[nfc] [lldb] Align `user_id_t` format to the current `DIERef` format

[jankratochvil]

Summary

Current user_id_t format is:
63{isDebugTypes} 62..32{dwo || 7fffffff} 31..0 {die_offset}
while current DIERef format is (I have made up the bit positions but the field widths do match):
63{m_section==isDebugTypes} 62{m_dwo_num_valid} 61..32{m_dwo_num} 31..0 {m_die_offset}

Proposing to change user_id_t to:
63{isDebugTypes} 62{dwo_is_valid} 61..32{dwo; 0 if !valid} 31..0 {die_offset}

There is no benefit of having 31-bits wide dwo_num in user_id_t when it gets converted to 30-bits width in DIERef.

This patch is for future DWZ patchset which extends the dwo_is_valid bit into a 2-bit field (normal, DWO, DWZ, DWZcommon) so that both user_id_t and DIERef can be changed then the same way.

It would be best to somehow unify user_id_t and DIERef but I do not plan to do that. user_id_t should probably remain a number for the Python API compatibility while there still needs to be some class with all the methods to access it.

SymbolFileDWARF::GetDwpSymbolFile() and SymbolFileDWARF::GetDIE use 0x3fffffff for DWP but that does not clash:
formerly:

• 31bits32..62:0x7fffffff = normal unit / not any DWO
• 31bits32..62:0x3fffffff = DWP
• 31bits32..62:others = DWO unit number

after this patch:

• bit62=0 30bits32..61:any = normal unit / not any DWO
• bit62=1 30bits32..61:0x3fffffff = DWP
• bit62=1 30bits32..61:others = DWO unit number

[labath]

Sounds good. I like how this makes the user ids small for the common case.

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

if (type_system != nullptr)
     type_system->GetDWARFParser()->EnsureAllDIEsInDeclContextHaveBeenParsed(
         decl_ctx);
 }
 
 user_id_t SymbolFileDWARF::GetUID(DIERef ref) {
   if (GetDebugMapSymfile())
     return GetID() | ref.die_offset();
 
-  return user_id_t(GetDwoNum().getValueOr(0x7fffffff)) << 32 |
-         ref.die_offset() |
+  lldbassert(GetDwoNum().getValueOr(0) <= 0x3fffffff);
+  return user_id_t(GetDwoNum().getValueOr(0)) << 32 | ref.die_offset() |
+         lldb::user_id_t(GetDwoNum().hasValue()) << 62 |
          lldb::user_id_t(ref.section() == DIERef::Section::DebugTypes) << 63;
 }
 
 llvm::Optional<SymbolFileDWARF::DecodedUID>
 SymbolFileDWARF::DecodeUID(lldb::user_id_t uid) {
   // This method can be called without going through the symbol vendor so we
   // need to lock the module.
   std::lock_guard<std::recursive_mutex> guard(GetModuleMutex());
SymbolFileDWARF::DecodeUID(lldb::user_id_t uid) {
   }
   dw_offset_t die_offset = uid;
   if (die_offset == DW_INVALID_OFFSET)
     return llvm::None;
 
   DIERef::Section section =
       uid >> 63 ? DIERef::Section::DebugTypes : DIERef::Section::DebugInfo;
 
-  llvm::Optional<uint32_t> dwo_num = uid >> 32 & 0x7fffffff;
-  if (*dwo_num == 0x7fffffff)
-    dwo_num = llvm::None;
+  llvm::Optional<uint32_t> dwo_num;
+  bool dwo_valid = uid >> 62 & 1;
+  if (dwo_valid)
+    dwo_num = uid >> 32 & 0x3fffffff;
 
   return DecodedUID{*this, {dwo_num, section, die_offset}};
 }
 
 DWARFDIE
 SymbolFileDWARF::GetDIE(lldb::user_id_t uid) {
   // This method can be called without going through the symbol vendor so we
   // need to lock the module.

lldb/test/Shell/SymbolFile/DWARF/DW_AT_low_pc-addrx.s

 # REQUIRES: x86
 
 # RUN: llvm-mc -triple=x86_64-pc-linux -filetype=obj %s > %t
 # RUN: %lldb %t -o "image lookup -v -s lookup_rnglists" \
 # RUN:   -o exit | FileCheck %s
 
 # CHECK-LABEL: image lookup -v -s lookup_rnglists
-# CHECK:  Function: id = {0x7fffffff0000002d}, name = "rnglists", range = [0x0000000000000000-0x0000000000000003)
-# CHECK:    Blocks: id = {0x7fffffff0000002d}, range = [0x00000000-0x00000003)
-# CHECK-NEXT:       id = {0x7fffffff00000043}, range = [0x00000001-0x00000002)
+# CHECK:  Function: id = {0x0000002d}, name = "rnglists", range = [0x0000000000000000-0x0000000000000003)
+# CHECK:    Blocks: id = {0x0000002d}, range = [0x00000000-0x00000003)
+# CHECK-NEXT:       id = {0x00000043}, range = [0x00000001-0x00000002)
 
         .text
 rnglists:
         nop
 .Lblock1_begin:
 lookup_rnglists:
         nop
 .Lblock1_end:

lldb/test/Shell/SymbolFile/DWARF/array-sizes.s

 # This tests a bug where we would incorrectly parse the size of an array if that
 # size happened to match an existing DIE offset. This happened because we
 # misinterpreted that value as a reference to a DIE specifying the VLA size even
 # though the form was a data form (as it should be).
 
