1//===- ConstantMerge.cpp - Merge duplicate global constants ---------------===//
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 file defines the interface to a pass that merges duplicate global
11// constants together into a single constant that is shared.  This is useful
12// because some passes (ie TraceValues) insert a lot of string constants into
13// the program, regardless of whether or not an existing string is available.
14//
15// Algorithm: ConstantMerge is designed to build up a map of available constants
16// and eliminate duplicates when it is initialized.
17//
18//===----------------------------------------------------------------------===//
19
20#include "llvm/Transforms/IPO/ConstantMerge.h"
21#include "llvm/ADT/DenseMap.h"
22#include "llvm/ADT/PointerIntPair.h"
23#include "llvm/ADT/SmallPtrSet.h"
24#include "llvm/ADT/Statistic.h"
25#include "llvm/IR/Constants.h"
26#include "llvm/IR/DataLayout.h"
27#include "llvm/IR/DerivedTypes.h"
28#include "llvm/IR/Module.h"
29#include "llvm/IR/Operator.h"
30#include "llvm/Pass.h"
31#include "llvm/Transforms/IPO.h"
32using namespace llvm;
33
34#define DEBUG_TYPE "constmerge"
35
36STATISTIC(NumMerged, "Number of global constants merged");
37
38/// Find values that are marked as llvm.used.
39static void FindUsedValues(GlobalVariable *LLVMUsed,
40                           SmallPtrSetImpl<const GlobalValue*> &UsedValues) {
41  if (!LLVMUsed) return;
42  ConstantArray *Inits = cast<ConstantArray>(LLVMUsed->getInitializer());
43
44  for (unsigned i = 0, e = Inits->getNumOperands(); i != e; ++i) {
45    Value *Operand = Inits->getOperand(i)->stripPointerCastsNoFollowAliases();
46    GlobalValue *GV = cast<GlobalValue>(Operand);
47    UsedValues.insert(GV);
48  }
49}
50
51// True if A is better than B.
52static bool IsBetterCanonical(const GlobalVariable &A,
53                              const GlobalVariable &B) {
54  if (!A.hasLocalLinkage() && B.hasLocalLinkage())
55    return true;
56
57  if (A.hasLocalLinkage() && !B.hasLocalLinkage())
58    return false;
59
60  return A.hasGlobalUnnamedAddr();
61}
62
63static unsigned getAlignment(GlobalVariable *GV) {
64  unsigned Align = GV->getAlignment();
65  if (Align)
66    return Align;
67  return GV->getParent()->getDataLayout().getPreferredAlignment(GV);
68}
69
70static bool mergeConstants(Module &M) {
71  // Find all the globals that are marked "used".  These cannot be merged.
72  SmallPtrSet<const GlobalValue*, 8> UsedGlobals;
73  FindUsedValues(M.getGlobalVariable("llvm.used"), UsedGlobals);
74  FindUsedValues(M.getGlobalVariable("llvm.compiler.used"), UsedGlobals);
75
76  // Map unique constants to globals.
77  DenseMap<Constant *, GlobalVariable *> CMap;
78
79  // Replacements - This vector contains a list of replacements to perform.
80  SmallVector<std::pair<GlobalVariable*, GlobalVariable*>, 32> Replacements;
81
82  bool MadeChange = false;
83
84  // Iterate constant merging while we are still making progress.  Merging two
85  // constants together may allow us to merge other constants together if the
86  // second level constants have initializers which point to the globals that
87  // were just merged.
88  while (1) {
89
90    // First: Find the canonical constants others will be merged with.
91    for (Module::global_iterator GVI = M.global_begin(), E = M.global_end();
92         GVI != E; ) {
93      GlobalVariable *GV = &*GVI++;
94
95      // If this GV is dead, remove it.
96      GV->removeDeadConstantUsers();
97      if (GV->use_empty() && GV->hasLocalLinkage()) {
98        GV->eraseFromParent();
99        continue;
100      }
101
102      // Only process constants with initializers in the default address space.
103      if (!GV->isConstant() || !GV->hasDefinitiveInitializer() ||
104          GV->getType()->getAddressSpace() != 0 || GV->hasSection() ||
105          // Don't touch values marked with attribute(used).
