[sanitizer] Improve FreeBSD ASLR detection
[llvm-project.git] / llvm / lib / TableGen / StringMatcher.cpp
blob7474c5dfe885c5709d884b992ae6d717377fafc9
1 //===- StringMatcher.cpp - Generate a matcher for input strings -----------===//
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 // This file implements the StringMatcher class.
11 //===----------------------------------------------------------------------===//
13 #include "llvm/TableGen/StringMatcher.h"
14 #include "llvm/ADT/StringRef.h"
15 #include "llvm/Support/raw_ostream.h"
16 #include <cassert>
17 #include <map>
18 #include <string>
19 #include <utility>
20 #include <vector>
22 using namespace llvm;
24 /// FindFirstNonCommonLetter - Find the first character in the keys of the
25 /// string pairs that is not shared across the whole set of strings. All
26 /// strings are assumed to have the same length.
27 static unsigned
28 FindFirstNonCommonLetter(const std::vector<const
29 StringMatcher::StringPair*> &Matches) {
30 assert(!Matches.empty());
31 for (unsigned i = 0, e = Matches[0]->first.size(); i != e; ++i) {
32 // Check to see if letter i is the same across the set.
33 char Letter = Matches[0]->first[i];
35 for (const StringMatcher::StringPair *Match : Matches)
36 if (Match->first[i] != Letter)
37 return i;
40 return Matches[0]->first.size();
43 /// EmitStringMatcherForChar - Given a set of strings that are known to be the
44 /// same length and whose characters leading up to CharNo are the same, emit
45 /// code to verify that CharNo and later are the same.
46 ///
47 /// \return - True if control can leave the emitted code fragment.
48 bool StringMatcher::EmitStringMatcherForChar(
49 const std::vector<const StringPair *> &Matches, unsigned CharNo,
50 unsigned IndentCount, bool IgnoreDuplicates) const {
51 assert(!Matches.empty() && "Must have at least one string to match!");
52 std::string Indent(IndentCount * 2 + 4, ' ');
54 // If we have verified that the entire string matches, we're done: output the
55 // matching code.
56 if (CharNo == Matches[0]->first.size()) {
57 if (Matches.size() > 1 && !IgnoreDuplicates)
58 report_fatal_error("Had duplicate keys to match on");
60 // If the to-execute code has \n's in it, indent each subsequent line.
61 StringRef Code = Matches[0]->second;
63 std::pair<StringRef, StringRef> Split = Code.split('\n');
64 OS << Indent << Split.first << "\t // \"" << Matches[0]->first << "\"\n";
66 Code = Split.second;
67 while (!Code.empty()) {
68 Split = Code.split('\n');
69 OS << Indent << Split.first << "\n";
70 Code = Split.second;
72 return false;
75 // Bucket the matches by the character we are comparing.
76 std::map<char, std::vector<const StringPair*>> MatchesByLetter;
78 for (const StringPair *Match : Matches)
79 MatchesByLetter[Match->first[CharNo]].push_back(Match);
81 // If we have exactly one bucket to match, see how many characters are common
82 // across the whole set and match all of them at once.
83 if (MatchesByLetter.size() == 1) {
84 unsigned FirstNonCommonLetter = FindFirstNonCommonLetter(Matches);
85 unsigned NumChars = FirstNonCommonLetter-CharNo;
87 // Emit code to break out if the prefix doesn't match.
88 if (NumChars == 1) {
89 // Do the comparison with if (Str[1] != 'f')
90 // FIXME: Need to escape general characters.
91 OS << Indent << "if (" << StrVariableName << "[" << CharNo << "] != '"
92 << Matches[0]->first[CharNo] << "')\n";
93 OS << Indent << " break;\n";
94 } else {
95 // Do the comparison with if memcmp(Str.data()+1, "foo", 3).
96 // FIXME: Need to escape general strings.
97 OS << Indent << "if (memcmp(" << StrVariableName << ".data()+" << CharNo
98 << ", \"" << Matches[0]->first.substr(CharNo, NumChars) << "\", "
99 << NumChars << ") != 0)\n";
100 OS << Indent << " break;\n";
103 return EmitStringMatcherForChar(Matches, FirstNonCommonLetter, IndentCount,
104 IgnoreDuplicates);
107 // Otherwise, we have multiple possible things, emit a switch on the
108 // character.
109 OS << Indent << "switch (" << StrVariableName << "[" << CharNo << "]) {\n";
110 OS << Indent << "default: break;\n";
112 for (const auto &LI : MatchesByLetter) {
113 // TODO: escape hard stuff (like \n) if we ever care about it.
114 OS << Indent << "case '" << LI.first << "':\t // " << LI.second.size()
115 << " string";
116 if (LI.second.size() != 1)
117 OS << 's';
118 OS << " to match.\n";
119 if (EmitStringMatcherForChar(LI.second, CharNo + 1, IndentCount + 1,
120 IgnoreDuplicates))
121 OS << Indent << " break;\n";
124 OS << Indent << "}\n";
125 return true;
128 /// Emit - Top level entry point.
130 void StringMatcher::Emit(unsigned Indent, bool IgnoreDuplicates) const {
131 // If nothing to match, just fall through.
132 if (Matches.empty()) return;
134 // First level categorization: group strings by length.
135 std::map<unsigned, std::vector<const StringPair*>> MatchesByLength;
137 for (const StringPair &Match : Matches)
138 MatchesByLength[Match.first.size()].push_back(&Match);
140 // Output a switch statement on length and categorize the elements within each
141 // bin.
142 OS.indent(Indent*2+2) << "switch (" << StrVariableName << ".size()) {\n";
143 OS.indent(Indent*2+2) << "default: break;\n";
145 for (const auto &LI : MatchesByLength) {
146 OS.indent(Indent * 2 + 2)
147 << "case " << LI.first << ":\t // " << LI.second.size() << " string"
148 << (LI.second.size() == 1 ? "" : "s") << " to match.\n";
149 if (EmitStringMatcherForChar(LI.second, 0, Indent, IgnoreDuplicates))
150 OS.indent(Indent*2+4) << "break;\n";
153 OS.indent(Indent*2+2) << "}\n";