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