Op); 1733 void AddPseudoTwoAddrDeps(); 1734 void PrescheduleNodesWithMultipleUses(); 1735 void CalculateSethiUllmanNumbers(); 1736}; 1737 1738template<class SF> 1739static SUnit *popFromQueueImpl(std::vector<SUnit*> &Q, SF &Picker) { 1740 std::vector<SUnit *>::iterator Best = Q.begin(); 1741 for (std::vector<SUnit *>::iterator I = llvm::next(Q.begin()), 1742 E = Q.end(); I != E; ++I) 1743 if (Picker(*Best, *I)) 1744 Best = I; 1745 SUnit *V = *Best; 1746 if (Best != prior(Q.end())) 1747 std::swap(*Best, Q.back()); 1748 Q.pop_back(); 1749 return V; 1750} 1751 1752template<class SF> 1753SUnit *popFromQueue(std::vector<SUnit*> &Q, SF &Picker, ScheduleDAG *DAG) { 1754#ifndef NDEBUG 1755 if (DAG->StressSched) { 1756 reverse_sort<SF> RPicker(Picker); 1757 return popFromQueueImpl(Q, RPicker); 1758 } 1759#endif 1760 (void)DAG; 1761 return popFromQueueImpl(Q, Picker); 1762} 1763 1764template<class SF> 1765class RegReductionPriorityQueue : public RegReductionPQBase { 1766 SF Picker; 1767 1768public: 1769 RegReductionPriorityQueue(MachineFunction &mf, 1770 bool tracksrp, 1771 bool srcorder, 1772 const TargetInstrInfo *tii, 1773 const TargetRegisterInfo *tri, 1774 const TargetLowering *tli) 1775 : RegReductionPQBase(mf, SF::HasReadyFilter, tracksrp, srcorder, 1776 tii, tri, tli), 1777 Picker(this) {} 1778 1779 bool isBottomUp() const { return SF::IsBottomUp; } 1780 1781 bool isReady(SUnit *U) const { 1782 return Picker.HasReadyFilter && Picker.isReady(U, getCurCycle()); 1783 } 1784 1785 SUnit *pop() { 1786 if (Queue.empty()) return NULL; 1787 1788 SUnit *V = popFromQueue(Queue, Picker, scheduleDAG); 1789 V->NodeQueueId = 0; 1790 return V; 1791 } 1792 1793#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) 1794 void dump(ScheduleDAG *DAG) const { 1795 // Emulate pop() without clobbering NodeQueueIds. 1796 std::vector<SUnit*> DumpQueue = Queue; 1797 SF DumpPicker = Picker; 1798 while (!DumpQueue.empty()) { 1799 SUnit *SU = popFromQueue(DumpQueue, DumpPicker, scheduleDAG); 1800 dbgs() << "Height " << SU->getHeight() << ": "; 1801 SU->dump(DAG); 1802 } 1803 } 1804#endif 1805}; 1806 1807typedef RegReductionPriorityQueue<bu_ls_rr_sort> 1808BURegReductionPriorityQueue; 1809 1810typedef RegReductionPriorityQueue<src_ls_rr_sort> 1811SrcRegReductionPriorityQueue; 1812 1813typedef RegReductionPriorityQueue<hybrid_ls_rr_sort> 1814HybridBURRPriorityQueue; 1815 1816typedef RegReductionPriorityQueue<ilp_ls_rr_sort> 1817ILPBURRPriorityQueue; 1818} // end anonymous namespace 1819 1820//===----------------------------------------------------------------------===// 1821// Static Node Priority for Register Pressure Reduction 1822//===----------------------------------------------------------------------===// 1823 1824// Check for special nodes that bypass scheduling heuristics. 1825// Currently this pushes TokenFactor nodes down, but may be used for other 1826// pseudo-ops as well. 1827// 1828// Return -1 to schedule right above left, 1 for left above right. 1829// Return 0 if no bias exists. 1830static int checkSpecialNodes(const SUnit *left, const SUnit *right) { 1831 bool LSchedLow = left->isScheduleLow; 1832 bool RSchedLow = right->isScheduleLow; 1833 if (LSchedLow != RSchedLow) 1834 return LSchedLow < RSchedLow ? 1 : -1; 1835 return 0; 1836} 1837 1838/// CalcNodeSethiUllmanNumber - Compute Sethi Ullman number. 1839/// Smaller number is the higher priority. 1840static unsigned 1841CalcNodeSethiUllmanNumber(const SUnit *SU, std::vector<unsigned> &SUNumbers) { 1842 unsigned &SethiUllmanNumber = SUNumbers[SU->NodeNum]; 1843 if (SethiUllmanNumber != 0) 1844 return SethiUllmanNumber; 1845 1846 unsigned Extra = 0; 1847 for (SUnit::const_pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 1848 I != E; ++I) { 1849 if (I->isCtrl()) continue; // ignore chain preds 1850 SUnit *PredSU = I->getSUnit(); 1851 unsigned PredSethiUllman = CalcNodeSethiUllmanNumber(PredSU, SUNumbers); 1852 if (PredSethiUllman > SethiUllmanNumber) { 1853 SethiUllmanNumber = PredSethiUllman; 1854 Extra = 0; 1855 } else if (PredSethiUllman == SethiUllmanNumber) 1856 ++Extra; 1857 } 1858 1859 SethiUllmanNumber += Extra; 1860 1861 if (SethiUllmanNumber == 0) 1862 SethiUllmanNumber = 1; 1863 1864 return SethiUllmanNumber; 1865} 1866 1867/// CalculateSethiUllmanNumbers - Calculate Sethi-Ullman numbers of all 1868/// scheduling units. 1869void RegReductionPQBase::CalculateSethiUllmanNumbers() { 1870 SethiUllmanNumbers.assign(SUnits->size(), 0); 1871 1872 for (unsigned i = 0, e = SUnits->size(); i != e; ++i) 1873 CalcNodeSethiUllmanNumber(&(*SUnits)[i], SethiUllmanNumbers); 1874} 1875 1876void RegReductionPQBase::addNode(const SUnit *SU) { 1877 unsigned SUSize = SethiUllmanNumbers.size(); 1878 if (SUnits->size() > SUSize) 1879 SethiUllmanNumbers.resize(SUSize*2, 0); 1880 CalcNodeSethiUllmanNumber(SU, SethiUllmanNumbers); 1881} 1882 1883void RegReductionPQBase::updateNode(const SUnit *SU) { 1884 SethiUllmanNumbers[SU->NodeNum] = 0; 1885 CalcNodeSethiUllmanNumber(SU, SethiUllmanNumbers); 1886} 1887 1888// Lower priority means schedule further down. For bottom-up scheduling, lower 1889// priority SUs are scheduled before higher priority SUs. 1890unsigned RegReductionPQBase::getNodePriority(const SUnit *SU) const { 1891 assert(SU->NodeNum < SethiUllmanNumbers.size()); 1892 unsigned Opc = SU->getNode() ? SU->getNode()->getOpcode() : 0; 1893 if (Opc == ISD::TokenFactor || Opc == ISD::CopyToReg) 1894 // CopyToReg should be close to its uses to facilitate coalescing and 1895 // avoid spilling. 1896 return 0; 1897 if (Opc == TargetOpcode::EXTRACT_SUBREG || 1898 Opc == TargetOpcode::SUBREG_TO_REG || 1899 Opc == TargetOpcode::INSERT_SUBREG) 1900 // EXTRACT_SUBREG, INSERT_SUBREG, and SUBREG_TO_REG nodes should be 1901 // close to their uses to facilitate coalescing. 1902 return 0; 1903 if (SU->NumSuccs == 0 && SU->NumPreds != 0) 1904 // If SU does not have a register use, i.e. it doesn't produce a value 1905 // that would be consumed (e.g. store), then it terminates a chain of 1906 // computation. Give it a large SethiUllman number so it will be 1907 // scheduled right before its predecessors that it doesn't lengthen 1908 // their live ranges. 