DWARFUnit.cpp revision 6948897e478cbd66626159776a8017b3c18579b9 
1//===-- DWARFUnit.cpp -----------------------------------------------------===//
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#include "llvm/DebugInfo/DWARF/DWARFUnit.h"
11#include "llvm/DebugInfo/DWARF/DWARFContext.h"
12#include "llvm/DebugInfo/DWARF/DWARFFormValue.h"
13#include "llvm/Support/Dwarf.h"
14#include "llvm/Support/Path.h"
15#include <cstdio>
16
17using namespace llvm;
18using namespace dwarf;
19
20void DWARFUnitSectionBase::parse(DWARFContext &C, const DWARFSection &Section) {
21  parseImpl(C, Section, C.getDebugAbbrev(), C.getRangeSection(),
22            C.getStringSection(), StringRef(), C.getAddrSection(),
23            C.isLittleEndian());
24}
25
26void DWARFUnitSectionBase::parseDWO(DWARFContext &C,
27                                    const DWARFSection &DWOSection) {
28  parseImpl(C, DWOSection, C.getDebugAbbrevDWO(), C.getRangeDWOSection(),
29            C.getStringDWOSection(), C.getStringOffsetDWOSection(),
30            C.getAddrSection(), C.isLittleEndian());
31}
32
33DWARFUnit::DWARFUnit(DWARFContext &DC, const DWARFSection &Section,
34                     const DWARFDebugAbbrev *DA, StringRef RS, StringRef SS,
35                     StringRef SOS, StringRef AOS, bool LE,
36                     const DWARFUnitSectionBase &UnitSection)
37    : Context(DC), InfoSection(Section), Abbrev(DA), RangeSection(RS),
38      StringSection(SS), StringOffsetSection(SOS), AddrOffsetSection(AOS),
39      isLittleEndian(LE), UnitSection(UnitSection) {
40  clear();
41}
42
43DWARFUnit::~DWARFUnit() {
44}
45
46bool DWARFUnit::getAddrOffsetSectionItem(uint32_t Index,
47                                                uint64_t &Result) const {
48  uint32_t Offset = AddrOffsetSectionBase + Index * AddrSize;
49  if (AddrOffsetSection.size() < Offset + AddrSize)
50    return false;
51  DataExtractor DA(AddrOffsetSection, isLittleEndian, AddrSize);
52  Result = DA.getAddress(&Offset);
53  return true;
54}
55
56bool DWARFUnit::getStringOffsetSectionItem(uint32_t Index,
57                                                  uint32_t &Result) const {
58  // FIXME: string offset section entries are 8-byte for DWARF64.
59  const uint32_t ItemSize = 4;
60  uint32_t Offset = Index * ItemSize;
61  if (StringOffsetSection.size() < Offset + ItemSize)
62    return false;
63  DataExtractor DA(StringOffsetSection, isLittleEndian, 0);
64  Result = DA.getU32(&Offset);
65  return true;
66}
67
68bool DWARFUnit::extractImpl(DataExtractor debug_info, uint32_t *offset_ptr) {
69  Length = debug_info.getU32(offset_ptr);
70  Version = debug_info.getU16(offset_ptr);
71  uint64_t AbbrOffset = debug_info.getU32(offset_ptr);
72  AddrSize = debug_info.getU8(offset_ptr);
73
74  bool LengthOK = debug_info.isValidOffset(getNextUnitOffset() - 1);
75  bool VersionOK = DWARFContext::isSupportedVersion(Version);
76  bool AddrSizeOK = AddrSize == 4 || AddrSize == 8;
77
78  if (!LengthOK || !VersionOK || !AddrSizeOK)
79    return false;
80
81  Abbrevs = Abbrev->getAbbreviationDeclarationSet(AbbrOffset);
82  return Abbrevs != nullptr;
83}
84
85bool DWARFUnit::extract(DataExtractor debug_info, uint32_t *offset_ptr) {
86  clear();
87
88  Offset = *offset_ptr;
89
90  if (debug_info.isValidOffset(*offset_ptr)) {
91    if (extractImpl(debug_info, offset_ptr))
92      return true;
93
94    // reset the offset to where we tried to parse from if anything went wrong
95    *offset_ptr = Offset;
96  }
97
98  return false;
99}
100
101bool DWARFUnit::extractRangeList(uint32_t RangeListOffset,
102                                        DWARFDebugRangeList &RangeList) const {
103  // Require that compile unit is extracted.
