lib/Analysis/DataStructure/EquivClassGraphs.cpp

   1 //===- EquivClassGraphs.cpp - Merge equiv-class graphs & inline bottom-up -===//
   2 //
   3 //                     The LLVM Compiler Infrastructure
   4 //
   5 // This file was developed by the LLVM research group and is distributed under
   6 // the University of Illinois Open Source License. See LICENSE.TXT for details.
   7 //
   8 //===----------------------------------------------------------------------===//
   9 //
  10 // This pass is the same as the complete bottom-up graphs, but
  11 // with functions partitioned into equivalence classes and a single merged
  12 // DS graph for all functions in an equivalence class.  After this merging,
  13 // graphs are inlined bottom-up on the SCCs of the final (CBU) call graph.
  14 //
  15 //===----------------------------------------------------------------------===//
  16
  17 #define DEBUG_TYPE "ECGraphs"
  18 #include "llvm/Analysis/DataStructure/DataStructure.h"
  19 #include "llvm/DerivedTypes.h"
  20 #include "llvm/Module.h"
  21 #include "llvm/Pass.h"
  22 #include "llvm/Analysis/DataStructure/DSGraph.h"
  23 #include "llvm/Support/CallSite.h"
  24 #include "llvm/Support/Debug.h"
  25 #include "llvm/ADT/SCCIterator.h"
  26 #include "llvm/ADT/Statistic.h"
  27 #include "llvm/ADT/EquivalenceClasses.h"
  28 #include "llvm/ADT/STLExtras.h"
  29 #include <iostream>
  30 using namespace llvm;
  31
  32 namespace {
  33   RegisterAnalysis<EquivClassGraphs> X("eqdatastructure",
  34                     "Equivalence-class Bottom-up Data Structure Analysis");
  35   Statistic<> NumEquivBUInlines("equivdatastructures",
  36                                 "Number of graphs inlined");
  37   Statistic<> NumFoldGraphInlines("Inline equiv-class graphs bottom up",
  38                                   "Number of graphs inlined");
  39 }
  40
  41 #ifndef NDEBUG
  42 template<typename GT>
  43 static void CheckAllGraphs(Module *M, GT &ECGraphs) {
  44   for (Module::iterator I = M->begin(), E = M->end(); I != E; ++I)
  45     if (!I->isExternal()) {
  46       DSGraph &G = ECGraphs.getDSGraph(*I);
  47       if (G.retnodes_begin()->first != I)
  48         continue;  // Only check a graph once.
  49
  50       DSGraph::NodeMapTy GlobalsGraphNodeMapping;
  51       G.computeGToGGMapping(GlobalsGraphNodeMapping);
  52     }
  53 }
  54 #endif
  55
  56 // getSomeCalleeForCallSite - Return any one callee function at a call site.
  57 //
  58 Function *EquivClassGraphs::getSomeCalleeForCallSite(const CallSite &CS) const{
  59   Function *thisFunc = CS.getCaller();
  60   assert(thisFunc && "getSomeCalleeForCallSite(): Not a valid call site?");
  61   DSGraph &DSG = getDSGraph(*thisFunc);
  62   DSNode *calleeNode = DSG.getNodeForValue(CS.getCalledValue()).getNode();
  63   std::map<DSNode*, Function *>::const_iterator I =
  64     OneCalledFunction.find(calleeNode);
  65   return (I == OneCalledFunction.end())? NULL : I->second;
  66 }
  67
  68 // runOnModule - Calculate the bottom up data structure graphs for each function
  69 // in the program.
  70 //
  71 bool EquivClassGraphs::runOnModule(Module &M) {
  72   CBU = &getAnalysis<CompleteBUDataStructures>();
  73   GlobalECs = CBU->getGlobalECs();
  74   DEBUG(CheckAllGraphs(&M, *CBU));
  75
  76   GlobalsGraph = new DSGraph(CBU->getGlobalsGraph(), GlobalECs);
  77   GlobalsGraph->setPrintAuxCalls();
  78
  79   ActualCallees = CBU->getActualCallees();
  80
  81   // Find equivalence classes of functions called from common call sites.
