Cherry-pick dsymutil DWARF 5 improvements

llvm/include/llvm/DWARFLinker/DWARFLinker.h

@@ -75,22 +75,22 @@ class AddressesMap {
 
 using Offset2UnitMap = DenseMap<uint64_t, CompileUnit *>;
 
-struct DebugAddrPool {
-  DenseMap<uint64_t, uint64_t> AddrIndexMap;
-  SmallVector<uint64_t> Addrs;
-
-  uint64_t getAddrIndex(uint64_t Addr) {
-    DenseMap<uint64_t, uint64_t>::iterator It = AddrIndexMap.find(Addr);
-    if (It == AddrIndexMap.end()) {
-      It = AddrIndexMap.insert(std::make_pair(Addr, Addrs.size())).first;
-      Addrs.push_back(Addr);
+struct DebugDieValuePool {
+  DenseMap<uint64_t, uint64_t> DieValueMap;
+  SmallVector<uint64_t> DieValues;
+
+  uint64_t getValueIndex(uint64_t Value) {
+    DenseMap<uint64_t, uint64_t>::iterator It = DieValueMap.find(Value);
+    if (It == DieValueMap.end()) {
+      It = DieValueMap.insert(std::make_pair(Value, DieValues.size())).first;
+      DieValues.push_back(Value);
     }
     return It->second;
   }
 
   void clear() {
-    AddrIndexMap.clear();
-    Addrs.clear();
+    DieValueMap.clear();
+    DieValues.clear();
   }
 };
 
@@ -110,6 +110,11 @@ class DwarfEmitter {
   /// Emit the string table described by \p Pool into .debug_str table.
   virtual void emitStrings(const NonRelocatableStringpool &Pool) = 0;
 
+  /// Emit the debug string offset table described by \p StringOffsets into the
+  /// .debug_str_offsets table.
+  virtual void emitStringOffsets(const SmallVector<uint64_t> &StringOffsets,
+                                 uint16_t TargetDWARFVersion) = 0;
+
   /// Emit the string table described by \p Pool into .debug_line_str table.
   virtual void emitLineStrings(const NonRelocatableStringpool &Pool) = 0;
 
@@ -137,10 +142,9 @@ class DwarfEmitter {
   virtual MCSymbol *emitDwarfDebugRangeListHeader(const CompileUnit &Unit) = 0;
 
   /// Emit debug ranges (.debug_ranges, .debug_rnglists) fragment.
-  virtual void
-  emitDwarfDebugRangeListFragment(const CompileUnit &Unit,
-                                  const AddressRanges &LinkedRanges,
-                                  PatchLocation Patch) = 0;
+  virtual void emitDwarfDebugRangeListFragment(
+      const CompileUnit &Unit, const AddressRanges &LinkedRanges,
+      PatchLocation Patch, DebugDieValuePool &AddrPool) = 0;
 
   /// Emit debug ranges (.debug_ranges, .debug_rnglists) footer.
   virtual void emitDwarfDebugRangeListFooter(const CompileUnit &Unit,
@@ -153,7 +157,7 @@ class DwarfEmitter {
   virtual void emitDwarfDebugLocListFragment(
       const CompileUnit &Unit,
       const DWARFLocationExpressionsVector &LinkedLocationExpression,
-      PatchLocation Patch, DebugAddrPool &AddrPool) = 0;
+      PatchLocation Patch, DebugDieValuePool &AddrPool) = 0;
 
   /// Emit debug locations (.debug_loc, .debug_loclists) footer.
   virtual void emitDwarfDebugLocListFooter(const CompileUnit &Unit,
@@ -604,6 +608,7 @@ class DWARFLinker {
                         DeclContextTree &ODRContexts,
                         OffsetsStringPool &DebugStrPool,
                         OffsetsStringPool &DebugLineStrPool,
+                        DebugDieValuePool &StringOffsetPool,
                         unsigned Indent = 0);
 
   unsigned shouldKeepDIE(AddressesMap &RelocMgr, const DWARFDie &DIE,
@@ -651,7 +656,8 @@ class DWARFLinker {
     DWARFFile &ObjFile;
     OffsetsStringPool &DebugStrPool;
     OffsetsStringPool &DebugLineStrPool;
-    DebugAddrPool AddrPool;
+    DebugDieValuePool &StringOffsetPool;
+    DebugDieValuePool AddrPool;
 
