[DWARFv5] Support DW_FORM_line_strp in llvm-dwarfdump.

This form is like DW_FORM_strp, but points to .debug_line_str instead
of .debug_str as the string section.  It's intended to be used from
the line-table header, and allows string-pooling of directory and
filenames across compilation units.

Differential Revision: https://reviews.llvm.org/D42553

llvm-svn: 323476

llvm/include/llvm/BinaryFormat/Dwarf.def

@@ -840,6 +840,7 @@ HANDLE_DWARF_SECTION(DebugAranges, ".debug_aranges", "debug-aranges")
 HANDLE_DWARF_SECTION(DebugInfo, ".debug_info", "debug-info")
 HANDLE_DWARF_SECTION(DebugTypes, ".debug_types", "debug-types")
 HANDLE_DWARF_SECTION(DebugLine, ".debug_line", "debug-line")
+HANDLE_DWARF_SECTION(DebugLineStr, ".debug_line_str", "debug-line-str")
 HANDLE_DWARF_SECTION(DebugLoc, ".debug_loc", "debug-loc")
 HANDLE_DWARF_SECTION(DebugFrame, ".debug_frame", "debug-frame")
 HANDLE_DWARF_SECTION(DebugMacro, ".debug_macro", "debug-macro")

llvm/include/llvm/DebugInfo/DWARF/DWARFCompileUnit.h

@@ -20,10 +20,11 @@ public:
   DWARFCompileUnit(DWARFContext &Context, const DWARFSection &Section,
                    const DWARFDebugAbbrev *DA, const DWARFSection *RS,
                    StringRef SS, const DWARFSection &SOS,
-                   const DWARFSection *AOS, const DWARFSection &LS, bool LE,
+                   const DWARFSection *AOS, const DWARFSection &LS,
-                   bool IsDWO, const DWARFUnitSectionBase &UnitSection,
+                   StringRef LSS, bool LE, bool IsDWO,
+                   const DWARFUnitSectionBase &UnitSection,
                    const DWARFUnitIndex::Entry *Entry)
-      : DWARFUnit(Context, Section, DA, RS, SS, SOS, AOS, LS, LE, IsDWO,
+      : DWARFUnit(Context, Section, DA, RS, SS, SOS, AOS, LS, LSS, LE, IsDWO,
                   UnitSection, Entry) {}
 
   // VTable anchor.

llvm/include/llvm/DebugInfo/DWARF/DWARFObject.h

@@ -42,6 +42,7 @@ public:
   virtual StringRef getDebugFrameSection() const { return ""; }
   virtual StringRef getEHFrameSection() const { return ""; }
   virtual const DWARFSection &getLineSection() const { return Dummy; }
+  virtual StringRef getLineStringSection() const { return ""; }
   virtual StringRef getStringSection() const { return ""; }
   virtual const DWARFSection &getRangeSection() const { return Dummy; }
   virtual StringRef getMacinfoSection() const { return ""; }

llvm/include/llvm/DebugInfo/DWARF/DWARFTypeUnit.h

@@ -32,10 +32,10 @@ public:
   DWARFTypeUnit(DWARFContext &Context, const DWARFSection &Section,
                 const DWARFDebugAbbrev *DA, const DWARFSection *RS,
                 StringRef SS, const DWARFSection &SOS, const DWARFSection *AOS,
-                const DWARFSection &LS, bool LE, bool IsDWO,
+                const DWARFSection &LS, StringRef LSS, bool LE, bool IsDWO,
                 const DWARFUnitSectionBase &UnitSection,
                 const DWARFUnitIndex::Entry *Entry)
-      : DWARFUnit(Context, Section, DA, RS, SS, SOS, AOS, LS, LE, IsDWO,
+      : DWARFUnit(Context, Section, DA, RS, SS, SOS, AOS, LS, LSS, LE, IsDWO,
                   UnitSection, Entry) {}
 