1909 return 0xffff; 1910 if (SU->NumPreds == 0 && SU->NumSuccs != 0) 1911 // If SU does not have a register def, schedule it close to its uses 1912 // because it does not lengthen any live ranges. 1913 return 0; 1914#if 1 1915 return SethiUllmanNumbers[SU->NodeNum]; 1916#else 1917 unsigned Priority = SethiUllmanNumbers[SU->NodeNum]; 1918 if (SU->isCallOp) { 1919 // FIXME: This assumes all of the defs are used as call operands. 1920 int NP = (int)Priority - SU->getNode()->getNumValues(); 1921 return (NP > 0) ? NP : 0; 1922 } 1923 return Priority; 1924#endif 1925} 1926 1927//===----------------------------------------------------------------------===// 1928// Register Pressure Tracking 1929//===----------------------------------------------------------------------===// 1930 1931void RegReductionPQBase::dumpRegPressure() const { 1932#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) 1933 for (TargetRegisterInfo::regclass_iterator I = TRI->regclass_begin(), 1934 E = TRI->regclass_end(); I != E; ++I) { 1935 const TargetRegisterClass *RC = *I; 1936 unsigned Id = RC->getID(); 1937 unsigned RP = RegPressure[Id]; 1938 if (!RP) continue; 1939 DEBUG(dbgs() << RC->getName() << ": " << RP << " / " << RegLimit[Id] 1940 << '\n'); 1941 } 1942#endif 1943} 1944 1945bool RegReductionPQBase::HighRegPressure(const SUnit *SU) const { 1946 if (!TLI) 1947 return false; 1948 1949 for (SUnit::const_pred_iterator I = SU->Preds.begin(),E = SU->Preds.end(); 1950 I != E; ++I) { 1951 if (I->isCtrl()) 1952 continue; 1953 SUnit *PredSU = I->getSUnit(); 1954 // NumRegDefsLeft is zero when enough uses of this node have been scheduled 1955 // to cover the number of registers defined (they are all live). 1956 if (PredSU->NumRegDefsLeft == 0) { 1957 continue; 1958 } 1959 for (ScheduleDAGSDNodes::RegDefIter RegDefPos(PredSU, scheduleDAG); 1960 RegDefPos.IsValid(); RegDefPos.Advance()) { 1961 unsigned RCId, Cost; 1962 GetCostForDef(RegDefPos, TLI, TII, TRI, RCId, Cost, MF); 1963 1964 if ((RegPressure[RCId] + Cost) >= RegLimit[RCId]) 1965 return true; 1966 } 1967 } 1968 return false; 1969} 1970 1971bool RegReductionPQBase::MayReduceRegPressure(SUnit *SU) const { 1972 const SDNode *N = SU->getNode(); 1973 1974 if (!N->isMachineOpcode() || !SU->NumSuccs) 1975 return false; 1976 1977 unsigned NumDefs = TII->get(N->getMachineOpcode()).getNumDefs(); 1978 for (unsigned i = 0; i != NumDefs; ++i) { 1979 MVT VT = N->getSimpleValueType(i); 1980 if (!N->hasAnyUseOfValue(i)) 1981 continue; 1982 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 1983 if (RegPressure[RCId] >= RegLimit[RCId]) 1984 return true; 1985 } 1986 return false; 1987} 1988 1989// Compute the register pressure contribution by this instruction by count up 1990// for uses that are not live and down for defs. Only count register classes 1991// that are already under high pressure. As a side effect, compute the number of 1992// uses of registers that are already live. 1993// 1994// FIXME: This encompasses the logic in HighRegPressure and MayReduceRegPressure 1995// so could probably be factored. 1996int RegReductionPQBase::RegPressureDiff(SUnit *SU, unsigned &LiveUses) const { 1997 LiveUses = 0; 1998 int PDiff = 0; 1999 for (SUnit::const_pred_iterator I = SU->Preds.begin(),E = SU->Preds.end(); 2000 I != E; ++I) { 2001 if (I->isCtrl()) 2002 continue; 2003 SUnit *PredSU = I->getSUnit(); 2004 // NumRegDefsLeft is zero when enough uses of this node have been scheduled 2005 // to cover the number of registers defined (they are all live). 2006 if (PredSU->NumRegDefsLeft == 0) { 2007 if (PredSU->getNode()->isMachineOpcode()) 2008 ++LiveUses; 2009 continue; 2010 } 2011 for (ScheduleDAGSDNodes::RegDefIter RegDefPos(PredSU, scheduleDAG); 2012 RegDefPos.IsValid(); RegDefPos.Advance()) { 2013 MVT VT = RegDefPos.GetValue(); 2014 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 2015 if (RegPressure[RCId] >= RegLimit[RCId]) 2016 ++PDiff; 2017 } 2018 } 2019 const SDNode *N = SU->getNode(); 2020 2021 if (!N || !N->isMachineOpcode() || !SU->NumSuccs) 2022 return PDiff; 2023 2024 unsigned NumDefs = TII->get(N->getMachineOpcode()).getNumDefs(); 2025 for (unsigned i = 0; i != NumDefs; ++i) { 2026 MVT VT = N->getSimpleValueType(i); 2027 if (!N->hasAnyUseOfValue(i)) 2028 continue; 2029 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 2030 if (RegPressure[RCId] >= RegLimit[RCId]) 2031 --PDiff; 2032 } 2033 return PDiff; 2034} 2035 2036void RegReductionPQBase::scheduledNode(SUnit *SU) { 2037 if (!TracksRegPressure) 2038 return; 2039 2040 if (!SU->getNode()) 2041 return; 2042 2043 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 2044 I != E; ++I) { 2045 if (I->isCtrl()) 2046 continue; 2047 SUnit *PredSU = I->getSUnit(); 2048 // NumRegDefsLeft is zero when enough uses of this node have been scheduled 2049 // to cover the number of registers defined (they are all live). 2050 if (PredSU->NumRegDefsLeft == 0) { 2051 continue; 2052 } 2053 // FIXME: The ScheduleDAG currently loses information about which of a 2054 // node's values is consumed by each dependence. Consequently, if the node 2055 // defines multiple register classes, we don't know which to pressurize 2056 // here. Instead the following loop consumes the register defs in an 2057 // arbitrary order. At least it handles the common case of clustered loads 2058 // to the same class. For precise liveness, each SDep needs to indicate the 2059 // result number. But that tightly couples the ScheduleDAG with the 2060 // SelectionDAG making updates tricky. A simpler hack would be to attach a 2061 // value type or register class to SDep. 2062 // 2063 // The most important aspect of register tracking is balancing the increase 2064 // here with the reduction further below. Note that this SU may use multiple 2065 // defs in PredSU. The can't be determined here, but we've already 2066 // compensated by reducing NumRegDefsLeft in PredSU during 2067 // ScheduleDAGSDNodes::AddSchedEdges. 