aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/DebugInfo/GSYM
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/lib/DebugInfo/GSYM')
-rw-r--r--llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp572
-rw-r--r--llvm/lib/DebugInfo/GSYM/FunctionInfo.cpp11
-rw-r--r--llvm/lib/DebugInfo/GSYM/GsymCreator.cpp61
-rw-r--r--llvm/lib/DebugInfo/GSYM/GsymReader.cpp153
-rw-r--r--llvm/lib/DebugInfo/GSYM/InlineInfo.cpp18
-rw-r--r--llvm/lib/DebugInfo/GSYM/LookupResult.cpp31
-rw-r--r--llvm/lib/DebugInfo/GSYM/ObjectFileTransformer.cpp116
-rw-r--r--llvm/lib/DebugInfo/GSYM/Range.cpp10
8 files changed, 928 insertions, 44 deletions
diff --git a/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp b/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp
new file mode 100644
index 000000000000..1e527ab3916e
--- /dev/null
+++ b/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp
@@ -0,0 +1,572 @@
+//===- DwarfTransformer.cpp -----------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include <thread>
+#include <unordered_set>
+
+#include "llvm/DebugInfo/DIContext.h"
+#include "llvm/DebugInfo/DWARF/DWARFContext.h"
+#include "llvm/Support/Error.h"
+#include "llvm/Support/ThreadPool.h"
+#include "llvm/Support/raw_ostream.h"
+
+#include "llvm/DebugInfo/GSYM/DwarfTransformer.h"
+#include "llvm/DebugInfo/GSYM/FunctionInfo.h"
+#include "llvm/DebugInfo/GSYM/GsymCreator.h"
+#include "llvm/DebugInfo/GSYM/GsymReader.h"
+#include "llvm/DebugInfo/GSYM/InlineInfo.h"
+
+using namespace llvm;
+using namespace gsym;
+
+struct llvm::gsym::CUInfo {
+ const DWARFDebugLine::LineTable *LineTable;
+ const char *CompDir;
+ std::vector<uint32_t> FileCache;
+ uint64_t Language = 0;
+ uint8_t AddrSize = 0;
+
+ CUInfo(DWARFContext &DICtx, DWARFCompileUnit *CU) {
+ LineTable = DICtx.getLineTableForUnit(CU);
+ CompDir = CU->getCompilationDir();
+ FileCache.clear();
+ if (LineTable)
+ FileCache.assign(LineTable->Prologue.FileNames.size() + 1, UINT32_MAX);
+ DWARFDie Die = CU->getUnitDIE();
+ Language = dwarf::toUnsigned(Die.find(dwarf::DW_AT_language), 0);
+ AddrSize = CU->getAddressByteSize();
+ }
+
+ /// Return true if Addr is the highest address for a given compile unit. The
+ /// highest address is encoded as -1, of all ones in the address. These high
+ /// addresses are used by some linkers to indicate that a function has been
+ /// dead stripped or didn't end up in the linked executable.
+ bool isHighestAddress(uint64_t Addr) const {
+ if (AddrSize == 4)
+ return Addr == UINT32_MAX;
+ else if (AddrSize == 8)
+ return Addr == UINT64_MAX;
+ return false;
+ }
+
+ /// Convert a DWARF compile unit file index into a GSYM global file index.
+ ///
+ /// Each compile unit in DWARF has its own file table in the line table
+ /// prologue. GSYM has a single large file table that applies to all files
+ /// from all of the info in a GSYM file. This function converts between the
+ /// two and caches and DWARF CU file index that has already been converted so
+ /// the first client that asks for a compile unit file index will end up
+ /// doing the conversion, and subsequent clients will get the cached GSYM
+ /// index.
+ uint32_t DWARFToGSYMFileIndex(GsymCreator &Gsym, uint32_t DwarfFileIdx) {
+ if (!LineTable)
+ return 0;
+ assert(DwarfFileIdx < FileCache.size());
+ uint32_t &GsymFileIdx = FileCache[DwarfFileIdx];
+ if (GsymFileIdx != UINT32_MAX)
+ return GsymFileIdx;
+ std::string File;
+ if (LineTable->getFileNameByIndex(
+ DwarfFileIdx, CompDir,
+ DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath, File))
+ GsymFileIdx = Gsym.insertFile(File);
+ else
+ GsymFileIdx = 0;
+ return GsymFileIdx;
+ }
+};
+
+
+static DWARFDie GetParentDeclContextDIE(DWARFDie &Die) {
+ if (DWARFDie SpecDie =
+ Die.getAttributeValueAsReferencedDie(dwarf::DW_AT_specification)) {
+ if (DWARFDie SpecParent = GetParentDeclContextDIE(SpecDie))
+ return SpecParent;
+ }
+ if (DWARFDie AbstDie =
+ Die.getAttributeValueAsReferencedDie(dwarf::DW_AT_abstract_origin)) {
+ if (DWARFDie AbstParent = GetParentDeclContextDIE(AbstDie))
+ return AbstParent;
+ }
+
+ // We never want to follow parent for inlined subroutine - that would
+ // give us information about where the function is inlined, not what
+ // function is inlined
+ if (Die.getTag() == dwarf::DW_TAG_inlined_subroutine)
+ return DWARFDie();
+
+ DWARFDie ParentDie = Die.getParent();
+ if (!ParentDie)
+ return DWARFDie();
+
+ switch (ParentDie.getTag()) {
+ case dwarf::DW_TAG_namespace:
+ case dwarf::DW_TAG_structure_type:
+ case dwarf::DW_TAG_union_type:
+ case dwarf::DW_TAG_class_type:
+ case dwarf::DW_TAG_subprogram:
+ return ParentDie; // Found parent decl context DIE
+ case dwarf::DW_TAG_lexical_block:
+ return GetParentDeclContextDIE(ParentDie);
+ default:
+ break;
+ }
+
+ return DWARFDie();
+}
+
+/// Get the GsymCreator string table offset for the qualified name for the
+/// DIE passed in. This function will avoid making copies of any strings in
+/// the GsymCreator when possible. We don't need to copy a string when the
+/// string comes from our .debug_str section or is an inlined string in the
+/// .debug_info. If we create a qualified name string in this function by
+/// combining multiple strings in the DWARF string table or info, we will make
+/// a copy of the string when we add it to the string table.