   uint32_t getHeaderSize() const override {

llvm/include/llvm/DebugInfo/DWARF/DWARFUnit.h

@@ -60,7 +60,8 @@ protected:
                          const DWARFDebugAbbrev *DA, const DWARFSection *RS,
                          StringRef SS, const DWARFSection &SOS,
                          const DWARFSection *AOS, const DWARFSection &LS,
-                         bool isLittleEndian, bool isDWO, bool Lazy) = 0;
+                         StringRef LSS, bool isLittleEndian, bool isDWO,
+                         bool Lazy) = 0;
 };
 
 const DWARFUnitIndex &getDWARFUnitIndex(DWARFContext &Context,
@@ -119,7 +120,7 @@ private:
   void parseImpl(DWARFContext &Context, const DWARFSection &Section,
                  const DWARFDebugAbbrev *DA, const DWARFSection *RS,
                  StringRef SS, const DWARFSection &SOS, const DWARFSection *AOS,
-                 const DWARFSection &LS, bool LE, bool IsDWO,
+                 const DWARFSection &LS, StringRef LSS, bool LE, bool IsDWO,
                  bool Lazy) override {
     if (Parsed)
       return;
@@ -133,7 +134,7 @@ private:
         if (!Data.isValidOffset(Offset))
           return nullptr;
         auto U = llvm::make_unique<UnitType>(
-            Context, Section, DA, RS, SS, SOS, AOS, LS, LE, IsDWO, *this,
+            Context, Section, DA, RS, SS, SOS, AOS, LS, LSS, LE, IsDWO, *this,
             Index ? Index->getFromOffset(Offset) : nullptr);
         if (!U->extract(Data, &Offset))
           return nullptr;
@@ -197,6 +198,7 @@ class DWARFUnit {
   const DWARFSection *RangeSection;
   uint32_t RangeSectionBase;
   const DWARFSection &LineSection;
+  StringRef LineStringSection;
   StringRef StringSection;
   const DWARFSection &StringOffsetSection;
   const DWARFSection *AddrOffsetSection;
@@ -293,7 +295,7 @@ public:
   DWARFUnit(DWARFContext &Context, const DWARFSection &Section,
             const DWARFDebugAbbrev *DA, const DWARFSection *RS, StringRef SS,
             const DWARFSection &SOS, const DWARFSection *AOS,
-            const DWARFSection &LS, bool LE, bool IsDWO,
+            const DWARFSection &LS, StringRef LSS, bool LE, bool IsDWO,
             const DWARFUnitSectionBase &UnitSection,
             const DWARFUnitIndex::Entry *IndexEntry = nullptr);
 
@@ -302,6 +304,7 @@ public:
   DWARFContext& getContext() const { return Context; }
 
   const DWARFSection &getLineSection() const { return LineSection; }
+  StringRef getLineStringSection() const { return LineStringSection; }
   StringRef getStringSection() const { return StringSection; }
   const DWARFSection &getStringOffsetSection() const {
     return StringOffsetSection;
@@ -322,11 +325,13 @@ public:
 
   DWARFDataExtractor getDebugInfoExtractor() const;
 
+  DataExtractor getLineStringExtractor() const {
+    return DataExtractor(LineStringSection, false, 0);
+  }
   DataExtractor getStringExtractor() const {
     return DataExtractor(StringSection, false, 0);
   }
 
-
   bool extract(DataExtractor debug_info, uint32_t* offset_ptr);
 
   /// extractRangeList - extracts the range list referenced by this compile

llvm/lib/DebugInfo/DWARF/DWARFContext.cpp

@@ -459,6 +459,16 @@ void DWARFContext::dump(
       strDWOOffset = offset;
     }
   }
+  if (shouldDump(Explicit, ".debug_line_str", DIDT_ID_DebugLineStr,
+                 DObj->getLineStringSection())) {
+    DataExtractor strData(DObj->getLineStringSection(), isLittleEndian(), 0);
+    uint32_t offset = 0;
+    uint32_t strOffset = 0;
+    while (const char *s = strData.getCStr(&offset)) {
+      OS << format("0x%8.8x: \"%s\"\n", strOffset, s);
+      strOffset = offset;
+    }
+  }
 
   if (shouldDump(Explicit, ".debug_ranges", DIDT_ID_DebugRanges,
                  DObj->getRangeSection().Data)) {
@@ -1193,6 +1203,7 @@ class DWARFObjInMemory final : public DWARFObject {
   StringRef CUIndexSection;
   StringRef GdbIndexSection;
   StringRef TUIndexSection;
+  StringRef LineStringSection;
 
