1 //===-- AMDILCFGStructurizer.cpp - CFG Structurizer -----------------------===// 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 /// \file 9 //==-----------------------------------------------------------------------===// 10 11 #include "AMDGPU.h" 12 #include "AMDGPUInstrInfo.h" 13 #include "AMDGPUSubtarget.h" 14 #include "R600InstrInfo.h" 15 #include "llvm/ADT/DepthFirstIterator.h" 16 #include "llvm/ADT/SCCIterator.h" 17 #include "llvm/ADT/SmallVector.h" 18 #include "llvm/ADT/Statistic.h" 19 #include "llvm/CodeGen/MachineDominators.h" 20 #include "llvm/CodeGen/MachineFunction.h" 21 #include "llvm/CodeGen/MachineFunctionAnalysis.h" 22 #include "llvm/CodeGen/MachineFunctionPass.h" 23 #include "llvm/CodeGen/MachineInstrBuilder.h" 24 #include "llvm/CodeGen/MachineJumpTableInfo.h" 25 #include "llvm/CodeGen/MachineLoopInfo.h" 26 #include "llvm/CodeGen/MachinePostDominators.h" 27 #include "llvm/CodeGen/MachineRegisterInfo.h" 28 #include "llvm/IR/Dominators.h" 29 #include "llvm/Support/Debug.h" 30 #include "llvm/Support/raw_ostream.h" 31 #include "llvm/Target/TargetInstrInfo.h" 32 #include "llvm/Target/TargetMachine.h" 33 #include <deque> 34 35 using namespace llvm; 36 37 #define DEBUG_TYPE "structcfg" 38 39 #define DEFAULT_VEC_SLOTS 8 40 41 // TODO: move-begin. 42 43 //===----------------------------------------------------------------------===// 44 // 45 // Statistics for CFGStructurizer. 46 // 47 //===----------------------------------------------------------------------===// 48 49 STATISTIC(numSerialPatternMatch, "CFGStructurizer number of serial pattern " 50 "matched"); 51 STATISTIC(numIfPatternMatch, "CFGStructurizer number of if pattern " 52 "matched"); 53 STATISTIC(numClonedBlock, "CFGStructurizer cloned blocks"); 54 STATISTIC(numClonedInstr, "CFGStructurizer cloned instructions"); 55 56 namespace llvm { 57 void initializeAMDGPUCFGStructurizerPass(PassRegistry&); 58 } 59 60 //===----------------------------------------------------------------------===// 61 // 62 // Miscellaneous utility for CFGStructurizer. 63 // 64 //===----------------------------------------------------------------------===// 65 namespace { 66 #define SHOWNEWINSTR(i) \ 67 DEBUG(dbgs() << "New instr: " << *i << "\n"); 68 69 #define SHOWNEWBLK(b, msg) \ 70 DEBUG( \ 71 dbgs() << msg << "BB" << b->getNumber() << "size " << b->size(); \ 72 dbgs() << "\n"; \ 73 ); 74 75 #define SHOWBLK_DETAIL(b, msg) \ 76 DEBUG( \ 77 if (b) { \ 78 dbgs() << msg << "BB" << b->getNumber() << "size " << b->size(); \ 79 b->print(dbgs()); \ 80 dbgs() << "\n"; \ 81 } \ 82 ); 83 84 #define INVALIDSCCNUM -1 85 86 template<class NodeT> 87 void ReverseVector(SmallVectorImpl<NodeT *> &Src) { 88 size_t sz = Src.size(); 89 for (size_t i = 0; i < sz/2; ++i) { 90 NodeT *t = Src[i]; 91 Src[i] = Src[sz - i - 1]; 92 Src[sz - i - 1] = t; 93 } 94 } 95 96 } // end anonymous namespace 97 98 //===----------------------------------------------------------------------===// 99 // 100 // supporting data structure for CFGStructurizer 101 // 102 //===----------------------------------------------------------------------===// 103 104 105 namespace { 106 107 class BlockInformation { 108 public: 109 bool IsRetired; 110 int SccNum; 111 BlockInformation() : IsRetired(false), SccNum(INVALIDSCCNUM) {} 112 }; 113 114 } // end anonymous namespace 115 116 //===----------------------------------------------------------------------===// 117 // 118 // CFGStructurizer 119 // 120 //===----------------------------------------------------------------------===// 121 122 namespace { 123 class AMDGPUCFGStructurizer : public MachineFunctionPass { 124 public: 125 typedef SmallVector<MachineBasicBlock *, 32> MBBVector; 126 typedef std::map<MachineBasicBlock *, BlockInformation *> MBBInfoMap; 127 typedef std::map<MachineLoop *, MachineBasicBlock *> LoopLandInfoMap; 128 129 enum PathToKind { 130 Not_SinglePath = 0, 131 SinglePath_InPath = 1, 132 SinglePath_NotInPath = 2 133 }; 134 135 static char ID; 136 137 AMDGPUCFGStructurizer() : 138 MachineFunctionPass(ID), TII(nullptr), TRI(nullptr) { 139 initializeAMDGPUCFGStructurizerPass(*PassRegistry::getPassRegistry()); 140 } 141 142 const char *getPassName() const override { 143 return "AMDGPU Control Flow Graph structurizer Pass"; 144 } 145 146 void getAnalysisUsage(AnalysisUsage &AU) const override { 147 AU.addPreserved<MachineFunctionAnalysis>(); 148 AU.addRequired<MachineFunctionAnalysis>(); 149 AU.addRequired<MachineDominatorTree>(); 150 AU.addRequired<MachinePostDominatorTree>(); 151 AU.addRequired<MachineLoopInfo>(); 152 } 153 154 /// Perform the CFG structurization 155 bool run(); 156 157 /// Perform the CFG preparation 158 /// This step will remove every unconditionnal/dead jump instructions and make 159 /// sure all loops have an exit block 160 bool prepare(); 161 162 bool runOnMachineFunction(MachineFunction &MF) override { 163 TII = MF.getSubtarget<R600Subtarget>().getInstrInfo(); 164 TRI = &TII->getRegisterInfo(); 165 DEBUG(MF.dump();); 166 OrderedBlks.clear(); 167 Visited.clear(); 168 FuncRep = &MF; 169 MLI = &getAnalysis<MachineLoopInfo>(); 170 DEBUG(dbgs() << "LoopInfo:\n"; PrintLoopinfo(*MLI);); 171 MDT = &getAnalysis<MachineDominatorTree>(); 172 DEBUG(MDT->print(dbgs(), (const llvm::Module*)nullptr);); 173 PDT = &getAnalysis<MachinePostDominatorTree>(); 174 DEBUG(PDT->print(dbgs());); 175 prepare(); 176 run(); 177 DEBUG(MF.dump();); 178 return true; 179 } 180 181 protected: 182 MachineDominatorTree *MDT; 183 MachinePostDominatorTree *PDT; 184 MachineLoopInfo *MLI; 185 const R600InstrInfo *TII; 186 const R600RegisterInfo *TRI; 187 188 // PRINT FUNCTIONS 189 /// Print the ordered Blocks. 190 void printOrderedBlocks() const { 191 size_t i = 0; 192 for (MBBVector::const_iterator iterBlk = OrderedBlks.begin(), 193 iterBlkEnd = OrderedBlks.end(); iterBlk != iterBlkEnd; ++iterBlk, ++i) { 194 dbgs() << "BB" << (*iterBlk)->getNumber(); 195 dbgs() << "(" << getSCCNum(*iterBlk) << "," << (*iterBlk)->size() << ")"; 196 if (i != 0 && i % 10 == 0) { 197 dbgs() << "\n"; 198 } else { 199 dbgs() << " "; 200 } 201 } 202 } 203 static void PrintLoopinfo(const MachineLoopInfo &LoopInfo) { 204 for (MachineLoop::iterator iter = LoopInfo.begin(), 205 iterEnd = LoopInfo.end(); iter != iterEnd; ++iter) { 206 (*iter)->print(dbgs(), 0); 207 } 208 } 209 210 // UTILITY FUNCTIONS 211 int getSCCNum(MachineBasicBlock *MBB) const; 212 MachineBasicBlock *getLoopLandInfo(MachineLoop *LoopRep) const; 213 bool hasBackEdge(MachineBasicBlock *MBB) const; 214 bool isRetiredBlock(MachineBasicBlock *MBB) const; 215 bool isActiveLoophead(MachineBasicBlock *MBB) const; 216 PathToKind singlePathTo(MachineBasicBlock *SrcMBB, MachineBasicBlock *DstMBB, 217 bool AllowSideEntry = true) const; 218 int countActiveBlock(MBBVector::const_iterator It, 219 MBBVector::const_iterator E) const; 220 bool needMigrateBlock(MachineBasicBlock *MBB) const; 221 222 // Utility Functions 223 void reversePredicateSetter(MachineBasicBlock::iterator I); 224 /// Compute the reversed DFS post order of Blocks 225 void orderBlocks(MachineFunction *MF); 226 227 // Function originally from CFGStructTraits 228 void insertInstrEnd(MachineBasicBlock *MBB, int NewOpcode, 229 const DebugLoc &DL = DebugLoc()); 230 MachineInstr *insertInstrBefore(MachineBasicBlock *MBB, int NewOpcode, 231 const DebugLoc &DL = DebugLoc()); 232 MachineInstr *insertInstrBefore(MachineBasicBlock::iterator I, int NewOpcode); 233 void insertCondBranchBefore(MachineBasicBlock::iterator I, int NewOpcode, 234 const DebugLoc &DL); 235 void insertCondBranchBefore(MachineBasicBlock *MBB, 236 MachineBasicBlock::iterator I, int NewOpcode, 237 int RegNum, const DebugLoc &DL); 238 static int getBranchNzeroOpcode(int OldOpcode); 239 static int getBranchZeroOpcode(int OldOpcode); 240 static int getContinueNzeroOpcode(int OldOpcode); 241 static int getContinueZeroOpcode(int OldOpcode); 242 static MachineBasicBlock *getTrueBranch(MachineInstr *MI); 243 static void setTrueBranch(MachineInstr *MI, MachineBasicBlock *MBB); 244 static MachineBasicBlock *getFalseBranch(MachineBasicBlock *MBB, 245 MachineInstr *MI); 246 static bool isCondBranch(MachineInstr *MI); 247 static bool isUncondBranch(MachineInstr *MI); 248 static DebugLoc getLastDebugLocInBB(MachineBasicBlock *MBB); 249 static MachineInstr *getNormalBlockBranchInstr(MachineBasicBlock *MBB); 250 /// The correct naming for this is getPossibleLoopendBlockBranchInstr. 