+static Optional<uint32_t> getQualifiedNameIndex(DWARFDie &Die,
+ uint64_t Language,
+ GsymCreator &Gsym) {
+ // If the dwarf has mangled name, use mangled name
+ if (auto LinkageName =
+ dwarf::toString(Die.findRecursively({dwarf::DW_AT_MIPS_linkage_name,
+ dwarf::DW_AT_linkage_name}),
+ nullptr))
+ return Gsym.insertString(LinkageName, /* Copy */ false);
+
+ StringRef ShortName(Die.getName(DINameKind::ShortName));
+ if (ShortName.empty())
+ return llvm::None;
+
+ // For C++ and ObjC, prepend names of all parent declaration contexts
+ if (!(Language == dwarf::DW_LANG_C_plus_plus ||
+ Language == dwarf::DW_LANG_C_plus_plus_03 ||
+ Language == dwarf::DW_LANG_C_plus_plus_11 ||
+ Language == dwarf::DW_LANG_C_plus_plus_14 ||
+ Language == dwarf::DW_LANG_ObjC_plus_plus ||
+ // This should not be needed for C, but we see C++ code marked as C
+ // in some binaries. This should hurt, so let's do it for C as well
+ Language == dwarf::DW_LANG_C))
+ return Gsym.insertString(ShortName, /* Copy */ false);
+
+ // Some GCC optimizations create functions with names ending with .isra.<num>
+ // or .part.<num> and those names are just DW_AT_name, not DW_AT_linkage_name
+ // If it looks like it could be the case, don't add any prefix
+ if (ShortName.startswith("_Z") &&
+ (ShortName.contains(".isra.") || ShortName.contains(".part.")))
+ return Gsym.insertString(ShortName, /* Copy */ false);
+
+ DWARFDie ParentDeclCtxDie = GetParentDeclContextDIE(Die);
+ if (ParentDeclCtxDie) {
+ std::string Name = ShortName.str();
+ while (ParentDeclCtxDie) {
+ StringRef ParentName(ParentDeclCtxDie.getName(DINameKind::ShortName));
+ if (!ParentName.empty()) {
+ // "lambda" names are wrapped in < >. Replace with { }
+ // to be consistent with demangled names and not to confuse with
+ // templates
+ if (ParentName.front() == '<' && ParentName.back() == '>')
+ Name = "{" + ParentName.substr(1, ParentName.size() - 2).str() + "}" +
+ "::" + Name;
+ else
+ Name = ParentName.str() + "::" + Name;
+ }
+ ParentDeclCtxDie = GetParentDeclContextDIE(ParentDeclCtxDie);
+ }
+ // Copy the name since we created a new name in a std::string.
+ return Gsym.insertString(Name, /* Copy */ true);
+ }
+ // Don't copy the name since it exists in the DWARF object file.
+ return Gsym.insertString(ShortName, /* Copy */ false);
+}
+
+static bool hasInlineInfo(DWARFDie Die, uint32_t Depth) {
+ bool CheckChildren = true;
+ switch (Die.getTag()) {
+ case dwarf::DW_TAG_subprogram:
+ // Don't look into functions within functions.
+ CheckChildren = Depth == 0;
+ break;
+ case dwarf::DW_TAG_inlined_subroutine:
+ return true;
+ default:
+ break;
+ }
+ if (!CheckChildren)
+ return false;
+ for (DWARFDie ChildDie : Die.children()) {
+ if (hasInlineInfo(ChildDie, Depth + 1))
+ return true;
+ }
+ return false;
+}
+
+static void parseInlineInfo(GsymCreator &Gsym, CUInfo &CUI, DWARFDie Die,
+ uint32_t Depth, FunctionInfo &FI,
+ InlineInfo &parent) {
+ if (!hasInlineInfo(Die, Depth))
+ return;
+
+ dwarf::Tag Tag = Die.getTag();
+ if (Tag == dwarf::DW_TAG_inlined_subroutine) {
+ // create new InlineInfo and append to parent.children
+ InlineInfo II;
+ DWARFAddressRange FuncRange =
+ DWARFAddressRange(FI.startAddress(), FI.endAddress());
+ Expected<DWARFAddressRangesVector> RangesOrError = Die.getAddressRanges();
+ if (RangesOrError) {
+ for (const DWARFAddressRange &Range : RangesOrError.get()) {
+ // Check that the inlined function is within the range of the function
+ // info, it might not be in case of split functions
+ if (FuncRange.LowPC <= Range.LowPC && Range.HighPC <= FuncRange.HighPC)
+ II.Ranges.insert(AddressRange(Range.LowPC, Range.HighPC));
+ }
+ }
+ if (II.Ranges.empty())
+ return;
+
+ if (auto NameIndex = getQualifiedNameIndex(Die, CUI.Language, Gsym))
+ II.Name = *NameIndex;
+ II.CallFile = CUI.DWARFToGSYMFileIndex(
+ Gsym, dwarf::toUnsigned(Die.find(dwarf::DW_AT_call_file), 0));
+ II.CallLine = dwarf::toUnsigned(Die.find(dwarf::DW_AT_call_line), 0);
+ // parse all children and append to parent
+ for (DWARFDie ChildDie : Die.children())
+ parseInlineInfo(Gsym, CUI, ChildDie, Depth + 1, FI, II);
+ parent.Children.emplace_back(std::move(II));
+ return;
+ }
+ if (Tag == dwarf::DW_TAG_subprogram || Tag == dwarf::DW_TAG_lexical_block) {
+ // skip this Die and just recurse down
+ for (DWARFDie ChildDie : Die.children())
+ parseInlineInfo(Gsym, CUI, ChildDie, Depth + 1, FI, parent);
+ }
+}
+
+static void convertFunctionLineTable(raw_ostream &Log, CUInfo &CUI,
+ DWARFDie Die, GsymCreator &Gsym,
+ FunctionInfo &FI) {
+ std::vector<uint32_t> RowVector;
+ const uint64_t StartAddress = FI.startAddress();
+ const uint64_t EndAddress = FI.endAddress();
+ const uint64_t RangeSize = EndAddress - StartAddress;
+ const object::SectionedAddress SecAddress{
+ StartAddress, object::SectionedAddress::UndefSection};
+
+
+ if (!CUI.LineTable->lookupAddressRange(SecAddress, RangeSize, RowVector)) {
+ // If we have a DW_TAG_subprogram but no line entries, fall back to using
+ // the DW_AT_decl_file an d DW_AT_decl_line if we have both attributes.
+ if (auto FileIdx =
+ dwarf::toUnsigned(Die.findRecursively({dwarf::DW_AT_decl_file}))) {
+ if (auto Line =
+ dwarf::toUnsigned(Die.findRecursively({dwarf::DW_AT_decl_line}))) {
+ LineEntry LE(StartAddress, CUI.DWARFToGSYMFileIndex(Gsym, *FileIdx),
+ *Line);
+ FI.OptLineTable = LineTable();
+ FI.OptLineTable->push(LE);
+ // LE.Addr = EndAddress;
+ // FI.OptLineTable->push(LE);
+ }
+ }
+ return;
+ }
+
+ FI.OptLineTable = LineTable();
+ DWARFDebugLine::Row PrevRow;
+ for (uint32_t RowIndex : RowVector) {
+ // Take file number and line/column from the row.