  82   // Fold the CBU graphs for all functions in an equivalence class.
  83   buildIndirectFunctionSets(M);
  84
  85   // Stack of functions used for Tarjan's SCC-finding algorithm.
  86   std::vector<DSGraph*> Stack;
  87   std::map<DSGraph*, unsigned> ValMap;
  88   unsigned NextID = 1;
  89
  90   Function *MainFunc = M.getMainFunction();
  91   if (MainFunc && !MainFunc->isExternal()) {
  92     processSCC(getOrCreateGraph(*MainFunc), Stack, NextID, ValMap);
  93   } else {
  94     std::cerr << "Fold Graphs: No 'main' function found!\n";
  95   }
  96
  97   for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
  98     if (!I->isExternal())
  99       processSCC(getOrCreateGraph(*I), Stack, NextID, ValMap);
 100
 101   DEBUG(CheckAllGraphs(&M, *this));
 102
 103   getGlobalsGraph().removeTriviallyDeadNodes();
 104   getGlobalsGraph().markIncompleteNodes(DSGraph::IgnoreGlobals);
 105
 106   // Merge the globals variables (not the calls) from the globals graph back
 107   // into the main function's graph so that the main function contains all of
 108   // the information about global pools and GV usage in the program.
 109   if (MainFunc && !MainFunc->isExternal()) {
 110     DSGraph &MainGraph = getOrCreateGraph(*MainFunc);
 111     const DSGraph &GG = *MainGraph.getGlobalsGraph();
 112     ReachabilityCloner RC(MainGraph, GG,
 113                           DSGraph::DontCloneCallNodes |
 114                           DSGraph::DontCloneAuxCallNodes);
 115
 116     // Clone the global nodes into this graph.
 117     for (DSScalarMap::global_iterator I = GG.getScalarMap().global_begin(),
 118            E = GG.getScalarMap().global_end(); I != E; ++I)
 119       if (isa<GlobalVariable>(*I))
 120         RC.getClonedNH(GG.getNodeForValue(*I));
 121
 122     MainGraph.maskIncompleteMarkers();
 123     MainGraph.markIncompleteNodes(DSGraph::MarkFormalArgs |
 124                                   DSGraph::IgnoreGlobals);
 125   }
 126
 127   // Final processing.  Note that dead node elimination may actually remove
 128   // globals from a function graph that are immediately used.  If there are no
 129   // scalars pointing to the node (e.g. because the only use is a direct store
 130   // to a scalar global) we have to make sure to rematerialize the globals back
 131   // into the graphs here, or clients will break!
 132   for (Module::global_iterator GI = M.global_begin(), E = M.global_end();
 133        GI != E; ++GI)
 134     // This only happens to first class typed globals.
 135     if (GI->getType()->getElementType()->isFirstClassType())
 136       for (Value::use_iterator UI = GI->use_begin(), E = GI->use_end();
 137            UI != E; ++UI)
 138         // This only happens to direct uses by instructions.
 139         if (Instruction *User = dyn_cast<Instruction>(*UI)) {
 140           DSGraph &DSG = getOrCreateGraph(*User->getParent()->getParent());
 141           if (!DSG.getScalarMap().count(GI)) {
 142             // If this global does not exist in the graph, but it is immediately
 143             // used by an instruction in the graph, clone it over from the
 144             // globals graph.
 145             ReachabilityCloner RC(DSG, *GlobalsGraph, 0);
 146             RC.getClonedNH(GlobalsGraph->getNodeForValue(GI));
 147           }
 148         }
 149
 150   return false;
 151 }
 152
 153
 154 // buildIndirectFunctionSets - Iterate over the module looking for indirect
 155 // calls to functions.  If a call site can invoke any functions [F1, F2... FN],
 156 // unify the N functions together in the FuncECs set.
