1 //===-- llvm-symbolizer.cpp - Simple addr2line-like symbolizer ------------===//
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 utility works much like "addr2line". It is able of transforming
10 // tuples (module name, module offset) to code locations (function name,
11 // file, line number, column number). It is targeted for compiler-rt tools
12 // (especially AddressSanitizer and ThreadSanitizer) that can use it
13 // to symbolize stack traces in their error reports.
14 //
15 //===----------------------------------------------------------------------===//
16 
17 #include "llvm/ADT/StringRef.h"
18 #include "llvm/DebugInfo/Symbolize/DIPrinter.h"
19 #include "llvm/DebugInfo/Symbolize/Symbolize.h"
20 #include "llvm/Support/COM.h"
21 #include "llvm/Support/CommandLine.h"
22 #include "llvm/Support/Debug.h"
23 #include "llvm/Support/FileSystem.h"
24 #include "llvm/Support/InitLLVM.h"
25 #include "llvm/Support/Path.h"
26 #include "llvm/Support/raw_ostream.h"
27 #include <cstdio>
28 #include <cstring>
29 #include <string>
30 
31 using namespace llvm;
32 using namespace symbolize;
33 
34 static cl::opt<bool>
35 ClUseSymbolTable("use-symbol-table", cl::init(true),
36                  cl::desc("Prefer names in symbol table to names "
37                           "in debug info"));
38 
39 static cl::opt<FunctionNameKind> ClPrintFunctions(
40     "functions", cl::init(FunctionNameKind::LinkageName),
41     cl::desc("Print function name for a given address"), cl::ValueOptional,
42     cl::values(clEnumValN(FunctionNameKind::None, "none", "omit function name"),
43                clEnumValN(FunctionNameKind::ShortName, "short",
44                           "print short function name"),
45                clEnumValN(FunctionNameKind::LinkageName, "linkage",
46                           "print function linkage name (default)"),
47                // Sentinel value for unspecified value.
48                clEnumValN(FunctionNameKind::LinkageName, "", "")));
49 static cl::alias ClPrintFunctionsShort("f", cl::desc("Alias for -functions"),
50                                        cl::NotHidden, cl::Grouping,
51                                        cl::aliasopt(ClPrintFunctions));
52 
53 static cl::opt<bool>
54     ClUseRelativeAddress("relative-address", cl::init(false),
55                          cl::desc("Interpret addresses as relative addresses"),
56                          cl::ReallyHidden);
57 
58 static cl::opt<bool>
59     ClPrintInlining("inlining", cl::init(true),
60                     cl::desc("Print all inlined frames for a given address"));
61 static cl::alias
62     ClPrintInliningAliasI("i", cl::desc("Alias for -inlining"),
63                           cl::NotHidden, cl::aliasopt(ClPrintInlining),
64                           cl::Grouping);
65 static cl::alias
66     ClPrintInliningAliasInlines("inlines", cl::desc("Alias for -inlining"),
67                                 cl::NotHidden, cl::aliasopt(ClPrintInlining));
68 
69 // -basenames, -s
70 static cl::opt<bool> ClBasenames("basenames", cl::init(false),
71                                  cl::desc("Strip directory names from paths"));
72 static cl::alias ClBasenamesShort("s", cl::desc("Alias for -basenames"),
73                                   cl::NotHidden, cl::aliasopt(ClBasenames));
74 
75 // -demangle, -C, -no-demangle
76 static cl::opt<bool>
77 ClDemangle("demangle", cl::init(true), cl::desc("Demangle function names"));
78 static cl::alias
79 ClDemangleShort("C", cl::desc("Alias for -demangle"),
80                 cl::NotHidden, cl::aliasopt(ClDemangle), cl::Grouping);
81 static cl::opt<bool>
82 ClNoDemangle("no-demangle", cl::init(false),
83              cl::desc("Don't demangle function names"));
84 
85 static cl::opt<std::string> ClDefaultArch("default-arch", cl::init(""),