+ const DWARFDebugLine::Row &Row = CUI.LineTable->Rows[RowIndex];
+ const uint32_t FileIdx = CUI.DWARFToGSYMFileIndex(Gsym, Row.File);
+ uint64_t RowAddress = Row.Address.Address;
+ // Watch out for a RowAddress that is in the middle of a line table entry
+ // in the DWARF. If we pass an address in between two line table entries
+ // we will get a RowIndex for the previous valid line table row which won't
+ // be contained in our function. This is usually a bug in the DWARF due to
+ // linker problems or LTO or other DWARF re-linking so it is worth emitting
+ // an error, but not worth stopping the creation of the GSYM.
+ if (!FI.Range.contains(RowAddress)) {
+ if (RowAddress < FI.Range.Start) {
+ Log << "error: DIE has a start address whose LowPC is between the "
+ "line table Row[" << RowIndex << "] with address "
+ << HEX64(RowAddress) << " and the next one.\n";
+ Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
+ RowAddress = FI.Range.Start;
+ } else {
+ continue;
+ }
+ }
+
+ LineEntry LE(RowAddress, FileIdx, Row.Line);
+ if (RowIndex != RowVector[0] && Row.Address < PrevRow.Address) {
+ // We have seen full duplicate line tables for functions in some
+ // DWARF files. Watch for those here by checking the the last
+ // row was the function's end address (HighPC) and that the
+ // current line table entry's address is the same as the first
+ // line entry we already have in our "function_info.Lines". If
+ // so break out after printing a warning.
+ auto FirstLE = FI.OptLineTable->first();
+ if (FirstLE && *FirstLE == LE) {
+ Log << "warning: duplicate line table detected for DIE:\n";
+ Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
+ } else {
+ // Print out (ignore if os == nulls as this is expensive)
+ Log << "error: line table has addresses that do not "
+ << "monotonically increase:\n";
+ for (uint32_t RowIndex2 : RowVector) {
+ CUI.LineTable->Rows[RowIndex2].dump(Log);
+ }
+ Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
+ }
+ break;
+ }
+
+ // Skip multiple line entries for the same file and line.
+ auto LastLE = FI.OptLineTable->last();
+ if (LastLE && LastLE->File == FileIdx && LastLE->Line == Row.Line)
+ continue;
+ // Only push a row if it isn't an end sequence. End sequence markers are
+ // included for the last address in a function or the last contiguous
+ // address in a sequence.
+ if (Row.EndSequence) {
+ // End sequence means that the next line entry could have a lower address
+ // that the previous entries. So we clear the previous row so we don't
+ // trigger the line table error about address that do not monotonically
+ // increase.
+ PrevRow = DWARFDebugLine::Row();
+ } else {
+ FI.OptLineTable->push(LE);
+ PrevRow = Row;
+ }
+ }
+ // If not line table rows were added, clear the line table so we don't encode
+ // on in the GSYM file.
+ if (FI.OptLineTable->empty())
+ FI.OptLineTable = llvm::None;
+}
+
+void DwarfTransformer::handleDie(raw_ostream &OS, CUInfo &CUI, DWARFDie Die) {
+ switch (Die.getTag()) {
+ case dwarf::DW_TAG_subprogram: {
+ Expected<DWARFAddressRangesVector> RangesOrError = Die.getAddressRanges();
+ if (!RangesOrError) {
+ consumeError(RangesOrError.takeError());
+ break;
+ }
+ const DWARFAddressRangesVector &Ranges = RangesOrError.get();
+ if (Ranges.empty())
+ break;
+ auto NameIndex = getQualifiedNameIndex(Die, CUI.Language, Gsym);
+ if (!NameIndex) {
+ OS << "error: function at " << HEX64(Die.getOffset())
+ << " has no name\n ";
+ Die.dump(OS, 0, DIDumpOptions::getForSingleDIE());
+ break;
+ }
+
+ // Create a function_info for each range
+ for (const DWARFAddressRange &Range : Ranges) {
+ // The low PC must be less than the high PC. Many linkers don't remove
+ // DWARF for functions that don't get linked into the final executable.
+ // If both the high and low pc have relocations, linkers will often set
+ // the address values for both to the same value to indicate the function
+ // has been remove. Other linkers have been known to set the one or both
+ // PC values to a UINT32_MAX for 4 byte addresses and UINT64_MAX for 8
+ // byte addresses to indicate the function isn't valid. The check below
+ // tries to watch for these cases and abort if it runs into them.
+ if (Range.LowPC >= Range.HighPC || CUI.isHighestAddress(Range.LowPC))
+ break;
+
+ // Many linkers can't remove DWARF and might set the LowPC to zero. Since
+ // high PC can be an offset from the low PC in more recent DWARF versions
+ // we need to watch for a zero'ed low pc which we do using
+ // ValidTextRanges below.
+ if (!Gsym.IsValidTextAddress(Range.LowPC)) {
+ // We expect zero and -1 to be invalid addresses in DWARF depending
+ // on the linker of the DWARF. This indicates a function was stripped
+ // and the debug info wasn't able to be stripped from the DWARF. If
+ // the LowPC isn't zero or -1, then we should emit an error.
+ if (Range.LowPC != 0) {
+ // Unexpected invalid address, emit an error
+ Log << "warning: DIE has an address range whose start address is "
+ "not in any executable sections (" <<
+ *Gsym.GetValidTextRanges() << ") and will not be processed:\n";
+ Die.dump(Log, 0, DIDumpOptions::getForSingleDIE());
+ }
+ break;
+ }
+
+ FunctionInfo FI;
+ FI.setStartAddress(Range.LowPC);
+ FI.setEndAddress(Range.HighPC);
+ FI.Name = *NameIndex;
+ if (CUI.LineTable) {
+ convertFunctionLineTable(OS, CUI, Die, Gsym, FI);
+ }
+ if (hasInlineInfo(Die, 0)) {
+ FI.Inline = InlineInfo();
+ FI.Inline->Name = *NameIndex;
+ FI.Inline->Ranges.insert(FI.Range);
+ parseInlineInfo(Gsym, CUI, Die, 0, FI, *FI.Inline);
+ }
+ Gsym.addFunctionInfo(std::move(FI));
+ }
+ } break;
+ default:
+ break;
+ }
+ for (DWARFDie ChildDie : Die.children())
+ handleDie(OS, CUI, ChildDie);
+}
+
+Error DwarfTransformer::convert(uint32_t NumThreads) {
+ size_t NumBefore = Gsym.getNumFunctionInfos();
+ if (NumThreads == 1) {
+ // Parse all DWARF data from this thread, use the same string/file table
+ // for everything
+ for (const auto &CU : DICtx.compile_units()) {
+ DWARFDie Die = CU->getUnitDIE(false);
+ CUInfo CUI(DICtx, dyn_cast<DWARFCompileUnit>(CU.get()));
+ handleDie(Log, CUI, Die);
+ }
+ } else {
+ // LLVM Dwarf parser is not thread-safe and we need to parse all DWARF up
+ // front before we start accessing any DIEs since there might be
+ // cross compile unit references in the DWARF. If we don't do this we can
+ // end up crashing.