2068 --PredSU->NumRegDefsLeft; 2069 unsigned SkipRegDefs = PredSU->NumRegDefsLeft; 2070 for (ScheduleDAGSDNodes::RegDefIter RegDefPos(PredSU, scheduleDAG); 2071 RegDefPos.IsValid(); RegDefPos.Advance(), --SkipRegDefs) { 2072 if (SkipRegDefs) 2073 continue; 2074 2075 unsigned RCId, Cost; 2076 GetCostForDef(RegDefPos, TLI, TII, TRI, RCId, Cost, MF); 2077 RegPressure[RCId] += Cost; 2078 break; 2079 } 2080 } 2081 2082 // We should have this assert, but there may be dead SDNodes that never 2083 // materialize as SUnits, so they don't appear to generate liveness. 2084 //assert(SU->NumRegDefsLeft == 0 && "not all regdefs have scheduled uses"); 2085 int SkipRegDefs = (int)SU->NumRegDefsLeft; 2086 for (ScheduleDAGSDNodes::RegDefIter RegDefPos(SU, scheduleDAG); 2087 RegDefPos.IsValid(); RegDefPos.Advance(), --SkipRegDefs) { 2088 if (SkipRegDefs > 0) 2089 continue; 2090 unsigned RCId, Cost; 2091 GetCostForDef(RegDefPos, TLI, TII, TRI, RCId, Cost, MF); 2092 if (RegPressure[RCId] < Cost) { 2093 // Register pressure tracking is imprecise. This can happen. But we try 2094 // hard not to let it happen because it likely results in poor scheduling. 2095 DEBUG(dbgs() << " SU(" << SU->NodeNum << ") has too many regdefs\n"); 2096 RegPressure[RCId] = 0; 2097 } 2098 else { 2099 RegPressure[RCId] -= Cost; 2100 } 2101 } 2102 dumpRegPressure(); 2103} 2104 2105void RegReductionPQBase::unscheduledNode(SUnit *SU) { 2106 if (!TracksRegPressure) 2107 return; 2108 2109 const SDNode *N = SU->getNode(); 2110 if (!N) return; 2111 2112 if (!N->isMachineOpcode()) { 2113 if (N->getOpcode() != ISD::CopyToReg) 2114 return; 2115 } else { 2116 unsigned Opc = N->getMachineOpcode(); 2117 if (Opc == TargetOpcode::EXTRACT_SUBREG || 2118 Opc == TargetOpcode::INSERT_SUBREG || 2119 Opc == TargetOpcode::SUBREG_TO_REG || 2120 Opc == TargetOpcode::REG_SEQUENCE || 2121 Opc == TargetOpcode::IMPLICIT_DEF) 2122 return; 2123 } 2124 2125 for (SUnit::pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 2126 I != E; ++I) { 2127 if (I->isCtrl()) 2128 continue; 2129 SUnit *PredSU = I->getSUnit(); 2130 // NumSuccsLeft counts all deps. Don't compare it with NumSuccs which only 2131 // counts data deps. 2132 if (PredSU->NumSuccsLeft != PredSU->Succs.size()) 2133 continue; 2134 const SDNode *PN = PredSU->getNode(); 2135 if (!PN->isMachineOpcode()) { 2136 if (PN->getOpcode() == ISD::CopyFromReg) { 2137 MVT VT = PN->getSimpleValueType(0); 2138 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 2139 RegPressure[RCId] += TLI->getRepRegClassCostFor(VT); 2140 } 2141 continue; 2142 } 2143 unsigned POpc = PN->getMachineOpcode(); 2144 if (POpc == TargetOpcode::IMPLICIT_DEF) 2145 continue; 2146 if (POpc == TargetOpcode::EXTRACT_SUBREG || 2147 POpc == TargetOpcode::INSERT_SUBREG || 2148 POpc == TargetOpcode::SUBREG_TO_REG) { 2149 MVT VT = PN->getSimpleValueType(0); 2150 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 2151 RegPressure[RCId] += TLI->getRepRegClassCostFor(VT); 2152 continue; 2153 } 2154 unsigned NumDefs = TII->get(PN->getMachineOpcode()).getNumDefs(); 2155 for (unsigned i = 0; i != NumDefs; ++i) { 2156 MVT VT = PN->getSimpleValueType(i); 2157 if (!PN->hasAnyUseOfValue(i)) 2158 continue; 2159 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 2160 if (RegPressure[RCId] < TLI->getRepRegClassCostFor(VT)) 2161 // Register pressure tracking is imprecise. This can happen. 2162 RegPressure[RCId] = 0; 2163 else 2164 RegPressure[RCId] -= TLI->getRepRegClassCostFor(VT); 2165 } 2166 } 2167 2168 // Check for isMachineOpcode() as PrescheduleNodesWithMultipleUses() 2169 // may transfer data dependencies to CopyToReg. 2170 if (SU->NumSuccs && N->isMachineOpcode()) { 2171 unsigned NumDefs = TII->get(N->getMachineOpcode()).getNumDefs(); 2172 for (unsigned i = NumDefs, e = N->getNumValues(); i != e; ++i) { 2173 MVT VT = N->getSimpleValueType(i); 2174 if (VT == MVT::Glue || VT == MVT::Other) 2175 continue; 2176 if (!N->hasAnyUseOfValue(i)) 2177 continue; 2178 unsigned RCId = TLI->getRepRegClassFor(VT)->getID(); 2179 RegPressure[RCId] += TLI->getRepRegClassCostFor(VT); 2180 } 2181 } 2182 2183 dumpRegPressure(); 2184} 2185 2186//===----------------------------------------------------------------------===// 2187// Dynamic Node Priority for Register Pressure Reduction 2188//===----------------------------------------------------------------------===// 2189 2190/// closestSucc - Returns the scheduled cycle of the successor which is 2191/// closest to the current cycle. 2192static unsigned closestSucc(const SUnit *SU) { 2193 unsigned MaxHeight = 0; 2194 for (SUnit::const_succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 2195 I != E; ++I) { 2196 if (I->isCtrl()) continue; // ignore chain succs 2197 unsigned Height = I->getSUnit()->getHeight(); 2198 // If there are bunch of CopyToRegs stacked up, they should be considered 2199 // to be at the same position. 2200 if (I->getSUnit()->getNode() && 2201 I->getSUnit()->getNode()->getOpcode() == ISD::CopyToReg) 2202 Height = closestSucc(I->getSUnit())+1; 2203 if (Height > MaxHeight) 2204 MaxHeight = Height; 2205 } 2206 return MaxHeight; 2207} 2208 2209/// calcMaxScratches - Returns an cost estimate of the worse case requirement 2210/// for scratch registers, i.e. number of data dependencies. 2211static unsigned calcMaxScratches(const SUnit *SU) { 2212 unsigned Scratches = 0; 2213 for (SUnit::const_pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 2214 I != E; ++I) { 2215 if (I->isCtrl()) continue; // ignore chain preds 2216 Scratches++; 2217 } 2218 return Scratches; 2219} 2220 2221/// hasOnlyLiveInOpers - Return true if SU has only value predecessors that are 2222/// CopyFromReg from a virtual register. 2223static bool hasOnlyLiveInOpers(const SUnit *SU) { 2224 bool RetVal = false; 2225 for (SUnit::const_pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 2226 I != E; ++I) { 2227 if (I->isCtrl()) continue; 2228 const SUnit *PredSU = I->getSUnit(); 2229 if (PredSU->getNode() && 2230 PredSU->getNode()->getOpcode() == ISD::CopyFromReg) { 2231 unsigned Reg = 2232 cast<RegisterSDNode>(PredSU->getNode()->getOperand(1))->getReg(); 2233 if (TargetRegisterInfo::isVirtualRegister(Reg)) { 2234 RetVal = true; 2235 continue; 2236 } 2237 } 2238 return false; 2239 } 2240 return RetVal; 2241} 2242 2243/// hasOnlyLiveOutUses - Return true if SU has only value successors that are 2244/// CopyToReg to a virtual register. This SU def is probably a liveout and 2245/// it has no other use. It should be scheduled closer to the terminator. 