106          UsedGlobals.count(GV))
107        continue;
108
109      // This transformation is legal for weak ODR globals in the sense it
110      // doesn't change semantics, but we really don't want to perform it
111      // anyway; it's likely to pessimize code generation, and some tools
112      // (like the Darwin linker in cases involving CFString) don't expect it.
113      if (GV->isWeakForLinker())
114        continue;
115
116      Constant *Init = GV->getInitializer();
117
118      // Check to see if the initializer is already known.
119      GlobalVariable *&Slot = CMap[Init];
120
121      // If this is the first constant we find or if the old one is local,
122      // replace with the current one. If the current is externally visible
123      // it cannot be replace, but can be the canonical constant we merge with.
124      if (!Slot || IsBetterCanonical(*GV, *Slot))
125        Slot = GV;
126    }
127
128    // Second: identify all globals that can be merged together, filling in
129    // the Replacements vector.  We cannot do the replacement in this pass
130    // because doing so may cause initializers of other globals to be rewritten,
131    // invalidating the Constant* pointers in CMap.
132    for (Module::global_iterator GVI = M.global_begin(), E = M.global_end();
133         GVI != E; ) {
134      GlobalVariable *GV = &*GVI++;
135
136      // Only process constants with initializers in the default address space.
137      if (!GV->isConstant() || !GV->hasDefinitiveInitializer() ||
138          GV->getType()->getAddressSpace() != 0 || GV->hasSection() ||
139          // Don't touch values marked with attribute(used).
140          UsedGlobals.count(GV))
141        continue;
142
143      // We can only replace constant with local linkage.
144      if (!GV->hasLocalLinkage())
145        continue;
146
147      Constant *Init = GV->getInitializer();
148
149      // Check to see if the initializer is already known.
150      GlobalVariable *Slot = CMap[Init];
151
152      if (!Slot || Slot == GV)
153        continue;
154
155      if (!Slot->hasGlobalUnnamedAddr() && !GV->hasGlobalUnnamedAddr())
156        continue;
157
158      if (!GV->hasGlobalUnnamedAddr())
159        Slot->setUnnamedAddr(GlobalValue::UnnamedAddr::None);
160
161      // Make all uses of the duplicate constant use the canonical version.
162      Replacements.push_back(std::make_pair(GV, Slot));
163    }
164
165    if (Replacements.empty())
166      return MadeChange;
167    CMap.clear();
168
169    // Now that we have figured out which replacements must be made, do them all
170    // now.  This avoid invalidating the pointers in CMap, which are unneeded
171    // now.
172    for (unsigned i = 0, e = Replacements.size(); i != e; ++i) {
173      // Bump the alignment if necessary.
174      if (Replacements[i].first->getAlignment() ||
175          Replacements[i].second->getAlignment()) {
176        Replacements[i].second->setAlignment(
177            std::max(getAlignment(Replacements[i].first),
178                     getAlignment(Replacements[i].second)));
179      }
180
181      // Eliminate any uses of the dead global.
182      Replacements[i].first->replaceAllUsesWith(Replacements[i].second);
183
184      // Delete the global value from the module.
185      assert(Replacements[i].first->hasLocalLinkage() &&
186             "Refusing to delete an externally visible global variable.");
187      Replacements[i].first->eraseFromParent();
188    }
189
190    NumMerged += Replacements.size();
191    Replacements.clear();
192  }
193}
194
195PreservedAnalyses ConstantMergePass::run(Module &M, ModuleAnalysisManager &) {
196  if (!mergeConstants(M))
197    return PreservedAnalyses::all();
198  return PreservedAnalyses::none();
199}
200
201namespace {
202struct ConstantMergeLegacyPass : public ModulePass {
203  static char ID; // Pass identification, replacement for typeid
204  ConstantMergeLegacyPass() : ModulePass(ID) {
205    initializeConstantMergeLegacyPassPass(*PassRegistry::getPassRegistry());
206  }
207
208  // For this pass, process all of the globals in the module, eliminating
209  // duplicate constants.
210  bool runOnModule(Module &M) {
211    if (skipModule(M))
212      return false;
213    return mergeConstants(M);
214  }
215};
216}
217
218char ConstantMergeLegacyPass::ID = 0;
219INITIALIZE_PASS(ConstantMergeLegacyPass, "constmerge",
220                "Merge Duplicate Global Constants", false, false)
221
222ModulePass *llvm::createConstantMergePass() {
223  return new ConstantMergeLegacyPass();
224}
225