 157 //
 158 void EquivClassGraphs::buildIndirectFunctionSets(Module &M) {
 159   const ActualCalleesTy& AC = CBU->getActualCallees();
 160
 161   // Loop over all of the indirect calls in the program.  If a call site can
 162   // call multiple different functions, we need to unify all of the callees into
 163   // the same equivalence class.
 164   Instruction *LastInst = 0;
 165   Function *FirstFunc = 0;
 166   for (ActualCalleesTy::const_iterator I=AC.begin(), E=AC.end(); I != E; ++I) {
 167     if (I->second->isExternal())
 168       continue;                         // Ignore functions we cannot modify
 169
 170     CallSite CS = CallSite::get(I->first);
 171
 172     if (CS.getCalledFunction()) {       // Direct call:
 173       FuncECs.insert(I->second);        // -- Make sure function has equiv class
 174       FirstFunc = I->second;            // -- First callee at this site
 175     } else {                            // Else indirect call
 176       // DEBUG(std::cerr << "CALLEE: " << I->second->getName()
 177       //       << " from : " << I->first);
 178       if (I->first != LastInst) {
 179         // This is the first callee from this call site.
 180         LastInst = I->first;
 181         FirstFunc = I->second;
 182         // Instead of storing the lastInst For Indirection call Sites we store
 183         // the DSNode for the function ptr arguemnt
 184         Function *thisFunc = LastInst->getParent()->getParent();
 185         DSGraph &TFG = CBU->getDSGraph(*thisFunc);
 186         DSNode *calleeNode = TFG.getNodeForValue(CS.getCalledValue()).getNode();
 187         OneCalledFunction[calleeNode] = FirstFunc;
 188         FuncECs.insert(I->second);
 189       } else {
 190         // This is not the first possible callee from a particular call site.
 191         // Union the callee in with the other functions.
 192         FuncECs.unionSets(FirstFunc, I->second);
 193 #ifndef NDEBUG
 194         Function *thisFunc = LastInst->getParent()->getParent();
 195         DSGraph &TFG = CBU->getDSGraph(*thisFunc);
 196         DSNode *calleeNode = TFG.getNodeForValue(CS.getCalledValue()).getNode();
 197         assert(OneCalledFunction.count(calleeNode) > 0 && "Missed a call?");
 198 #endif
 199       }
 200     }
 201
 202     // Now include all functions that share a graph with any function in the
 203     // equivalence class.  More precisely, if F is in the class, and G(F) is
 204     // its graph, then we include all other functions that are also in G(F).
 205     // Currently, that is just the functions in the same call-graph-SCC as F.
 206     //
 207     DSGraph& funcDSGraph = CBU->getDSGraph(*I->second);
 208     for (DSGraph::retnodes_iterator RI = funcDSGraph.retnodes_begin(),
 209            RE = funcDSGraph.retnodes_end(); RI != RE; ++RI)
 210       FuncECs.unionSets(FirstFunc, RI->first);
 211   }
 212
 213   // Now that all of the equivalences have been built, merge the graphs for
 214   // each equivalence class.
 215   //
 216   DEBUG(std::cerr << "\nIndirect Function Equivalence Sets:\n");
 217   for (EquivalenceClasses<Function*>::iterator EQSI = FuncECs.begin(), E =
 218          FuncECs.end(); EQSI != E; ++EQSI) {
 219     if (!EQSI->isLeader()) continue;
 220
 221     EquivalenceClasses<Function*>::member_iterator SI =
 222       FuncECs.member_begin(EQSI);
 223     assert(SI != FuncECs.member_end() && "Empty equiv set??");
 224     EquivalenceClasses<Function*>::member_iterator SN = SI;
 225     ++SN;
 226     if (SN == FuncECs.member_end())
 227       continue;   // Single function equivalence set, no merging to do.