     /// Allocator used for all the DIEValue objects.
     BumpPtrAllocator &DIEAlloc;
@@ -669,10 +675,12 @@ class DWARFLinker {
               BumpPtrAllocator &DIEAlloc,
               std::vector<std::unique_ptr<CompileUnit>> &CompileUnits,
               bool Update, OffsetsStringPool &DebugStrPool,
-              OffsetsStringPool &DebugLineStrPool)
+              OffsetsStringPool &DebugLineStrPool,
+              DebugDieValuePool &StringOffsetPool)
         : Linker(Linker), Emitter(Emitter), ObjFile(ObjFile),
           DebugStrPool(DebugStrPool), DebugLineStrPool(DebugLineStrPool),
-          DIEAlloc(DIEAlloc), CompileUnits(CompileUnits), Update(Update) {}
+          StringOffsetPool(StringOffsetPool), DIEAlloc(DIEAlloc),
+          CompileUnits(CompileUnits), Update(Update) {}
 
     /// Recursively clone \p InputDIE into an tree of DIE objects
     /// where useless (as decided by lookForDIEsToKeep()) bits have been
@@ -733,6 +741,9 @@ class DWARFLinker {
       /// Is this DIE only a declaration?
       bool IsDeclaration = false;
 
+      /// Is there a DW_AT_str_offsets_base in the CU?
+      bool AttrStrOffsetBaseSeen = false;
+
       AttributesInfo() = default;
     };
 
@@ -821,7 +832,8 @@ class DWARFLinker {
 
   /// Compute and emit debug ranges(.debug_aranges, .debug_ranges,
   /// .debug_rnglists) for \p Unit, patch the attributes referencing it.
-  void generateUnitRanges(CompileUnit &Unit, const DWARFFile &File) const;
+  void generateUnitRanges(CompileUnit &Unit, const DWARFFile &File,
+                          DebugDieValuePool &AddrPool) const;
 
   /// Emit the accelerator entries for \p Unit.
   void emitAcceleratorEntriesForUnit(CompileUnit &Unit);

llvm/include/llvm/DWARFLinker/DWARFStreamer.h

@@ -78,6 +78,11 @@ class DwarfStreamer : public DwarfEmitter {
   /// Emit the string table described by \p Pool into .debug_str table.
   void emitStrings(const NonRelocatableStringpool &Pool) override;
 
+  /// Emit the debug string offset table described by \p StringOffsets into the
+  /// .debug_str_offsets table.
+  void emitStringOffsets(const SmallVector<uint64_t> &StringOffset,
+                         uint16_t TargetDWARFVersion) override;
+
   /// Emit the string table described by \p Pool into .debug_line_str table.
   void emitLineStrings(const NonRelocatableStringpool &Pool) override;
 
@@ -95,7 +100,8 @@ class DwarfStreamer : public DwarfEmitter {
   /// Emit debug ranges(.debug_ranges, .debug_rnglists) fragment.
   void emitDwarfDebugRangeListFragment(const CompileUnit &Unit,
                                        const AddressRanges &LinkedRanges,
-                                       PatchLocation Patch) override;
+                                       PatchLocation Patch,
+                                       DebugDieValuePool &AddrPool) override;
 
   /// Emit debug ranges(.debug_ranges, .debug_rnglists) footer.
   void emitDwarfDebugRangeListFooter(const CompileUnit &Unit,
@@ -119,7 +125,7 @@ class DwarfStreamer : public DwarfEmitter {
   void emitDwarfDebugLocListFragment(
       const CompileUnit &Unit,
       const DWARFLocationExpressionsVector &LinkedLocationExpression,
-      PatchLocation Patch, DebugAddrPool &AddrPool) override;
+      PatchLocation Patch, DebugDieValuePool &AddrPool) override;
 
