|
173
|
1 //===- DwarfTransformer.cpp -----------------------------------------------===//
|
|
|
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 #include <thread>
|
|
|
10 #include <unordered_set>
|
|
|
11
|
|
|
12 #include "llvm/DebugInfo/DIContext.h"
|
|
236
|
13 #include "llvm/DebugInfo/DWARF/DWARFCompileUnit.h"
|
|
173
|
14 #include "llvm/DebugInfo/DWARF/DWARFContext.h"
|
|
|
15 #include "llvm/Support/Error.h"
|
|
|
16 #include "llvm/Support/ThreadPool.h"
|
|
|
17 #include "llvm/Support/raw_ostream.h"
|
|
|
18
|
|
|
19 #include "llvm/DebugInfo/GSYM/DwarfTransformer.h"
|
|
|
20 #include "llvm/DebugInfo/GSYM/FunctionInfo.h"
|
|
|
21 #include "llvm/DebugInfo/GSYM/GsymCreator.h"
|
|
|
22 #include "llvm/DebugInfo/GSYM/GsymReader.h"
|
|
|
23 #include "llvm/DebugInfo/GSYM/InlineInfo.h"
|
|
|
24
|
|
|
25 using namespace llvm;
|
|
|
26 using namespace gsym;
|
|
|
27
|
|
|
28 struct llvm::gsym::CUInfo {
|
|
|
29 const DWARFDebugLine::LineTable *LineTable;
|
|
|
30 const char *CompDir;
|
|
|
31 std::vector<uint32_t> FileCache;
|
|
|
32 uint64_t Language = 0;
|
|
|
33 uint8_t AddrSize = 0;
|
|
|
34
|
|
|
35 CUInfo(DWARFContext &DICtx, DWARFCompileUnit *CU) {
|
|
|
36 LineTable = DICtx.getLineTableForUnit(CU);
|
|
|
37 CompDir = CU->getCompilationDir();
|
|
|
38 FileCache.clear();
|
|
|
39 if (LineTable)
|
|
|
40 FileCache.assign(LineTable->Prologue.FileNames.size() + 1, UINT32_MAX);
|
|
|
41 DWARFDie Die = CU->getUnitDIE();
|
|
|
42 Language = dwarf::toUnsigned(Die.find(dwarf::DW_AT_language), 0);
|
|
|
43 AddrSize = CU->getAddressByteSize();
|
|
|
44 }
|
|
|
45
|
|
|
46 /// Return true if Addr is the highest address for a given compile unit. The
|
|
|
47 /// highest address is encoded as -1, of all ones in the address. These high
|
|
|
48 /// addresses are used by some linkers to indicate that a function has been
|
|
|
49 /// dead stripped or didn't end up in the linked executable.
|
|
|
50 bool isHighestAddress(uint64_t Addr) const {
|
|
|
51 if (AddrSize == 4)
|
|
|
52 return Addr == UINT32_MAX;
|
|
|
53 else if (AddrSize == 8)
|
|
|
54 return Addr == UINT64_MAX;
|
|
|
55 return false;
|
|
|
56 }
|
|
|
57
|
|
|
58 /// Convert a DWARF compile unit file index into a GSYM global file index.
|
|
|
59 ///
|
|
|
60 /// Each compile unit in DWARF has its own file table in the line table
|
|
|
61 /// prologue. GSYM has a single large file table that applies to all files
|
|
|
62 /// from all of the info in a GSYM file. This function converts between the
|
|
|
63 /// two and caches and DWARF CU file index that has already been converted so
|
|
|
64 /// the first client that asks for a compile unit file index will end up
|
|
|
65 /// doing the conversion, and subsequent clients will get the cached GSYM
|
|
|
66 /// index.