251 /// 252 /// BB with backward-edge could have move instructions after the branch 253 /// instruction. Such move instruction "belong to" the loop backward-edge. 254 MachineInstr *getLoopendBlockBranchInstr(MachineBasicBlock *MBB); 255 static MachineInstr *getReturnInstr(MachineBasicBlock *MBB); 256 static bool isReturnBlock(MachineBasicBlock *MBB); 257 static void cloneSuccessorList(MachineBasicBlock *DstMBB, 258 MachineBasicBlock *SrcMBB) ; 259 static MachineBasicBlock *clone(MachineBasicBlock *MBB); 260 /// MachineBasicBlock::ReplaceUsesOfBlockWith doesn't serve the purpose 261 /// because the AMDGPU instruction is not recognized as terminator fix this 262 /// and retire this routine 263 void replaceInstrUseOfBlockWith(MachineBasicBlock *SrcMBB, 264 MachineBasicBlock *OldMBB, MachineBasicBlock *NewBlk); 265 static void wrapup(MachineBasicBlock *MBB); 266 267 268 int patternMatch(MachineBasicBlock *MBB); 269 int patternMatchGroup(MachineBasicBlock *MBB); 270 int serialPatternMatch(MachineBasicBlock *MBB); 271 int ifPatternMatch(MachineBasicBlock *MBB); 272 int loopendPatternMatch(); 273 int mergeLoop(MachineLoop *LoopRep); 274 275 /// return true iff src1Blk->succ_size() == 0 && src1Blk and src2Blk are in 276 /// the same loop with LoopLandInfo without explicitly keeping track of 277 /// loopContBlks and loopBreakBlks, this is a method to get the information. 278 bool isSameloopDetachedContbreak(MachineBasicBlock *Src1MBB, 279 MachineBasicBlock *Src2MBB); 280 int handleJumpintoIf(MachineBasicBlock *HeadMBB, 281 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB); 282 int handleJumpintoIfImp(MachineBasicBlock *HeadMBB, 283 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB); 284 int improveSimpleJumpintoIf(MachineBasicBlock *HeadMBB, 285 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB, 286 MachineBasicBlock **LandMBBPtr); 287 void showImproveSimpleJumpintoIf(MachineBasicBlock *HeadMBB, 288 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB, 289 MachineBasicBlock *LandMBB, bool Detail = false); 290 int cloneOnSideEntryTo(MachineBasicBlock *PreMBB, 291 MachineBasicBlock *SrcMBB, MachineBasicBlock *DstMBB); 292 void mergeSerialBlock(MachineBasicBlock *DstMBB, 293 MachineBasicBlock *SrcMBB); 294 295 void mergeIfthenelseBlock(MachineInstr *BranchMI, 296 MachineBasicBlock *MBB, MachineBasicBlock *TrueMBB, 297 MachineBasicBlock *FalseMBB, MachineBasicBlock *LandMBB); 298 void mergeLooplandBlock(MachineBasicBlock *DstMBB, 299 MachineBasicBlock *LandMBB); 300 void mergeLoopbreakBlock(MachineBasicBlock *ExitingMBB, 301 MachineBasicBlock *LandMBB); 302 void settleLoopcontBlock(MachineBasicBlock *ContingMBB, 303 MachineBasicBlock *ContMBB); 304 /// normalizeInfiniteLoopExit change 305 /// B1: 306 /// uncond_br LoopHeader 307 /// 308 /// to 309 /// B1: 310 /// cond_br 1 LoopHeader dummyExit 311 /// and return the newly added dummy exit block 312 MachineBasicBlock *normalizeInfiniteLoopExit(MachineLoop *LoopRep); 313 void removeUnconditionalBranch(MachineBasicBlock *MBB); 314 /// Remove duplicate branches instructions in a block. 315 /// For instance 316 /// B0: 317 /// cond_br X B1 B2 318 /// cond_br X B1 B2 319 /// is transformed to 320 /// B0: 321 /// cond_br X B1 B2 322 void removeRedundantConditionalBranch(MachineBasicBlock *MBB); 323 void addDummyExitBlock(SmallVectorImpl<MachineBasicBlock *> &RetMBB); 324 void removeSuccessor(MachineBasicBlock *MBB); 325 MachineBasicBlock *cloneBlockForPredecessor(MachineBasicBlock *MBB, 326 MachineBasicBlock *PredMBB); 327 void migrateInstruction(MachineBasicBlock *SrcMBB, 328 MachineBasicBlock *DstMBB, MachineBasicBlock::iterator I); 329 void recordSccnum(MachineBasicBlock *MBB, int SCCNum); 330 void retireBlock(MachineBasicBlock *MBB); 331 332 333 private: 334 MBBInfoMap BlockInfoMap; 335 LoopLandInfoMap LLInfoMap; 336 std::map<MachineLoop *, bool> Visited; 337 MachineFunction *FuncRep; 338 SmallVector<MachineBasicBlock *, DEFAULT_VEC_SLOTS> OrderedBlks; 339 }; 340 341 int AMDGPUCFGStructurizer::getSCCNum(MachineBasicBlock *MBB) const { 342 MBBInfoMap::const_iterator It = BlockInfoMap.find(MBB); 343 if (It == BlockInfoMap.end()) 344 return INVALIDSCCNUM; 345 return (*It).second->SccNum; 346 } 347 348 MachineBasicBlock *AMDGPUCFGStructurizer::getLoopLandInfo(MachineLoop *LoopRep) 349 const { 350 LoopLandInfoMap::const_iterator It = LLInfoMap.find(LoopRep); 351 if (It == LLInfoMap.end()) 352 return nullptr; 353 return (*It).second; 354 } 355 356 bool AMDGPUCFGStructurizer::hasBackEdge(MachineBasicBlock *MBB) const { 357 MachineLoop *LoopRep = MLI->getLoopFor(MBB); 358 if (!LoopRep) 359 return false; 360 MachineBasicBlock *LoopHeader = LoopRep->getHeader(); 361 return MBB->isSuccessor(LoopHeader); 362 } 363 364 bool AMDGPUCFGStructurizer::isRetiredBlock(MachineBasicBlock *MBB) const { 365 MBBInfoMap::const_iterator It = BlockInfoMap.find(MBB); 366 if (It == BlockInfoMap.end()) 367 return false; 368 return (*It).second->IsRetired; 369 } 370 371 bool AMDGPUCFGStructurizer::isActiveLoophead(MachineBasicBlock *MBB) const { 372 MachineLoop *LoopRep = MLI->getLoopFor(MBB); 373 while (LoopRep && LoopRep->getHeader() == MBB) { 374 MachineBasicBlock *LoopLand = getLoopLandInfo(LoopRep); 375 if(!LoopLand) 376 return true; 377 if (!isRetiredBlock(LoopLand)) 378 return true; 379 LoopRep = LoopRep->getParentLoop(); 380 } 381 return false; 382 } 383 AMDGPUCFGStructurizer::PathToKind AMDGPUCFGStructurizer::singlePathTo( 384 MachineBasicBlock *SrcMBB, MachineBasicBlock *DstMBB, 385 bool AllowSideEntry) const { 386 assert(DstMBB); 387 if (SrcMBB == DstMBB) 388 return SinglePath_InPath; 389 while (SrcMBB && SrcMBB->succ_size() == 1) { 390 SrcMBB = *SrcMBB->succ_begin(); 391 if (SrcMBB == DstMBB) 392 return SinglePath_InPath; 393 if (!AllowSideEntry && SrcMBB->pred_size() > 1) 394 return Not_SinglePath; 395 } 396 if (SrcMBB && SrcMBB->succ_size()==0) 397 return SinglePath_NotInPath; 398 return Not_SinglePath; 399 } 400 401 int AMDGPUCFGStructurizer::countActiveBlock(MBBVector::const_iterator It, 402 MBBVector::const_iterator E) const { 403 int Count = 0; 404 while (It != E) { 405 if (!isRetiredBlock(*It)) 406 ++Count; 407 ++It; 408 } 409 return Count; 410 } 411 412 bool AMDGPUCFGStructurizer::needMigrateBlock(MachineBasicBlock *MBB) const { 413 