2246static bool hasOnlyLiveOutUses(const SUnit *SU) { 2247 bool RetVal = false; 2248 for (SUnit::const_succ_iterator I = SU->Succs.begin(), E = SU->Succs.end(); 2249 I != E; ++I) { 2250 if (I->isCtrl()) continue; 2251 const SUnit *SuccSU = I->getSUnit(); 2252 if (SuccSU->getNode() && SuccSU->getNode()->getOpcode() == ISD::CopyToReg) { 2253 unsigned Reg = 2254 cast<RegisterSDNode>(SuccSU->getNode()->getOperand(1))->getReg(); 2255 if (TargetRegisterInfo::isVirtualRegister(Reg)) { 2256 RetVal = true; 2257 continue; 2258 } 2259 } 2260 return false; 2261 } 2262 return RetVal; 2263} 2264 2265// Set isVRegCycle for a node with only live in opers and live out uses. Also 2266// set isVRegCycle for its CopyFromReg operands. 2267// 2268// This is only relevant for single-block loops, in which case the VRegCycle 2269// node is likely an induction variable in which the operand and target virtual 2270// registers should be coalesced (e.g. pre/post increment values). Setting the 2271// isVRegCycle flag helps the scheduler prioritize other uses of the same 2272// CopyFromReg so that this node becomes the virtual register "kill". This 2273// avoids interference between the values live in and out of the block and 2274// eliminates a copy inside the loop. 2275static void initVRegCycle(SUnit *SU) { 2276 if (DisableSchedVRegCycle) 2277 return; 2278 2279 if (!hasOnlyLiveInOpers(SU) || !hasOnlyLiveOutUses(SU)) 2280 return; 2281 2282 DEBUG(dbgs() << "VRegCycle: SU(" << SU->NodeNum << ")\n"); 2283 2284 SU->isVRegCycle = true; 2285 2286 for (SUnit::const_pred_iterator I = SU->Preds.begin(), E = SU->Preds.end(); 2287 I != E; ++I) { 2288 if (I->isCtrl()) continue; 2289 I->getSUnit()->isVRegCycle = true; 2290 } 2291} 2292 2293// After scheduling the definition of a VRegCycle, clear the isVRegCycle flag of 2294// CopyFromReg operands. We should no longer penalize other uses of this VReg. 2295static void resetVRegCycle(SUnit *SU) { 2296 if (!SU->isVRegCycle) 2297 return; 2298 2299 for (SUnit::const_pred_iterator I = SU->Preds.begin(),E = SU->Preds.end(); 2300 I != E; ++I) { 2301 if (I->isCtrl()) continue; // ignore chain preds 2302 SUnit *PredSU = I->getSUnit(); 2303 if (PredSU->isVRegCycle) { 2304 assert(PredSU->getNode()->getOpcode() == ISD::CopyFromReg && 2305 "VRegCycle def must be CopyFromReg"); 2306 I->getSUnit()->isVRegCycle = 0; 2307 } 2308 } 2309} 2310 2311// Return true if this SUnit uses a CopyFromReg node marked as a VRegCycle. This 2312// means a node that defines the VRegCycle has not been scheduled yet. 2313static bool hasVRegCycleUse(const SUnit *SU) { 2314 // If this SU also defines the VReg, don't hoist it as a "use". 2315 if (SU->isVRegCycle) 2316 return false; 2317 2318 for (SUnit::const_pred_iterator I = SU->Preds.begin(),E = SU->Preds.end(); 2319 I != E; ++I) { 2320 if (I->isCtrl()) continue; // ignore chain preds 2321 if (I->getSUnit()->isVRegCycle && 2322 I->getSUnit()->getNode()->getOpcode() == ISD::CopyFromReg) { 2323 DEBUG(dbgs() << " VReg cycle use: SU (" << SU->NodeNum << ")\n"); 2324 return true; 2325 } 2326 } 2327 return false; 2328} 2329 2330// Check for either a dependence (latency) or resource (hazard) stall. 2331// 2332// Note: The ScheduleHazardRecognizer interface requires a non-const SU. 2333static bool BUHasStall(SUnit *SU, int Height, RegReductionPQBase *SPQ) { 2334 if ((int)SPQ->getCurCycle() < Height) return true; 2335 if (SPQ->getHazardRec()->getHazardType(SU, 0) 2336 != ScheduleHazardRecognizer::NoHazard) 2337 return true; 2338 return false; 2339} 2340 2341// Return -1 if left has higher priority, 1 if right has higher priority. 2342// Return 0 if latency-based priority is equivalent. 2343static int BUCompareLatency(SUnit *left, SUnit *right, bool checkPref, 2344 RegReductionPQBase *SPQ) { 2345 // Scheduling an instruction that uses a VReg whose postincrement has not yet 2346 // been scheduled will induce a copy. Model this as an extra cycle of latency. 2347 int LPenalty = hasVRegCycleUse(left) ? 1 : 0; 2348 int RPenalty = hasVRegCycleUse(right) ? 1 : 0; 2349 int LHeight = (int)left->getHeight() + LPenalty; 2350 int RHeight = (int)right->getHeight() + RPenalty; 2351 2352 bool LStall = (!checkPref || left->SchedulingPref == Sched::ILP) && 2353 BUHasStall(left, LHeight, SPQ); 2354 bool RStall = (!checkPref || right->SchedulingPref == Sched::ILP) && 2355 BUHasStall(right, RHeight, SPQ); 2356 2357 // If scheduling one of the node will cause a pipeline stall, delay it. 2358 // If scheduling either one of the node will cause a pipeline stall, sort 2359 // them according to their height. 2360 if (LStall) { 2361 if (!RStall) 2362 return 1; 2363 if (LHeight != RHeight) 2364 return LHeight > RHeight ? 1 : -1; 2365 } else if (RStall) 2366 return -1; 2367 2368 // If either node is scheduling for latency, sort them by height/depth 2369 // and latency. 2370 if (!checkPref || (left->SchedulingPref == Sched::ILP || 2371 right->SchedulingPref == Sched::ILP)) { 2372 // If neither instruction stalls (!LStall && !RStall) and HazardRecognizer 2373 // is enabled, grouping instructions by cycle, then its height is already 2374 // covered so only its depth matters. We also reach this point if both stall 2375 // but have the same height. 2376 if (!SPQ->getHazardRec()->isEnabled()) { 2377 if (LHeight != RHeight) 2378 return LHeight > RHeight ? 1 : -1; 2379 } 2380 int LDepth = left->getDepth() - LPenalty; 2381 int RDepth = right->getDepth() - RPenalty; 2382 if (LDepth != RDepth) { 2383 DEBUG(dbgs() << " Comparing latency of SU (" << left->NodeNum 2384 << ") depth " << LDepth << " vs SU (" << right->NodeNum 2385 << ") depth " << RDepth << "\n"); 2386 return LDepth < RDepth ? 1 : -1; 2387 } 2388 if (left->Latency != right->Latency) 2389 return left->Latency > right->Latency ? 