|
|
|
67 uint32_t DWARFToGSYMFileIndex(GsymCreator &Gsym, uint32_t DwarfFileIdx) {
|
|
|
68 if (!LineTable)
|
|
|
69 return 0;
|
|
|
70 assert(DwarfFileIdx < FileCache.size());
|
|
|
71 uint32_t &GsymFileIdx = FileCache[DwarfFileIdx];
|
|
|
72 if (GsymFileIdx != UINT32_MAX)
|
|
|
73 return GsymFileIdx;
|
|
|
74 std::string File;
|
|
|
75 if (LineTable->getFileNameByIndex(
|
|
|
76 DwarfFileIdx, CompDir,
|
|
|
77 DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath, File))
|
|
|
78 GsymFileIdx = Gsym.insertFile(File);
|
|
|
79 else
|
|
|
80 GsymFileIdx = 0;
|
|
|
81 return GsymFileIdx;
|
|
|
82 }
|
|
|
83 };
|
|
|
84
|
|
|
85
|
|
|
86 static DWARFDie GetParentDeclContextDIE(DWARFDie &Die) {
|
|
|
87 if (DWARFDie SpecDie =
|
|
|
88 Die.getAttributeValueAsReferencedDie(dwarf::DW_AT_specification)) {
|
|
|
89 if (DWARFDie SpecParent = GetParentDeclContextDIE(SpecDie))
|
|
|
90 return SpecParent;
|
|
|
91 }
|
|
|
92 if (DWARFDie AbstDie =
|
|
|
93 Die.getAttributeValueAsReferencedDie(dwarf::DW_AT_abstract_origin)) {
|
|
|
94 if (DWARFDie AbstParent = GetParentDeclContextDIE(AbstDie))
|
|
|
95 return AbstParent;
|
|
|
96 }
|
|
|
97
|
|
|
98 // We never want to follow parent for inlined subroutine - that would
|
|
|
99 // give us information about where the function is inlined, not what
|
|
|
100 // function is inlined
|
|
|
101 if (Die.getTag() == dwarf::DW_TAG_inlined_subroutine)
|
|
|
102 return DWARFDie();
|
|
|
103
|
|
|
104 DWARFDie ParentDie = Die.getParent();
|
|
|
105 if (!ParentDie)
|
|
|
106 return DWARFDie();
|
|
|
107
|
|
|
108 switch (ParentDie.getTag()) {
|
|
|
109 case dwarf::DW_TAG_namespace:
|
|
|
110 case dwarf::DW_TAG_structure_type:
|
|
|
111 case dwarf::DW_TAG_union_type:
|
|
|
112 case dwarf::DW_TAG_class_type:
|
|
|
113 case dwarf::DW_TAG_subprogram:
|
|
|
114 return ParentDie; // Found parent decl context DIE
|
|
|
115 case dwarf::DW_TAG_lexical_block:
|
|
|
116 return GetParentDeclContextDIE(ParentDie);
|
|
|
117 default:
|
|
|
118 break;
|
|
|
119 }
|
|
|
120
|
|
|
121 return DWARFDie();
|
|
|
122 }
|
|
|
123
|
|
|
124 /// Get the GsymCreator string table offset for the qualified name for the
|
|
|
125 /// DIE passed in. This function will avoid making copies of any strings in
|
|
|
126 /// the GsymCreator when possible. We don't need to copy a string when the
|
|
|
127 /// string comes from our .debug_str section or is an inlined string in the
|
|
|
128 /// .debug_info. If we create a qualified name string in this function by
|
|
|
129 /// combining multiple strings in the DWARF string table or info, we will make
|
|
|
130 /// a copy of the string when we add it to the string table.
|
|
|
131 static Optional<uint32_t> getQualifiedNameIndex(DWARFDie &Die,
|
|
|
132 uint64_t Language,
|
|
|
133 GsymCreator &Gsym) {
|
|
|
134 // If the dwarf has mangled name, use mangled name
|
|
|
135 if (auto LinkageName =
|
|
|
136 dwarf::toString(Die.findRecursively({dwarf::DW_AT_MIPS_linkage_name,
|
|
|
137 dwarf::DW_AT_linkage_name}),
|
|
|
138 nullptr))
|
|
|
139 return Gsym.insertString(LinkageName, /* Copy */ false);
|
|
|
140
|
|
|
141 StringRef ShortName(Die.getName(DINameKind::ShortName));
|
|
|
142 if (ShortName.empty())
|
|
|
143 return llvm::None;
|
|
|
144
|
|
|
145 // For C++ and ObjC, prepend names of all parent declaration contexts
|
|
|
146 if (!(Language == dwarf::DW_LANG_C_plus_plus ||
|
|
|
147 Language == dwarf::DW_LANG_C_plus_plus_03 ||
|
|
|
148 Language == dwarf::DW_LANG_C_plus_plus_11 ||
|
|
|
149 Language == dwarf::DW_LANG_C_plus_plus_14 ||
|
|
|
150 Language == dwarf::DW_LANG_ObjC_plus_plus ||
|
|
|
151 // This should not be needed for C, but we see C++ code marked as C
|
|
|
152 // in some binaries. This should hurt, so let's do it for C as well
|
|
|
153 Language == dwarf::DW_LANG_C))
|
|
|
154 return Gsym.insertString(ShortName, /* Copy */ false);
|
|
|
155
|
|
|
156 // Some GCC optimizations create functions with names ending with .isra.<num>
|
|
|
157 // or .part.<num> and those names are just DW_AT_name, not DW_AT_linkage_name
|
|
|
158 // If it looks like it could be the case, don't add any prefix
|
|
|
159 if (ShortName.startswith("_Z") &&
|
|
|
160 (ShortName.contains(".isra.") || ShortName.contains(".part.")))
