1//===-- SpecialCaseList.cpp - special case list for sanitizers ------------===//
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 is a utility class for instrumentation passes (like AddressSanitizer
10// or ThreadSanitizer) to avoid instrumenting some functions or global
11// variables, or to instrument some functions or global variables in a specific
12// way, based on a user-supplied list.
13//
14//===----------------------------------------------------------------------===//
15
16#include "llvm/Support/SpecialCaseList.h"
17#include "llvm/Support/LineIterator.h"
18#include "llvm/Support/MemoryBuffer.h"
19#include "llvm/Support/VirtualFileSystem.h"
20#include <stdio.h>
21#include <string>
22#include <system_error>
23#include <utility>
24
25namespace llvm {
26
27Error SpecialCaseList::Matcher::insert(StringRef Pattern, unsigned LineNumber,
28 bool UseGlobs) {
29 if (Pattern.empty())
30 return createStringError(EC: errc::invalid_argument,
31 S: Twine("Supplied ") +
32 (UseGlobs ? "glob" : "regex") + " was blank");
33
34 if (!UseGlobs) {
35 // Replace * with .*
36 auto Regexp = Pattern.str();
37 for (size_t pos = 0; (pos = Regexp.find(c: '*', pos: pos)) != std::string::npos;
38 pos += strlen(s: ".*")) {
39 Regexp.replace(pos: pos, n1: strlen(s: "*"), s: ".*");
40 }
41
42 Regexp = (Twine("^(") + StringRef(Regexp) + ")$").str();
43
44 // Check that the regexp is valid.
45 Regex CheckRE(Regexp);
46 std::string REError;
47 if (!CheckRE.isValid(Error&: REError))
48 return createStringError(EC: errc::invalid_argument, S: REError);
49
50 RegExes.emplace_back(args: std::make_pair(
51 x: std::make_unique<Regex>(args: std::move(CheckRE)), y&: LineNumber));
52
53 return Error::success();
54 }
55
56 auto [It, DidEmplace] = Globs.try_emplace(Key: Pattern);
57 if (DidEmplace) {
58 // We must be sure to use the string in the map rather than the provided
59 // reference which could be destroyed before match() is called
60 Pattern = It->getKey();
61 auto &Pair = It->getValue();
62 if (auto Err = GlobPattern::create(Pat: Pattern, /*MaxSubPatterns=*/1024)
63 .moveInto(Value&: Pair.first))
64 return Err;
65 Pair.second = LineNumber;
66 }
67 return Error::success();
68}
69
70unsigned SpecialCaseList::Matcher::match(StringRef Query) const {
71 for (const auto &[Pattern, Pair] : Globs)
72 if (Pair.first.match(S: Query))
73 return Pair.second;
74 for (const auto &[Regex, LineNumber] : RegExes)
75 if (Regex->match(String: Query))
76 return LineNumber;
77 return 0;
78}
79
80// TODO: Refactor this to return Expected<...>
81std::unique_ptr<SpecialCaseList>
82SpecialCaseList::create(const std::vector<std::string> &Paths,
83 llvm::vfs::FileSystem &FS, std::string &Error) {
84 std::unique_ptr<SpecialCaseList> SCL(new SpecialCaseList());
85 if (SCL->createInternal(Paths, VFS&: FS, Error))
86 return SCL;
87 return nullptr;
88}
89
90std::unique_ptr<SpecialCaseList> SpecialCaseList::create(const MemoryBuffer *MB,
91 std::string &Error) {
92 std::unique_ptr<SpecialCaseList> SCL(new SpecialCaseList());
93 if (SCL->createInternal(MB, Error))
94 return SCL;
95 return nullptr;
96}
97
98std::unique_ptr<SpecialCaseList>
99SpecialCaseList::createOrDie(const std::vector<std::string> &Paths,
100 llvm::vfs::FileSystem &FS) {
101 std::string Error;
102 if (auto SCL = create(Paths, FS, Error))
103 return SCL;
104 report_fatal_error(reason: Twine(Error));
105}
106
107bool SpecialCaseList::createInternal(const std::vector<std::string> &Paths,
108 vfs::FileSystem &VFS, std::string &Error) {
109 for (const auto &Path : Paths) {
110 ErrorOr<std::unique_ptr<MemoryBuffer>> FileOrErr =
111 VFS.getBufferForFile(Name: Path);
112 if (std::error_code EC = FileOrErr.getError()) {
113 Error = (Twine("can't open file '") + Path + "': " + EC.message()).str();
114 return false;
115 }
116 std::string ParseError;
117 if (!parse(MB: FileOrErr.get().get(), Error&: ParseError)) {
118 Error = (Twine("error parsing file '") + Path + "': " + ParseError).str();