unsigned BlockSizeThreshold = 30; 414 unsigned CloneInstrThreshold = 100; 415 bool MultiplePreds = MBB && (MBB->pred_size() > 1); 416 417 if(!MultiplePreds) 418 return false; 419 unsigned BlkSize = MBB->size(); 420 return ((BlkSize > BlockSizeThreshold) && 421 (BlkSize * (MBB->pred_size() - 1) > CloneInstrThreshold)); 422 } 423 424 void AMDGPUCFGStructurizer::reversePredicateSetter( 425 MachineBasicBlock::iterator I) { 426 assert(static_cast<MachineInstr *>(I) && "Expected valid iterator"); 427 for (;; --I) { 428 if (I->getOpcode() == AMDGPU::PRED_X) { 429 switch (static_cast<MachineInstr *>(I)->getOperand(2).getImm()) { 430 case OPCODE_IS_ZERO_INT: 431 static_cast<MachineInstr *>(I)->getOperand(2) 432 .setImm(OPCODE_IS_NOT_ZERO_INT); 433 return; 434 case OPCODE_IS_NOT_ZERO_INT: 435 static_cast<MachineInstr *>(I)->getOperand(2) 436 .setImm(OPCODE_IS_ZERO_INT); 437 return; 438 case OPCODE_IS_ZERO: 439 static_cast<MachineInstr *>(I)->getOperand(2) 440 .setImm(OPCODE_IS_NOT_ZERO); 441 return; 442 case OPCODE_IS_NOT_ZERO: 443 static_cast<MachineInstr *>(I)->getOperand(2) 444 .setImm(OPCODE_IS_ZERO); 445 return; 446 default: 447 llvm_unreachable("PRED_X Opcode invalid!"); 448 } 449 } 450 } 451 } 452 453 void AMDGPUCFGStructurizer::insertInstrEnd(MachineBasicBlock *MBB, 454 int NewOpcode, const DebugLoc &DL) { 455 MachineInstr *MI = 456 MBB->getParent()->CreateMachineInstr(TII->get(NewOpcode), DL); 457 MBB->push_back(MI); 458 //assume the instruction doesn't take any reg operand ... 459 SHOWNEWINSTR(MI); 460 } 461 462 MachineInstr *AMDGPUCFGStructurizer::insertInstrBefore(MachineBasicBlock *MBB, 463 int NewOpcode, 464 const DebugLoc &DL) { 465 MachineInstr *MI = 466 MBB->getParent()->CreateMachineInstr(TII->get(NewOpcode), DL); 467 if (MBB->begin() != MBB->end()) 468 MBB->insert(MBB->begin(), MI); 469 else 470 MBB->push_back(MI); 471 SHOWNEWINSTR(MI); 472 return MI; 473 } 474 475 MachineInstr *AMDGPUCFGStructurizer::insertInstrBefore( 476 MachineBasicBlock::iterator I, int NewOpcode) { 477 MachineInstr *OldMI = &(*I); 478 MachineBasicBlock *MBB = OldMI->getParent(); 479 MachineInstr *NewMBB = 480 MBB->getParent()->CreateMachineInstr(TII->get(NewOpcode), DebugLoc()); 481 MBB->insert(I, NewMBB); 482 //assume the instruction doesn't take any reg operand ... 483 SHOWNEWINSTR(NewMBB); 484 return NewMBB; 485 } 486 487 void AMDGPUCFGStructurizer::insertCondBranchBefore( 488 MachineBasicBlock::iterator I, int NewOpcode, const DebugLoc &DL) { 489 MachineInstr *OldMI = &(*I); 490 MachineBasicBlock *MBB = OldMI->getParent(); 491 MachineFunction *MF = MBB->getParent(); 492 MachineInstr *NewMI = MF->CreateMachineInstr(TII->get(NewOpcode), DL); 493 MBB->insert(I, NewMI); 494 MachineInstrBuilder MIB(*MF, NewMI); 495 MIB.addReg(OldMI->getOperand(1).getReg(), false); 496 SHOWNEWINSTR(NewMI); 497 //erase later oldInstr->eraseFromParent(); 498 } 499 500 void AMDGPUCFGStructurizer::insertCondBranchBefore( 501 MachineBasicBlock *blk, MachineBasicBlock::iterator I, int NewOpcode, 502 int RegNum, const DebugLoc &DL) { 503 MachineFunction *MF = blk->getParent(); 504 MachineInstr *NewInstr = MF->CreateMachineInstr(TII->get(NewOpcode), DL); 505 //insert before 506 blk->insert(I, NewInstr); 507 MachineInstrBuilder(*MF, NewInstr).addReg(RegNum, false); 508 SHOWNEWINSTR(NewInstr); 509 } 510 511 int AMDGPUCFGStructurizer::getBranchNzeroOpcode(int OldOpcode) { 512 switch(OldOpcode) { 513 case AMDGPU::JUMP_COND: 514 case AMDGPU::JUMP: return AMDGPU::IF_PREDICATE_SET; 515 case AMDGPU::BRANCH_COND_i32: 516 case AMDGPU::BRANCH_COND_f32: return AMDGPU::IF_LOGICALNZ_f32; 517 default: llvm_unreachable("internal error"); 518 } 519 return -1; 520 } 521 522 int AMDGPUCFGStructurizer::getBranchZeroOpcode(int OldOpcode) { 523 switch(OldOpcode) { 524 case AMDGPU::JUMP_COND: 525 case AMDGPU::JUMP: return AMDGPU::IF_PREDICATE_SET; 526 case AMDGPU::BRANCH_COND_i32: 527 case AMDGPU::BRANCH_COND_f32: return AMDGPU::IF_LOGICALZ_f32; 528 default: llvm_unreachable("internal error"); 529 } 530 return -1; 531 } 532 533 int AMDGPUCFGStructurizer::getContinueNzeroOpcode(int OldOpcode) { 534 switch(OldOpcode) { 535 case AMDGPU::JUMP_COND: 536 case AMDGPU::JUMP: return AMDGPU::CONTINUE_LOGICALNZ_i32; 537 default: llvm_unreachable("internal error"); 538 }; 539 return -1; 540 } 541 542 int AMDGPUCFGStructurizer::getContinueZeroOpcode(int OldOpcode) { 543 switch(OldOpcode) { 544 case AMDGPU::JUMP_COND: 545 case AMDGPU::JUMP: return AMDGPU::CONTINUE_LOGICALZ_i32; 546 default: llvm_unreachable("internal error"); 547 } 548 return -1; 549 } 550 551 MachineBasicBlock *AMDGPUCFGStructurizer::getTrueBranch(MachineInstr *MI) { 552 return MI->getOperand(0).getMBB(); 553 } 554 555 void AMDGPUCFGStructurizer::setTrueBranch(MachineInstr *MI, 556 MachineBasicBlock *MBB) { 557 MI->getOperand(0).setMBB(MBB); 558 } 559 560 MachineBasicBlock * 561 AMDGPUCFGStructurizer::getFalseBranch(MachineBasicBlock *MBB, 562 MachineInstr *MI) { 563 assert(MBB->succ_size() == 2); 564 MachineBasicBlock *TrueBranch = getTrueBranch(MI); 565 MachineBasicBlock::succ_iterator It = MBB->succ_begin(); 566 MachineBasicBlock::succ_iterator Next = It; 567 ++Next; 568 return (*It == TrueBranch) ? *Next : *It; 569 } 570 571 bool AMDGPUCFGStructurizer::isCondBranch(MachineInstr *MI) { 572 switch (MI->getOpcode()) { 573 case AMDGPU::JUMP_COND: 574 case AMDGPU::BRANCH_COND_i32: 575 case AMDGPU::BRANCH_COND_f32: return true; 576 default: 577 return false; 578 } 579 return false; 580 } 581 582 bool AMDGPUCFGStructurizer::isUncondBranch(MachineInstr *MI) { 583 switch (MI->getOpcode()) { 584 case AMDGPU::JUMP: 585 case AMDGPU::BRANCH: 586 return true; 587 default: 588 return false; 589 } 590 return false; 591 } 592 593 DebugLoc AMDGPUCFGStructurizer::getLastDebugLocInBB(MachineBasicBlock *MBB) { 594 //get DebugLoc from the first MachineBasicBlock instruction with debug info 595 DebugLoc DL; 596 for (MachineBasicBlock::iterator It = MBB->begin(); It != MBB->end(); 597 ++It) { 598 MachineInstr *instr = &(*It); 599 if (instr->getDebugLoc()) 600 DL = instr->getDebugLoc(); 601 } 602 return DL; 603 } 604 605 MachineInstr *AMDGPUCFGStructurizer::getNormalBlockBranchInstr( 606 MachineBasicBlock *MBB) { 607 MachineBasicBlock::reverse_iterator It = MBB->rbegin(); 608 MachineInstr *MI = &*It; 609 if (MI && (isCondBranch(MI) || isUncondBranch(MI))) 610 return MI; 611 return nullptr; 612 } 613 614 MachineInstr *AMDGPUCFGStructurizer::getLoopendBlockBranchInstr( 615 MachineBasicBlock *MBB) { 616 for (MachineBasicBlock::reverse_iterator It = MBB->rbegin(), E = MBB->rend(); 617 It != E; ++It) { 618 // FIXME: Simplify 619 MachineInstr *MI = &*It; 620 if (MI) { 621 if (isCondBranch(MI) || isUncondBranch(MI)) 622 return MI; 623 else if (!TII->isMov(MI->getOpcode())) 624 break; 625 } 626 } 627 return nullptr; 628 } 629 630 MachineInstr *AMDGPUCFGStructurizer::getReturnInstr(MachineBasicBlock *MBB) { 631 MachineBasicBlock::reverse_iterator It = MBB->rbegin(); 632 if (It != MBB->rend()) { 633 MachineInstr *instr = &(*It); 634 if (instr->getOpcode() == AMDGPU::RETURN) 635 return instr; 636 } 637 return nullptr; 638 } 639 640 bool AMDGPUCFGStructurizer::isReturnBlock(MachineBasicBlock *MBB) { 641 MachineInstr *MI = getReturnInstr(MBB); 642 bool IsReturn = (MBB->succ_size() == 0); 643 if (MI) 644 assert(IsReturn); 645 else if (IsReturn) 646 DEBUG( 647 