|
|
|
161 return Gsym.insertString(ShortName, /* Copy */ false);
|
|
|
162
|
|
|
163 DWARFDie ParentDeclCtxDie = GetParentDeclContextDIE(Die);
|
|
|
164 if (ParentDeclCtxDie) {
|
|
|
165 std::string Name = ShortName.str();
|
|
|
166 while (ParentDeclCtxDie) {
|
|
|
167 StringRef ParentName(ParentDeclCtxDie.getName(DINameKind::ShortName));
|
|
|
168 if (!ParentName.empty()) {
|
|
|
169 // "lambda" names are wrapped in < >. Replace with { }
|
|
|
170 // to be consistent with demangled names and not to confuse with
|
|
|
171 // templates
|
|
|
172 if (ParentName.front() == '<' && ParentName.back() == '>')
|
|
|
173 Name = "{" + ParentName.substr(1, ParentName.size() - 2).str() + "}" +
|
|
|
174 "::" + Name;
|
|
|
175 else
|
|
|
176 Name = ParentName.str() + "::" + Name;
|
|
|
177 }
|
|
|
178 ParentDeclCtxDie = GetParentDeclContextDIE(ParentDeclCtxDie);
|
|
|
179 }
|
|
|
180 // Copy the name since we created a new name in a std::string.
|
|
|
181 return Gsym.insertString(Name, /* Copy */ true);
|
|
|
182 }
|
|
|
183 // Don't copy the name since it exists in the DWARF object file.
|
|
|
184 return Gsym.insertString(ShortName, /* Copy */ false);
|
|
|
185 }
|
|
|
186
|
|
|
187 static bool hasInlineInfo(DWARFDie Die, uint32_t Depth) {
|
|
|
188 bool CheckChildren = true;
|
|
|
189 switch (Die.getTag()) {
|
|
|
190 case dwarf::DW_TAG_subprogram:
|
|
|
191 // Don't look into functions within functions.
|
|
|
192 CheckChildren = Depth == 0;
|
|
|
193 break;
|
|
|
194 case dwarf::DW_TAG_inlined_subroutine:
|
|
|
195 return true;
|
|
|
196 default:
|
|
|
197 break;
|
|
|
198 }
|
|
|
199 if (!CheckChildren)
|
|
|
200 return false;
|
|
|
201 for (DWARFDie ChildDie : Die.children()) {
|
|
|
202 if (hasInlineInfo(ChildDie, Depth + 1))
|
|
|
203 return true;
|
|
|
204 }
|
|
|
205 return false;
|
|
|
206 }
|
|
|
207
|
|
|
208 static void parseInlineInfo(GsymCreator &Gsym, CUInfo &CUI, DWARFDie Die,
|
|
|
209 uint32_t Depth, FunctionInfo &FI,
|
|
|
210 InlineInfo &parent) {
|
|
|
211 if (!hasInlineInfo(Die, Depth))
|
|
|
212 return;
|
|
|
213
|
|
|
214 dwarf::Tag Tag = Die.getTag();
|
|
|
215 if (Tag == dwarf::DW_TAG_inlined_subroutine) {
|
|
|
216 // create new InlineInfo and append to parent.children
|
|
|
217 InlineInfo II;
|
|
|
218 DWARFAddressRange FuncRange =
|
|
|
219 DWARFAddressRange(FI.startAddress(), FI.endAddress());
|
|
|
220 Expected<DWARFAddressRangesVector> RangesOrError = Die.getAddressRanges();
|
|
|
221 if (RangesOrError) {
|
|
|
222 for (const DWARFAddressRange &Range : RangesOrError.get()) {
|
|
|
223 // Check that the inlined function is within the range of the function
|
|
|
224 // info, it might not be in case of split functions
|
|
|
225 if (FuncRange.LowPC <= Range.LowPC && Range.HighPC <= FuncRange.HighPC)
|
|
|
226 II.Ranges.insert(AddressRange(Range.LowPC, Range.HighPC));
|
|
|
227 }
|
|
|
228 }
|
|
|
229 if (II.Ranges.empty())
|
|
|
230 return;
|
|
|
231
|
|
|