+
+ // We need to call getAbbreviations sequentially first so that getUnitDIE()
+ // only works with its local data.
+ for (const auto &CU : DICtx.compile_units())
+ CU->getAbbreviations();
+
+ // Now parse all DIEs in case we have cross compile unit references in a
+ // thread pool.
+ ThreadPool pool(hardware_concurrency(NumThreads));
+ for (const auto &CU : DICtx.compile_units())
+ pool.async([&CU]() { CU->getUnitDIE(false /*CUDieOnly*/); });
+ pool.wait();
+
+ // Now convert all DWARF to GSYM in a thread pool.
+ std::mutex LogMutex;
+ for (const auto &CU : DICtx.compile_units()) {
+ DWARFDie Die = CU->getUnitDIE(false /*CUDieOnly*/);
+ if (Die) {
+ CUInfo CUI(DICtx, dyn_cast<DWARFCompileUnit>(CU.get()));
+ pool.async([this, CUI, &LogMutex, Die]() mutable {
+ std::string ThreadLogStorage;
+ raw_string_ostream ThreadOS(ThreadLogStorage);
+ handleDie(ThreadOS, CUI, Die);
+ ThreadOS.flush();
+ if (!ThreadLogStorage.empty()) {
+ // Print ThreadLogStorage lines into an actual stream under a lock
+ std::lock_guard<std::mutex> guard(LogMutex);
+ Log << ThreadLogStorage;
+ }
+ });
+ }
+ }
+ pool.wait();
+ }
+ size_t FunctionsAddedCount = Gsym.getNumFunctionInfos() - NumBefore;
+ Log << "Loaded " << FunctionsAddedCount << " functions from DWARF.\n";
+ return Error::success();
+}
+
+llvm::Error DwarfTransformer::verify(StringRef GsymPath) {
+ Log << "Verifying GSYM file \"" << GsymPath << "\":\n";
+
+ auto Gsym = GsymReader::openFile(GsymPath);
+ if (!Gsym)
+ return Gsym.takeError();
+
+ auto NumAddrs = Gsym->getNumAddresses();
+ DILineInfoSpecifier DLIS(
+ DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath,
+ DILineInfoSpecifier::FunctionNameKind::LinkageName);
+ std::string gsymFilename;
+ for (uint32_t I = 0; I < NumAddrs; ++I) {
+ auto FuncAddr = Gsym->getAddress(I);
+ if (!FuncAddr)
+ return createStringError(std::errc::invalid_argument,
+ "failed to extract address[%i]", I);
+
+ auto FI = Gsym->getFunctionInfo(*FuncAddr);
+ if (!FI)
+ return createStringError(std::errc::invalid_argument,
+ "failed to extract function info for address 0x%"
+ PRIu64, *FuncAddr);
+
+ for (auto Addr = *FuncAddr; Addr < *FuncAddr + FI->size(); ++Addr) {
+ const object::SectionedAddress SectAddr{
+ Addr, object::SectionedAddress::UndefSection};
+ auto LR = Gsym->lookup(Addr);
+ if (!LR)
+ return LR.takeError();
+
+ auto DwarfInlineInfos =
+ DICtx.getInliningInfoForAddress(SectAddr, DLIS);
+ uint32_t NumDwarfInlineInfos = DwarfInlineInfos.getNumberOfFrames();
+ if (NumDwarfInlineInfos == 0) {
+ DwarfInlineInfos.addFrame(
+ DICtx.getLineInfoForAddress(SectAddr, DLIS));
+ }
+
+ // Check for 1 entry that has no file and line info
+ if (NumDwarfInlineInfos == 1 &&
+ DwarfInlineInfos.getFrame(0).FileName == "<invalid>") {
+ DwarfInlineInfos = DIInliningInfo();
+ NumDwarfInlineInfos = 0;
+ }
+ if (NumDwarfInlineInfos > 0 &&
+ NumDwarfInlineInfos != LR->Locations.size()) {
+ Log << "error: address " << HEX64(Addr) << " has "
+ << NumDwarfInlineInfos << " DWARF inline frames and GSYM has "
+ << LR->Locations.size() << "\n";
+ Log << " " << NumDwarfInlineInfos << " DWARF frames:\n";
+ for (size_t Idx = 0; Idx < NumDwarfInlineInfos; ++Idx) {
+ const auto dii = DwarfInlineInfos.getFrame(Idx);
+ Log << " [" << Idx << "]: " << dii.FunctionName << " @ "
+ << dii.FileName << ':' << dii.Line << '\n';
+ }
+ Log << " " << LR->Locations.size() << " GSYM frames:\n";
+ for (size_t Idx = 0, count = LR->Locations.size();
+ Idx < count; ++Idx) {
+ const auto &gii = LR->Locations[Idx];
+ Log << " [" << Idx << "]: " << gii.Name << " @ " << gii.Dir
+ << '/' << gii.Base << ':' << gii.Line << '\n';
+ }
+ DwarfInlineInfos = DICtx.getInliningInfoForAddress(SectAddr, DLIS);
+ Gsym->dump(Log, *FI);
+ continue;
+ }
+
+ for (size_t Idx = 0, count = LR->Locations.size(); Idx < count;
+ ++Idx) {
+ const auto &gii = LR->Locations[Idx];
+ if (Idx < NumDwarfInlineInfos) {
+ const auto dii = DwarfInlineInfos.getFrame(Idx);
+ gsymFilename = LR->getSourceFile(Idx);
+ // Verify function name
+ if (dii.FunctionName.find(gii.Name.str()) != 0)
+ Log << "error: address " << HEX64(Addr) << " DWARF function \""
+ << dii.FunctionName.c_str()
+ << "\" doesn't match GSYM function \"" << gii.Name << "\"\n";
+ // Verify source file path
+ if (dii.FileName != gsymFilename)
+ Log << "error: address " << HEX64(Addr) << " DWARF path \""
+ << dii.FileName.c_str() << "\" doesn't match GSYM path \""
+ << gsymFilename.c_str() << "\"\n";
+ // Verify source file line
+ if (dii.Line != gii.Line)
+ Log << "error: address " << HEX64(Addr) << " DWARF line "
+ << dii.Line << " != GSYM line " << gii.Line << "\n";
+ }
+ }
+ }
+ }
+ return Error::success();
+}
diff --git a/llvm/lib/DebugInfo/GSYM/FunctionInfo.cpp b/llvm/lib/DebugInfo/GSYM/FunctionInfo.cpp
index 6731a8b27443..cef1b9498c5c 100644
--- a/llvm/lib/DebugInfo/GSYM/FunctionInfo.cpp
+++ b/llvm/lib/DebugInfo/GSYM/FunctionInfo.cpp
@@ -25,8 +25,11 @@ enum InfoType : uint32_t {
};
raw_ostream &llvm::gsym::operator<<(raw_ostream &OS, const FunctionInfo &FI) {
- OS << '[' << HEX64(FI.Range.Start) << '-' << HEX64(FI.Range.End) << "): "
- << "Name=" << HEX32(FI.Name) << '\n' << FI.OptLineTable << FI.Inline;
+ OS << FI.Range << ": " << "Name=" << HEX32(FI.Name) << '\n';
+ if (FI.OptLineTable)
+ OS << FI.OptLineTable << '\n';
+ if (FI.Inline)
+ OS << FI.Inline << '\n';
return OS;
}
@@ -167,7 +170,7 @@ llvm::Expected<LookupResult> FunctionInfo::lookup(DataExtractor &Data,
// This function will be called with the result of a binary search of the
// address table, we must still make sure the address does not fall into a
// gap between functions are after the last function.
