[lldb] Add the ability to load DWARF64 .debug_str_offsets tables for DWARF32 DWARF units in .dwp files in LLDB.

lldb/source/Plugins/SymbolFile/DWARF/DWARFUnit.cpp

@@ -348,6 +348,10 @@ void DWARFUnit::ExtractDIEsRWLocked() {
 void DWARFUnit::SetDwoStrOffsetsBase() {
   lldb::offset_t baseOffset = 0;
 
+  // Size of offset for .debug_str_offsets is same as DWARF offset byte size
+  // of the DWARFUnit as a default. We might override this if below if needed.
+  m_str_offset_size = m_header.getDwarfOffsetByteSize();
+
   if (const llvm::DWARFUnitIndex::Entry *entry = m_header.getIndexEntry()) {
     if (const auto *contribution =
             entry->getContribution(llvm::DW_SECT_STR_OFFSETS))
@@ -357,14 +361,17 @@ void DWARFUnit::SetDwoStrOffsetsBase() {
   }
 
   if (GetVersion() >= 5) {
-    const DWARFDataExtractor &strOffsets =
-        GetSymbolFileDWARF().GetDWARFContext().getOrLoadStrOffsetsData();
-    uint64_t length = strOffsets.GetU32(&baseOffset);
-    if (length == 0xffffffff)
-      length = strOffsets.GetU64(&baseOffset);
-
+    const llvm::DWARFDataExtractor &strOffsets = GetSymbolFileDWARF()
+                                                     .GetDWARFContext()
+                                                     .getOrLoadStrOffsetsData()
+                                                     .GetAsLLVMDWARF();
+
+    uint64_t length;
+    llvm::dwarf::DwarfFormat format;
+    std::tie(length, format) = strOffsets.getInitialLength(&baseOffset);
+    m_str_offset_size = format == llvm::dwarf::DwarfFormat::DWARF64 ? 8 : 4;
     // Check version.
-    if (strOffsets.GetU16(&baseOffset) < 5)
+    if (strOffsets.getU16(&baseOffset) < 5)
       return;
 
     // Skip padding.
@@ -409,7 +416,16 @@ void DWARFUnit::AddUnitDIE(const DWARFDebugInfoEntry &cu_die) {
       SetRangesBase(form_value.Unsigned());
       break;
     case DW_AT_str_offsets_base:
+      // When we have a DW_AT_str_offsets_base attribute, it points us to the
+      // first string offset for this DWARFUnit which is after the string
+      // offsets table header. In this case we use the DWARF32/DWARF64 of the
+      // DWARFUnit to determine the string offset byte size. DWO files do not
+      // use this attribute and they point to the start of the string offsets
+      // table header which can be used to determine the DWARF32/DWARF64 status
+      // of the string table. See SetDwoStrOffsetsBase() for now it figures out
+      // the m_str_offset_size value that should be used.
       SetStrOffsetsBase(form_value.Unsigned());
+      m_str_offset_size = m_header.getDwarfOffsetByteSize();
       break;
     case DW_AT_low_pc:
       SetBaseAddress(form_value.Address());
@@ -1079,10 +1095,9 @@ uint32_t DWARFUnit::GetHeaderByteSize() const { return m_header.getSize(); }
 
 std::optional<uint64_t>
 DWARFUnit::GetStringOffsetSectionItem(uint32_t index) const {
-  lldb::offset_t offset =
-      GetStrOffsetsBase() + index * m_header.getDwarfOffsetByteSize();
+  lldb::offset_t offset = GetStrOffsetsBase() + index * m_str_offset_size;
   return m_dwarf.GetDWARFContext().getOrLoadStrOffsetsData().GetMaxU64(
-      &offset, m_header.getDwarfOffsetByteSize());
+      &offset, m_str_offset_size);
 }
 
 llvm::Expected<llvm::DWARFAddressRangesVector>

lldb/source/Plugins/SymbolFile/DWARF/DWARFUnit.h

@@ -364,6 +364,7 @@ class DWARFUnit : public DWARFExpression::Delegate, public UserID {
   dw_offset_t m_line_table_offset = DW_INVALID_OFFSET;
 
   dw_offset_t m_str_offsets_base = 0; // Value of DW_AT_str_offsets_base.
+  dw_offset_t m_str_offset_size = 4;  // Size in bytes of a string offset.
 