232 if (auto NameIndex = getQualifiedNameIndex(Die, CUI.Language, Gsym))
|
|
|
233 II.Name = *NameIndex;
|
|
|
234 II.CallFile = CUI.DWARFToGSYMFileIndex(
|
|
|
235 Gsym, dwarf::toUnsigned(Die.find(dwarf::DW_AT_call_file), 0));
|
|
|
236 II.CallLine = dwarf::toUnsigned(Die.find(dwarf::DW_AT_call_line), 0);
|
|
|
237 // parse all children and append to parent
|
|
|
238 for (DWARFDie ChildDie : Die.children())
|
|
|
239 parseInlineInfo(Gsym, CUI, ChildDie, Depth + 1, FI, II);
|
|
|
240 parent.Children.emplace_back(std::move(II));
|
|
|
241 return;
|
|
|
242 }
|
|
|
243 if (Tag == dwarf::DW_TAG_subprogram || Tag == dwarf::DW_TAG_lexical_block) {
|
|
|
244 // skip this Die and just recurse down
|
|
|
245 for (DWARFDie ChildDie : Die.children())
|
|
|
246 parseInlineInfo(Gsym, CUI, ChildDie, Depth + 1, FI, parent);
|
|
|
247 }
|
|
|
248 }
|
|
|
249
|
|
|
250 static void convertFunctionLineTable(raw_ostream &Log, CUInfo &CUI,
|
|
|
251 DWARFDie Die, GsymCreator &Gsym,
|
|
|
252 FunctionInfo &FI) {
|
|
|
253 std::vector<uint32_t> RowVector;
|
|
|
254 const uint64_t StartAddress = FI.startAddress();
|
|
|
255 const uint64_t EndAddress = FI.endAddress();
|
|
|
256 const uint64_t RangeSize = EndAddress - StartAddress;
|
|
|
257 const object::SectionedAddress SecAddress{
|
|
|
258 StartAddress, object::SectionedAddress::UndefSection};
|
|
|
259
|
|
|
260
|
|
|
261 if (!CUI.LineTable->lookupAddressRange(SecAddress, RangeSize, RowVector)) {
|
|
|
262 // If we have a DW_TAG_subprogram but no line entries, fall back to using
|
|
|
263 // the DW_AT_decl_file an d DW_AT_decl_line if we have both attributes.
|
|
236
|
264 std::string FilePath = Die.getDeclFile(
|
|
|
265 DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath);
|
|
|
266 if (FilePath.empty())
|
|
|
267 return;
|
|
|
268 if (auto Line =
|
|
|
269 dwarf::toUnsigned(Die.findRecursively({dwarf::DW_AT_decl_line}))) {
|
|
|
270 LineEntry LE(StartAddress, Gsym.insertFile(FilePath), *Line);
|
|
|
271 FI.OptLineTable = LineTable();
|
|
|
272 FI.OptLineTable->push(LE);
|
|
173
|
273 }
|
|
|
274 return;
|
|
|
275 }
|
|
|
276
|
|
|
277 FI.OptLineTable = LineTable();
|
|
|
278 DWARFDebugLine::Row PrevRow;
|
|
|
279 for (uint32_t RowIndex : RowVector) {
|
|
|
280 // Take file number and line/column from the row.
|
|
|
281 const DWARFDebugLine::Row &Row = CUI.LineTable->Rows[RowIndex];
|
|
|
282 const uint32_t FileIdx = CUI.DWARFToGSYMFileIndex(Gsym, Row.File);
|
|
|
283 uint64_t RowAddress = Row.Address.Address;
|
|
|
284 // Watch out for a RowAddress that is in the middle of a line table entry
|
|
|
285 // in the DWARF. If we pass an address in between two line table entries
|
|
|
286 // we will get a RowIndex for the previous valid line table row which won't
|
|
|
287 // be contained in our function. This is usually a bug in the DWARF due to
|
|
|
288 // linker problems or LTO or other DWARF re-linking so it is worth emitting
|
|
|
289 // an error, but not worth stopping the creation of the GSYM.