dbgs() << "BB" << MBB->getNumber() 648 <<" is return block without RETURN instr\n";); 649 return IsReturn; 650 } 651 652 void AMDGPUCFGStructurizer::cloneSuccessorList(MachineBasicBlock *DstMBB, 653 MachineBasicBlock *SrcMBB) { 654 for (MachineBasicBlock::succ_iterator It = SrcMBB->succ_begin(), 655 iterEnd = SrcMBB->succ_end(); It != iterEnd; ++It) 656 DstMBB->addSuccessor(*It); // *iter's predecessor is also taken care of 657 } 658 659 MachineBasicBlock *AMDGPUCFGStructurizer::clone(MachineBasicBlock *MBB) { 660 MachineFunction *Func = MBB->getParent(); 661 MachineBasicBlock *NewMBB = Func->CreateMachineBasicBlock(); 662 Func->push_back(NewMBB); //insert to function 663 for (const MachineInstr &It : *MBB) 664 NewMBB->push_back(Func->CloneMachineInstr(&It)); 665 return NewMBB; 666 } 667 668 void AMDGPUCFGStructurizer::replaceInstrUseOfBlockWith( 669 MachineBasicBlock *SrcMBB, MachineBasicBlock *OldMBB, 670 MachineBasicBlock *NewBlk) { 671 MachineInstr *BranchMI = getLoopendBlockBranchInstr(SrcMBB); 672 if (BranchMI && isCondBranch(BranchMI) && 673 getTrueBranch(BranchMI) == OldMBB) 674 setTrueBranch(BranchMI, NewBlk); 675 } 676 677 void AMDGPUCFGStructurizer::wrapup(MachineBasicBlock *MBB) { 678 assert((!MBB->getParent()->getJumpTableInfo() 679 || MBB->getParent()->getJumpTableInfo()->isEmpty()) 680 && "found a jump table"); 681 682 //collect continue right before endloop 683 SmallVector<MachineInstr *, DEFAULT_VEC_SLOTS> ContInstr; 684 MachineBasicBlock::iterator Pre = MBB->begin(); 685 MachineBasicBlock::iterator E = MBB->end(); 686 MachineBasicBlock::iterator It = Pre; 687 while (It != E) { 688 if (Pre->getOpcode() == AMDGPU::CONTINUE 689 && It->getOpcode() == AMDGPU::ENDLOOP) 690 ContInstr.push_back(&*Pre); 691 Pre = It; 692 ++It; 693 } 694 695 //delete continue right before endloop 696 for (unsigned i = 0; i < ContInstr.size(); ++i) 697 ContInstr[i]->eraseFromParent(); 698 699 // TODO to fix up jump table so later phase won't be confused. if 700 // (jumpTableInfo->isEmpty() == false) { need to clean the jump table, but 701 // there isn't such an interface yet. alternatively, replace all the other 702 // blocks in the jump table with the entryBlk //} 703 704 } 705 706 707 bool AMDGPUCFGStructurizer::prepare() { 708 bool Changed = false; 709 710 //FIXME: if not reducible flow graph, make it so ??? 711 712 DEBUG(dbgs() << "AMDGPUCFGStructurizer::prepare\n";); 713 714 orderBlocks(FuncRep); 715 716 SmallVector<MachineBasicBlock *, DEFAULT_VEC_SLOTS> RetBlks; 717 718 // Add an ExitBlk to loop that don't have one 719 for (MachineLoopInfo::iterator It = MLI->begin(), 720 E = MLI->end(); It != E; ++It) { 721 MachineLoop *LoopRep = (*It); 722 MBBVector ExitingMBBs; 723 LoopRep->getExitingBlocks(ExitingMBBs); 724 725 if (ExitingMBBs.size() == 0) { 726 MachineBasicBlock* DummyExitBlk = normalizeInfiniteLoopExit(LoopRep); 727 if (DummyExitBlk) 728 RetBlks.push_back(DummyExitBlk); 729 } 730 } 731 732 // Remove unconditional branch instr. 733 // Add dummy exit block iff there are multiple returns. 734 for (SmallVectorImpl<MachineBasicBlock *>::const_iterator 735 It = OrderedBlks.begin(), E = OrderedBlks.end(); It != E; ++It) { 736 MachineBasicBlock *MBB = *It; 737 removeUnconditionalBranch(MBB); 738 removeRedundantConditionalBranch(MBB); 739 if (isReturnBlock(MBB)) { 740 RetBlks.push_back(MBB); 741 } 742 assert(MBB->succ_size() <= 2); 743 } 744 745 if (RetBlks.size() >= 2) { 746 addDummyExitBlock(RetBlks); 747 Changed = true; 748 } 749 750 return Changed; 751 } 752 753 bool AMDGPUCFGStructurizer::run() { 754 755 //Assume reducible CFG... 756 DEBUG(dbgs() << "AMDGPUCFGStructurizer::run\n"); 757 758 #ifdef STRESSTEST 759 //Use the worse block ordering to test the algorithm. 760 ReverseVector(orderedBlks); 761 #endif 762 763 DEBUG(dbgs() << "Ordered blocks:\n"; printOrderedBlocks();); 764 int NumIter = 0; 765 bool Finish = false; 766 MachineBasicBlock *MBB; 767 bool MakeProgress = false; 768 int NumRemainedBlk = countActiveBlock(OrderedBlks.begin(), 769 OrderedBlks.end()); 770 771 do { 772 ++NumIter; 773 DEBUG( 774 dbgs() << "numIter = " << NumIter 775 << ", numRemaintedBlk = " << NumRemainedBlk << "\n"; 776 ); 777 778 SmallVectorImpl<MachineBasicBlock *>::const_iterator It = 779 OrderedBlks.begin(); 780 SmallVectorImpl<MachineBasicBlock *>::const_iterator E = 781 OrderedBlks.end(); 782 783 SmallVectorImpl<MachineBasicBlock *>::const_iterator SccBeginIter = 784 It; 785 MachineBasicBlock *SccBeginMBB = nullptr; 786 int SccNumBlk = 0; // The number of active blocks, init to a 787 // maximum possible number. 788 int SccNumIter; // Number of iteration in this SCC. 789 790 while (It != E) { 791 MBB = *It; 792 793 if (!SccBeginMBB) { 794 SccBeginIter = It; 795 SccBeginMBB = MBB; 796 SccNumIter = 0; 797 SccNumBlk = NumRemainedBlk; // Init to maximum possible number. 798 DEBUG( 799 dbgs() << "start processing SCC" << getSCCNum(SccBeginMBB); 800 dbgs() << "\n"; 801 ); 802 } 803 804 if (!isRetiredBlock(MBB)) 805 patternMatch(MBB); 806 807 ++It; 808 809 bool ContNextScc = true; 810 if (It == E 811 || getSCCNum(SccBeginMBB) != getSCCNum(*It)) { 812 // Just finish one scc. 813 ++SccNumIter; 814 int sccRemainedNumBlk = countActiveBlock(SccBeginIter, It); 815 if (sccRemainedNumBlk != 1 && sccRemainedNumBlk >= SccNumBlk) { 816 DEBUG( 817 dbgs() << "Can't reduce SCC " << getSCCNum(MBB) 818 << ", sccNumIter = " << SccNumIter; 819 dbgs() << "doesn't make any progress\n"; 820 ); 821 ContNextScc = true; 822 } else if (sccRemainedNumBlk != 1 && sccRemainedNumBlk < SccNumBlk) { 823 SccNumBlk = sccRemainedNumBlk; 824 It = SccBeginIter; 825 ContNextScc = false; 826 DEBUG( 827 dbgs() << "repeat processing SCC" << getSCCNum(MBB) 828 << "sccNumIter = " << SccNumIter << '\n'; 829 ); 830 } else { 831 // Finish the current scc. 832 ContNextScc = true; 833 } 834 } else { 835 // Continue on next component in the current scc. 836 ContNextScc = false; 837 } 838 839 if (ContNextScc) 840 SccBeginMBB = nullptr; 841 } //while, "one iteration" over the function. 842 843 MachineBasicBlock *EntryMBB = 844 &*GraphTraits<MachineFunction *>::nodes_begin(FuncRep); 845 if (EntryMBB->succ_size() == 0) { 846 Finish = true; 847 DEBUG( 848 dbgs() << "Reduce to one block\n"; 849 ); 850 } else { 851 int NewnumRemainedBlk 852 = countActiveBlock(OrderedBlks.begin(), OrderedBlks.end()); 853 // consider cloned blocks ?? 854 if (NewnumRemainedBlk == 1 || NewnumRemainedBlk < NumRemainedBlk) { 855 MakeProgress = true; 856 NumRemainedBlk = NewnumRemainedBlk; 857 } else { 858 MakeProgress = false; 859 DEBUG( 860 dbgs() << "No progress\n"; 861 ); 862 } 863 } 864 } while (!Finish && MakeProgress); 865 866 // Misc wrap up to maintain the consistency of the Function representation. 867 wrapup(&*GraphTraits<MachineFunction *>::nodes_begin(FuncRep)); 868 869 // Detach retired Block, release memory. 870 for (MBBInfoMap::iterator It = BlockInfoMap.begin(), E = BlockInfoMap.end(); 871 It != E; ++It) { 872 if ((*It).second && (*It).second->IsRetired) { 873 assert(((*It).first)->getNumber() != -1); 874 DEBUG( 875 dbgs() << "Erase BB" << ((*It).first)->getNumber() << "\n"; 876 ); 877 (*It).first->eraseFromParent(); //Remove from the parent Function. 