   SmallVector<SmallString<32>, 4> UncompressedSections;
 
@@ -1215,6 +1226,7 @@ class DWARFObjInMemory final : public DWARFObject {
         .Case("debug_cu_index", &CUIndexSection)
         .Case("debug_tu_index", &TUIndexSection)
         .Case("gdb_index", &GdbIndexSection)
+        .Case("debug_line_str", &LineStringSection)
         // Any more debug info sections go here.
         .Default(nullptr);
   }
@@ -1433,6 +1445,7 @@ public:
   const DWARFSection &getStringOffsetSection() const override {
     return StringOffsetSection;
   }
+  StringRef getLineStringSection() const override { return LineStringSection; }
 
   // Sections for DWARF5 split dwarf proposal.
   const DWARFSection &getInfoDWOSection() const override {

llvm/lib/DebugInfo/DWARF/DWARFFormValue.cpp

@@ -264,6 +264,7 @@ bool DWARFFormValue::isFormClass(DWARFFormValue::FormClass FC) const {
   case DW_FORM_strx2:
   case DW_FORM_strx3:
   case DW_FORM_strx4:
+  case DW_FORM_line_strp:
     return (FC == FC_String);
   case DW_FORM_implicit_const:
     return (FC == FC_Constant);
@@ -582,6 +583,11 @@ Optional<const char *> DWARFFormValue::getAsCString() const {
   if (Form == DW_FORM_GNU_strp_alt || U == nullptr)
     return None;
   uint32_t Offset = Value.uval;
+  if (Form == DW_FORM_line_strp) {
+    if (const char *Str = U->getLineStringExtractor().getCStr(&Offset))
+      return Str;
+    return None;
+  }
   if (Form == DW_FORM_GNU_str_index || Form == DW_FORM_strx ||
       Form == DW_FORM_strx1 || Form == DW_FORM_strx2 || Form == DW_FORM_strx3 ||
       Form == DW_FORM_strx4) {

llvm/lib/DebugInfo/DWARF/DWARFUnit.cpp

@@ -34,8 +34,8 @@ void DWARFUnitSectionBase::parse(DWARFContext &C, const DWARFSection &Section) {
   const DWARFObject &D = C.getDWARFObj();
   parseImpl(C, Section, C.getDebugAbbrev(), &D.getRangeSection(),
             D.getStringSection(), D.getStringOffsetSection(),
-            &D.getAddrSection(), D.getLineSection(), D.isLittleEndian(), false,
+            &D.getAddrSection(), D.getLineSection(), D.getLineStringSection(),
-            false);
+            D.isLittleEndian(), false, false);
 }
 
 void DWARFUnitSectionBase::parseDWO(DWARFContext &C,
@@ -43,20 +43,21 @@ void DWARFUnitSectionBase::parseDWO(DWARFContext &C,
   const DWARFObject &D = C.getDWARFObj();
   parseImpl(C, DWOSection, C.getDebugAbbrevDWO(), &D.getRangeDWOSection(),
             D.getStringDWOSection(), D.getStringOffsetDWOSection(),
-            &D.getAddrSection(), D.getLineDWOSection(), C.isLittleEndian(),
+            &D.getAddrSection(), D.getLineDWOSection(), StringRef(),
-            true, Lazy);
+            C.isLittleEndian(), true, Lazy);
 }
 