- if (Addr >= LR.FuncRange.End)
+ if (LR.FuncRange.size() > 0 && !LR.FuncRange.contains(Addr))
return createStringError(std::errc::io_error,
"address 0x%" PRIx64 " is not in GSYM", Addr);
@@ -220,6 +223,7 @@ llvm::Expected<LookupResult> FunctionInfo::lookup(DataExtractor &Data,
// location as best we can and return.
SourceLocation SrcLoc;
SrcLoc.Name = LR.FuncName;
+ SrcLoc.Offset = Addr - FuncAddr;
LR.Locations.push_back(SrcLoc);
return LR;
}
@@ -232,6 +236,7 @@ llvm::Expected<LookupResult> FunctionInfo::lookup(DataExtractor &Data,
SourceLocation SrcLoc;
SrcLoc.Name = LR.FuncName;
+ SrcLoc.Offset = Addr - FuncAddr;
SrcLoc.Dir = GR.getString(LineEntryFile->Dir);
SrcLoc.Base = GR.getString(LineEntryFile->Base);
SrcLoc.Line = LineEntry->Line;
diff --git a/llvm/lib/DebugInfo/GSYM/GsymCreator.cpp b/llvm/lib/DebugInfo/GSYM/GsymCreator.cpp
index f371426f2010..7d9b72c6283d 100644
--- a/llvm/lib/DebugInfo/GSYM/GsymCreator.cpp
+++ b/llvm/lib/DebugInfo/GSYM/GsymCreator.cpp
@@ -29,7 +29,13 @@ uint32_t GsymCreator::insertFile(StringRef Path,
llvm::sys::path::Style Style) {
llvm::StringRef directory = llvm::sys::path::parent_path(Path, Style);
llvm::StringRef filename = llvm::sys::path::filename(Path, Style);
- FileEntry FE(insertString(directory), insertString(filename));
+ // We must insert the strings first, then call the FileEntry constructor.
+ // If we inline the insertString() function call into the constructor, the
+ // call order is undefined due to parameter lists not having any ordering
+ // requirements.
+ const uint32_t Dir = insertString(directory);
+ const uint32_t Base = insertString(filename);
+ FileEntry FE(Dir, Base);
std::lock_guard<std::recursive_mutex> Guard(Mutex);
const auto NextIndex = Files.size();
@@ -62,7 +68,8 @@ llvm::Error GsymCreator::encode(FileWriter &O) const {
if (Funcs.size() > UINT32_MAX)
return createStringError(std::errc::invalid_argument,
"too many FunctionInfos");
- const uint64_t MinAddr = Funcs.front().startAddress();
+
+ const uint64_t MinAddr = BaseAddress ? *BaseAddress : Funcs.front().startAddress();
const uint64_t MaxAddr = Funcs.back().startAddress();
const uint64_t AddrDelta = MaxAddr - MinAddr;
Header Hdr;
@@ -73,7 +80,7 @@ llvm::Error GsymCreator::encode(FileWriter &O) const {
Hdr.BaseAddress = MinAddr;
Hdr.NumAddresses = static_cast<uint32_t>(Funcs.size());
Hdr.StrtabOffset = 0; // We will fix this up later.
- Hdr.StrtabOffset = 0; // We will fix this up later.
+ Hdr.StrtabSize = 0; // We will fix this up later.
memset(Hdr.UUID, 0, sizeof(Hdr.UUID));
if (UUID.size() > sizeof(Hdr.UUID))
return createStringError(std::errc::invalid_argument,
@@ -203,9 +210,8 @@ llvm::Error GsymCreator::finalize(llvm::raw_ostream &OS) {
// that have debug info are last in the sort.
if (*Prev == *Curr) {
// FunctionInfo entries match exactly (range, lines, inlines)
- OS << "warning: duplicate function info entries, removing "
- "duplicate:\n"
- << *Curr << '\n';
+ OS << "warning: duplicate function info entries for range: "
+ << Curr->Range << '\n';
Curr = Funcs.erase(Prev);
} else {
if (!Prev->hasRichInfo() && Curr->hasRichInfo()) {
@@ -239,20 +245,43 @@ llvm::Error GsymCreator::finalize(llvm::raw_ostream &OS) {
Prev = Curr++;
}
+ // If our last function info entry doesn't have a size and if we have valid
+ // text ranges, we should set the size of the last entry since any search for
+ // a high address might match our last entry. By fixing up this size, we can
+ // help ensure we don't cause lookups to always return the last symbol that
+ // has no size when doing lookups.
+ if (!Funcs.empty() && Funcs.back().Range.size() == 0 && ValidTextRanges) {
+ if (auto Range = ValidTextRanges->getRangeThatContains(
+ Funcs.back().Range.Start)) {
+ Funcs.back().Range.End = Range->End;
+ }
+ }
OS << "Pruned " << NumBefore - Funcs.size() << " functions, ended with "
<< Funcs.size() << " total\n";
return Error::success();
}
-uint32_t GsymCreator::insertString(StringRef S) {
- std::lock_guard<std::recursive_mutex> Guard(Mutex);
+uint32_t GsymCreator::insertString(StringRef S, bool Copy) {
if (S.empty())
return 0;
+ std::lock_guard<std::recursive_mutex> Guard(Mutex);
+ if (Copy) {
+ // We need to provide backing storage for the string if requested
+ // since StringTableBuilder stores references to strings. Any string
+ // that comes from a section in an object file doesn't need to be
+ // copied, but any string created by code will need to be copied.