86                                           cl::desc("Default architecture "
87                                                    "(for multi-arch objects)"));
88 
89 // -obj, -exe, -e
90 static cl::opt<std::string>
91 ClBinaryName("obj", cl::init(""),
92              cl::desc("Path to object file to be symbolized (if not provided, "
93                       "object file should be specified for each input line)"));
94 static cl::alias
95 ClBinaryNameAliasExe("exe", cl::desc("Alias for -obj"),
96                      cl::NotHidden, cl::aliasopt(ClBinaryName));
97 static cl::alias
98 ClBinaryNameAliasE("e", cl::desc("Alias for -obj"),
99                    cl::NotHidden, cl::aliasopt(ClBinaryName));
100 
101 
102 static cl::opt<std::string>
103     ClDwpName("dwp", cl::init(""),
104               cl::desc("Path to DWP file to be use for any split CUs"));
105 
106 static cl::list<std::string>
107 ClDsymHint("dsym-hint", cl::ZeroOrMore,
108            cl::desc("Path to .dSYM bundles to search for debug info for the "
109                     "object files"));
110 
111 // -print-address, -addresses, -a
112 static cl::opt<bool>
113 ClPrintAddress("print-address", cl::init(false),
114                cl::desc("Show address before line information"));
115 static cl::alias
116 ClPrintAddressAliasAddresses("addresses", cl::desc("Alias for -print-address"),
117                              cl::NotHidden, cl::aliasopt(ClPrintAddress));
118 static cl::alias
119 ClPrintAddressAliasA("a", cl::desc("Alias for -print-address"),
120                      cl::NotHidden, cl::aliasopt(ClPrintAddress), cl::Grouping);
121 
122 // -pretty-print, -p
123 static cl::opt<bool>
124     ClPrettyPrint("pretty-print", cl::init(false),
125                   cl::desc("Make the output more human friendly"));
126 static cl::alias ClPrettyPrintShort("p", cl::desc("Alias for -pretty-print"),
127                                     cl::NotHidden,
128                                     cl::aliasopt(ClPrettyPrint), cl::Grouping);
129 
130 static cl::opt<int> ClPrintSourceContextLines(
131     "print-source-context-lines", cl::init(0),
132     cl::desc("Print N number of source file context"));
133 
134 static cl::opt<bool> ClVerbose("verbose", cl::init(false),
135                                cl::desc("Print verbose line info"));
136 
137 // -adjust-vma
138 static cl::opt<unsigned long long>
139     ClAdjustVMA("adjust-vma", cl::init(0), cl::value_desc("offset"),
140                 cl::desc("Add specified offset to object file addresses"));
141 
142 static cl::list<std::string> ClInputAddresses(cl::Positional,
143                                               cl::desc("<input addresses>..."),
144                                               cl::ZeroOrMore);
145 
146 static cl::opt<std::string>
147     ClFallbackDebugPath("fallback-debug-path", cl::init(""),
148                         cl::desc("Fallback path for debug binaries."));
149 
150 template<typename T>
151 static bool error(Expected<T> &ResOrErr) {
152   if (ResOrErr)
153     return false;
154   logAllUnhandledErrors(ResOrErr.takeError(), errs(),
155                         "LLVMSymbolizer: error reading file: ");
156   return true;
157 }
158 
159 static bool parseCommand(StringRef InputString, bool &IsData,
160                          std::string &ModuleName, uint64_t &ModuleOffset) {
161   const char kDelimiters[] = " \n\r";
162   ModuleName = "";
163   if (InputString.consume_front("CODE ")) {
164     IsData = false;
165   } else if (InputString.consume_front("DATA ")) {
166     IsData = true;
167   } else {
168     // If no cmd, assume it's CODE.
169     IsData = false;
170   }
171   const char *pos = InputString.data();
172   // Skip delimiters and parse input filename (if needed).