   std::optional<llvm::DWARFDebugRnglistTable> m_rnglist_table;
   bool m_rnglist_table_done = false;

lldb/test/Shell/SymbolFile/DWARF/Inputs/dwp-str-offsets-dwarf64-dwp.yaml

@@ -0,0 +1,44 @@
+--- !ELF
+FileHeader:
+  Class:           ELFCLASS64
+  Data:            ELFDATA2LSB
+  Type:            ET_REL
+  Machine:         EM_X86_64
+  SectionHeaderStringTable: .strtab
+Sections:
+  - Name:            .debug_abbrev.dwo
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_EXCLUDE ]
+    AddressAlign:    0x1
+    Content:         01110125251305032576250000022E01111B1206401803253A0B3B0B49133F190000030500021803253A0B3B0B4913000004240003253E0B0B0B0000050F00491300000626004913000000
+  - Name:            .debug_str.dwo
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_EXCLUDE, SHF_MERGE, SHF_STRINGS ]
+    AddressAlign:    0x1
+    EntSize:         0x1
+    Content:         6D61696E00696E74006172676300617267760063686172004170706C6520636C616E672076657273696F6E2031372E302E302028636C616E672D313730302E342E342E3129006D61696E2E6D696E696D616C2E637070006D61696E2E6D696E696D616C2E64776F00
+  - Name:            .debug_str_offsets.dwo
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_EXCLUDE ]
+    AddressAlign:    0x1
+    Content:         'FFFFFFFF4400000000000000050000000000000000000000050000000000000009000000000000000E000000000000001300000000000000180000000000000046000000000000005700000000000000'
+  - Name:            .debug_info.dwo
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_EXCLUDE ]
+    AddressAlign:    0x1
+    Content:         54000000050005080000000099E97383BBC6980B0105210006070200160000000156000001400000000302917802000140000000030291700300014400000000040105040549000000054E00000006530000000404060100
+  - Name:            .debug_cu_index
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         05000000030000000100000002000000000000000000000099E97383BBC6980B0000000001000000010000000300000006000000000000000000000000000000580000004B00000028000000
+  - Type:            SectionHeaderTable
+    Sections:
+      - Name:            .strtab
+      - Name:            .debug_abbrev.dwo
+      - Name:            .debug_str.dwo
+      - Name:            .debug_str_offsets.dwo
+      - Name:            .debug_info.dwo
+      - Name:            .debug_cu_index
+      - Name:            .symtab
+Symbols:         []
+...

lldb/test/Shell/SymbolFile/DWARF/Inputs/dwp-str-offsets-dwarf64-exe.yaml

@@ -0,0 +1,100 @@
+--- !ELF
+FileHeader:
+  Class:           ELFCLASS64
+  Data:            ELFDATA2LSB
+  Type:            ET_EXEC
+  Machine:         EM_X86_64
+ProgramHeaders:
+  - Type:            PT_PHDR
+    Flags:           [ PF_R ]
+    VAddr:           0x200040
+    Align:           0x8
+    Offset:          0x40
+  - Type:            PT_LOAD
+    Flags:           [ PF_R ]
+    FirstSec:        .eh_frame
+    LastSec:         .eh_frame
+    VAddr:           0x200000
+    Align:           0x1000
+    Offset:          0x0
+  - Type:            PT_LOAD
+    Flags:           [ PF_X, PF_R ]
+    FirstSec:        .text
+    LastSec:         .text
+    VAddr:           0x201160
+    Align:           0x1000
+    Offset:          0x160
+  - Type:            PT_GNU_STACK
+    Flags:           [ PF_W, PF_R ]
+    Align:           0x0
+    Offset:          0x0
+Sections:
+  - Name:            .eh_frame
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_ALLOC ]
+    Address:         0x200120
+    AddressAlign:    0x8
+    Content:         1400000000000000017A5200017810011B0C0708900100001C0000001C000000201000001600000000410E108602430D06510C070800000000000000
+  - Name:            .text
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_ALLOC, SHF_EXECINSTR ]
+    Address:         0x201160
+    AddressAlign:    0x10
+    Content:         554889E5C745FC00000000897DF8488975F031C05DC3
+  - Name:            .debug_abbrev
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         014A00101772171B25B442197625111B12067317000000
+  - Name:            .debug_info
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         24000000050004080000000099E97383BBC6980B0100000000080000000001001600000008000000
+  - Name:            .debug_str_offsets
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         0C000000050000000000000002000000
+  - Name:            .debug_gnu_pubnames
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         18000000020000000000280000001A000000306D61696E0000000000
+  - Name:            .debug_gnu_pubtypes
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         '21000000020000000000280000004000000090696E74005300000090636861720000000000'
+  - Name:            .comment
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_MERGE, SHF_STRINGS ]
+    AddressAlign:    0x1
+    EntSize:         0x1
+    Content:         004170706C6520636C616E672076657273696F6E2031372E302E302028636C616E672D313730302E342E342E3129004C696E6B65723A204C4C442032322E302E30202868747470733A2F2F6769746875622E636F6D2F636C6179626F72672F6C6C766D2D70726F6A6563742E67697420613234333130363863303837656463303938393330303934343864343162356138336361303363392900
+  - Name:            .debug_line
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x1
+    Content:         5A0000000500080037000000010101FB0E0D00010101010000000100000101011F010000000003011F020F051E0102000000004E0649A10A56ED4D381C49A3DB4F6825040000090260112000000000000105030A0821060B2E0202000101
+  - Name:            .debug_line_str
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_MERGE, SHF_STRINGS ]
+    AddressAlign:    0x1
+    EntSize:         0x1
+    Content:         2E006D61696E2E6D696E696D616C2E63707000
+Symbols:
+  - Name:            main.minimal.cpp
+    Type:            STT_FILE
+    Index:           SHN_ABS
+  - Name:            main
+    Type:            STT_FUNC
+    Section:         .text
+    Binding:         STB_GLOBAL
+    Value:           0x201160
+    Size:            0x16
+DWARF:
+  debug_str:
+    - .
+    - main.minimal.dwo
+  debug_addr:
+    - Length:          0xC
+      Version:         0x5
+      AddressSize:     0x8
+      Entries:
+        - Address:         0x201160
+...