104  assert(DieArray.size() > 0);
105  DataExtractor RangesData(RangeSection, isLittleEndian, AddrSize);
106  uint32_t ActualRangeListOffset = RangeSectionBase + RangeListOffset;
107  return RangeList.extract(RangesData, &ActualRangeListOffset);
108}
109
110void DWARFUnit::clear() {
111  Offset = 0;
112  Length = 0;
113  Version = 0;
114  Abbrevs = nullptr;
115  AddrSize = 0;
116  BaseAddr = 0;
117  RangeSectionBase = 0;
118  AddrOffsetSectionBase = 0;
119  clearDIEs(false);
120  DWO.reset();
121}
122
123const char *DWARFUnit::getCompilationDir() {
124  extractDIEsIfNeeded(true);
125  if (DieArray.empty())
126    return nullptr;
127  return DieArray[0].getAttributeValueAsString(this, DW_AT_comp_dir, nullptr);
128}
129
130uint64_t DWARFUnit::getDWOId() {
131  extractDIEsIfNeeded(true);
132  const uint64_t FailValue = -1ULL;
133  if (DieArray.empty())
134    return FailValue;
135  return DieArray[0]
136      .getAttributeValueAsUnsignedConstant(this, DW_AT_GNU_dwo_id, FailValue);
137}
138
139void DWARFUnit::setDIERelations() {
140  if (DieArray.size() <= 1)
141    return;
142
143  std::vector<DWARFDebugInfoEntryMinimal *> ParentChain;
144  DWARFDebugInfoEntryMinimal *SiblingChain = nullptr;
145  for (auto &DIE : DieArray) {
146    if (SiblingChain) {
147      SiblingChain->setSibling(&DIE);
148    }
149    if (const DWARFAbbreviationDeclaration *AbbrDecl =
150            DIE.getAbbreviationDeclarationPtr()) {
151      // Normal DIE.
152      if (AbbrDecl->hasChildren()) {
153        ParentChain.push_back(&DIE);
154        SiblingChain = nullptr;
155      } else {
156        SiblingChain = &DIE;
157      }
158    } else {
159      // NULL entry terminates the sibling chain.
160      SiblingChain = ParentChain.back();
161      ParentChain.pop_back();
162    }
163  }
164  assert(SiblingChain == nullptr || SiblingChain == &DieArray[0]);
165  assert(ParentChain.empty());
166}
167
168void DWARFUnit::extractDIEsToVector(
169    bool AppendCUDie, bool AppendNonCUDies,
170    std::vector<DWARFDebugInfoEntryMinimal> &Dies) const {
171  if (!AppendCUDie && !AppendNonCUDies)
172    return;
173
174  // Set the offset to that of the first DIE and calculate the start of the
175  // next compilation unit header.
176  uint32_t DIEOffset = Offset + getHeaderSize();
177  uint32_t NextCUOffset = getNextUnitOffset();
178  DWARFDebugInfoEntryMinimal DIE;
179  uint32_t Depth = 0;
180  bool IsCUDie = true;
181
182  while (DIEOffset < NextCUOffset && DIE.extractFast(this, &DIEOffset)) {
183    if (IsCUDie) {
184      if (AppendCUDie)
185        Dies.push_back(DIE);
186      if (!AppendNonCUDies)
187        break;
188      // The average bytes per DIE entry has been seen to be
189      // around 14-20 so let's pre-reserve the needed memory for
190      // our DIE entries accordingly.