1 : -1; 2390 } 2391 return 0; 2392} 2393 2394static bool BURRSort(SUnit *left, SUnit *right, RegReductionPQBase *SPQ) { 2395 // Schedule physical register definitions close to their use. This is 2396 // motivated by microarchitectures that can fuse cmp+jump macro-ops. But as 2397 // long as shortening physreg live ranges is generally good, we can defer 2398 // creating a subtarget hook. 2399 if (!DisableSchedPhysRegJoin) { 2400 bool LHasPhysReg = left->hasPhysRegDefs; 2401 bool RHasPhysReg = right->hasPhysRegDefs; 2402 if (LHasPhysReg != RHasPhysReg) { 2403 #ifndef NDEBUG 2404 static const char *const PhysRegMsg[] = { " has no physreg", 2405 " defines a physreg" }; 2406 #endif 2407 DEBUG(dbgs() << " SU (" << left->NodeNum << ") " 2408 << PhysRegMsg[LHasPhysReg] << " SU(" << right->NodeNum << ") " 2409 << PhysRegMsg[RHasPhysReg] << "\n"); 2410 return LHasPhysReg < RHasPhysReg; 2411 } 2412 } 2413 2414 // Prioritize by Sethi-Ulmann number and push CopyToReg nodes down. 2415 unsigned LPriority = SPQ->getNodePriority(left); 2416 unsigned RPriority = SPQ->getNodePriority(right); 2417 2418 // Be really careful about hoisting call operands above previous calls. 2419 // Only allows it if it would reduce register pressure. 2420 if (left->isCall && right->isCallOp) { 2421 unsigned RNumVals = right->getNode()->getNumValues(); 2422 RPriority = (RPriority > RNumVals) ? (RPriority - RNumVals) : 0; 2423 } 2424 if (right->isCall && left->isCallOp) { 2425 unsigned LNumVals = left->getNode()->getNumValues(); 2426 LPriority = (LPriority > LNumVals) ? (LPriority - LNumVals) : 0; 2427 } 2428 2429 if (LPriority != RPriority) 2430 return LPriority > RPriority; 2431 2432 // One or both of the nodes are calls and their sethi-ullman numbers are the 2433 // same, then keep source order. 2434 if (left->isCall || right->isCall) { 2435 unsigned LOrder = SPQ->getNodeOrdering(left); 2436 unsigned ROrder = SPQ->getNodeOrdering(right); 2437 2438 // Prefer an ordering where the lower the non-zero order number, the higher 2439 // the preference. 2440 if ((LOrder || ROrder) && LOrder != ROrder) 2441 return LOrder != 0 && (LOrder < ROrder || ROrder == 0); 2442 } 2443 2444 // Try schedule def + use closer when Sethi-Ullman numbers are the same. 2445 // e.g. 2446 // t1 = op t2, c1 2447 // t3 = op t4, c2 2448 // 2449 // and the following instructions are both ready. 2450 // t2 = op c3 2451 // t4 = op c4 2452 // 2453 // Then schedule t2 = op first. 2454 // i.e. 2455 // t4 = op c4 2456 // t2 = op c3 2457 // t1 = op t2, c1 2458 // t3 = op t4, c2 2459 // 2460 // This creates more short live intervals. 2461 unsigned LDist = closestSucc(left); 2462 unsigned RDist = closestSucc(right); 2463 if (LDist != RDist) 2464 return LDist < RDist; 2465 2466 // How many registers becomes live when the node is scheduled. 2467 unsigned LScratch = calcMaxScratches(left); 2468 unsigned RScratch = calcMaxScratches(right); 2469 if (LScratch != RScratch) 2470 return LScratch > RScratch; 2471 2472 // Comparing latency against a call makes little sense unless the node 2473 // is register pressure-neutral. 2474 if ((left->isCall && RPriority > 0) || (right->isCall && LPriority > 0)) 2475 return (left->NodeQueueId > right->NodeQueueId); 2476 2477 // Do not compare latencies when one or both of the nodes are calls. 2478 if (!DisableSchedCycles && 2479 !(left->isCall || right->isCall)) { 2480 int result = BUCompareLatency(left, right, false /*checkPref*/, SPQ); 2481 if (result != 0) 2482 return result > 0; 2483 } 2484 else { 2485 if (left->getHeight() != right->getHeight()) 2486 return left->getHeight() > right->getHeight(); 2487 2488 if (left->getDepth() != right->getDepth()) 2489 return left->getDepth() < right->getDepth(); 2490 } 2491 2492 assert(left->NodeQueueId && right->NodeQueueId && 2493 "NodeQueueId cannot be zero"); 2494 return (left->NodeQueueId > right->NodeQueueId); 2495} 2496 2497// Bottom up 2498bool bu_ls_rr_sort::operator()(SUnit *left, SUnit *right) const { 2499 if (int res = checkSpecialNodes(left, right)) 2500 return res > 0; 2501 2502 return BURRSort(left, right, SPQ); 2503} 2504 2505// Source order, otherwise bottom up. 2506bool src_ls_rr_sort::operator()(SUnit *left, SUnit *right) const { 2507 if (int res = checkSpecialNodes(left, right)) 2508 return res > 0; 2509 2510 unsigned LOrder = SPQ->getNodeOrdering(left); 2511 unsigned ROrder = SPQ->getNodeOrdering(right); 2512 2513 // Prefer an ordering where the lower the non-zero order number, the higher 2514 // the preference. 2515 if ((LOrder || ROrder) && LOrder != ROrder) 2516 return LOrder != 0 && (LOrder < ROrder || ROrder == 0); 2517 2518 return BURRSort(left, right, SPQ); 2519} 2520 2521// If the time between now and when the instruction will be ready can cover 2522// the spill code, then avoid adding it to the ready queue. This gives long 2523// stalls highest priority and allows hoisting across calls. It should also 2524// speed up processing the available queue. 2525bool hybrid_ls_rr_sort::isReady(SUnit *SU, unsigned CurCycle) const { 2526 static const unsigned ReadyDelay = 3; 2527 2528 if (SPQ->MayReduceRegPressure(SU)) return true; 2529 2530 if (SU->getHeight() > (CurCycle + ReadyDelay)) return false; 2531 2532 if (SPQ->getHazardRec()->getHazardType(SU, -ReadyDelay) 2533 != ScheduleHazardRecognizer::NoHazard) 2534 return false; 2535 2536 return true; 2537} 2538 2539// Return true if right should be scheduled with higher priority than left. 2540bool hybrid_ls_rr_sort::operator()(SUnit *left, SUnit *right) const { 2541 if (int res = checkSpecialNodes(left, right)) 2542 return res > 0; 2543 2544 if (left->isCall || right->isCall) 2545 // No way to compute latency of calls. 2546 return BURRSort(left, right, SPQ); 2547 2548 bool LHigh = SPQ->HighRegPressure(left); 2549 bool RHigh = SPQ->HighRegPressure(right); 2550 // Avoid causing spills. If register pressure is high, schedule for 2551 // register pressure reduction. 2552 if (LHigh && !RHigh) { 2553 DEBUG(dbgs() << " pressure SU(" << left->NodeNum << ") > SU(" 2554 << right->NodeNum << ")\n"); 2555 return true; 2556 } 2557 else if (!LHigh && RHigh) { 2558 DEBUG(dbgs() << " pressure SU(" << right->NodeNum << ") > SU(" 2559 << left->NodeNum << ")\n"); 2560 return false; 2561 } 2562 if (!LHigh && !RHigh) { 2563 int result = BUCompareLatency(left, right, true /*checkPref*/, SPQ); 2564 if (result != 0) 2565 return result > 0; 2566 } 2567 return BURRSort(left, right, SPQ); 2568} 2569 2570// Schedule as many instructions in each cycle as possible. So don't make an 2571// instruction available unless it is ready in the current cycle. 