lldb/test/Shell/SymbolFile/DWARF/dwp-str-offsets-dwarf64.test

@@ -0,0 +1,25 @@
+# This test verifies that LLDB can read a .dwp file that has a DWARF32 compile
+# unit that has a DWARF64 .debug_str_offsets table. This will be needed for
+# llvm-dwp changes that will start emitting 64 bit line tables for compile units
+# that have strings in the .debug_str section whose string offsets exceed 4GB.
+# By having a 64 bit .debug_str_offsets table, we can emit .debug_str sections
+# with no size limits.
+
+# RUN: yaml2obj %S/Inputs/dwp-str-offsets-dwarf64-exe.yaml > %t
+# RUN: yaml2obj %S/Inputs/dwp-str-offsets-dwarf64-dwp.yaml > %t.dwp
+# RUN: %lldb %t -b \
+# RUN:      -o 'image lookup --verbose --address 0x0000000000201172' | \
+# RUN:      FileCheck %s
+
+# CHECK:      (lldb) image lookup --verbose --address 0x0000000000201172
+# CHECK-NEXT:      Address: dwp-str-offsets-dwarf64.test.tmp[0x0000000000201172] (dwp-str-offsets-dwarf64.test.tmp.PT_LOAD[1]..text + 18)
+# CHECK-NEXT:      Summary: dwp-str-offsets-dwarf64.test.tmp`main + 18 at main.minimal.cpp:2:3
+# CHECK-NEXT:       Module: file = "{{.*}}/dwp-str-offsets-dwarf64.test.tmp", arch = "x86_64"
+# CHECK-NEXT:  CompileUnit: id = {0x00000000}, file = "main.minimal.cpp", language = "<not loaded>"
+# CHECK-NEXT:     Function: id = {0x7fffff000000001a}, name = "main", range = [0x0000000000201160-0x0000000000201176)
+# CHECK-NEXT:     FuncType: id = {0x7fffff000000001a}, byte-size = 0, decl = main.minimal.cpp:1, compiler_type = "int (int, const char **)"
+# CHECK-NEXT:       Blocks: id = {0x7fffff000000001a}, range = [0x00201160-0x00201176)
+# CHECK-NEXT:    LineEntry: [0x0000000000201172-0x0000000000201174): main.minimal.cpp:2:3
+# CHECK-NEXT:       Symbol: id = {0x00000002}, range = [0x0000000000201160-0x0000000000201176), name="main"
+# CHECK-NEXT:     Variable: id = {0x7fffff0000000029}, name = "argc", type = "int", valid ranges = <block>, location = DW_OP_fbreg -8, decl = main.minimal.cpp:1
+# CHECK-NEXT:     Variable: id = {0x7fffff0000000034}, name = "argv", type = "const char **", valid ranges = <block>, location = DW_OP_fbreg -16, decl = main.minimal.cpp:1