191      Dies.reserve(Dies.size() + getDebugInfoSize() / 14);
192      IsCUDie = false;
193    } else {
194      Dies.push_back(DIE);
195    }
196
197    if (const DWARFAbbreviationDeclaration *AbbrDecl =
198            DIE.getAbbreviationDeclarationPtr()) {
199      // Normal DIE
200      if (AbbrDecl->hasChildren())
201        ++Depth;
202    } else {
203      // NULL DIE.
204      if (Depth > 0)
205        --Depth;
206      if (Depth == 0)
207        break;  // We are done with this compile unit!
208    }
209  }
210
211  // Give a little bit of info if we encounter corrupt DWARF (our offset
212  // should always terminate at or before the start of the next compilation
213  // unit header).
214  if (DIEOffset > NextCUOffset)
215    fprintf(stderr, "warning: DWARF compile unit extends beyond its "
216                    "bounds cu 0x%8.8x at 0x%8.8x'\n", getOffset(), DIEOffset);
217}
218
219size_t DWARFUnit::extractDIEsIfNeeded(bool CUDieOnly) {
220  if ((CUDieOnly && DieArray.size() > 0) ||
221      DieArray.size() > 1)
222    return 0; // Already parsed.
223
224  bool HasCUDie = DieArray.size() > 0;
225  extractDIEsToVector(!HasCUDie, !CUDieOnly, DieArray);
226
227  if (DieArray.empty())
228    return 0;
229
230  // If CU DIE was just parsed, copy several attribute values from it.
231  if (!HasCUDie) {
232    uint64_t BaseAddr =
233        DieArray[0].getAttributeValueAsAddress(this, DW_AT_low_pc, -1ULL);
234    if (BaseAddr == -1ULL)
235      BaseAddr = DieArray[0].getAttributeValueAsAddress(this, DW_AT_entry_pc, 0);
236    setBaseAddress(BaseAddr);
237    AddrOffsetSectionBase = DieArray[0].getAttributeValueAsSectionOffset(
238        this, DW_AT_GNU_addr_base, 0);
239    RangeSectionBase = DieArray[0].getAttributeValueAsSectionOffset(
240        this, DW_AT_ranges_base, 0);
241    // Don't fall back to DW_AT_GNU_ranges_base: it should be ignored for
242    // skeleton CU DIE, so that DWARF users not aware of it are not broken.
243  }
244
245  setDIERelations();
246  return DieArray.size();
247}
248
249DWARFUnit::DWOHolder::DWOHolder(StringRef DWOPath)
250    : DWOFile(), DWOContext(), DWOU(nullptr) {
251  auto Obj = object::ObjectFile::createObjectFile(DWOPath);
252  if (!Obj)
253    return;
254  DWOFile = std::move(Obj.get());
255  DWOContext.reset(
256      cast<DWARFContext>(new DWARFContextInMemory(*DWOFile.getBinary())));
257  if (DWOContext->getNumDWOCompileUnits() > 0)
258    DWOU = DWOContext->getDWOCompileUnitAtIndex(0);
259}
260
261bool DWARFUnit::parseDWO() {
262  if (DWO.get())
263    return false;
264  extractDIEsIfNeeded(true);
265  if (DieArray.empty())
266    return false;
267  const char *DWOFileName =
268      DieArray[0].getAttributeValueAsString(this, DW_AT_GNU_dwo_name, nullptr);
269  if (!DWOFileName)
270    return false;
271  const char *CompilationDir =
272      DieArray[0].getAttributeValueAsString(this, DW_AT_comp_dir, nullptr);
273  SmallString<16> AbsolutePath;
274  if (sys::path::is_relative(DWOFileName) && CompilationDir != nullptr) {
275    sys::path::append(AbsolutePath, CompilationDir);
276  }
277  sys::path::append(AbsolutePath, DWOFileName);
278  DWO = llvm::make_unique<DWOHolder>(AbsolutePath);
279  DWARFUnit *DWOCU = DWO->getUnit();
280  // Verify that compile unit in .dwo file is valid.