+ // This allows GsymCreator to be really fast when parsing DWARF and
+ // other object files as most strings don't need to be copied.
+ CachedHashStringRef CHStr(S);
+ if (!StrTab.contains(CHStr))
+ S = StringStorage.insert(S).first->getKey();
+ }
return StrTab.add(S);
}
void GsymCreator::addFunctionInfo(FunctionInfo &&FI) {
std::lock_guard<std::recursive_mutex> Guard(Mutex);
+ Ranges.insert(FI.Range);
Funcs.emplace_back(FI);
}
@@ -273,3 +302,19 @@ void GsymCreator::forEachFunctionInfo(
break;
}
}
+
+size_t GsymCreator::getNumFunctionInfos() const{
+ std::lock_guard<std::recursive_mutex> Guard(Mutex);
+ return Funcs.size();
+}
+
+bool GsymCreator::IsValidTextAddress(uint64_t Addr) const {
+ if (ValidTextRanges)
+ return ValidTextRanges->contains(Addr);
+ return true; // No valid text ranges has been set, so accept all ranges.
+}
+
+bool GsymCreator::hasFunctionInfoForAddress(uint64_t Addr) const {
+ std::lock_guard<std::recursive_mutex> Guard(Mutex);
+ return Ranges.contains(Addr);
+}
diff --git a/llvm/lib/DebugInfo/GSYM/GsymReader.cpp b/llvm/lib/DebugInfo/GSYM/GsymReader.cpp
index b4f3f2052ae7..2ad18bf63d5d 100644
--- a/llvm/lib/DebugInfo/GSYM/GsymReader.cpp
+++ b/llvm/lib/DebugInfo/GSYM/GsymReader.cpp
@@ -225,20 +225,33 @@ Optional<uint64_t> GsymReader::getAddressInfoOffset(size_t Index) const {
Expected<uint64_t>
GsymReader::getAddressIndex(const uint64_t Addr) const {
- if (Addr < Hdr->BaseAddress)
- return createStringError(std::errc::invalid_argument,
- "address 0x%" PRIx64 " not in GSYM", Addr);
- const uint64_t AddrOffset = Addr - Hdr->BaseAddress;
- switch (Hdr->AddrOffSize) {
- case 1: return getAddressOffsetIndex<uint8_t>(AddrOffset);
- case 2: return getAddressOffsetIndex<uint16_t>(AddrOffset);
- case 4: return getAddressOffsetIndex<uint32_t>(AddrOffset);
- case 8: return getAddressOffsetIndex<uint64_t>(AddrOffset);
- default: break;
+ if (Addr >= Hdr->BaseAddress) {
+ const uint64_t AddrOffset = Addr - Hdr->BaseAddress;
+ Optional<uint64_t> AddrOffsetIndex;
+ switch (Hdr->AddrOffSize) {
+ case 1:
+ AddrOffsetIndex = getAddressOffsetIndex<uint8_t>(AddrOffset);
+ break;
+ case 2:
+ AddrOffsetIndex = getAddressOffsetIndex<uint16_t>(AddrOffset);
+ break;
+ case 4:
+ AddrOffsetIndex = getAddressOffsetIndex<uint32_t>(AddrOffset);
+ break;
+ case 8:
+ AddrOffsetIndex = getAddressOffsetIndex<uint64_t>(AddrOffset);
+ break;
+ default:
+ return createStringError(std::errc::invalid_argument,
+ "unsupported address offset size %u",
+ Hdr->AddrOffSize);
+ }
+ if (AddrOffsetIndex)
+ return *AddrOffsetIndex;
}
return createStringError(std::errc::invalid_argument,
- "unsupported address offset size %u",
- Hdr->AddrOffSize);
+ "address 0x%" PRIx64 " is not in GSYM", Addr);
+
}
llvm::Expected<FunctionInfo> GsymReader::getFunctionInfo(uint64_t Addr) const {
@@ -255,7 +268,7 @@ llvm::Expected<FunctionInfo> GsymReader::getFunctionInfo(uint64_t Addr) const {
if (ExpectedFI->Range.contains(Addr) || ExpectedFI->Range.size() == 0)
return ExpectedFI;
return createStringError(std::errc::invalid_argument,
- "address 0x%" PRIx64 " not in GSYM", Addr);
+ "address 0x%" PRIx64 " is not in GSYM", Addr);
}
}
return createStringError(std::errc::invalid_argument,
@@ -277,3 +290,117 @@ llvm::Expected<LookupResult> GsymReader::lookup(uint64_t Addr) const {
"failed to extract address[%" PRIu64 "]",
*AddressIndex);
}
+
+void GsymReader::dump(raw_ostream &OS) {
+ const auto &Header = getHeader();
+ // Dump the GSYM header.
+ OS << Header << "\n";
+ // Dump the address table.
+ OS << "Address Table:\n";
+ OS << "INDEX OFFSET";
+
+ switch (Hdr->AddrOffSize) {
+ case 1: OS << "8 "; break;
+ case 2: OS << "16"; break;
+ case 4: OS << "32"; break;
+ case 8: OS << "64"; break;
+ default: OS << "??"; break;
+ }
+ OS << " (ADDRESS)\n";
+ OS << "====== =============================== \n";
+ for (uint32_t I = 0; I < Header.NumAddresses; ++I) {
+ OS << format("[%4u] ", I);
+ switch (Hdr->AddrOffSize) {
+ case 1: OS << HEX8(getAddrOffsets<uint8_t>()[I]); break;
+ case 2: OS << HEX16(getAddrOffsets<uint16_t>()[I]); break;
+ case 4: OS << HEX32(getAddrOffsets<uint32_t>()[I]); break;
+ case 8: OS << HEX32(getAddrOffsets<uint64_t>()[I]); break;
+ default: break;
+ }
+ OS << " (" << HEX64(*getAddress(I)) << ")\n";
+ }
+ // Dump the address info offsets table.
+ OS << "\nAddress Info Offsets:\n";
+ OS << "INDEX Offset\n";
+ OS << "====== ==========\n";
+ for (uint32_t I = 0; I < Header.NumAddresses; ++I)
+ OS << format("[%4u] ", I) << HEX32(AddrInfoOffsets[I]) << "\n";
+ // Dump the file table.