878 } 879 delete (*It).second; 880 } 881 BlockInfoMap.clear(); 882 LLInfoMap.clear(); 883 884 if (!Finish) { 885 DEBUG(FuncRep->viewCFG()); 886 report_fatal_error("IRREDUCIBLE_CFG"); 887 } 888 889 return true; 890 } 891 892 893 894 void AMDGPUCFGStructurizer::orderBlocks(MachineFunction *MF) { 895 int SccNum = 0; 896 MachineBasicBlock *MBB; 897 for (scc_iterator<MachineFunction *> It = scc_begin(MF); !It.isAtEnd(); 898 ++It, ++SccNum) { 899 const std::vector<MachineBasicBlock *> &SccNext = *It; 900 for (std::vector<MachineBasicBlock *>::const_iterator 901 blockIter = SccNext.begin(), blockEnd = SccNext.end(); 902 blockIter != blockEnd; ++blockIter) { 903 MBB = *blockIter; 904 OrderedBlks.push_back(MBB); 905 recordSccnum(MBB, SccNum); 906 } 907 } 908 909 //walk through all the block in func to check for unreachable 910 typedef GraphTraits<MachineFunction *> GTM; 911 MachineFunction::iterator It = GTM::nodes_begin(MF), E = GTM::nodes_end(MF); 912 for (; It != E; ++It) { 913 MachineBasicBlock *MBB = &(*It); 914 SccNum = getSCCNum(MBB); 915 if (SccNum == INVALIDSCCNUM) 916 dbgs() << "unreachable block BB" << MBB->getNumber() << "\n"; 917 } 918 } 919 920 int AMDGPUCFGStructurizer::patternMatch(MachineBasicBlock *MBB) { 921 int NumMatch = 0; 922 int CurMatch; 923 924 DEBUG( 925 dbgs() << "Begin patternMatch BB" << MBB->getNumber() << "\n"; 926 ); 927 928 while ((CurMatch = patternMatchGroup(MBB)) > 0) 929 NumMatch += CurMatch; 930 931 DEBUG( 932 dbgs() << "End patternMatch BB" << MBB->getNumber() 933 << ", numMatch = " << NumMatch << "\n"; 934 ); 935 936 return NumMatch; 937 } 938 939 int AMDGPUCFGStructurizer::patternMatchGroup(MachineBasicBlock *MBB) { 940 int NumMatch = 0; 941 NumMatch += loopendPatternMatch(); 942 NumMatch += serialPatternMatch(MBB); 943 NumMatch += ifPatternMatch(MBB); 944 return NumMatch; 945 } 946 947 948 int AMDGPUCFGStructurizer::serialPatternMatch(MachineBasicBlock *MBB) { 949 if (MBB->succ_size() != 1) 950 return 0; 951 952 MachineBasicBlock *childBlk = *MBB->succ_begin(); 953 if (childBlk->pred_size() != 1 || isActiveLoophead(childBlk)) 954 return 0; 955 956 mergeSerialBlock(MBB, childBlk); 957 ++numSerialPatternMatch; 958 return 1; 959 } 960 961 int AMDGPUCFGStructurizer::ifPatternMatch(MachineBasicBlock *MBB) { 962 //two edges 963 if (MBB->succ_size() != 2) 964 return 0; 965 if (hasBackEdge(MBB)) 966 return 0; 967 MachineInstr *BranchMI = getNormalBlockBranchInstr(MBB); 968 if (!BranchMI) 969 return 0; 970 971 assert(isCondBranch(BranchMI)); 972 int NumMatch = 0; 973 974 MachineBasicBlock *TrueMBB = getTrueBranch(BranchMI); 975 NumMatch += serialPatternMatch(TrueMBB); 976 NumMatch += ifPatternMatch(TrueMBB); 977 MachineBasicBlock *FalseMBB = getFalseBranch(MBB, BranchMI); 978 NumMatch += serialPatternMatch(FalseMBB); 979 NumMatch += ifPatternMatch(FalseMBB); 980 MachineBasicBlock *LandBlk; 981 int Cloned = 0; 982 983 assert (!TrueMBB->succ_empty() || !FalseMBB->succ_empty()); 984 // TODO: Simplify 985 if (TrueMBB->succ_size() == 1 && FalseMBB->succ_size() == 1 986 && *TrueMBB->succ_begin() == *FalseMBB->succ_begin()) { 987 // Diamond pattern 988 LandBlk = *TrueMBB->succ_begin(); 989 } else if (TrueMBB->succ_size() == 1 && *TrueMBB->succ_begin() == FalseMBB) { 990 // Triangle pattern, false is empty 991 LandBlk = FalseMBB; 992 FalseMBB = nullptr; 993 } else if (FalseMBB->succ_size() == 1 994 && *FalseMBB->succ_begin() == TrueMBB) { 995 // Triangle pattern, true is empty 996 // We reverse the predicate to make a triangle, empty false pattern; 997 std::swap(TrueMBB, FalseMBB); 998 reversePredicateSetter(MBB->end()); 999 LandBlk = FalseMBB; 1000 FalseMBB = nullptr; 1001 } else if (FalseMBB->succ_size() == 1 1002 && isSameloopDetachedContbreak(TrueMBB, FalseMBB)) { 1003 LandBlk = *FalseMBB->succ_begin(); 1004 } else if (TrueMBB->succ_size() == 1 1005 && isSameloopDetachedContbreak(FalseMBB, TrueMBB)) { 1006 LandBlk = *TrueMBB->succ_begin(); 1007 } else { 1008 return NumMatch + handleJumpintoIf(MBB, TrueMBB, FalseMBB); 1009 } 1010 1011 // improveSimpleJumpinfoIf can handle the case where landBlk == NULL but the 1012 // new BB created for landBlk==NULL may introduce new challenge to the 1013 // reduction process. 1014 if (LandBlk && 1015 ((TrueMBB && TrueMBB->pred_size() > 1) 1016 || (FalseMBB && FalseMBB->pred_size() > 1))) { 1017 Cloned += improveSimpleJumpintoIf(MBB, TrueMBB, FalseMBB, &LandBlk); 1018 } 1019 1020 if (TrueMBB && TrueMBB->pred_size() > 1) { 1021 TrueMBB = cloneBlockForPredecessor(TrueMBB, MBB); 1022 ++Cloned; 1023 } 1024 1025 if (FalseMBB && FalseMBB->pred_size() > 1) { 1026 FalseMBB = cloneBlockForPredecessor(FalseMBB, MBB); 1027 ++Cloned; 1028 } 1029 1030 mergeIfthenelseBlock(BranchMI, MBB, TrueMBB, FalseMBB, LandBlk); 1031 1032 ++numIfPatternMatch; 1033 1034 numClonedBlock += Cloned; 1035 1036 return 1 + Cloned + NumMatch; 1037 } 1038 1039 int AMDGPUCFGStructurizer::loopendPatternMatch() { 1040 std::deque<MachineLoop *> NestedLoops; 1041 for (auto &It: *MLI) 1042 for (MachineLoop *ML : depth_first(It)) 1043 NestedLoops.push_front(ML); 1044 1045 if (NestedLoops.size() == 0) 1046 return 0; 1047 1048 // Process nested loop outside->inside (we did push_front), 1049 // so "continue" to a outside loop won't be mistaken as "break" 1050 // of the current loop. 1051 int Num = 0; 1052 for (MachineLoop *ExaminedLoop : NestedLoops) { 1053 if (ExaminedLoop->getNumBlocks() == 0 || Visited[ExaminedLoop]) 1054 continue; 1055 DEBUG(dbgs() << "Processing:\n"; ExaminedLoop->dump();); 1056 int NumBreak = mergeLoop(ExaminedLoop); 1057 if (NumBreak == -1) 1058 break; 1059 Num += NumBreak; 1060 } 1061 return Num; 1062 } 1063 1064 int AMDGPUCFGStructurizer::mergeLoop(MachineLoop *LoopRep) { 1065 MachineBasicBlock *LoopHeader = LoopRep->getHeader(); 1066 MBBVector ExitingMBBs; 1067 LoopRep->getExitingBlocks(ExitingMBBs); 1068 assert(!ExitingMBBs.empty() && "Infinite Loop not supported"); 1069 DEBUG(dbgs() << "Loop has " << ExitingMBBs.size() << " exiting blocks\n";); 1070 // We assume a single ExitBlk 1071 MBBVector ExitBlks; 1072 LoopRep->getExitBlocks(ExitBlks); 1073 SmallPtrSet<MachineBasicBlock *, 2> ExitBlkSet; 1074 for (unsigned i = 0, e = ExitBlks.size(); i < e; ++i) 1075 ExitBlkSet.insert(ExitBlks[i]); 1076 assert(ExitBlkSet.size() == 1); 1077 MachineBasicBlock *ExitBlk = *ExitBlks.begin(); 1078 assert(ExitBlk && "Loop has several exit block"); 1079 MBBVector LatchBlks; 1080 typedef GraphTraits<Inverse<MachineBasicBlock*> > InvMBBTraits; 1081 InvMBBTraits::ChildIteratorType PI = InvMBBTraits::child_begin(LoopHeader), 1082 PE = InvMBBTraits::child_end(LoopHeader); 1083 for (; PI != PE; PI++) { 1084 if (LoopRep->contains(*PI)) 1085 LatchBlks.push_back(*PI); 1086 } 1087 1088 for (unsigned i = 0, e = ExitingMBBs.size(); i < e; ++i) 1089 mergeLoopbreakBlock(ExitingMBBs[i], ExitBlk); 1090 for (unsigned i = 0, e = LatchBlks.size(); i < e; ++i) 1091 settleLoopcontBlock(LatchBlks[i], LoopHeader); 1092 int Match = 0; 1093 do { 1094 Match = 0; 1095 Match += serialPatternMatch(LoopHeader); 1096 Match += ifPatternMatch(LoopHeader); 1097 } while (Match > 0); 1098 mergeLooplandBlock(LoopHeader, ExitBlk); 1099 MachineLoop *ParentLoop = LoopRep->getParentLoop(); 1100 if (ParentLoop) 1101 MLI->changeLoopFor(LoopHeader, ParentLoop); 1102 else 1103 MLI->removeBlock(LoopHeader); 1104 Visited[LoopRep] = true; 1105 return 1; 1106 } 1107 1108 bool AMDGPUCFGStructurizer::isSameloopDetachedContbreak( 1109 MachineBasicBlock *Src1MBB, MachineBasicBlock *Src2MBB) { 1110 if (Src1MBB->succ_size() == 