2572bool ilp_ls_rr_sort::isReady(SUnit *SU, unsigned CurCycle) const { 2573 if (SU->getHeight() > CurCycle) return false; 2574 2575 if (SPQ->getHazardRec()->getHazardType(SU, 0) 2576 != ScheduleHazardRecognizer::NoHazard) 2577 return false; 2578 2579 return true; 2580} 2581 2582static bool canEnableCoalescing(SUnit *SU) { 2583 unsigned Opc = SU->getNode() ? SU->getNode()->getOpcode() : 0; 2584 if (Opc == ISD::TokenFactor || Opc == ISD::CopyToReg) 2585 // CopyToReg should be close to its uses to facilitate coalescing and 2586 // avoid spilling. 2587 return true; 2588 2589 if (Opc == TargetOpcode::EXTRACT_SUBREG || 2590 Opc == TargetOpcode::SUBREG_TO_REG || 2591 Opc == TargetOpcode::INSERT_SUBREG) 2592 // EXTRACT_SUBREG, INSERT_SUBREG, and SUBREG_TO_REG nodes should be 2593 // close to their uses to facilitate coalescing. 2594 return true; 2595 2596 if (SU->NumPreds == 0 && SU->NumSuccs != 0) 2597 // If SU does not have a register def, schedule it close to its uses 2598 // because it does not lengthen any live ranges. 2599 return true; 2600 2601 return false; 2602} 2603 2604// list-ilp is currently an experimental scheduler that allows various 2605// heuristics to be enabled prior to the normal register reduction logic. 2606bool ilp_ls_rr_sort::operator()(SUnit *left, SUnit *right) const { 2607 if (int res = checkSpecialNodes(left, right)) 2608 return res > 0; 2609 2610 if (left->isCall || right->isCall) 2611 // No way to compute latency of calls. 2612 return BURRSort(left, right, SPQ); 2613 2614 unsigned LLiveUses = 0, RLiveUses = 0; 2615 int LPDiff = 0, RPDiff = 0; 2616 if (!DisableSchedRegPressure || !DisableSchedLiveUses) { 2617 LPDiff = SPQ->RegPressureDiff(left, LLiveUses); 2618 RPDiff = SPQ->RegPressureDiff(right, RLiveUses); 2619 } 2620 if (!DisableSchedRegPressure && LPDiff != RPDiff) { 2621 DEBUG(dbgs() << "RegPressureDiff SU(" << left->NodeNum << "): " << LPDiff 2622 << " != SU(" << right->NodeNum << "): " << RPDiff << "\n"); 2623 return LPDiff > RPDiff; 2624 } 2625 2626 if (!DisableSchedRegPressure && (LPDiff > 0 || RPDiff > 0)) { 2627 bool LReduce = canEnableCoalescing(left); 2628 bool RReduce = canEnableCoalescing(right); 2629 if (LReduce && !RReduce) return false; 2630 if (RReduce && !LReduce) return true; 2631 } 2632 2633 if (!DisableSchedLiveUses && (LLiveUses != RLiveUses)) { 2634 DEBUG(dbgs() << "Live uses SU(" << left->NodeNum << "): " << LLiveUses 2635 << " != SU(" << right->NodeNum << "): " << RLiveUses << "\n"); 2636 return LLiveUses < RLiveUses; 2637 } 2638 2639 if (!DisableSchedStalls) { 2640 bool LStall = BUHasStall(left, left->getHeight(), SPQ); 2641 bool RStall = BUHasStall(right, right->getHeight(), SPQ); 2642 if (LStall != RStall) 2643 return left->getHeight() > right->getHeight(); 2644 } 2645 2646 if (!DisableSchedCriticalPath) { 2647 int spread = (int)left->getDepth() - (int)right->getDepth(); 2648 if (std::abs(spread) > MaxReorderWindow) { 2649 DEBUG(dbgs() << "Depth of SU(" << left->NodeNum << "): " 2650 << left->getDepth() << " != SU(" << right->NodeNum << "): " 2651 << right->getDepth() << "\n"); 2652 return left->getDepth() < right->getDepth(); 2653 } 2654 } 2655 2656 if (!DisableSchedHeight && left->getHeight() != right->getHeight()) { 2657 int spread = (int)left->getHeight() - (int)right->getHeight(); 2658 if (std::abs(spread) > MaxReorderWindow) 2659 return left->getHeight() > right->getHeight(); 2660 } 2661 2662 return BURRSort(left, right, SPQ); 2663} 2664 2665void RegReductionPQBase::initNodes(std::vector<SUnit> &sunits) { 2666 SUnits = &sunits; 2667 // Add pseudo dependency edges for two-address nodes. 2668 if (!Disable2AddrHack) 2669 AddPseudoTwoAddrDeps(); 2670 // Reroute edges to nodes with multiple uses. 2671 if (!TracksRegPressure && !SrcOrder) 2672 PrescheduleNodesWithMultipleUses(); 2673 // Calculate node priorities. 2674 CalculateSethiUllmanNumbers(); 2675 2676 // For single block loops, mark nodes that look like canonical IV increments. 2677 if (scheduleDAG->BB->isSuccessor(scheduleDAG->BB)) { 2678 for (unsigned i = 0, e = sunits.size(); i != e; ++i) { 2679 initVRegCycle(&sunits[i]); 2680 } 2681 } 2682} 2683 2684//===----------------------------------------------------------------------===// 2685// Preschedule for Register Pressure 2686//===----------------------------------------------------------------------===// 2687 2688bool RegReductionPQBase::canClobber(const SUnit *SU, const SUnit *Op) { 2689 if (SU->isTwoAddress) { 2690 unsigned Opc = SU->getNode()->getMachineOpcode(); 2691 const MCInstrDesc &MCID = TII->get(Opc); 2692 unsigned NumRes = MCID.getNumDefs(); 2693 unsigned NumOps = MCID.getNumOperands() - NumRes; 2694 for (unsigned i = 0; i != NumOps; ++i) { 2695 if (MCID.getOperandConstraint(i+NumRes, MCOI::TIED_TO) != -1) { 2696 SDNode *DU = SU->getNode()->getOperand(i).getNode(); 2697 if (DU->getNodeId() != -1 && 2698 Op->OrigNode == &(*SUnits)[DU->getNodeId()]) 2699 return true; 2700 } 2701 } 2702 } 2703 return false; 2704} 2705 2706/// canClobberReachingPhysRegUse - True if SU would clobber one of it's 2707/// successor's explicit physregs whose definition can reach DepSU. 2708/// i.e. DepSU should not be scheduled above SU. 2709static bool canClobberReachingPhysRegUse(const SUnit *DepSU, const SUnit *SU, 2710 ScheduleDAGRRList *scheduleDAG, 2711 const TargetInstrInfo *TII, 2712 const TargetRegisterInfo *TRI) { 2713 const uint16_t *ImpDefs 2714 = TII->get(SU->getNode()->getMachineOpcode()).getImplicitDefs(); 2715 const uint32_t *RegMask = getNodeRegMask(SU->getNode()); 2716 if(!ImpDefs && !RegMask) 2717 return false; 2718 2719 for (SUnit::const_succ_iterator SI = SU->Succs.begin(), SE = SU->Succs.end(); 2720 SI != SE; ++SI) { 2721 SUnit *SuccSU = SI->getSUnit(); 2722 for (SUnit::const_pred_iterator PI = SuccSU->Preds.begin(), 2723 PE = SuccSU->Preds.end(); PI != PE; ++PI) { 2724 if (!PI->isAssignedRegDep()) 2725 continue; 2726 2727 if (RegMask && MachineOperand::clobbersPhysReg(RegMask, PI->getReg()) && 2728 scheduleDAG->IsReachable(DepSU, PI->getSUnit())) 2729 return true; 2730 2731 if (ImpDefs) 2732 for (const uint16_t *ImpDef = ImpDefs; *ImpDef; ++ImpDef) 2733 // Return true if SU clobbers this physical register use and the 2734 // definition of the register reaches from DepSU. IsReachable queries 2735 // a topological forward sort of the DAG (following the successors). 