   /// Emit debug ranges(.debug_loc, .debug_loclists) footer.
   void emitDwarfDebugLocListFooter(const CompileUnit &Unit,
@@ -217,7 +223,8 @@ class DwarfStreamer : public DwarfEmitter {
   /// Emit piece of .debug_rnglists for \p LinkedRanges.
   void emitDwarfDebugRngListsTableFragment(const CompileUnit &Unit,
                                            const AddressRanges &LinkedRanges,
-                                           PatchLocation Patch);
+                                           PatchLocation Patch,
+                                           DebugDieValuePool &AddrPool);
 
   /// Emit piece of .debug_loc for \p LinkedRanges.
   void emitDwarfDebugLocTableFragment(
@@ -229,7 +236,7 @@ class DwarfStreamer : public DwarfEmitter {
   void emitDwarfDebugLocListsTableFragment(
       const CompileUnit &Unit,
       const DWARFLocationExpressionsVector &LinkedLocationExpression,
-      PatchLocation Patch, DebugAddrPool &AddrPool);
+      PatchLocation Patch, DebugDieValuePool &AddrPool);
 
   /// \defgroup Line table emission
   /// @{
@@ -288,6 +295,7 @@ class DwarfStreamer : public DwarfEmitter {
   uint64_t MacInfoSectionSize = 0;
   uint64_t MacroSectionSize = 0;
   uint64_t AddrSectionSize = 0;
+  uint64_t StrOffsetSectionSize = 0;
 
   /// Keep track of emitted CUs and their Unique ID.
   struct EmittedUnit {

llvm/lib/DWARFLinker/DWARFLinker.cpp

@@ -1014,32 +1014,36 @@ void DWARFLinker::assignAbbrev(DIEAbbrev &Abbrev) {
 unsigned DWARFLinker::DIECloner::cloneStringAttribute(DIE &Die,
                                                       AttributeSpec AttrSpec,
                                                       const DWARFFormValue &Val,
-                                                      const DWARFUnit &,
+                                                      const DWARFUnit &U,
                                                       AttributesInfo &Info) {
   std::optional<const char *> String = dwarf::toString(Val);
   if (!String)
     return 0;
-
   DwarfStringPoolEntryRef StringEntry;
   if (AttrSpec.Form == dwarf::DW_FORM_line_strp) {
     StringEntry = DebugLineStrPool.getEntry(*String);
   } else {
     StringEntry = DebugStrPool.getEntry(*String);
-
     // Update attributes info.
     if (AttrSpec.Attr == dwarf::DW_AT_name)
       Info.Name = StringEntry;
     else if (AttrSpec.Attr == dwarf::DW_AT_MIPS_linkage_name ||
              AttrSpec.Attr == dwarf::DW_AT_linkage_name)
       Info.MangledName = StringEntry;
-
+    if (U.getVersion() >= 5) {
+      // Switch everything to DW_FORM_strx strings.
+      auto StringOffsetIndex =
+          StringOffsetPool.getValueIndex(StringEntry.getOffset());
+      return Die
+          .addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr),
+                    dwarf::DW_FORM_strx, DIEInteger(StringOffsetIndex))
+          ->sizeOf(U.getFormParams());
+    }
     // Switch everything to out of line strings.
     AttrSpec.Form = dwarf::DW_FORM_strp;
   }
-
   Die.addValue(DIEAlloc, dwarf::Attribute(AttrSpec.Attr), AttrSpec.Form,
                DIEInteger(StringEntry.getOffset()));
-
   return 4;
 }
 
@@ -1359,7 +1363,7 @@ unsigned DWARFLinker::DIECloner::cloneAddressAttribute(
     return Unit.getOrigUnit().getAddressByteSize();
   }
 
-  auto AddrIndex = AddrPool.getAddrIndex(*Addr);
+  auto AddrIndex = AddrPool.getValueIndex(*Addr);
 
   return Die
       .addValue(DIEAlloc, static_cast<dwarf::Attribute>(AttrSpec.Attr),
@@ -1391,6 +1395,17 @@ unsigned DWARFLinker::DIECloner::cloneScalarAttribute(
     }
   }
 