119 return false;
120 }
121 }
122 return true;
123}
124
125bool SpecialCaseList::createInternal(const MemoryBuffer *MB,
126 std::string &Error) {
127 if (!parse(MB, Error))
128 return false;
129 return true;
130}
131
132Expected<SpecialCaseList::Section *>
133SpecialCaseList::addSection(StringRef SectionStr, unsigned LineNo,
134 bool UseGlobs) {
135 auto [It, DidEmplace] = Sections.try_emplace(Key: SectionStr);
136 auto &Section = It->getValue();
137 if (DidEmplace)
138 if (auto Err = Section.SectionMatcher->insert(Pattern: SectionStr, LineNumber: LineNo, UseGlobs))
139 return createStringError(EC: errc::invalid_argument,
140 S: "malformed section at line " + Twine(LineNo) +
141 ": '" + SectionStr +
142 "': " + toString(E: std::move(Err)));
143 return &Section;
144}
145
146bool SpecialCaseList::parse(const MemoryBuffer *MB, std::string &Error) {
147 Section *CurrentSection;
148 if (auto Err = addSection(SectionStr: "*", LineNo: 1).moveInto(Value&: CurrentSection)) {
149 Error = toString(E: std::move(Err));
150 return false;
151 }
152
153 // In https://reviews.llvm.org/D154014 we added glob support and planned to
154 // remove regex support in patterns. We temporarily support the original
155 // behavior using regexes if "#!special-case-list-v1" is the first line of the
156 // file. For more details, see
157 // https://discourse.llvm.org/t/use-glob-instead-of-regex-for-specialcaselists/71666
158 bool UseGlobs = !MB->getBuffer().starts_with(Prefix: "#!special-case-list-v1\n");
159
160 for (line_iterator LineIt(*MB, /*SkipBlanks=*/true, /*CommentMarker=*/'#');
161 !LineIt.is_at_eof(); LineIt++) {
162 unsigned LineNo = LineIt.line_number();
163 StringRef Line = LineIt->trim();
164 if (Line.empty())
165 continue;
166
167 // Save section names
168 if (Line.starts_with(Prefix: "[")) {
169 if (!Line.ends_with(Suffix: "]")) {
170 Error =
171 ("malformed section header on line " + Twine(LineNo) + ": " + Line)
172 .str();
173 return false;
174 }
175
176 if (auto Err = addSection(SectionStr: Line.drop_front().drop_back(), LineNo, UseGlobs)
177 .moveInto(Value&: CurrentSection)) {
178 Error = toString(E: std::move(Err));
179 return false;
180 }
181 continue;
182 }
183
184 // Get our prefix and unparsed glob.
185 auto [Prefix, Postfix] = Line.split(Separator: ":");
186 if (Postfix.empty()) {
187 // Missing ':' in the line.
188 Error = ("malformed line " + Twine(LineNo) + ": '" + Line + "'").str();
189 return false;
190 }
191
192 auto [Pattern, Category] = Postfix.split(Separator: "=");
193 auto &Entry = CurrentSection->Entries[Prefix][Category];
194 if (auto Err = Entry.insert(Pattern, LineNumber: LineNo, UseGlobs)) {
195 Error =
196 (Twine("malformed ") + (UseGlobs ? "glob" : "regex") + " in line " +
197 Twine(LineNo) + ": '" + Pattern + "': " + toString(E: std::move(Err)))
198 .str();
199 return false;
200 }
201 }
202 return true;
203}
204
205SpecialCaseList::~SpecialCaseList() = default;
206
207bool SpecialCaseList::inSection(StringRef Section, StringRef Prefix,
208 StringRef Query, StringRef Category) const {
209 return inSectionBlame(Section, Prefix, Query, Category);
210}
211
212unsigned SpecialCaseList::inSectionBlame(StringRef Section, StringRef Prefix,
213 StringRef Query,
214 StringRef Category) const {
215 for (const auto &It : Sections) {
216 const auto &S = It.getValue();
217 if (S.SectionMatcher->match(Query: Section)) {
218 unsigned Blame = inSectionBlame(Entries: S.Entries, Prefix, Query, Category);
219 if (Blame)
220 return Blame;
221 }
222 }
223 return 0;
224}
225
226unsigned SpecialCaseList::inSectionBlame(const SectionEntries &Entries,
227 StringRef Prefix, StringRef Query,
228 StringRef Category) const {
229 SectionEntries::const_iterator I = Entries.find(Key: Prefix);
230 if (I == Entries.end()) return 0;
231 StringMap<Matcher>::const_iterator II = I->second.find(Key: Category);
232 if (II == I->second.end()) return 0;
233
234 return II->getValue().match(Query);
235}
236
237} // namespace llvm
238