+ OS << "\nFiles:\n";
+ OS << "INDEX DIRECTORY BASENAME PATH\n";
+ OS << "====== ========== ========== ==============================\n";
+ for (uint32_t I = 0; I < Files.size(); ++I) {
+ OS << format("[%4u] ", I) << HEX32(Files[I].Dir) << ' '
+ << HEX32(Files[I].Base) << ' ';
+ dump(OS, getFile(I));
+ OS << "\n";
+ }
+ OS << "\n" << StrTab << "\n";
+
+ for (uint32_t I = 0; I < Header.NumAddresses; ++I) {
+ OS << "FunctionInfo @ " << HEX32(AddrInfoOffsets[I]) << ": ";
+ if (auto FI = getFunctionInfo(*getAddress(I)))
+ dump(OS, *FI);
+ else
+ logAllUnhandledErrors(FI.takeError(), OS, "FunctionInfo:");
+ }
+}
+
+void GsymReader::dump(raw_ostream &OS, const FunctionInfo &FI) {
+ OS << FI.Range << " \"" << getString(FI.Name) << "\"\n";
+ if (FI.OptLineTable)
+ dump(OS, *FI.OptLineTable);
+ if (FI.Inline)
+ dump(OS, *FI.Inline);
+}
+
+void GsymReader::dump(raw_ostream &OS, const LineTable &LT) {
+ OS << "LineTable:\n";
+ for (auto &LE: LT) {
+ OS << " " << HEX64(LE.Addr) << ' ';
+ if (LE.File)
+ dump(OS, getFile(LE.File));
+ OS << ':' << LE.Line << '\n';
+ }
+}
+
+void GsymReader::dump(raw_ostream &OS, const InlineInfo &II, uint32_t Indent) {
+ if (Indent == 0)
+ OS << "InlineInfo:\n";
+ else
+ OS.indent(Indent);
+ OS << II.Ranges << ' ' << getString(II.Name);
+ if (II.CallFile != 0) {
+ if (auto File = getFile(II.CallFile)) {
+ OS << " called from ";
+ dump(OS, File);
+ OS << ':' << II.CallLine;
+ }
+ }
+ OS << '\n';
+ for (const auto &ChildII: II.Children)
+ dump(OS, ChildII, Indent + 2);
+}
+
+void GsymReader::dump(raw_ostream &OS, Optional<FileEntry> FE) {
+ if (FE) {
+ // IF we have the file from index 0, then don't print anything
+ if (FE->Dir == 0 && FE->Base == 0)
+ return;
+ StringRef Dir = getString(FE->Dir);
+ StringRef Base = getString(FE->Base);
+ if (!Dir.empty()) {
+ OS << Dir;
+ if (Dir.contains('\\') && !Dir.contains('/'))
+ OS << '\\';
+ else
+ OS << '/';
+ }
+ if (!Base.empty()) {
+ OS << Base;
+ }
+ if (!Dir.empty() || !Base.empty())
+ return;
+ }
+ OS << "<invalid-file>";
+}
diff --git a/llvm/lib/DebugInfo/GSYM/InlineInfo.cpp b/llvm/lib/DebugInfo/GSYM/InlineInfo.cpp
index 1b8c974fdcd2..21679b1b78aa 100644
--- a/llvm/lib/DebugInfo/GSYM/InlineInfo.cpp
+++ b/llvm/lib/DebugInfo/GSYM/InlineInfo.cpp
@@ -142,13 +142,17 @@ static bool lookup(const GsymReader &GR, DataExtractor &Data, uint64_t &Offset,
return false;
}
- SourceLocation SrcLoc;
- SrcLoc.Name = SrcLocs.back().Name;
- SrcLoc.Dir = GR.getString(CallFile->Dir);
- SrcLoc.Base = GR.getString(CallFile->Base);
- SrcLoc.Line = Inline.CallLine;
- SrcLocs.back().Name = GR.getString(Inline.Name);
- SrcLocs.push_back(SrcLoc);
+ if (CallFile->Dir || CallFile->Base) {
+ SourceLocation SrcLoc;
+ SrcLoc.Name = SrcLocs.back().Name;
+ SrcLoc.Offset = SrcLocs.back().Offset;
+ SrcLoc.Dir = GR.getString(CallFile->Dir);
+ SrcLoc.Base = GR.getString(CallFile->Base);
+ SrcLoc.Line = Inline.CallLine;
+ SrcLocs.back().Name = GR.getString(Inline.Name);
+ SrcLocs.back().Offset = Addr - Inline.Ranges[0].Start;
+ SrcLocs.push_back(SrcLoc);
+ }
return true;
}
diff --git a/llvm/lib/DebugInfo/GSYM/LookupResult.cpp b/llvm/lib/DebugInfo/GSYM/LookupResult.cpp
index c54b166b2887..8a624226b1d3 100644
--- a/llvm/lib/DebugInfo/GSYM/LookupResult.cpp
+++ b/llvm/lib/DebugInfo/GSYM/LookupResult.cpp
@@ -21,7 +21,7 @@ std::string LookupResult::getSourceFile(uint32_t Index) const {
if (Index < Locations.size()) {
if (!Locations[Index].Dir.empty()) {
if (Locations[Index].Base.empty()) {
- Fullpath = Locations[Index].Dir;
+ Fullpath = std::string(Locations[Index].Dir);
} else {
llvm::SmallString<64> Storage;
llvm::sys::path::append(Storage, Locations[Index].Dir,
@@ -29,25 +29,30 @@ std::string LookupResult::getSourceFile(uint32_t Index) const {
Fullpath.assign(Storage.begin(), Storage.end());
}
} else if (!Locations[Index].Base.empty())
- Fullpath = Locations[Index].Base;
+ Fullpath = std::string(Locations[Index].Base);
}
return Fullpath;
}
raw_ostream &llvm::gsym::operator<<(raw_ostream &OS, const SourceLocation &SL) {
- OS << SL.Name << " @ ";
- if (!SL.Dir.empty()) {
- OS << SL.Dir;
- if (SL.Dir.contains('\\') and not SL.Dir.contains('/'))
- OS << '\\';
+ OS << SL.Name;
+ if (SL.Offset > 0)
+ OS << " + " << SL.Offset;
+ if (SL.Dir.size() || SL.Base.size()) {
+ OS << " @ ";
+ if (!SL.Dir.empty()) {
+ OS << SL.Dir;
+ if (SL.Dir.contains('\\') and not SL.Dir.contains('/'))
+ OS << '\\';
+ else
+ OS << '/';
+ }
+ if (SL.Base.empty())
+ OS << "<invalid-file>";
else
- OS << '/';
+ OS << SL.Base;
+ OS << ':' << SL.Line;
}
- if (SL.Base.empty())
- OS << "<invalid-file>";
- else
- OS << SL.Base;
- OS << ':' << SL.Line;
return OS;
}