0) { 1111 MachineLoop *LoopRep = MLI->getLoopFor(Src1MBB); 1112 if (LoopRep&& LoopRep == MLI->getLoopFor(Src2MBB)) { 1113 MachineBasicBlock *&TheEntry = LLInfoMap[LoopRep]; 1114 if (TheEntry) { 1115 DEBUG( 1116 dbgs() << "isLoopContBreakBlock yes src1 = BB" 1117 << Src1MBB->getNumber() 1118 << " src2 = BB" << Src2MBB->getNumber() << "\n"; 1119 ); 1120 return true; 1121 } 1122 } 1123 } 1124 return false; 1125 } 1126 1127 int AMDGPUCFGStructurizer::handleJumpintoIf(MachineBasicBlock *HeadMBB, 1128 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB) { 1129 int Num = handleJumpintoIfImp(HeadMBB, TrueMBB, FalseMBB); 1130 if (Num == 0) { 1131 DEBUG( 1132 dbgs() << "handleJumpintoIf swap trueBlk and FalseBlk" << "\n"; 1133 ); 1134 Num = handleJumpintoIfImp(HeadMBB, FalseMBB, TrueMBB); 1135 } 1136 return Num; 1137 } 1138 1139 int AMDGPUCFGStructurizer::handleJumpintoIfImp(MachineBasicBlock *HeadMBB, 1140 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB) { 1141 int Num = 0; 1142 MachineBasicBlock *DownBlk; 1143 1144 //trueBlk could be the common post dominator 1145 DownBlk = TrueMBB; 1146 1147 DEBUG( 1148 dbgs() << "handleJumpintoIfImp head = BB" << HeadMBB->getNumber() 1149 << " true = BB" << TrueMBB->getNumber() 1150 << ", numSucc=" << TrueMBB->succ_size() 1151 << " false = BB" << FalseMBB->getNumber() << "\n"; 1152 ); 1153 1154 while (DownBlk) { 1155 DEBUG( 1156 dbgs() << "check down = BB" << DownBlk->getNumber(); 1157 ); 1158 1159 if (singlePathTo(FalseMBB, DownBlk) == SinglePath_InPath) { 1160 DEBUG( 1161 dbgs() << " working\n"; 1162 ); 1163 1164 Num += cloneOnSideEntryTo(HeadMBB, TrueMBB, DownBlk); 1165 Num += cloneOnSideEntryTo(HeadMBB, FalseMBB, DownBlk); 1166 1167 numClonedBlock += Num; 1168 Num += serialPatternMatch(*HeadMBB->succ_begin()); 1169 Num += serialPatternMatch(*std::next(HeadMBB->succ_begin())); 1170 Num += ifPatternMatch(HeadMBB); 1171 assert(Num > 0); 1172 1173 break; 1174 } 1175 DEBUG( 1176 dbgs() << " not working\n"; 1177 ); 1178 DownBlk = (DownBlk->succ_size() == 1) ? (*DownBlk->succ_begin()) : nullptr; 1179 } // walk down the postDomTree 1180 1181 return Num; 1182 } 1183 1184 void AMDGPUCFGStructurizer::showImproveSimpleJumpintoIf( 1185 MachineBasicBlock *HeadMBB, MachineBasicBlock *TrueMBB, 1186 MachineBasicBlock *FalseMBB, MachineBasicBlock *LandMBB, bool Detail) { 1187 dbgs() << "head = BB" << HeadMBB->getNumber() 1188 << " size = " << HeadMBB->size(); 1189 if (Detail) { 1190 dbgs() << "\n"; 1191 HeadMBB->print(dbgs()); 1192 dbgs() << "\n"; 1193 } 1194 1195 if (TrueMBB) { 1196 dbgs() << ", true = BB" << TrueMBB->getNumber() << " size = " 1197 << TrueMBB->size() << " numPred = " << TrueMBB->pred_size(); 1198 if (Detail) { 1199 dbgs() << "\n"; 1200 TrueMBB->print(dbgs()); 1201 dbgs() << "\n"; 1202 } 1203 } 1204 if (FalseMBB) { 1205 dbgs() << ", false = BB" << FalseMBB->getNumber() << " size = " 1206 << FalseMBB->size() << " numPred = " << FalseMBB->pred_size(); 1207 if (Detail) { 1208 dbgs() << "\n"; 1209 FalseMBB->print(dbgs()); 1210 dbgs() << "\n"; 1211 } 1212 } 1213 if (LandMBB) { 1214 dbgs() << ", land = BB" << LandMBB->getNumber() << " size = " 1215 << LandMBB->size() << " numPred = " << LandMBB->pred_size(); 1216 if (Detail) { 1217 dbgs() << "\n"; 1218 LandMBB->print(dbgs()); 1219 dbgs() << "\n"; 1220 } 1221 } 1222 1223 dbgs() << "\n"; 1224 } 1225 1226 int AMDGPUCFGStructurizer::improveSimpleJumpintoIf(MachineBasicBlock *HeadMBB, 1227 MachineBasicBlock *TrueMBB, MachineBasicBlock *FalseMBB, 1228 MachineBasicBlock **LandMBBPtr) { 1229 bool MigrateTrue = false; 1230 bool MigrateFalse = false; 1231 1232 MachineBasicBlock *LandBlk = *LandMBBPtr; 1233 1234 assert((!TrueMBB || TrueMBB->succ_size() <= 1) 1235 && (!FalseMBB || FalseMBB->succ_size() <= 1)); 1236 1237 if (TrueMBB == FalseMBB) 1238 return 0; 1239 1240 MigrateTrue = needMigrateBlock(TrueMBB); 1241 MigrateFalse = needMigrateBlock(FalseMBB); 1242 1243 if (!MigrateTrue && !MigrateFalse) 1244 return 0; 1245 1246 // If we need to migrate either trueBlk and falseBlk, migrate the rest that 1247 // have more than one predecessors. without doing this, its predecessor 1248 // rather than headBlk will have undefined value in initReg. 1249 if (!MigrateTrue && TrueMBB && TrueMBB->pred_size() > 1) 1250 MigrateTrue = true; 1251 if (!MigrateFalse && FalseMBB && FalseMBB->pred_size() > 1) 1252 MigrateFalse = true; 1253 1254 DEBUG( 1255 dbgs() << "before improveSimpleJumpintoIf: "; 1256 showImproveSimpleJumpintoIf(HeadMBB, TrueMBB, FalseMBB, LandBlk, 0); 1257 ); 1258 1259 // org: headBlk => if () {trueBlk} else {falseBlk} => landBlk 1260 // 1261 // new: headBlk => if () {initReg = 1; org trueBlk branch} else 1262 // {initReg = 0; org falseBlk branch } 1263 // => landBlk => if (initReg) {org trueBlk} else {org falseBlk} 1264 // => org landBlk 1265 // if landBlk->pred_size() > 2, put the about if-else inside 1266 // if (initReg !=2) {...} 1267 // 1268 // add initReg = initVal to headBlk 1269 1270 const TargetRegisterClass * I32RC = TRI->getCFGStructurizerRegClass(MVT::i32); 1271 if (!MigrateTrue || !MigrateFalse) { 1272 // XXX: We have an opportunity here to optimize the "branch into if" case 1273 // here. Branch into if looks like this: 1274 // entry 1275 // / | 1276 // diamond_head branch_from 1277 // / \ | 1278 // diamond_false diamond_true 1279 // \ / 1280 // done 1281 // 1282 // The diamond_head block begins the "if" and the diamond_true block 1283 // is the block being "branched into". 1284 // 1285 // If MigrateTrue is true, then TrueBB is the block being "branched into" 1286 // and if MigrateFalse is true, then FalseBB is the block being 1287 // "branched into" 1288 // 1289 // Here is the pseudo code for how I think the optimization should work: 1290 // 1. Insert MOV GPR0, 0 before the branch instruction in diamond_head. 1291 // 2. Insert MOV GPR0, 1 before the branch instruction in branch_from. 1292 // 3. Move the branch instruction from diamond_head into its own basic 1293 // block (new_block). 1294 // 4. Add an unconditional branch from diamond_head to new_block 1295 // 5. Replace the branch instruction in branch_from with an unconditional 1296 // branch to new_block. If branch_from has multiple predecessors, then 1297 // we need to replace the True/False block in the branch 1298 // instruction instead of replacing it. 1299 // 6. Change the condition of the branch instruction in new_block from 1300 // COND to (COND || GPR0) 1301 // 1302 // In order insert these MOV instruction, we will need to use the 1303 // RegisterScavenger. Usually liveness stops being tracked during 1304 // the late machine optimization passes, however if we implement 1305 // bool TargetRegisterInfo::requiresRegisterScavenging( 1306 // const MachineFunction &MF) 1307 // and have it return true, liveness will be tracked correctly 1308 // by generic optimization passes. We will also need to make sure that 1309 // all of our target-specific passes that run after regalloc and before 1310 // the CFGStructurizer track liveness and we will need to modify this pass 1311 // to correctly track liveness. 