2736 if (TRI->regsOverlap(*ImpDef, PI->getReg()) && 2737 scheduleDAG->IsReachable(DepSU, PI->getSUnit())) 2738 return true; 2739 } 2740 } 2741 return false; 2742} 2743 2744/// canClobberPhysRegDefs - True if SU would clobber one of SuccSU's 2745/// physical register defs. 2746static bool canClobberPhysRegDefs(const SUnit *SuccSU, const SUnit *SU, 2747 const TargetInstrInfo *TII, 2748 const TargetRegisterInfo *TRI) { 2749 SDNode *N = SuccSU->getNode(); 2750 unsigned NumDefs = TII->get(N->getMachineOpcode()).getNumDefs(); 2751 const uint16_t *ImpDefs = TII->get(N->getMachineOpcode()).getImplicitDefs(); 2752 assert(ImpDefs && "Caller should check hasPhysRegDefs"); 2753 for (const SDNode *SUNode = SU->getNode(); SUNode; 2754 SUNode = SUNode->getGluedNode()) { 2755 if (!SUNode->isMachineOpcode()) 2756 continue; 2757 const uint16_t *SUImpDefs = 2758 TII->get(SUNode->getMachineOpcode()).getImplicitDefs(); 2759 const uint32_t *SURegMask = getNodeRegMask(SUNode); 2760 if (!SUImpDefs && !SURegMask) 2761 continue; 2762 for (unsigned i = NumDefs, e = N->getNumValues(); i != e; ++i) { 2763 EVT VT = N->getValueType(i); 2764 if (VT == MVT::Glue || VT == MVT::Other) 2765 continue; 2766 if (!N->hasAnyUseOfValue(i)) 2767 continue; 2768 unsigned Reg = ImpDefs[i - NumDefs]; 2769 if (SURegMask && MachineOperand::clobbersPhysReg(SURegMask, Reg)) 2770 return true; 2771 if (!SUImpDefs) 2772 continue; 2773 for (;*SUImpDefs; ++SUImpDefs) { 2774 unsigned SUReg = *SUImpDefs; 2775 if (TRI->regsOverlap(Reg, SUReg)) 2776 return true; 2777 } 2778 } 2779 } 2780 return false; 2781} 2782 2783/// PrescheduleNodesWithMultipleUses - Nodes with multiple uses 2784/// are not handled well by the general register pressure reduction 2785/// heuristics. When presented with code like this: 2786/// 2787/// N 2788/// / | 2789/// / | 2790/// U store 2791/// | 2792/// ... 2793/// 2794/// the heuristics tend to push the store up, but since the 2795/// operand of the store has another use (U), this would increase 2796/// the length of that other use (the U->N edge). 2797/// 2798/// This function transforms code like the above to route U's 2799/// dependence through the store when possible, like this: 2800/// 2801/// N 2802/// || 2803/// || 2804/// store 2805/// | 2806/// U 2807/// | 2808/// ... 2809/// 2810/// This results in the store being scheduled immediately 2811/// after N, which shortens the U->N live range, reducing 2812/// register pressure. 2813/// 2814void RegReductionPQBase::PrescheduleNodesWithMultipleUses() { 2815 // Visit all the nodes in topological order, working top-down. 2816 for (unsigned i = 0, e = SUnits->size(); i != e; ++i) { 2817 SUnit *SU = &(*SUnits)[i]; 2818 // For now, only look at nodes with no data successors, such as stores. 2819 // These are especially important, due to the heuristics in 2820 // getNodePriority for nodes with no data successors. 2821 if (SU->NumSuccs != 0) 2822 continue; 2823 // For now, only look at nodes with exactly one data predecessor. 2824 if (SU->NumPreds != 1) 2825 continue; 2826 // Avoid prescheduling copies to virtual registers, which don't behave 2827 // like other nodes from the perspective of scheduling heuristics. 2828 if (SDNode *N = SU->getNode()) 2829 if (N->getOpcode() == ISD::CopyToReg && 2830 TargetRegisterInfo::isVirtualRegister 2831 (cast<RegisterSDNode>(N->getOperand(1))->getReg())) 2832 continue; 2833 2834 // Locate the single data predecessor. 2835 SUnit *PredSU = 0; 2836 for (SUnit::const_pred_iterator II = SU->Preds.begin(), 2837 EE = SU->Preds.end(); II != EE; ++II) 2838 if (!II->isCtrl()) { 2839 PredSU = II->getSUnit(); 2840 break; 2841 } 2842 assert(PredSU); 2843 2844 // Don't rewrite edges that carry physregs, because that requires additional 2845 // support infrastructure. 2846 if (PredSU->hasPhysRegDefs) 2847 continue; 2848 // Short-circuit the case where SU is PredSU's only data successor. 2849 if (PredSU->NumSuccs == 1) 2850 continue; 2851 // Avoid prescheduling to copies from virtual registers, which don't behave 2852 // like other nodes from the perspective of scheduling heuristics. 2853 if (SDNode *N = SU->getNode()) 2854 if (N->getOpcode() == ISD::CopyFromReg && 2855 TargetRegisterInfo::isVirtualRegister 2856 (cast<RegisterSDNode>(N->getOperand(1))->getReg())) 2857 continue; 2858 2859 // Perform checks on the successors of PredSU. 2860 for (SUnit::const_succ_iterator II = PredSU->Succs.begin(), 2861 EE = PredSU->Succs.end(); II != EE; ++II) { 2862 SUnit *PredSuccSU = II->getSUnit(); 2863 if (PredSuccSU == SU) continue; 2864 // If PredSU has another successor with no data successors, for 2865 // now don't attempt to choose either over the other. 2866 if (PredSuccSU->NumSuccs == 0) 2867 goto outer_loop_continue; 2868 // Don't break physical register dependencies. 2869 if (SU->hasPhysRegClobbers && PredSuccSU->hasPhysRegDefs) 2870 if (canClobberPhysRegDefs(PredSuccSU, SU, TII, TRI)) 2871 goto outer_loop_continue; 2872 // Don't introduce graph cycles. 2873 if (scheduleDAG->IsReachable(SU, PredSuccSU)) 2874 goto outer_loop_continue; 2875 } 2876 2877 // Ok, the transformation is safe and the heuristics suggest it is 2878 // profitable. Update the graph. 