 228
 229     Function* LF = *SI;
 230
 231 #ifndef NDEBUG
 232     DEBUG(std::cerr <<"  Equivalence set for leader " << LF->getName() <<" = ");
 233     for (SN = SI; SN != FuncECs.member_end(); ++SN)
 234       DEBUG(std::cerr << " " << (*SN)->getName() << "," );
 235     DEBUG(std::cerr << "\n");
 236 #endif
 237
 238     // This equiv class has multiple functions: merge their graphs.  First,
 239     // clone the CBU graph for the leader and make it the common graph for the
 240     // equivalence graph.
 241     DSGraph &MergedG = getOrCreateGraph(*LF);
 242
 243     // Record the argument nodes for use in merging later below.
 244     std::vector<DSNodeHandle> ArgNodes;
 245
 246     for (Function::arg_iterator AI = LF->arg_begin(), E = LF->arg_end();
 247          AI != E; ++AI)
 248       if (DS::isPointerType(AI->getType()))
 249         ArgNodes.push_back(MergedG.getNodeForValue(AI));
 250
 251     // Merge in the graphs of all other functions in this equiv. class.  Note
 252     // that two or more functions may have the same graph, and it only needs
 253     // to be merged in once.
 254     std::set<DSGraph*> GraphsMerged;
 255     GraphsMerged.insert(&CBU->getDSGraph(*LF));
 256
 257     for (++SI; SI != FuncECs.member_end(); ++SI) {
 258       Function *F = *SI;
 259       DSGraph *&FG = DSInfo[F];
 260
 261       DSGraph &CBUGraph = CBU->getDSGraph(*F);
 262       if (GraphsMerged.insert(&CBUGraph).second) {
 263         // Record the "folded" graph for the function.
 264         for (DSGraph::retnodes_iterator I = CBUGraph.retnodes_begin(),
 265                E = CBUGraph.retnodes_end(); I != E; ++I) {
 266           assert(DSInfo[I->first] == 0 && "Graph already exists for Fn!");
 267           DSInfo[I->first] = &MergedG;
 268         }
 269
 270         // Clone this member of the equivalence class into MergedG.
 271         MergedG.cloneInto(CBUGraph);
 272       }
 273
 274       // Merge the return nodes of all functions together.
 275       MergedG.getReturnNodes()[LF].mergeWith(MergedG.getReturnNodes()[F]);
 276
 277       // Merge the function arguments with all argument nodes found so far.
 278       // If there are extra function args, add them to the vector of argNodes
 279       Function::arg_iterator AI2 = F->arg_begin(), AI2end = F->arg_end();
 280       for (unsigned arg = 0, numArgs = ArgNodes.size();
 281            arg != numArgs && AI2 != AI2end; ++AI2, ++arg)
 282         if (DS::isPointerType(AI2->getType()))
 283           ArgNodes[arg].mergeWith(MergedG.getNodeForValue(AI2));
 284
 285       for ( ; AI2 != AI2end; ++AI2)
 286         if (DS::isPointerType(AI2->getType()))
 287           ArgNodes.push_back(MergedG.getNodeForValue(AI2));
 288       DEBUG(MergedG.AssertGraphOK());
 289     }
 290   }
 291   DEBUG(std::cerr << "\n");
 292 }
 293
 294
 295 DSGraph &EquivClassGraphs::getOrCreateGraph(Function &F) {
 296   // Has the graph already been created?
 297   DSGraph *&Graph = DSInfo[&F];
 298   if (Graph) return *Graph;
 299
 300   DSGraph &CBUGraph = CBU->getDSGraph(F);
 301
 302   // Copy the CBU graph...
 303   Graph = new DSGraph(CBUGraph, GlobalECs);   // updates the map via reference
 304   Graph->setGlobalsGraph(&getGlobalsGraph());
 305   Graph->setPrintAuxCalls();
 306
 307   // Make sure to update the DSInfo map for all functions in the graph!