281  if (!DWOCU || DWOCU->getDWOId() != getDWOId()) {
282    DWO.reset();
283    return false;
284  }
285  // Share .debug_addr and .debug_ranges section with compile unit in .dwo
286  DWOCU->setAddrOffsetSection(AddrOffsetSection, AddrOffsetSectionBase);
287  uint32_t DWORangesBase = DieArray[0].getRangesBaseAttribute(this, 0);
288  DWOCU->setRangesSection(RangeSection, DWORangesBase);
289  return true;
290}
291
292void DWARFUnit::clearDIEs(bool KeepCUDie) {
293  if (DieArray.size() > (unsigned)KeepCUDie) {
294    // std::vectors never get any smaller when resized to a smaller size,
295    // or when clear() or erase() are called, the size will report that it
296    // is smaller, but the memory allocated remains intact (call capacity()
297    // to see this). So we need to create a temporary vector and swap the
298    // contents which will cause just the internal pointers to be swapped
299    // so that when temporary vector goes out of scope, it will destroy the
300    // contents.
301    std::vector<DWARFDebugInfoEntryMinimal> TmpArray;
302    DieArray.swap(TmpArray);
303    // Save at least the compile unit DIE
304    if (KeepCUDie)
305      DieArray.push_back(TmpArray.front());
306  }
307}
308
309void DWARFUnit::collectAddressRanges(DWARFAddressRangesVector &CURanges) {
310  const auto *U = getUnitDIE();
311  if (U == nullptr)
312    return;
313  // First, check if unit DIE describes address ranges for the whole unit.
314  const auto &CUDIERanges = U->getAddressRanges(this);
315  if (!CUDIERanges.empty()) {
316    CURanges.insert(CURanges.end(), CUDIERanges.begin(), CUDIERanges.end());
317    return;
318  }
319
320  // This function is usually called if there in no .debug_aranges section
321  // in order to produce a compile unit level set of address ranges that
322  // is accurate. If the DIEs weren't parsed, then we don't want all dies for
323  // all compile units to stay loaded when they weren't needed. So we can end
324  // up parsing the DWARF and then throwing them all away to keep memory usage
325  // down.
326  const bool ClearDIEs = extractDIEsIfNeeded(false) > 1;
327  DieArray[0].collectChildrenAddressRanges(this, CURanges);
328
329  // Collect address ranges from DIEs in .dwo if necessary.
330  bool DWOCreated = parseDWO();
331  if (DWO.get())
332    DWO->getUnit()->collectAddressRanges(CURanges);
333  if (DWOCreated)
334    DWO.reset();
335
336  // Keep memory down by clearing DIEs if this generate function
337  // caused them to be parsed.
338  if (ClearDIEs)
339    clearDIEs(true);
340}
341
342const DWARFDebugInfoEntryMinimal *
343DWARFUnit::getSubprogramForAddress(uint64_t Address) {
344  extractDIEsIfNeeded(false);
345  for (const DWARFDebugInfoEntryMinimal &DIE : DieArray) {
346    if (DIE.isSubprogramDIE() &&
347        DIE.addressRangeContainsAddress(this, Address)) {
348      return &DIE;
349    }
350  }
351  return nullptr;
352}
353
354DWARFDebugInfoEntryInlinedChain
355DWARFUnit::getInlinedChainForAddress(uint64_t Address) {
356  // First, find a subprogram that contains the given address (the root
357  // of inlined chain).
358  const DWARFUnit *ChainCU = nullptr;
359  const DWARFDebugInfoEntryMinimal *SubprogramDIE =
360      getSubprogramForAddress(Address);
361  if (SubprogramDIE) {
362    ChainCU = this;
363  } else {
364    // Try to look for subprogram DIEs in the DWO file.
365    parseDWO();
366    if (DWO.get()) {
367      SubprogramDIE = DWO->getUnit()->getSubprogramForAddress(Address);
368      if (SubprogramDIE)
369        ChainCU = DWO->getUnit();
370    }
371  }
372
373  // Get inlined chain rooted at this subprogram DIE.
374  if (!SubprogramDIE)
375    return DWARFDebugInfoEntryInlinedChain();
376  return SubprogramDIE->getInlinedChainForAddress(ChainCU, Address);
377}
378