1312 // 1313 // After the above changes, the new CFG should look like this: 1314 // entry 1315 // / | 1316 // diamond_head branch_from 1317 // \ / 1318 // new_block 1319 // / | 1320 // diamond_false diamond_true 1321 // \ / 1322 // done 1323 // 1324 // Without this optimization, we are forced to duplicate the diamond_true 1325 // block and we will end up with a CFG like this: 1326 // 1327 // entry 1328 // / | 1329 // diamond_head branch_from 1330 // / \ | 1331 // diamond_false diamond_true diamond_true (duplicate) 1332 // \ / | 1333 // done --------------------| 1334 // 1335 // Duplicating diamond_true can be very costly especially if it has a 1336 // lot of instructions. 1337 return 0; 1338 } 1339 1340 int NumNewBlk = 0; 1341 1342 bool LandBlkHasOtherPred = (LandBlk->pred_size() > 2); 1343 1344 //insert AMDGPU::ENDIF to avoid special case "input landBlk == NULL" 1345 MachineBasicBlock::iterator I = insertInstrBefore(LandBlk, AMDGPU::ENDIF); 1346 1347 if (LandBlkHasOtherPred) { 1348 report_fatal_error("Extra register needed to handle CFG"); 1349 unsigned CmpResReg = 1350 HeadMBB->getParent()->getRegInfo().createVirtualRegister(I32RC); 1351 report_fatal_error("Extra compare instruction needed to handle CFG"); 1352 insertCondBranchBefore(LandBlk, I, AMDGPU::IF_PREDICATE_SET, 1353 CmpResReg, DebugLoc()); 1354 } 1355 1356 // XXX: We are running this after RA, so creating virtual registers will 1357 // cause an assertion failure in the PostRA scheduling pass. 1358 unsigned InitReg = 1359 HeadMBB->getParent()->getRegInfo().createVirtualRegister(I32RC); 1360 insertCondBranchBefore(LandBlk, I, AMDGPU::IF_PREDICATE_SET, InitReg, 1361 DebugLoc()); 1362 1363 if (MigrateTrue) { 1364 migrateInstruction(TrueMBB, LandBlk, I); 1365 // need to uncondionally insert the assignment to ensure a path from its 1366 // predecessor rather than headBlk has valid value in initReg if 1367 // (initVal != 1). 1368 report_fatal_error("Extra register needed to handle CFG"); 1369 } 1370 insertInstrBefore(I, AMDGPU::ELSE); 1371 1372 if (MigrateFalse) { 1373 migrateInstruction(FalseMBB, LandBlk, I); 1374 // need to uncondionally insert the assignment to ensure a path from its 1375 // predecessor rather than headBlk has valid value in initReg if 1376 // (initVal != 0) 1377 report_fatal_error("Extra register needed to handle CFG"); 1378 } 1379 1380 if (LandBlkHasOtherPred) { 1381 // add endif 1382 insertInstrBefore(I, AMDGPU::ENDIF); 1383 1384 // put initReg = 2 to other predecessors of landBlk 1385 for (MachineBasicBlock::pred_iterator PI = LandBlk->pred_begin(), 1386 PE = LandBlk->pred_end(); PI != PE; ++PI) { 1387 MachineBasicBlock *MBB = *PI; 1388 if (MBB != TrueMBB && MBB != FalseMBB) 1389 report_fatal_error("Extra register needed to handle CFG"); 1390 } 1391 } 1392 DEBUG( 1393 dbgs() << "result from improveSimpleJumpintoIf: "; 1394 showImproveSimpleJumpintoIf(HeadMBB, TrueMBB, FalseMBB, LandBlk, 0); 1395 ); 1396 1397 // update landBlk 1398 *LandMBBPtr = LandBlk; 1399 1400 return NumNewBlk; 1401 } 1402 1403 void AMDGPUCFGStructurizer::mergeSerialBlock(MachineBasicBlock *DstMBB, 1404 MachineBasicBlock *SrcMBB) { 1405 DEBUG( 1406 dbgs() << "serialPattern BB" << DstMBB->getNumber() 1407 << " <= BB" << SrcMBB->getNumber() << "\n"; 1408 ); 1409 DstMBB->splice(DstMBB->end(), SrcMBB, SrcMBB->begin(), SrcMBB->end()); 1410 1411 DstMBB->removeSuccessor(SrcMBB, true); 1412 cloneSuccessorList(DstMBB, SrcMBB); 1413 1414 removeSuccessor(SrcMBB); 1415 MLI->removeBlock(SrcMBB); 1416 retireBlock(SrcMBB); 1417 } 1418 1419 void AMDGPUCFGStructurizer::mergeIfthenelseBlock(MachineInstr *BranchMI, 1420 MachineBasicBlock *MBB, MachineBasicBlock *TrueMBB, 1421 MachineBasicBlock *FalseMBB, MachineBasicBlock *LandMBB) { 1422 assert (TrueMBB); 1423 DEBUG( 1424 dbgs() << "ifPattern BB" << MBB->getNumber(); 1425 dbgs() << "{ "; 1426 if (TrueMBB) { 1427 dbgs() << "BB" << TrueMBB->getNumber(); 1428 } 1429 dbgs() << " } else "; 1430 dbgs() << "{ "; 1431 if (FalseMBB) { 1432 dbgs() << "BB" << FalseMBB->getNumber(); 1433 } 1434 dbgs() << " }\n "; 1435 dbgs() << "landBlock: "; 1436 if (!LandMBB) { 1437 dbgs() << "NULL"; 1438 } else { 1439 dbgs() << "BB" << LandMBB->getNumber(); 1440 } 1441 dbgs() << "\n"; 1442 ); 1443 1444 int OldOpcode = BranchMI->getOpcode(); 1445 DebugLoc BranchDL = BranchMI->getDebugLoc(); 1446 1447 // transform to 1448 // if cond 1449 // trueBlk 1450 // else 1451 // falseBlk 1452 // endif 1453 // landBlk 1454 1455 MachineBasicBlock::iterator I = BranchMI; 1456 insertCondBranchBefore(I, getBranchNzeroOpcode(OldOpcode), 1457 BranchDL); 1458 1459 if (TrueMBB) { 1460 MBB->splice(I, TrueMBB, TrueMBB->begin(), TrueMBB->end()); 1461 MBB->removeSuccessor(TrueMBB, true); 1462 if (LandMBB && TrueMBB->succ_size()!=0) 1463 TrueMBB->removeSuccessor(LandMBB, true); 1464 retireBlock(TrueMBB); 1465 MLI->removeBlock(TrueMBB); 1466 } 1467 1468 if (FalseMBB) { 1469 insertInstrBefore(I, AMDGPU::ELSE); 1470 MBB->splice(I, FalseMBB, FalseMBB->begin(), 1471 FalseMBB->end()); 1472 MBB->removeSuccessor(FalseMBB, true); 1473 if (LandMBB && FalseMBB->succ_size() != 0) 1474 FalseMBB->removeSuccessor(LandMBB, true); 1475 retireBlock(FalseMBB); 1476 MLI->removeBlock(FalseMBB); 1477 } 1478 insertInstrBefore(I, AMDGPU::ENDIF); 1479 1480 BranchMI->eraseFromParent(); 1481 1482 if (LandMBB && TrueMBB && FalseMBB) 1483 MBB->addSuccessor(LandMBB); 1484 1485 } 1486 1487 void AMDGPUCFGStructurizer::mergeLooplandBlock(MachineBasicBlock *DstBlk, 1488 MachineBasicBlock *LandMBB) { 1489 DEBUG(dbgs() << "loopPattern header = BB" << DstBlk->getNumber() 1490 << " land = BB" << LandMBB->getNumber() << "\n";); 1491 1492 insertInstrBefore(DstBlk, AMDGPU::WHILELOOP, DebugLoc()); 1493 insertInstrEnd(DstBlk, AMDGPU::ENDLOOP, DebugLoc()); 1494 DstBlk->replaceSuccessor(DstBlk, LandMBB); 1495 } 1496 1497 1498 void AMDGPUCFGStructurizer::mergeLoopbreakBlock(MachineBasicBlock *ExitingMBB, 1499 MachineBasicBlock *LandMBB) { 1500 DEBUG(dbgs() << "loopbreakPattern exiting = BB" << ExitingMBB->getNumber() 1501 << " land = BB" << LandMBB->getNumber() << "\n";); 1502 MachineInstr *BranchMI = getLoopendBlockBranchInstr(ExitingMBB); 1503 assert(BranchMI && isCondBranch(BranchMI)); 1504 DebugLoc DL = BranchMI->getDebugLoc(); 1505 MachineBasicBlock *TrueBranch = getTrueBranch(BranchMI); 1506 MachineBasicBlock::iterator I = BranchMI; 1507 if (TrueBranch != LandMBB) 1508 reversePredicateSetter(I); 1509 insertCondBranchBefore(ExitingMBB, I, AMDGPU::IF_PREDICATE_SET, AMDGPU::PREDICATE_BIT, DL); 1510 insertInstrBefore(I, AMDGPU::BREAK); 1511 insertInstrBefore(I, AMDGPU::ENDIF); 1512 //now branchInst can be erase safely 1513 BranchMI->eraseFromParent(); 1514 //now take care of successors, retire blocks 1515 ExitingMBB->removeSuccessor(LandMBB, true); 1516 } 1517 1518 void AMDGPUCFGStructurizer::settleLoopcontBlock(MachineBasicBlock *ContingMBB, 1519 MachineBasicBlock *ContMBB) { 1520 DEBUG(dbgs() << "settleLoopcontBlock conting = BB" 1521 << ContingMBB->getNumber() 1522 << ", cont = BB" << ContMBB->getNumber() << "\n";); 1523 1524 MachineInstr *MI = getLoopendBlockBranchInstr(ContingMBB); 1525 if (MI) { 1526 assert(isCondBranch(MI)); 1527 MachineBasicBlock::iterator I = MI; 1528 MachineBasicBlock *TrueBranch = getTrueBranch(MI); 1529 int OldOpcode = MI->getOpcode(); 1530 DebugLoc DL = MI->getDebugLoc(); 1531 1532 bool UseContinueLogical = ((&*ContingMBB->rbegin()) == MI); 1533 1534 if (!UseContinueLogical) { 1535 int BranchOpcode = 1536 TrueBranch == ContMBB ? getBranchNzeroOpcode(OldOpcode) : 1537 getBranchZeroOpcode(OldOpcode); 1538 insertCondBranchBefore(I, BranchOpcode, DL); 1539 // insertEnd to ensure phi-moves, if exist, go before the continue-instr. 