2879 DEBUG(dbgs() << " Prescheduling SU #" << SU->NodeNum 2880 << " next to PredSU #" << PredSU->NodeNum 2881 << " to guide scheduling in the presence of multiple uses\n"); 2882 for (unsigned i = 0; i != PredSU->Succs.size(); ++i) { 2883 SDep Edge = PredSU->Succs[i]; 2884 assert(!Edge.isAssignedRegDep()); 2885 SUnit *SuccSU = Edge.getSUnit(); 2886 if (SuccSU != SU) { 2887 Edge.setSUnit(PredSU); 2888 scheduleDAG->RemovePred(SuccSU, Edge); 2889 scheduleDAG->AddPred(SU, Edge); 2890 Edge.setSUnit(SU); 2891 scheduleDAG->AddPred(SuccSU, Edge); 2892 --i; 2893 } 2894 } 2895 outer_loop_continue:; 2896 } 2897} 2898 2899/// AddPseudoTwoAddrDeps - If two nodes share an operand and one of them uses 2900/// it as a def&use operand. Add a pseudo control edge from it to the other 2901/// node (if it won't create a cycle) so the two-address one will be scheduled 2902/// first (lower in the schedule). If both nodes are two-address, favor the 2903/// one that has a CopyToReg use (more likely to be a loop induction update). 2904/// If both are two-address, but one is commutable while the other is not 2905/// commutable, favor the one that's not commutable. 2906void RegReductionPQBase::AddPseudoTwoAddrDeps() { 2907 for (unsigned i = 0, e = SUnits->size(); i != e; ++i) { 2908 SUnit *SU = &(*SUnits)[i]; 2909 if (!SU->isTwoAddress) 2910 continue; 2911 2912 SDNode *Node = SU->getNode(); 2913 if (!Node || !Node->isMachineOpcode() || SU->getNode()->getGluedNode()) 2914 continue; 2915 2916 bool isLiveOut = hasOnlyLiveOutUses(SU); 2917 unsigned Opc = Node->getMachineOpcode(); 2918 const MCInstrDesc &MCID = TII->get(Opc); 2919 unsigned NumRes = MCID.getNumDefs(); 2920 unsigned NumOps = MCID.getNumOperands() - NumRes; 2921 for (unsigned j = 0; j != NumOps; ++j) { 2922 if (MCID.getOperandConstraint(j+NumRes, MCOI::TIED_TO) == -1) 2923 continue; 2924 SDNode *DU = SU->getNode()->getOperand(j).getNode(); 2925 if (DU->getNodeId() == -1) 2926 continue; 2927 const SUnit *DUSU = &(*SUnits)[DU->getNodeId()]; 2928 if (!DUSU) continue; 2929 for (SUnit::const_succ_iterator I = DUSU->Succs.begin(), 2930 E = DUSU->Succs.end(); I != E; ++I) { 2931 if (I->isCtrl()) continue; 2932 SUnit *SuccSU = I->getSUnit(); 2933 if (SuccSU == SU) 2934 continue; 2935 // Be conservative. Ignore if nodes aren't at roughly the same 2936 // depth and height. 2937 if (SuccSU->getHeight() < SU->getHeight() && 2938 (SU->getHeight() - SuccSU->getHeight()) > 1) 2939 continue; 2940 // Skip past COPY_TO_REGCLASS nodes, so that the pseudo edge 2941 // constrains whatever is using the copy, instead of the copy 2942 // itself. In the case that the copy is coalesced, this 2943 // preserves the intent of the pseudo two-address heurietics. 2944 while (SuccSU->Succs.size() == 1 && 2945 SuccSU->getNode()->isMachineOpcode() && 2946 SuccSU->getNode()->getMachineOpcode() == 2947 TargetOpcode::COPY_TO_REGCLASS) 2948 SuccSU = SuccSU->Succs.front().getSUnit(); 2949 // Don't constrain non-instruction nodes. 2950 if (!SuccSU->getNode() || !SuccSU->getNode()->isMachineOpcode()) 2951 continue; 2952 // Don't constrain nodes with physical register defs if the 2953 // predecessor can clobber them. 2954 if (SuccSU->hasPhysRegDefs && SU->hasPhysRegClobbers) { 2955 if (canClobberPhysRegDefs(SuccSU, SU, TII, TRI)) 2956 continue; 2957 } 2958 // Don't constrain EXTRACT_SUBREG, INSERT_SUBREG, and SUBREG_TO_REG; 2959 // these may be coalesced away. We want them close to their uses. 2960 unsigned SuccOpc = SuccSU->getNode()->getMachineOpcode(); 2961 if (SuccOpc == TargetOpcode::EXTRACT_SUBREG || 2962 SuccOpc == TargetOpcode::INSERT_SUBREG || 2963 SuccOpc == TargetOpcode::SUBREG_TO_REG) 2964 continue; 2965 if (!canClobberReachingPhysRegUse(SuccSU, SU, scheduleDAG, TII, TRI) && 2966 (!canClobber(SuccSU, DUSU) || 2967 (isLiveOut && !hasOnlyLiveOutUses(SuccSU)) || 2968 (!SU->isCommutable && SuccSU->isCommutable)) && 2969 !scheduleDAG->IsReachable(SuccSU, SU)) { 2970 DEBUG(dbgs() << " Adding a pseudo-two-addr edge from SU #" 2971 << SU->NodeNum << " to SU #" << SuccSU->NodeNum << "\n"); 2972 scheduleDAG->AddPred(SU, SDep(SuccSU, SDep::Artificial)); 2973 } 2974 } 2975 } 2976 } 2977} 2978 2979//===----------------------------------------------------------------------===// 2980// Public Constructor Functions 2981//===----------------------------------------------------------------------===// 2982 2983llvm::ScheduleDAGSDNodes * 2984llvm::createBURRListDAGScheduler(SelectionDAGISel *IS, 2985 CodeGenOpt::Level OptLevel) { 2986 const TargetMachine &TM = IS->TM; 2987 const TargetInstrInfo *TII = TM.getInstrInfo(); 2988 const TargetRegisterInfo *TRI = TM.getRegisterInfo(); 2989 2990 BURegReductionPriorityQueue *PQ = 2991 new BURegReductionPriorityQueue(*IS->MF, false, false, TII, TRI, 0); 2992 ScheduleDAGRRList *SD = new ScheduleDAGRRList(*IS->MF, false, PQ, OptLevel); 2993 PQ->setScheduleDAG(SD); 2994 return SD; 2995} 2996 2997llvm::ScheduleDAGSDNodes * 2998llvm::createSourceListDAGScheduler(SelectionDAGISel *IS, 2999 CodeGenOpt::Level OptLevel) { 3000 const TargetMachine &TM = IS->TM; 3001 const TargetInstrInfo *TII = TM.getInstrInfo(); 3002 const TargetRegisterInfo *TRI = TM.getRegisterInfo(); 3003 3004 SrcRegReductionPriorityQueue *PQ = 3005 new SrcRegReductionPriorityQueue(*IS->MF, false, true, TII, TRI, 0); 3006 ScheduleDAGRRList *SD = new ScheduleDAGRRList(*IS->MF, false, PQ, OptLevel); 3007 PQ->setScheduleDAG(SD); 3008 return SD; 3009} 3010 3011llvm::ScheduleDAGSDNodes * 3012llvm::createHybridListDAGScheduler(SelectionDAGISel *IS, 3013 CodeGenOpt::Level OptLevel) { 3014 const TargetMachine &TM = IS->TM; 3015 const TargetInstrInfo *TII = TM.getInstrInfo(); 3016 const TargetRegisterInfo *TRI = TM.getRegisterInfo(); 3017 const TargetLowering *TLI = IS->getTargetLowering(); 3018 3019 HybridBURRPriorityQueue *PQ = 3020 new HybridBURRPriorityQueue(*IS->MF, true, false, TII, TRI, TLI); 3021 3022 ScheduleDAGRRList *SD = new ScheduleDAGRRList(*IS->MF, true, PQ, OptLevel); 3023 PQ->setScheduleDAG(SD); 3024 return SD; 3025} 3026 3027llvm::ScheduleDAGSDNodes * 3028llvm::createILPListDAGScheduler(SelectionDAGISel *IS, 3029 CodeGenOpt::Level OptLevel) { 3030 const TargetMachine &TM = IS->TM; 3031 const TargetInstrInfo *TII = TM.getInstrInfo(); 3032 const TargetRegisterInfo *TRI = TM.getRegisterInfo(); 3033 const TargetLowering *TLI = IS->getTargetLowering(); 3034 3035 ILPBURRPriorityQueue *PQ = 3036 new ILPBURRPriorityQueue(*IS->MF, true, false, TII, TRI, TLI); 3037 ScheduleDAGRRList *SD = new ScheduleDAGRRList(*IS->MF, true, PQ, OptLevel); 3038 PQ->setScheduleDAG(SD); 3039 return SD; 3040} 3041