|
|
|
290 if (!FI.Range.contains(RowAddress)) {
|
|
236
|
291 if (RowAddress < FI.Range.start()) {
|
|
173
|
292 Log << "error: DIE has a start address whose LowPC is between the "
|
|
|
293 "line table Row[" << RowIndex << "] with address "
|
|
|
294 << HEX64(RowAddress) << " and the next one.\n";
|
|
|
295 Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
|
|
236
|
296 RowAddress = FI.Range.start();
|
|
173
|
297 } else {
|
|
|
298 continue;
|
|
|
299 }
|
|
|
300 }
|
|
|
301
|
|
|
302 LineEntry LE(RowAddress, FileIdx, Row.Line);
|
|
|
303 if (RowIndex != RowVector[0] && Row.Address < PrevRow.Address) {
|
|
|
304 // We have seen full duplicate line tables for functions in some
|
|
|
305 // DWARF files. Watch for those here by checking the the last
|
|
|
306 // row was the function's end address (HighPC) and that the
|
|
|
307 // current line table entry's address is the same as the first
|
|
|
308 // line entry we already have in our "function_info.Lines". If
|
|
|
309 // so break out after printing a warning.
|
|
|
310 auto FirstLE = FI.OptLineTable->first();
|
|
|
311 if (FirstLE && *FirstLE == LE) {
|
|
221
|
312 if (!Gsym.isQuiet()) {
|
|
|
313 Log << "warning: duplicate line table detected for DIE:\n";
|
|
|
314 Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
|
|
|
315 }
|
|
173
|
316 } else {
|
|
|
317 // Print out (ignore if os == nulls as this is expensive)
|
|
|
318 Log << "error: line table has addresses that do not "
|
|
|
319 << "monotonically increase:\n";
|
|
|
320 for (uint32_t RowIndex2 : RowVector) {
|
|
|
321 CUI.LineTable->Rows[RowIndex2].dump(Log);
|
|
|
322 }
|
|
|
323 Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
|
|
|
324 }
|
|
|
325 break;
|
|
|
326 }
|
|
|
327
|
|
|
328 // Skip multiple line entries for the same file and line.
|
|
|
329 auto LastLE = FI.OptLineTable->last();
|
|
|
330 if (LastLE && LastLE->File == FileIdx && LastLE->Line == Row.Line)
|
|
|
331 continue;
|
|
|
332 // Only push a row if it isn't an end sequence. End sequence markers are
|
|
|
333 // included for the last address in a function or the last contiguous
|
|
|
334 // address in a sequence.
|
|
|
335 if (Row.EndSequence) {
|
|
|
336 // End sequence means that the next line entry could have a lower address
|
|
|
337 // that the previous entries. So we clear the previous row so we don't
|
|
|
338 // trigger the line table error about address that do not monotonically
|
|
|
339 // increase.
|
|
|
340 PrevRow = DWARFDebugLine::Row();
|
|
|
341 } else {
|
|
|
342 FI.OptLineTable->push(LE);
|
|
|
343 PrevRow = Row;
|
|
|
344 }
|
|
|
345 }
|
|
|
346 // If not line table rows were added, clear the line table so we don't encode
|
|
|
347 // on in the GSYM file.
|
|
|
348 if (FI.OptLineTable->empty())
|
|
|
349 FI.OptLineTable = llvm::None;
|
|
|
350 }
|
|
|
351
|
|
|
352 void DwarfTransformer::handleDie(raw_ostream &OS, CUInfo &CUI, DWARFDie Die) {
|
|
|
353 switch (Die.getTag()) {
|
|
|
354 case dwarf::DW_TAG_subprogram: {
|
|
|
355 Expected<DWARFAddressRangesVector> RangesOrError = Die.getAddressRanges();
|
|
|
356 if (!RangesOrError) {
|
|
|
357 consumeError(RangesOrError.takeError());
|
|
|
358 break;
|
|
|
359 }
|
|
|
360 const DWARFAddressRangesVector &Ranges = RangesOrError.get();
|
|
|
361 if (Ranges.empty())
|
|
|
362 break;
|
|
|
363 auto NameIndex = getQualifiedNameIndex(Die, CUI.Language, Gsym);
|
|
|
364 if (!NameIndex) {
|
|
|
365 OS << "error: function at " << HEX64(Die.getOffset())
|
|
|
366 << " has no name\n ";
|
|
|
367 Die.dump(OS, 0, DIDumpOptions::getForSingleDIE());
|
|
|
368 break;
|
|
|
369 }
|
|
|
370
|
|
|
371 // Create a function_info for each range
|
|
|
372 for (const DWARFAddressRange &Range : Ranges) {
|
|
|
373 // The low PC must be less than the high PC. Many linkers don't remove
|
|
|
374 // DWARF for functions that don't get linked into the final executable.