1540 insertInstrEnd(ContingMBB, AMDGPU::CONTINUE, DL); 1541 insertInstrEnd(ContingMBB, AMDGPU::ENDIF, DL); 1542 } else { 1543 int BranchOpcode = 1544 TrueBranch == ContMBB ? getContinueNzeroOpcode(OldOpcode) : 1545 getContinueZeroOpcode(OldOpcode); 1546 insertCondBranchBefore(I, BranchOpcode, DL); 1547 } 1548 1549 MI->eraseFromParent(); 1550 } else { 1551 // if we've arrived here then we've already erased the branch instruction 1552 // travel back up the basic block to see the last reference of our debug 1553 // location we've just inserted that reference here so it should be 1554 // representative insertEnd to ensure phi-moves, if exist, go before the 1555 // continue-instr. 1556 insertInstrEnd(ContingMBB, AMDGPU::CONTINUE, 1557 getLastDebugLocInBB(ContingMBB)); 1558 } 1559 } 1560 1561 int AMDGPUCFGStructurizer::cloneOnSideEntryTo(MachineBasicBlock *PreMBB, 1562 MachineBasicBlock *SrcMBB, MachineBasicBlock *DstMBB) { 1563 int Cloned = 0; 1564 assert(PreMBB->isSuccessor(SrcMBB)); 1565 while (SrcMBB && SrcMBB != DstMBB) { 1566 assert(SrcMBB->succ_size() == 1); 1567 if (SrcMBB->pred_size() > 1) { 1568 SrcMBB = cloneBlockForPredecessor(SrcMBB, PreMBB); 1569 ++Cloned; 1570 } 1571 1572 PreMBB = SrcMBB; 1573 SrcMBB = *SrcMBB->succ_begin(); 1574 } 1575 1576 return Cloned; 1577 } 1578 1579 MachineBasicBlock * 1580 AMDGPUCFGStructurizer::cloneBlockForPredecessor(MachineBasicBlock *MBB, 1581 MachineBasicBlock *PredMBB) { 1582 assert(PredMBB->isSuccessor(MBB) && 1583 "succBlk is not a prececessor of curBlk"); 1584 1585 MachineBasicBlock *CloneMBB = clone(MBB); //clone instructions 1586 replaceInstrUseOfBlockWith(PredMBB, MBB, CloneMBB); 1587 //srcBlk, oldBlk, newBlk 1588 1589 PredMBB->replaceSuccessor(MBB, CloneMBB); 1590 1591 // add all successor to cloneBlk 1592 cloneSuccessorList(CloneMBB, MBB); 1593 1594 numClonedInstr += MBB->size(); 1595 1596 DEBUG( 1597 dbgs() << "Cloned block: " << "BB" 1598 << MBB->getNumber() << "size " << MBB->size() << "\n"; 1599 ); 1600 1601 SHOWNEWBLK(CloneMBB, "result of Cloned block: "); 1602 1603 return CloneMBB; 1604 } 1605 1606 void AMDGPUCFGStructurizer::migrateInstruction(MachineBasicBlock *SrcMBB, 1607 MachineBasicBlock *DstMBB, MachineBasicBlock::iterator I) { 1608 MachineBasicBlock::iterator SpliceEnd; 1609 //look for the input branchinstr, not the AMDGPU branchinstr 1610 MachineInstr *BranchMI = getNormalBlockBranchInstr(SrcMBB); 1611 if (!BranchMI) { 1612 DEBUG( 1613 dbgs() << "migrateInstruction don't see branch instr\n" ; 1614 ); 1615 SpliceEnd = SrcMBB->end(); 1616 } else { 1617 DEBUG(dbgs() << "migrateInstruction see branch instr: " << *BranchMI); 1618 SpliceEnd = BranchMI; 1619 } 1620 DEBUG( 1621 dbgs() << "migrateInstruction before splice dstSize = " << DstMBB->size() 1622 << "srcSize = " << SrcMBB->size() << "\n"; 1623 ); 1624 1625 //splice insert before insertPos 1626 DstMBB->splice(I, SrcMBB, SrcMBB->begin(), SpliceEnd); 1627 1628 DEBUG( 1629 dbgs() << "migrateInstruction after splice dstSize = " << DstMBB->size() 1630 << "srcSize = " << SrcMBB->size() << '\n'; 1631 ); 1632 } 1633 1634 MachineBasicBlock * 1635 AMDGPUCFGStructurizer::normalizeInfiniteLoopExit(MachineLoop* LoopRep) { 1636 MachineBasicBlock *LoopHeader = LoopRep->getHeader(); 1637 MachineBasicBlock *LoopLatch = LoopRep->getLoopLatch(); 1638 1639 if (!LoopHeader || !LoopLatch) 1640 return nullptr; 1641 MachineInstr *BranchMI = getLoopendBlockBranchInstr(LoopLatch); 1642 // Is LoopRep an infinite loop ? 1643 if (!BranchMI || !isUncondBranch(BranchMI)) 1644 return nullptr; 1645 1646 MachineBasicBlock *DummyExitBlk = FuncRep->CreateMachineBasicBlock(); 1647 FuncRep->push_back(DummyExitBlk); //insert to function 1648 SHOWNEWBLK(DummyExitBlk, "DummyExitBlock to normalize infiniteLoop: "); 1649 DEBUG(dbgs() << "Old branch instr: " << *BranchMI << "\n";); 1650 LLVMContext &Ctx = LoopHeader->getParent()->getFunction()->getContext(); 1651 Ctx.emitError("Extra register needed to handle CFG"); 1652 return nullptr; 1653 } 1654 1655 void AMDGPUCFGStructurizer::removeUnconditionalBranch(MachineBasicBlock *MBB) { 1656 MachineInstr *BranchMI; 1657 1658 // I saw two unconditional branch in one basic block in example 1659 // test_fc_do_while_or.c need to fix the upstream on this to remove the loop. 1660 while ((BranchMI = getLoopendBlockBranchInstr(MBB)) 1661 && isUncondBranch(BranchMI)) { 1662 DEBUG(dbgs() << "Removing uncond branch instr: " << *BranchMI); 1663 BranchMI->eraseFromParent(); 1664 } 1665 } 1666 1667 void AMDGPUCFGStructurizer::removeRedundantConditionalBranch( 1668 MachineBasicBlock *MBB) { 1669 if (MBB->succ_size() != 2) 1670 return; 1671 MachineBasicBlock *MBB1 = *MBB->succ_begin(); 1672 MachineBasicBlock *MBB2 = *std::next(MBB->succ_begin()); 1673 if (MBB1 != MBB2) 1674 return; 1675 1676 MachineInstr *BranchMI = getNormalBlockBranchInstr(MBB); 1677 assert(BranchMI && isCondBranch(BranchMI)); 1678 DEBUG(dbgs() << "Removing unneeded cond branch instr: " << *BranchMI); 1679 BranchMI->eraseFromParent(); 1680 SHOWNEWBLK(MBB1, "Removing redundant successor"); 1681 MBB->removeSuccessor(MBB1, true); 1682 } 1683 1684 void AMDGPUCFGStructurizer::addDummyExitBlock( 1685 SmallVectorImpl<MachineBasicBlock*> &RetMBB) { 1686 MachineBasicBlock *DummyExitBlk = FuncRep->CreateMachineBasicBlock(); 1687 FuncRep->push_back(DummyExitBlk); //insert to function 1688 insertInstrEnd(DummyExitBlk, AMDGPU::RETURN); 1689 1690 for (SmallVectorImpl<MachineBasicBlock *>::iterator It = RetMBB.begin(), 1691 E = RetMBB.end(); It != E; ++It) { 1692 MachineBasicBlock *MBB = *It; 1693 MachineInstr *MI = getReturnInstr(MBB); 1694 if (MI) 1695 MI->eraseFromParent(); 1696 MBB->addSuccessor(DummyExitBlk); 1697 DEBUG( 1698 dbgs() << "Add dummyExitBlock to BB" << MBB->getNumber() 1699 << " successors\n"; 1700 ); 1701 } 1702 SHOWNEWBLK(DummyExitBlk, "DummyExitBlock: "); 1703 } 1704 1705 void AMDGPUCFGStructurizer::removeSuccessor(MachineBasicBlock *MBB) { 1706 while (MBB->succ_size()) 1707 MBB->removeSuccessor(*MBB->succ_begin()); 1708 } 1709 1710 void AMDGPUCFGStructurizer::recordSccnum(MachineBasicBlock *MBB, 1711 int SccNum) { 1712 BlockInformation *&srcBlkInfo = BlockInfoMap[MBB]; 1713 if (!srcBlkInfo) 1714 srcBlkInfo = new BlockInformation(); 1715 srcBlkInfo->SccNum = SccNum; 1716 } 1717 1718 void AMDGPUCFGStructurizer::retireBlock(MachineBasicBlock *MBB) { 1719 DEBUG( 1720 dbgs() << "Retiring BB" << MBB->getNumber() << "\n"; 1721 ); 1722 1723 BlockInformation *&SrcBlkInfo = BlockInfoMap[MBB]; 1724 1725 if (!SrcBlkInfo) 1726 SrcBlkInfo = new BlockInformation(); 1727 1728 SrcBlkInfo->IsRetired = true; 1729 assert(MBB->succ_size() == 0 && MBB->pred_size() == 0 1730 && "can't retire block yet"); 1731 } 1732 1733 char AMDGPUCFGStructurizer::ID = 0; 1734 1735 } // end anonymous namespace 1736 1737 1738 INITIALIZE_PASS_BEGIN(AMDGPUCFGStructurizer, "amdgpustructurizer", 1739 "AMDGPU CFG Structurizer", false, false) 1740 INITIALIZE_PASS_DEPENDENCY(MachineDominatorTree) 1741 INITIALIZE_PASS_DEPENDENCY(MachinePostDominatorTree) 1742 INITIALIZE_PASS_DEPENDENCY(MachineLoopInfo) 1743 INITIALIZE_PASS_END(AMDGPUCFGStructurizer, "amdgpustructurizer", 1744 "AMDGPU CFG Structurizer", false, false) 1745 1746 FunctionPass *llvm::createAMDGPUCFGStructurizerPass() { 1747 return new AMDGPUCFGStructurizer(); 1748 } 1749