 DWARFUnit::DWARFUnit(DWARFContext &DC, const DWARFSection &Section,
                      const DWARFDebugAbbrev *DA, const DWARFSection *RS,
                      StringRef SS, const DWARFSection &SOS,
-                     const DWARFSection *AOS, const DWARFSection &LS, bool LE,
+                     const DWARFSection *AOS, const DWARFSection &LS,
-                     bool IsDWO, const DWARFUnitSectionBase &UnitSection,
+                     StringRef LSS, bool LE, bool IsDWO,
+                     const DWARFUnitSectionBase &UnitSection,
                      const DWARFUnitIndex::Entry *IndexEntry)
     : Context(DC), InfoSection(Section), Abbrev(DA), RangeSection(RS),
-      LineSection(LS), StringSection(SS), StringOffsetSection(SOS),
+      LineSection(LS), LineStringSection(LSS), StringSection(SS),
-      AddrOffsetSection(AOS), isLittleEndian(LE), isDWO(IsDWO),
+      StringOffsetSection(SOS), AddrOffsetSection(AOS), isLittleEndian(LE),
-      UnitSection(UnitSection), IndexEntry(IndexEntry) {
+      isDWO(IsDWO), UnitSection(UnitSection), IndexEntry(IndexEntry) {
   clear();
 }
 

llvm/lib/DebugInfo/DWARF/DWARFVerifier.cpp

@@ -282,8 +282,8 @@ bool DWARFVerifier::handleDebugInfo() {
             DCtx, DObj.getInfoSection(), DCtx.getDebugAbbrev(),
             &DObj.getRangeSection(), DObj.getStringSection(),
             DObj.getStringOffsetSection(), &DObj.getAppleObjCSection(),
-            DObj.getLineSection(), DCtx.isLittleEndian(), false, TUSection,
+            DObj.getLineSection(), DObj.getLineStringSection(),
-            nullptr));
+            DCtx.isLittleEndian(), false, TUSection, nullptr));
         break;
       }
       case dwarf::DW_UT_skeleton:
@@ -297,8 +297,8 @@ bool DWARFVerifier::handleDebugInfo() {
             DCtx, DObj.getInfoSection(), DCtx.getDebugAbbrev(),
             &DObj.getRangeSection(), DObj.getStringSection(),
             DObj.getStringOffsetSection(), &DObj.getAppleObjCSection(),
-            DObj.getLineSection(), DCtx.isLittleEndian(), false, CUSection,
+            DObj.getLineSection(), DObj.getLineStringSection(),
-            nullptr));
+            DCtx.isLittleEndian(), false, CUSection, nullptr));
         break;
       }
       default: { llvm_unreachable("Invalid UnitType."); }

llvm/test/DebugInfo/X86/dwarfdump-header.s

@@ -300,18 +300,18 @@ LH_5_params:
         # File table format
         .byte   3               # Three elements per file entry
         .byte   1               # DW_LNCT_path
-        .byte   0x08            # DW_FORM_string
+        .byte   0x1f            # DW_FORM_line_strp (-> .debug_line_str)
         .byte   2               # DW_LNCT_directory_index
         .byte   0x0b            # DW_FORM_data1
         .byte   5               # DW_LNCT_MD5
         .byte   0x1e            # DW_FORM_data16
         # File table entries
         .byte   2               # Two files
-        .asciz "File5a"
+        .long   lstr_LT_5a
         .byte   0
         .quad   0x7766554433221100
         .quad   0xffeeddccbbaa9988
-        .asciz "File5b"
+        .long   lstr_LT_5b
         .byte   1
         .quad   0x8899aabbccddeeff
         .quad   0x0011223344556677
@@ -332,6 +332,12 @@ LH_5_end:
 # CHECK: file_names[  2]    1 ffeeddccbbaa99887766554433221100 File5b{{$}}
 # CHECK-NOT: file_names
 
+        .section .debug_line_str,"MS",@progbits,1
+lstr_LT_5a:
+        .asciz "File5a"
+lstr_LT_5b:
+        .asciz "File5b"
+
 	.section .debug_line.dwo,"",@progbits
 # CHECK-LABEL: .debug_line.dwo