 308   for (DSGraph::retnodes_iterator I = Graph->retnodes_begin();
 309        I != Graph->retnodes_end(); ++I)
 310     if (I->first != &F) {
 311       DSGraph *&FG = DSInfo[I->first];
 312       assert(FG == 0 && "Merging function in SCC twice?");
 313       FG = Graph;
 314     }
 315
 316   return *Graph;
 317 }
 318
 319
 320 unsigned EquivClassGraphs::
 321 processSCC(DSGraph &FG, std::vector<DSGraph*> &Stack, unsigned &NextID,
 322            std::map<DSGraph*, unsigned> &ValMap) {
 323   std::map<DSGraph*, unsigned>::iterator It = ValMap.lower_bound(&FG);
 324   if (It != ValMap.end() && It->first == &FG)
 325     return It->second;
 326
 327   DEBUG(std::cerr << "    ProcessSCC for function " << FG.getFunctionNames()
 328                   << "\n");
 329
 330   unsigned Min = NextID++, MyID = Min;
 331   ValMap[&FG] = Min;
 332   Stack.push_back(&FG);
 333
 334   // The edges out of the current node are the call site targets...
 335   for (DSGraph::fc_iterator CI = FG.fc_begin(), CE = FG.fc_end();
 336        CI != CE; ++CI) {
 337     Instruction *Call = CI->getCallSite().getInstruction();
 338
 339     // Loop over all of the actually called functions...
 340     for (callee_iterator I = callee_begin(Call), E = callee_end(Call);
 341          I != E; ++I)
 342       if (!I->second->isExternal()) {
 343         // Process the callee as necessary.
 344         unsigned M = processSCC(getOrCreateGraph(*I->second),
 345                                 Stack, NextID, ValMap);
 346         if (M < Min) Min = M;
 347       }
 348   }
 349
 350   assert(ValMap[&FG] == MyID && "SCC construction assumption wrong!");
 351   if (Min != MyID)
 352     return Min;         // This is part of a larger SCC!
 353
 354   // If this is a new SCC, process it now.
 355   bool MergedGraphs = false;
 356   while (Stack.back() != &FG) {
 357     DSGraph *NG = Stack.back();
 358     ValMap[NG] = ~0U;
 359
 360     // If the SCC found is not the same as those found in CBU, make sure to
 361     // merge the graphs as appropriate.
 362     FG.cloneInto(*NG);
 363
 364     // Update the DSInfo map and delete the old graph...
 365     for (DSGraph::retnodes_iterator I = NG->retnodes_begin();
 366          I != NG->retnodes_end(); ++I)
 367       DSInfo[I->first] = &FG;
 368
 369     // Remove NG from the ValMap since the pointer may get recycled.
 370     ValMap.erase(NG);
 371     delete NG;
 372     MergedGraphs = true;
 373     Stack.pop_back();
 374   }
 375
 376   // Clean up the graph before we start inlining a bunch again.
 377   if (MergedGraphs)
 378     FG.removeTriviallyDeadNodes();
 379
 380   Stack.pop_back();
 381
 382   processGraph(FG);
 383   ValMap[&FG] = ~0U;
 384   return MyID;
 385 }
 386
 387
 388 /// processGraph - Process the CBU graphs for the program in bottom-up order on
 389 /// the SCC of the __ACTUAL__ call graph.  This builds final folded CBU graphs.
 390 void EquivClassGraphs::processGraph(DSGraph &G) {
 391   DEBUG(std::cerr << "    ProcessGraph for function "
 392                   << G.getFunctionNames() << "\n");
 393
 394   hash_set<Instruction*> calls;
 395
 396   // Else we need to inline some callee graph.  Visit all call sites.
 397   // The edges out of the current node are the call site targets...