|
|
|
375 // If both the high and low pc have relocations, linkers will often set
|
|
|
376 // the address values for both to the same value to indicate the function
|
|
|
377 // has been remove. Other linkers have been known to set the one or both
|
|
|
378 // PC values to a UINT32_MAX for 4 byte addresses and UINT64_MAX for 8
|
|
|
379 // byte addresses to indicate the function isn't valid. The check below
|
|
|
380 // tries to watch for these cases and abort if it runs into them.
|
|
|
381 if (Range.LowPC >= Range.HighPC || CUI.isHighestAddress(Range.LowPC))
|
|
|
382 break;
|
|
|
383
|
|
|
384 // Many linkers can't remove DWARF and might set the LowPC to zero. Since
|
|
|
385 // high PC can be an offset from the low PC in more recent DWARF versions
|
|
|
386 // we need to watch for a zero'ed low pc which we do using
|
|
|
387 // ValidTextRanges below.
|
|
|
388 if (!Gsym.IsValidTextAddress(Range.LowPC)) {
|
|
|
389 // We expect zero and -1 to be invalid addresses in DWARF depending
|
|
|
390 // on the linker of the DWARF. This indicates a function was stripped
|
|
|
391 // and the debug info wasn't able to be stripped from the DWARF. If
|
|
|
392 // the LowPC isn't zero or -1, then we should emit an error.
|
|
|
393 if (Range.LowPC != 0) {
|
|
221
|
394 if (!Gsym.isQuiet()) {
|
|
|
395 // Unexpected invalid address, emit a warning
|
|
236
|
396 OS << "warning: DIE has an address range whose start address is "
|
|
|
397 "not in any executable sections ("
|
|
|
398 << *Gsym.GetValidTextRanges()
|
|
|
399 << ") and will not be processed:\n";
|
|
|
400 Die.dump(OS, 0, DIDumpOptions::getForSingleDIE());
|
|
221
|
401 }
|
|
173
|
402 }
|
|
|
403 break;
|
|
|
404 }
|
|
|
405
|
|
|
406 FunctionInfo FI;
|
|
236
|
407 FI.Range = {Range.LowPC, Range.HighPC};
|
|
173
|
408 FI.Name = *NameIndex;
|
|
|
409 if (CUI.LineTable) {
|
|
|
410 convertFunctionLineTable(OS, CUI, Die, Gsym, FI);
|
|
|
411 }
|
|
|
412 if (hasInlineInfo(Die, 0)) {
|
|
|
413 FI.Inline = InlineInfo();
|
|
|
414 FI.Inline->Name = *NameIndex;
|
|
|
415 FI.Inline->Ranges.insert(FI.Range);
|
|
|
416 parseInlineInfo(Gsym, CUI, Die, 0, FI, *FI.Inline);
|
|
|
417 }
|
|
|
418 Gsym.addFunctionInfo(std::move(FI));
|
|
|
419 }
|
|
|
420 } break;
|
|
|
421 default:
|
|
|
422 break;
|
|
|
423 }
|
|
|
424 for (DWARFDie ChildDie : Die.children())
|
|
|
425 handleDie(OS, CUI, ChildDie);
|
|
|
426 }
|
|
|
427
|
|
|
428 Error DwarfTransformer::convert(uint32_t NumThreads) {
|
|
|
429 size_t NumBefore = Gsym.getNumFunctionInfos();
|
|
236
|
430 auto getDie = [&](DWARFUnit &DwarfUnit) -> DWARFDie {
|
|
|
431 DWARFDie ReturnDie = DwarfUnit.getUnitDIE(false);
|
|
|
432 if (llvm::Optional<uint64_t> DWOId = DwarfUnit.getDWOId()) {
|
|
|
433 DWARFUnit *DWOCU = DwarfUnit.getNonSkeletonUnitDIE(false).getDwarfUnit();
|
|
|
434 if (!DWOCU->isDWOUnit()) {
|
|
|
435 std::string DWOName = dwarf::toString(
|
|
|
436 DwarfUnit.getUnitDIE().find(
|
|
|
437 {dwarf::DW_AT_dwo_name, dwarf::DW_AT_GNU_dwo_name}),
|
|
|
438 "");
|
|
|
439 Log << "warning: Unable to retrieve DWO .debug_info section for "
|
|
|
440 << DWOName << "\n";
|
|
|
441 } else {
|
|
|
442 ReturnDie = DWOCU->getUnitDIE(false);
|
|
|
443 }
|
|
|
444 }
|
|
|
445 return ReturnDie;
|
|
|
446 };
|
|
173
|
447 if (NumThreads == 1) {
|
|
|
448 // Parse all DWARF data from this thread, use the same string/file table
|
|
|
449 // for everything
|
|
|
450 for (const auto &CU : DICtx.compile_units()) {
|
|
236
|
451 DWARFDie Die = getDie(*CU);
|
|
173
|
452 CUInfo CUI(DICtx, dyn_cast<DWARFCompileUnit>(CU.get()));
|
|
|
453 handleDie(Log, CUI, Die);
|
|
|
454 }
|
|
|
455 } else {
|
|
|
456 // LLVM Dwarf parser is not thread-safe and we need to parse all DWARF up
|
|
|
457 // front before we start accessing any DIEs since there might be
|
|
|
458 // cross compile unit references in the DWARF. If we don't do this we can
|
|
|
459 // end up crashing.
