lib/MC/StringTableBuilder.cpp

   1 //===- StringTableBuilder.cpp - String table building utility -------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #include "llvm/MC/StringTableBuilder.h"
  10 #include "llvm/ADT/CachedHashString.h"
  11 #include "llvm/ADT/SmallString.h"
  12 #include "llvm/ADT/StringRef.h"
  13 #include "llvm/BinaryFormat/COFF.h"
  14 #include "llvm/Support/Endian.h"
  15 #include "llvm/Support/MathExtras.h"
  16 #include "llvm/Support/raw_ostream.h"
  17 #include <cassert>
  18 #include <cstddef>
  19 #include <cstdint>
  20 #include <cstring>
  21 #include <utility>
  22 #include <vector>
  23
  24 using namespace llvm;
  25
  26 StringTableBuilder::~StringTableBuilder() = default;
  27
  28 void StringTableBuilder::initSize() {
  29   // Account for leading bytes in table so that offsets returned from add are
  30   // correct.
  31   switch (K) {
  32   case RAW:
  33   case DWARF:
  34     Size = 0;
  35     break;
  36   case MachO:
  37   case ELF:
  38     // Start the table with a NUL byte.
  39     Size = 1;
  40     break;
  41   case WinCOFF:
  42     // Make room to write the table size later.
  43     Size = 4;
  44     break;
  45   }
  46 }
  47
  48 StringTableBuilder::StringTableBuilder(Kind K, unsigned Alignment)
  49     : K(K), Alignment(Alignment) {
  50   initSize();
  51 }
  52
  53 void StringTableBuilder::write(raw_ostream &OS) const {
  54   assert(isFinalized());
  55   SmallString<0> Data;
  56   Data.resize(getSize());
  57   write((uint8_t *)Data.data());
  58   OS << Data;
  59 }
  60
  61 using StringPair = std::pair<CachedHashStringRef, size_t>;
  62
  63 void StringTableBuilder::write(uint8_t *Buf) const {
  64   assert(isFinalized());
  65   for (const StringPair &P : StringIndexMap) {
  66     StringRef Data = P.first.val();
  67     if (!Data.empty())
  68       memcpy(Buf + P.second, Data.data(), Data.size());
  69   }
  70   if (K != WinCOFF)
  71     return;
  72   support::endian::write32le(Buf, Size);
  73 }
  74
  75 // Returns the character at Pos from end of a string.
  76 static int charTailAt(StringPair *P, size_t Pos) {
  77   StringRef S = P->first.val();
  78   if (Pos >= S.size())
  79     return -1;
  80   return (unsigned char)S[S.size() - Pos - 1];
  81 }
  82
  83 // Three-way radix quicksort. This is much faster than std::sort with strcmp
  84 // because it does not compare characters that we already know the same.
  85 static void multikeySort(MutableArrayRef<StringPair *> Vec, int Pos) {
  86 tailcall:
  87   if (Vec.size() <= 1)
  88     return;
  89
  90   // Partition items so that items in [0, I) are greater than the pivot,
  91   // [I, J) are the same as the pivot, and [J, Vec.size()) are less than
  92   // the pivot.
  93   int Pivot = charTailAt(Vec[0], Pos);
  94   size_t I = 0;
  95   size_t J = Vec.size();
  96   for (size_t K = 1; K < J;) {
  97     int C = charTailAt(Vec[K], Pos);
  98     if (C > Pivot)
  99       std::swap(Vec[I++], Vec[K++]);
 100     else if (C < Pivot)
 101       std::swap(Vec[--J], Vec[K]);
 102     else
 103       K++;
 104   }
 105
 106   multikeySort(Vec.slice(0, I), Pos);
 107   multikeySort(Vec.slice(J), Pos);
 108
 109   // multikeySort(Vec.slice(I, J - I), Pos + 1), but with
 110   // tail call optimization.
 111   if (Pivot != -1) {
 112     Vec = Vec.slice(I, J - I);
 113     ++Pos;
 114     goto tailcall;
 115   }
 116 }
 117
 118 void StringTableBuilder::finalize() {
 119   assert(K != DWARF);
 120   finalizeStringTable(/*Optimize=*/true);
 121 }
 122
 123 void StringTableBuilder::finalizeInOrder() {
 124   finalizeStringTable(/*Optimize=*/false);
 125 }
 126
 127 void StringTableBuilder::finalizeStringTable(bool Optimize) {
 128   Finalized = true;
 129
 130   if (Optimize) {
 131     std::vector<StringPair *> Strings;
 132     Strings.reserve(StringIndexMap.size());
 133     for (StringPair &P : StringIndexMap)
 134       Strings.push_back(&P);
 135
 136     multikeySort(Strings, 0);
 137     initSize();
 138
 139     StringRef Previous;
 140     for (StringPair *P : Strings) {
 141       StringRef S = P->first.val();
 142       if (Previous.endswith(S)) {
 143         size_t Pos = Size - S.size() - (K != RAW);
 144         if (!(Pos & (Alignment - 1))) {
 145           P->second = Pos;
 146           continue;
 147         }
 148       }
 149
 150       Size = alignTo(Size, Alignment);
 151       P->second = Size;
 152
 153       Size += S.size();
 154       if (K != RAW)
 155         ++Size;
 156       Previous = S;
 157     }
 158   }
 159
 160   if (K == MachO)
 161     Size = alignTo(Size, 4); // Pad to multiple of 4.
 162
 163   // The first byte in an ELF string table must be null, according to the ELF
 164   // specification. In 'initSize()' we reserved the first byte to hold null for
 165   // this purpose and here we actually add the string to allow 'getOffset()' to
 166   // be called on an empty string.
 167   if (K == ELF)
 168     StringIndexMap[CachedHashStringRef("")] = 0;
 169 }
 170
 171 void StringTableBuilder::clear() {
 172   Finalized = false;
 173   StringIndexMap.clear();
 174 }
 175
 176 size_t StringTableBuilder::getOffset(CachedHashStringRef S) const {
 177   assert(isFinalized());
 178   auto I = StringIndexMap.find(S);
 179   assert(I != StringIndexMap.end() && "String is not in table!");
 180   return I->second;
 181 }
 182
 183 size_t StringTableBuilder::add(CachedHashStringRef S) {
 184   if (K == WinCOFF)
 185     assert(S.size() > COFF::NameSize && "Short string in COFF string table!");
 186
 187   assert(!isFinalized());
 188   auto P = StringIndexMap.insert(std::make_pair(S, 0));
 189   if (P.second) {
 190     size_t Start = alignTo(Size, Alignment);
 191     P.first->second = Start;
 192     Size = Start + S.size() + (K != RAW);
 193   }
 194   return P.first->second;
 195 }