173   if (ClBinaryName.empty()) {
174     pos += strspn(pos, kDelimiters);
175     if (*pos == '"' || *pos == '\'') {
176       char quote = *pos;
177       pos++;
178       const char *end = strchr(pos, quote);
179       if (!end)
180         return false;
181       ModuleName = std::string(pos, end - pos);
182       pos = end + 1;
183     } else {
184       int name_length = strcspn(pos, kDelimiters);
185       ModuleName = std::string(pos, name_length);
186       pos += name_length;
187     }
188   } else {
189     ModuleName = ClBinaryName;
190   }
191   // Skip delimiters and parse module offset.
192   pos += strspn(pos, kDelimiters);
193   int offset_length = strcspn(pos, kDelimiters);
194   return !StringRef(pos, offset_length).getAsInteger(0, ModuleOffset);
195 }
196 
197 static void symbolizeInput(StringRef InputString, LLVMSymbolizer &Symbolizer,
198                            DIPrinter &Printer) {
199   bool IsData = false;
200   std::string ModuleName;
201   uint64_t Offset = 0;
202   if (!parseCommand(StringRef(InputString), IsData, ModuleName, Offset)) {
203     outs() << InputString;
204     return;
205   }
206 
207   if (ClPrintAddress) {
208     outs() << "0x";
209     outs().write_hex(Offset);
210     StringRef Delimiter = ClPrettyPrint ? ": " : "\n";
211     outs() << Delimiter;
212   }
213   Offset -= ClAdjustVMA;
214   if (IsData) {
215     auto ResOrErr = Symbolizer.symbolizeData(
216         ModuleName, {Offset, object::SectionedAddress::UndefSection});
217     Printer << (error(ResOrErr) ? DIGlobal() : ResOrErr.get());
218   } else if (ClPrintInlining) {
219     auto ResOrErr = Symbolizer.symbolizeInlinedCode(
220         ModuleName, {Offset, object::SectionedAddress::UndefSection},
221         ClDwpName);
222     Printer << (error(ResOrErr) ? DIInliningInfo() : ResOrErr.get());
223   } else {
224     auto ResOrErr = Symbolizer.symbolizeCode(
225         ModuleName, {Offset, object::SectionedAddress::UndefSection},
226         ClDwpName);
227     Printer << (error(ResOrErr) ? DILineInfo() : ResOrErr.get());
228   }
229   outs() << "\n";
230   outs().flush();
231 }
232 
233 int main(int argc, char **argv) {
234   InitLLVM X(argc, argv);
235 
236   llvm::sys::InitializeCOMRAII COM(llvm::sys::COMThreadingMode::MultiThreaded);
237   cl::ParseCommandLineOptions(argc, argv, "llvm-symbolizer\n");
238 
239   // If both --demangle and --no-demangle are specified then pick the last one.
240   if (ClNoDemangle.getPosition() > ClDemangle.getPosition())
241     ClDemangle = !ClNoDemangle;
242 
243   LLVMSymbolizer::Options Opts(ClPrintFunctions, ClUseSymbolTable, ClDemangle,
244                                ClUseRelativeAddress, ClDefaultArch,
245                                ClFallbackDebugPath);
246 
247   for (const auto &hint : ClDsymHint) {
248     if (sys::path::extension(hint) == ".dSYM") {
249       Opts.DsymHints.push_back(hint);
250     } else {
251       errs() << "Warning: invalid dSYM hint: \"" << hint <<
252                 "\" (must have the '.dSYM' extension).\n";
253     }
254   }
255   LLVMSymbolizer Symbolizer(Opts);
256 
257   DIPrinter Printer(outs(), ClPrintFunctions != FunctionNameKind::None,
258                     ClPrettyPrint, ClPrintSourceContextLines, ClVerbose,
259                     ClBasenames);
260 
261   if (ClInputAddresses.empty()) {
262     const int kMaxInputStringLength = 1024;
263     char InputString[kMaxInputStringLength];
264 
265     while (fgets(InputString, sizeof(InputString), stdin))
266       symbolizeInput(InputString, Symbolizer, Printer);
267   } else {
268     for (StringRef Address : ClInputAddresses)
269       symbolizeInput(Address, Symbolizer, Printer);
270   }
271 
272   return 0;
273 }
274