|
|
|
460
|
|
|
461 // We need to call getAbbreviations sequentially first so that getUnitDIE()
|
|
|
462 // only works with its local data.
|
|
|
463 for (const auto &CU : DICtx.compile_units())
|
|
|
464 CU->getAbbreviations();
|
|
|
465
|
|
|
466 // Now parse all DIEs in case we have cross compile unit references in a
|
|
|
467 // thread pool.
|
|
|
468 ThreadPool pool(hardware_concurrency(NumThreads));
|
|
|
469 for (const auto &CU : DICtx.compile_units())
|
|
|
470 pool.async([&CU]() { CU->getUnitDIE(false /*CUDieOnly*/); });
|
|
|
471 pool.wait();
|
|
|
472
|
|
|
473 // Now convert all DWARF to GSYM in a thread pool.
|
|
|
474 std::mutex LogMutex;
|
|
|
475 for (const auto &CU : DICtx.compile_units()) {
|
|
236
|
476 DWARFDie Die = getDie(*CU);
|
|
173
|
477 if (Die) {
|
|
|
478 CUInfo CUI(DICtx, dyn_cast<DWARFCompileUnit>(CU.get()));
|
|
|
479 pool.async([this, CUI, &LogMutex, Die]() mutable {
|
|
|
480 std::string ThreadLogStorage;
|
|
|
481 raw_string_ostream ThreadOS(ThreadLogStorage);
|
|
|
482 handleDie(ThreadOS, CUI, Die);
|
|
|
483 ThreadOS.flush();
|
|
|
484 if (!ThreadLogStorage.empty()) {
|
|
|
485 // Print ThreadLogStorage lines into an actual stream under a lock
|
|
|
486 std::lock_guard<std::mutex> guard(LogMutex);
|
|
|
487 Log << ThreadLogStorage;
|
|
|
488 }
|
|
|
489 });
|
|
|
490 }
|
|
|
491 }
|
|
|
492 pool.wait();
|
|
|
493 }
|
|
|
494 size_t FunctionsAddedCount = Gsym.getNumFunctionInfos() - NumBefore;
|
|
|
495 Log << "Loaded " << FunctionsAddedCount << " functions from DWARF.\n";
|
|
|
496 return Error::success();
|
|
|
497 }
|
|
|
498
|
|
|
499 llvm::Error DwarfTransformer::verify(StringRef GsymPath) {
|
|
|
500 Log << "Verifying GSYM file \"" << GsymPath << "\":\n";
|
|
|
501
|
|
|
502 auto Gsym = GsymReader::openFile(GsymPath);
|
|
|
503 if (!Gsym)
|
|
|
504 return Gsym.takeError();
|
|
|
505
|
|
|
506 auto NumAddrs = Gsym->getNumAddresses();
|
|
|
507 DILineInfoSpecifier DLIS(
|
|
|
508 DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath,
|
|
|
509 DILineInfoSpecifier::FunctionNameKind::LinkageName);
|
|
|
510 std::string gsymFilename;
|
|
|
511 for (uint32_t I = 0; I < NumAddrs; ++I) {
|
|
|
512 auto FuncAddr = Gsym->getAddress(I);
|
|
|
513 if (!FuncAddr)
|
|
|
514 return createStringError(std::errc::invalid_argument,
|
|
|
515 "failed to extract address[%i]", I);
|
|
|
516
|
|
|
517 auto FI = Gsym->getFunctionInfo(*FuncAddr);
|
|
|
518 if (!FI)
|
|
|
519 return createStringError(std::errc::invalid_argument,
|
|
|
520 "failed to extract function info for address 0x%"
|
|
|
521 PRIu64, *FuncAddr);
|
|
|
522
|
|
|
523 for (auto Addr = *FuncAddr; Addr < *FuncAddr + FI->size(); ++Addr) {
|
|
|
524 const object::SectionedAddress SectAddr{
|
|
|
525 Addr, object::SectionedAddress::UndefSection};
|
|
|
526 auto LR = Gsym->lookup(Addr);
|
|
|
527 if (!LR)
|
|
|
528 return LR.takeError();
