1 //===-- SpecialCaseList.cpp - special case list for sanitizers ------------===//
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 is a utility class for instrumentation passes (like AddressSanitizer
11 // or ThreadSanitizer) to avoid instrumenting some functions or global
12 // variables, or to instrument some functions or global variables in a specific
13 // way, based on a user-supplied list.
14 //
15 //===----------------------------------------------------------------------===//
16 
17 #include "llvm/Support/SpecialCaseList.h"
18 #include "llvm/ADT/STLExtras.h"
19 #include "llvm/ADT/SmallVector.h"
20 #include "llvm/ADT/StringExtras.h"
21 #include "llvm/ADT/StringSet.h"
22 #include "llvm/Support/MemoryBuffer.h"
23 #include "llvm/Support/Regex.h"
24 #include "llvm/Support/raw_ostream.h"
25 #include <string>
26 #include <system_error>
27 #include <utility>
28 
29 namespace llvm {
30 
31 /// Represents a set of regular expressions.  Regular expressions which are
32 /// "literal" (i.e. no regex metacharacters) are stored in Strings, while all
33 /// others are represented as a single pipe-separated regex in RegEx.  The
34 /// reason for doing so is efficiency; StringSet is much faster at matching
35 /// literal strings than Regex.
36 struct SpecialCaseList::Entry {
37   Entry() {}
38   Entry(Entry &&Other)
39       : Strings(std::move(Other.Strings)), RegEx(std::move(Other.RegEx)) {}
40 
41   StringSet<> Strings;
42   std::unique_ptr<Regex> RegEx;
43 
44   bool match(StringRef Query) const {
45     return Strings.count(Query) || (RegEx && RegEx->match(Query));
46   }
47 };
48 
49 SpecialCaseList::SpecialCaseList() : Entries(), Regexps(), IsCompiled(false) {}
50 
51 std::unique_ptr<SpecialCaseList>
52 SpecialCaseList::create(const std::vector<std::string> &Paths,
53                         std::string &Error) {
54   std::unique_ptr<SpecialCaseList> SCL(new SpecialCaseList());
55   for (auto Path : Paths) {
56     ErrorOr<std::unique_ptr<MemoryBuffer>> FileOrErr =
57         MemoryBuffer::getFile(Path);
58     if (std::error_code EC = FileOrErr.getError()) {
59       Error = (Twine("can't open file '") + Path + "': " + EC.message()).str();
60       return nullptr;
61     }
62     std::string ParseError;
63     if (!SCL->parse(FileOrErr.get().get(), ParseError)) {
64       Error = (Twine("error parsing file '") + Path + "': " + ParseError).str();
65       return nullptr;
66     }
67   }
68   SCL->compile();
69   return SCL;
70 }
71 
72 std::unique_ptr<SpecialCaseList> SpecialCaseList::create(const MemoryBuffer *MB,
73                                                          std::string &Error) {
74   std::unique_ptr<SpecialCaseList> SCL(new SpecialCaseList());
75   if (!SCL->parse(MB, Error))
76     return nullptr;
77   SCL->compile();
78   return SCL;
79 }
80 
81 std::unique_ptr<SpecialCaseList>
82 SpecialCaseList::createOrDie(const std::vector<std::string> &Paths) {
83   std::string Error;
84   if (auto SCL = create(Paths, Error))
85     return SCL;
86   report_fatal_error(Error);
87 }
88 
89 bool SpecialCaseList::parse(const MemoryBuffer *MB, std::string &Error) {
90   // Iterate through each line in the blacklist file.
91   SmallVector<StringRef, 16> Lines;
92   SplitString(MB->getBuffer(), Lines, "\n\r");
93   int LineNo = 1;
94   for (auto I = Lines.begin(), E = Lines.end(); I != E; ++I, ++LineNo) {
95     // Ignore empty lines and lines starting with "#"
96     if (I->empty() || I->startswith("#"))
97       continue;
98     // Get our prefix and unparsed regexp.
99     std::pair<StringRef, StringRef> SplitLine = I->split(":");
100     StringRef Prefix = SplitLine.first;
101     if (SplitLine.second.empty()) {
102       // Missing ':' in the line.
103       Error = (Twine("malformed line ") + Twine(LineNo) + ": '" +
104                SplitLine.first + "'").str();
105       return false;
106     }
107 
108     std::pair<StringRef, StringRef> SplitRegexp = SplitLine.second.split("=");
109     std::string Regexp = SplitRegexp.first;
110     StringRef Category = SplitRegexp.second;
111 
112     // See if we can store Regexp in Strings.
113     if (Regex::isLiteralERE(Regexp)) {
114       Entries[Prefix][Category].Strings.insert(Regexp);
115       continue;
116     }
117 
118     // Replace * with .*
119     for (size_t pos = 0; (pos = Regexp.find("*", pos)) != std::string::npos;
120          pos += strlen(".*")) {
121       Regexp.replace(pos, strlen("*"), ".*");
122     }
123 
124     // Check that the regexp is valid.
125     Regex CheckRE(Regexp);
126     std::string REError;
127     if (!CheckRE.isValid(REError)) {
128       Error = (Twine("malformed regex in line ") + Twine(LineNo) + ": '" +
129                SplitLine.second + "': " + REError).str();
130       return false;
131     }
132 
133     // Add this regexp into the proper group by its prefix.
134     if (!Regexps[Prefix][Category].empty())
135       Regexps[Prefix][Category] += "|";
136     Regexps[Prefix][Category] += "^" + Regexp + "$";
137   }
138   return true;
139 }
140 
141 void SpecialCaseList::compile() {
142   assert(!IsCompiled && "compile() should only be called once");
143   // Iterate through each of the prefixes, and create Regexs for them.
144   for (StringMap<StringMap<std::string>>::const_iterator I = Regexps.begin(),
145                                                          E = Regexps.end();
146        I != E; ++I) {
147     for (StringMap<std::string>::const_iterator II = I->second.begin(),
148                                                 IE = I->second.end();
149          II != IE; ++II) {
150       Entries[I->getKey()][II->getKey()].RegEx.reset(new Regex(II->getValue()));
151     }
152   }
153   Regexps.clear();
154   IsCompiled = true;
155 }
156 
157 SpecialCaseList::~SpecialCaseList() {}
158 
159 bool SpecialCaseList::inSection(StringRef Section, StringRef Query,
160                                 StringRef Category) const {
161   assert(IsCompiled && "SpecialCaseList::compile() was not called!");
162   StringMap<StringMap<Entry> >::const_iterator I = Entries.find(Section);
163   if (I == Entries.end()) return false;
164   StringMap<Entry>::const_iterator II = I->second.find(Category);
165   if (II == I->second.end()) return false;
166 
167   return II->getValue().match(Query);
168 }
169 
170 }  // namespace llvm
171