 398   unsigned i = 0;
 399   for (DSGraph::fc_iterator CI = G.fc_begin(), CE = G.fc_end(); CI != CE;
 400        ++CI, ++i) {
 401     const DSCallSite &CS = *CI;
 402     Instruction *TheCall = CS.getCallSite().getInstruction();
 403
 404     assert(calls.insert(TheCall).second &&
 405            "Call instruction occurs multiple times in graph??");
 406
 407     if (CS.getRetVal().isNull() && CS.getNumPtrArgs() == 0)
 408       continue;
 409
 410     // Inline the common callee graph into the current graph, if the callee
 411     // graph has not changed.  Note that all callees should have the same
 412     // graph so we only need to do this once.
 413     //
 414     DSGraph* CalleeGraph = NULL;
 415     callee_iterator I = callee_begin(TheCall), E = callee_end(TheCall);
 416     unsigned TNum, Num;
 417
 418     // Loop over all potential callees to find the first non-external callee.
 419     for (TNum = 0, Num = std::distance(I, E); I != E; ++I, ++TNum)
 420       if (!I->second->isExternal())
 421         break;
 422
 423     // Now check if the graph has changed and if so, clone and inline it.
 424     if (I != E) {
 425       Function *CalleeFunc = I->second;
 426
 427       // Merge the callee's graph into this graph, if not already the same.
 428       // Callees in the same equivalence class (which subsumes those
 429       // in the same SCCs) have the same graph.  Note that all recursion
 430       // including self-recursion have been folded in the equiv classes.
 431       //
 432       CalleeGraph = &getOrCreateGraph(*CalleeFunc);
 433       if (CalleeGraph != &G) {
 434         ++NumFoldGraphInlines;
 435         G.mergeInGraph(CS, *CalleeFunc, *CalleeGraph,
 436                        DSGraph::StripAllocaBit |
 437                        DSGraph::DontCloneCallNodes |
 438                        DSGraph::DontCloneAuxCallNodes);
 439         DEBUG(std::cerr << "    Inlining graph [" << i << "/"
 440               << G.getFunctionCalls().size()-1
 441               << ":" << TNum << "/" << Num-1 << "] for "
 442               << CalleeFunc->getName() << "["
 443               << CalleeGraph->getGraphSize() << "+"
 444               << CalleeGraph->getAuxFunctionCalls().size()
 445               << "] into '" /*<< G.getFunctionNames()*/ << "' ["
 446               << G.getGraphSize() << "+" << G.getAuxFunctionCalls().size()
 447               << "]\n");
 448       }
 449     }
 450
 451 #ifndef NDEBUG
 452     // Now loop over the rest of the callees and make sure they have the
 453     // same graph as the one inlined above.
 454     if (CalleeGraph)
 455       for (++I, ++TNum; I != E; ++I, ++TNum)
 456         if (!I->second->isExternal())
 457           assert(CalleeGraph == &getOrCreateGraph(*I->second) &&
 458                  "Callees at a call site have different graphs?");
 459 #endif
 460   }
 461
 462   // Recompute the Incomplete markers.
 463   G.maskIncompleteMarkers();
 464   G.markIncompleteNodes(DSGraph::MarkFormalArgs);
 465
 466   // Delete dead nodes.  Treat globals that are unreachable but that can
 467   // reach live nodes as live.
 468   G.removeDeadNodes(DSGraph::KeepUnreachableGlobals);
 469
 470   // When this graph is finalized, clone the globals in the graph into the
 471   // globals graph to make sure it has everything, from all graphs.
 472   ReachabilityCloner RC(*G.getGlobalsGraph(), G, DSGraph::StripAllocaBit);
 473
 474   // Clone everything reachable from globals in the function graph into the
 475   // globals graph.
 476   DSScalarMap &MainSM = G.getScalarMap();
 477   for (DSScalarMap::global_iterator I = MainSM.global_begin(),
 478          E = MainSM.global_end(); I != E; ++I)
 479     RC.getClonedNH(MainSM[*I]);
 480
 481   DEBUG(std::cerr << "  -- DONE ProcessGraph for function "
 482                   << G.getFunctionNames() << "\n");
 483 }