+  if (AttrSpec.Attr == dwarf::DW_AT_str_offsets_base) {
+    // DWARFLinker generates common .debug_str_offsets table used for all
+    // compile units. The offset to the common .debug_str_offsets table is 8 on
+    // DWARF32.
+    Info.AttrStrOffsetBaseSeen = true;
+    return Die
+        .addValue(DIEAlloc, dwarf::DW_AT_str_offsets_base,
+                  dwarf::DW_FORM_sec_offset, DIEInteger(8))
+        ->sizeOf(Unit.getOrigUnit().getFormParams());
+  }
+
   if (LLVM_UNLIKELY(Linker.Options.Update)) {
     if (auto OptionalValue = Val.getAsUnsignedConstant())
       Value = *OptionalValue;
@@ -1608,9 +1623,6 @@ shouldSkipAttribute(bool Update,
     // Since DW_AT_rnglists_base is used for only DW_FORM_rnglistx the
     // DW_AT_rnglists_base is removed.
     return !Update;
-  case dwarf::DW_AT_str_offsets_base:
-    // FIXME: Use the string offset table with Dwarf 5.
-    return true;
   case dwarf::DW_AT_loclists_base:
     // In case !Update the .debug_addr table is not generated/preserved.
     // Thus instead of DW_FORM_loclistx the DW_FORM_sec_offset is used.
@@ -1777,6 +1789,14 @@ DIE *DWARFLinker::DIECloner::cloneDIE(const DWARFDie &InputDIE,
     }
   }
 
+  if (Unit.getOrigUnit().getVersion() >= 5 && !AttrInfo.AttrStrOffsetBaseSeen &&
+      Die->getTag() == dwarf::DW_TAG_compile_unit) {
+    // No DW_AT_str_offsets_base seen, add it to the DIE.
+    Die->addValue(DIEAlloc, dwarf::DW_AT_str_offsets_base,
+                  dwarf::DW_FORM_sec_offset, DIEInteger(8));
+    OutOffset += 4;
+  }
+
   DIEAbbrev NewAbbrev = Die->generateAbbrev();
   if (HasChildren)
     NewAbbrev.setChildrenFlag(dwarf::DW_CHILDREN_yes);
@@ -1812,8 +1832,8 @@ DIE *DWARFLinker::DIECloner::cloneDIE(const DWARFDie &InputDIE,
 /// Patch the input object file relevant debug_ranges or debug_rnglists
 /// entries and emit them in the output file. Update the relevant attributes
 /// to point at the new entries.
-void DWARFLinker::generateUnitRanges(CompileUnit &Unit,
-                                     const DWARFFile &File) const {
+void DWARFLinker::generateUnitRanges(CompileUnit &Unit, const DWARFFile &File,
+                                     DebugDieValuePool &AddrPool) const {
   if (LLVM_UNLIKELY(Options.Update))
     return;
 
@@ -1866,14 +1886,14 @@ void DWARFLinker::generateUnitRanges(CompileUnit &Unit,
       }
 
       // Emit linked ranges.
-      TheDwarfEmitter->emitDwarfDebugRangeListFragment(Unit, LinkedRanges,
-                                                       AttributePatch);
+      TheDwarfEmitter->emitDwarfDebugRangeListFragment(
+          Unit, LinkedRanges, AttributePatch, AddrPool);
     }
 
     // Emit ranges for Unit AT_ranges attribute.
     if (UnitRngListAttribute.has_value())
       TheDwarfEmitter->emitDwarfDebugRangeListFragment(
-          Unit, LinkedFunctionRanges, *UnitRngListAttribute);
+          Unit, LinkedFunctionRanges, *UnitRngListAttribute, AddrPool);
 
     // Emit ranges footer.
     TheDwarfEmitter->emitDwarfDebugRangeListFooter(Unit, EndLabel);
@@ -1955,13 +1975,14 @@ void DWARFLinker::DIECloner::emitDebugAddrSection(
   if (DwarfVersion < 5)
     return;
 
-  if (AddrPool.Addrs.empty())
+  if (AddrPool.DieValues.empty())
     return;
 