diff --git a/llvm/lib/DebugInfo/GSYM/ObjectFileTransformer.cpp b/llvm/lib/DebugInfo/GSYM/ObjectFileTransformer.cpp
new file mode 100644
index 000000000000..ad35aefe7774
--- /dev/null
+++ b/llvm/lib/DebugInfo/GSYM/ObjectFileTransformer.cpp
@@ -0,0 +1,116 @@
+//===- ObjectFileTransformer.cpp --------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include <unordered_set>
+
+#include "llvm/Object/ELFObjectFile.h"
+#include "llvm/Object/MachOUniversal.h"
+#include "llvm/Object/ObjectFile.h"
+#include "llvm/Support/DataExtractor.h"
+#include "llvm/Support/raw_ostream.h"
+
+#include "llvm/DebugInfo/GSYM/ObjectFileTransformer.h"
+#include "llvm/DebugInfo/GSYM/GsymCreator.h"
+
+using namespace llvm;
+using namespace gsym;
+
+constexpr uint32_t NT_GNU_BUILD_ID_TAG = 0x03;
+
+static std::vector<uint8_t> getUUID(const object::ObjectFile &Obj) {
+ // Extract the UUID from the object file
+ std::vector<uint8_t> UUID;
+ if (auto *MachO = dyn_cast<object::MachOObjectFile>(&Obj)) {
+ const ArrayRef<uint8_t> MachUUID = MachO->getUuid();
+ if (!MachUUID.empty())
+ UUID.assign(MachUUID.data(), MachUUID.data() + MachUUID.size());
+ } else if (isa<object::ELFObjectFileBase>(&Obj)) {
+ const StringRef GNUBuildID(".note.gnu.build-id");
+ for (const object::SectionRef &Sect : Obj.sections()) {
+ Expected<StringRef> SectNameOrErr = Sect.getName();
+ if (!SectNameOrErr) {
+ consumeError(SectNameOrErr.takeError());
+ continue;
+ }
+ StringRef SectName(*SectNameOrErr);
+ if (SectName != GNUBuildID)
+ continue;
+ StringRef BuildIDData;
+ Expected<StringRef> E = Sect.getContents();
+ if (E)
+ BuildIDData = *E;
+ else {
+ consumeError(E.takeError());
+ continue;
+ }
+ DataExtractor Decoder(BuildIDData, Obj.makeTriple().isLittleEndian(), 8);
+ uint64_t Offset = 0;
+ const uint32_t NameSize = Decoder.getU32(&Offset);
+ const uint32_t PayloadSize = Decoder.getU32(&Offset);
+ const uint32_t PayloadType = Decoder.getU32(&Offset);
+ StringRef Name(Decoder.getFixedLengthString(&Offset, NameSize));
+ if (Name == "GNU" && PayloadType == NT_GNU_BUILD_ID_TAG) {
+ Offset = alignTo(Offset, 4);
+ StringRef UUIDBytes(Decoder.getBytes(&Offset, PayloadSize));
+ if (!UUIDBytes.empty()) {
+ auto Ptr = reinterpret_cast<const uint8_t *>(UUIDBytes.data());
+ UUID.assign(Ptr, Ptr + UUIDBytes.size());
+ }
+ }
+ }
+ }
+ return UUID;
+}
+
+llvm::Error ObjectFileTransformer::convert(const object::ObjectFile &Obj,
+ raw_ostream &Log,
+ GsymCreator &Gsym) {
+ using namespace llvm::object;
+
+ const bool IsMachO = isa<MachOObjectFile>(&Obj);
+ const bool IsELF = isa<ELFObjectFileBase>(&Obj);
+
+ // Read build ID.
+ Gsym.setUUID(getUUID(Obj));
+
+ // Parse the symbol table.
+ size_t NumBefore = Gsym.getNumFunctionInfos();
+ for (const object::SymbolRef &Sym : Obj.symbols()) {
+ Expected<SymbolRef::Type> SymType = Sym.getType();
+ if (!SymType) {
+ consumeError(SymType.takeError());
+ continue;
+ }
+ Expected<uint64_t> AddrOrErr = Sym.getValue();
+ if (!AddrOrErr)
+ // TODO: Test this error.
+ return AddrOrErr.takeError();
+
+ if (SymType.get() != SymbolRef::Type::ST_Function ||
+ !Gsym.IsValidTextAddress(*AddrOrErr) ||
+ Gsym.hasFunctionInfoForAddress(*AddrOrErr))
+ continue;
+ // Function size for MachO files will be 0
+ constexpr bool NoCopy = false;
+ const uint64_t size = IsELF ? ELFSymbolRef(Sym).getSize() : 0;
+ Expected<StringRef> Name = Sym.getName();
+ if (!Name) {
+ logAllUnhandledErrors(Name.takeError(), Log, "ObjectFileTransformer: ");
+ continue;
+ }
+ // Remove the leading '_' character in any symbol names if there is one
+ // for mach-o files.
+ if (IsMachO)
+ Name->consume_front("_");
+ Gsym.addFunctionInfo(
+ FunctionInfo(*AddrOrErr, size, Gsym.insertString(*Name, NoCopy)));
+ }
+ size_t FunctionsAddedCount = Gsym.getNumFunctionInfos() - NumBefore;
+ Log << "Loaded " << FunctionsAddedCount << " functions from symbol table.\n";
+ return Error::success();
+}
diff --git a/llvm/lib/DebugInfo/GSYM/Range.cpp b/llvm/lib/DebugInfo/GSYM/Range.cpp
index f78101e49bf8..044ddb8ba1ba 100644
--- a/llvm/lib/DebugInfo/GSYM/Range.cpp
+++ b/llvm/lib/DebugInfo/GSYM/Range.cpp
@@ -53,6 +53,16 @@ bool AddressRanges::contains(AddressRange Range) const {
return Range.End <= It[-1].End;
}
+Optional<AddressRange>
+AddressRanges::getRangeThatContains(uint64_t Addr) const {
+ auto It = std::partition_point(
+ Ranges.begin(), Ranges.end(),
+ [=](const AddressRange &R) { return R.Start <= Addr; });
+ if (It != Ranges.begin() && Addr < It[-1].End)
+ return It[-1];
+ return llvm::None;
+}
+
raw_ostream &llvm::gsym::operator<<(raw_ostream &OS, const AddressRange &R) {
return OS << '[' << HEX64(R.Start) << " - " << HEX64(R.End) << ")";
}
generated by cgit v1.2.3 (git 2.25.1) at 2025-07-09 15:54:55 +0000