|
|
|
529
|
|
|
530 auto DwarfInlineInfos =
|
|
|
531 DICtx.getInliningInfoForAddress(SectAddr, DLIS);
|
|
|
532 uint32_t NumDwarfInlineInfos = DwarfInlineInfos.getNumberOfFrames();
|
|
|
533 if (NumDwarfInlineInfos == 0) {
|
|
|
534 DwarfInlineInfos.addFrame(
|
|
|
535 DICtx.getLineInfoForAddress(SectAddr, DLIS));
|
|
|
536 }
|
|
|
537
|
|
|
538 // Check for 1 entry that has no file and line info
|
|
|
539 if (NumDwarfInlineInfos == 1 &&
|
|
|
540 DwarfInlineInfos.getFrame(0).FileName == "<invalid>") {
|
|
|
541 DwarfInlineInfos = DIInliningInfo();
|
|
|
542 NumDwarfInlineInfos = 0;
|
|
|
543 }
|
|
|
544 if (NumDwarfInlineInfos > 0 &&
|
|
|
545 NumDwarfInlineInfos != LR->Locations.size()) {
|
|
|
546 Log << "error: address " << HEX64(Addr) << " has "
|
|
|
547 << NumDwarfInlineInfos << " DWARF inline frames and GSYM has "
|
|
|
548 << LR->Locations.size() << "\n";
|
|
|
549 Log << " " << NumDwarfInlineInfos << " DWARF frames:\n";
|
|
|
550 for (size_t Idx = 0; Idx < NumDwarfInlineInfos; ++Idx) {
|
|
236
|
551 const auto &dii = DwarfInlineInfos.getFrame(Idx);
|
|
173
|
552 Log << " [" << Idx << "]: " << dii.FunctionName << " @ "
|
|
|
553 << dii.FileName << ':' << dii.Line << '\n';
|
|
|
554 }
|
|
|
555 Log << " " << LR->Locations.size() << " GSYM frames:\n";
|
|
|
556 for (size_t Idx = 0, count = LR->Locations.size();
|
|
|
557 Idx < count; ++Idx) {
|
|
|
558 const auto &gii = LR->Locations[Idx];
|
|
|
559 Log << " [" << Idx << "]: " << gii.Name << " @ " << gii.Dir
|
|
|
560 << '/' << gii.Base << ':' << gii.Line << '\n';
|
|
|
561 }
|
|
|
562 DwarfInlineInfos = DICtx.getInliningInfoForAddress(SectAddr, DLIS);
|
|
|
563 Gsym->dump(Log, *FI);
|
|
|
564 continue;
|
|
|
565 }
|
|
|
566
|
|
|
567 for (size_t Idx = 0, count = LR->Locations.size(); Idx < count;
|
|
|
568 ++Idx) {
|
|
|
569 const auto &gii = LR->Locations[Idx];
|
|
|
570 if (Idx < NumDwarfInlineInfos) {
|
|
236
|
571 const auto &dii = DwarfInlineInfos.getFrame(Idx);
|
|
173
|
572 gsymFilename = LR->getSourceFile(Idx);
|
|
|
573 // Verify function name
|
|
|
574 if (dii.FunctionName.find(gii.Name.str()) != 0)
|
|
|
575 Log << "error: address " << HEX64(Addr) << " DWARF function \""
|
|
|
576 << dii.FunctionName.c_str()
|
|
|
577 << "\" doesn't match GSYM function \"" << gii.Name << "\"\n";
|
|
|
578 // Verify source file path
|
|
|
579 if (dii.FileName != gsymFilename)
|
|
|
580 Log << "error: address " << HEX64(Addr) << " DWARF path \""
|
|
|
581 << dii.FileName.c_str() << "\" doesn't match GSYM path \""
|
|
|
582 << gsymFilename.c_str() << "\"\n";
|
|
|
583 // Verify source file line
|
|
|
584 if (dii.Line != gii.Line)
|
|
|
585 Log << "error: address " << HEX64(Addr) << " DWARF line "
|
|
|
586 << dii.Line << " != GSYM line " << gii.Line << "\n";
|
|
|
587 }
|
|
|
588 }
|
|
|
589 }
|
|
|
590 }
|
|
|
591 return Error::success();
|
|
|
592 }
|