   MCSymbol *EndLabel = Emitter->emitDwarfDebugAddrsHeader(Unit);
   patchAddrBase(*Unit.getOutputUnitDIE(),
                 DIEInteger(Emitter->getDebugAddrSectionSize()));
-  Emitter->emitDwarfDebugAddrs(AddrPool.Addrs, Unit.getOrigUnit().getAddressByteSize());
+  Emitter->emitDwarfDebugAddrs(AddrPool.DieValues,
+                               Unit.getOrigUnit().getAddressByteSize());
   Emitter->emitDwarfDebugAddrsFooter(Unit, EndLabel);
 }
 
@@ -2515,7 +2536,7 @@ uint64_t DWARFLinker::DIECloner::cloneAllCompileUnits(
       if (LLVM_UNLIKELY(Linker.Options.Update))
         continue;
 
-      Linker.generateUnitRanges(*CurrentUnit, File);
+      Linker.generateUnitRanges(*CurrentUnit, File, AddrPool);
 
       auto ProcessExpr = [&](SmallVectorImpl<uint8_t> &SrcBytes,
                              SmallVectorImpl<uint8_t> &OutBytes,
@@ -2612,6 +2633,7 @@ Error DWARFLinker::link() {
   // reproducibility.
   OffsetsStringPool DebugStrPool(StringsTranslator, true);
   OffsetsStringPool DebugLineStrPool(StringsTranslator, false);
+  DebugDieValuePool StringOffsetPool;
 
   // ODR Contexts for the optimize.
   DeclContextTree ODRContexts;
@@ -2675,7 +2697,7 @@ Error DWARFLinker::link() {
 
     for (auto &CU : OptContext.ModuleUnits) {
       if (Error Err = cloneModuleUnit(OptContext, CU, ODRContexts, DebugStrPool,
-                                      DebugLineStrPool))
+                                      DebugLineStrPool, StringOffsetPool))
         reportWarning(toString(std::move(Err)), CU.File);
     }
   }
@@ -2772,7 +2794,7 @@ Error DWARFLinker::link() {
       SizeByObject[OptContext.File.FileName].Output =
           DIECloner(*this, TheDwarfEmitter.get(), OptContext.File, DIEAlloc,
                     OptContext.CompileUnits, Options.Update, DebugStrPool,
-                    DebugLineStrPool)
+                    DebugLineStrPool, StringOffsetPool)
               .cloneAllCompileUnits(*OptContext.File.Dwarf, OptContext.File,
                                     OptContext.File.Dwarf->isLittleEndian());
     }
@@ -2789,6 +2811,8 @@ Error DWARFLinker::link() {
     if (TheDwarfEmitter != nullptr) {
       TheDwarfEmitter->emitAbbrevs(Abbreviations, Options.TargetDWARFVersion);
       TheDwarfEmitter->emitStrings(DebugStrPool);
+      TheDwarfEmitter->emitStringOffsets(StringOffsetPool.DieValues,
+                                         Options.TargetDWARFVersion);
       TheDwarfEmitter->emitLineStrings(DebugLineStrPool);
       for (AccelTableKind TableKind : Options.AccelTables) {
         switch (TableKind) {
@@ -2905,6 +2929,7 @@ Error DWARFLinker::cloneModuleUnit(LinkContext &Context, RefModuleUnit &Unit,
                                    DeclContextTree &ODRContexts,
                                    OffsetsStringPool &DebugStrPool,
                                    OffsetsStringPool &DebugLineStrPool,
+                                   DebugDieValuePool &StringOffsetPool,
                                    unsigned Indent) {
   assert(Unit.Unit.get() != nullptr);
 
@@ -2931,7 +2956,7 @@ Error DWARFLinker::cloneModuleUnit(LinkContext &Context, RefModuleUnit &Unit,
   CompileUnits.emplace_back(std::move(Unit.Unit));
   assert(TheDwarfEmitter);
   DIECloner(*this, TheDwarfEmitter.get(), Unit.File, DIEAlloc, CompileUnits,
-            Options.Update, DebugStrPool, DebugLineStrPool)
+            Options.Update, DebugStrPool, DebugLineStrPool, StringOffsetPool)
       .cloneAllCompileUnits(*Unit.File.Dwarf, Unit.File,
                             Unit.File.Dwarf->isLittleEndian());
   return Error::success();