 # REQUIRES: lld, x86
 
 # RUN: llvm-mc -triple x86_64-pc-linux %s -filetype=obj > %t.o
 # RUN: ld.lld %t.o -o %t
 # RUN: lldb-test symbols %t | FileCheck %s
 
-# CHECK: Variable{0x7fffffff0000001e}, name = "X"
-# CHECK-SAME: type = {7fffffff00000033} 0x{{[0-9A-F]*}} (char [56])
+# CHECK: Variable{0x0000001e}, name = "X"
+# CHECK-SAME: type = {0000000000000033} 0x{{[0-9A-F]*}} (char [56])
 
 
 # Generated from "char X[47];"
 # The array size was modified by hand.
 
   .text
   .file "-"
   .file 1 "/tmp" "<stdin>"

lldb/test/Shell/SymbolFile/DWARF/debug-types-address-ranges.s

 # Check address lookup works correctly in the presence of type units.
 # Specifically check that we don't use the line table pointed to by the
 # DW_AT_stmt_list of the type unit (which used only for the file names) to
 # compute address range for the type unit as type units don't describe any
 # addresses. The addresses should always resolve to the relevant compile units.
 
 # REQUIRES: lld, x86
 
 # RUN: llvm-mc -dwarf-version=5 -triple x86_64-pc-linux %s -filetype=obj >%t.o
 # RUN: ld.lld %t.o -o %t -image-base=0x47000 -z separate-code
 # RUN: %lldb %t -o "image lookup -a 0x48000 -v" -o exit | FileCheck %s
 
 # CHECK:   CompileUnit: id = {0x00000001}, file = "/tmp/a.cc", language = "c++"
-# CHECK:      Function: id = {0x7fffffff0000006a}, name = "::_start({{.*}})", range = [0x0000000000048000-0x000000000004800c)
+# CHECK:      Function: id = {0x0000006a}, name = "::_start({{.*}})", range = [0x0000000000048000-0x000000000004800c)
 # CHECK:     LineEntry: [0x0000000000048000-0x000000000004800a): /tmp/a.cc:4
 # CHECK:        Symbol: id = {0x00000002}, range = [0x0000000000048000-0x000000000004800c), name="_start"
-# CHECK:      Variable: id = {0x7fffffff00000075}, name = "v1", {{.*}} decl = a.cc:4
-# CHECK:      Variable: id = {0x7fffffff00000080}, name = "v2", {{.*}} decl = a.cc:4
+# CHECK:      Variable: id = {0x00000075}, name = "v1", {{.*}} decl = a.cc:4
+# CHECK:      Variable: id = {0x00000080}, name = "v2", {{.*}} decl = a.cc:4
 
 
 # Output generated via
 # clang -g -fdebug-types-section -gdwarf-5 -S
 # from
 # enum E1 { e1 };
 # enum E2 { e2 };
 # extern "C" void _start(E1 v1, E2 v2) {}

lldb/test/Shell/SymbolFile/DWARF/debug-types-dwarf5.s

 # RUN: llvm-mc -triple=x86_64-pc-linux -filetype=obj %s > %t
 # RUN: %lldb %t -o "image lookup -v -s f1" -o exit | FileCheck %s
 
-# CHECK:  Function: id = {0x7fffffff0000003c}, name = "f1", range = [0x0000000000000000-0x0000000000000001)
-# CHECK:    Blocks: id = {0x7fffffff0000003c}, range = [0x00000000-0x00000001)
+# CHECK:  Function: id = {0x0000003c}, name = "f1", range = [0x0000000000000000-0x0000000000000001)
+# CHECK:    Blocks: id = {0x0000003c}, range = [0x00000000-0x00000001)
 
 
         .text
         .globl  f1
         .type   f1,@function
 f1:
         nop
 .Lfunc_end0:

lldb/test/Shell/SymbolFile/DWARF/debug_ranges-missing-section.s

 # REQUIRES: x86
 
 # RUN: llvm-mc -triple=x86_64-pc-linux -filetype=obj %s > %t
 # RUN: %lldb %t -o "image lookup -v -s lookup_ranges" -o exit 2>&1 | FileCheck %s
 
 # CHECK: DIE has DW_AT_ranges(0x47) attribute, but range extraction failed (No debug_ranges section),
-# CHECK:  Function: id = {0x7fffffff0000001c}, name = "ranges", range = [0x0000000000000000-0x0000000000000004)
-# CHECK:    Blocks: id = {0x7fffffff0000001c}, range = [0x00000000-0x00000004)
+# CHECK:  Function: id = {0x0000001c}, name = "ranges", range = [0x0000000000000000-0x0000000000000004)
+# CHECK:    Blocks: id = {0x0000001c}, range = [0x00000000-0x00000004)
 
         .text
         .p2align 12
         .globl  ranges
         .type   ranges,@function
 ranges:                                    # @ranges
         nop
 lookup_ranges:

lldb/test/Shell/SymbolFile/DWARF/debug_ranges.s

 # REQUIRES: x86
 
 # RUN: llvm-mc -triple=x86_64-pc-linux -filetype=obj %s > %t
 # RUN: %lldb %t -o "image lookup -v -s lookup_ranges" -o exit | FileCheck %s
 
-# CHECK:  Function: id = {0x7fffffff0000002b}, name = "ranges", range = [0x0000000000000000-0x0000000000000004)
-# CHECK:    Blocks: id = {0x7fffffff0000002b}, range = [0x00000000-0x00000004)
-# CHECK-NEXT:       id = {0x7fffffff0000003f}, ranges = [0x00000001-0x00000002)[0x00000003-0x00000004)
+# CHECK:  Function: id = {0x0000002b}, name = "ranges", range = [0x0000000000000000-0x0000000000000004)
+# CHECK:    Blocks: id = {0x0000002b}, range = [0x00000000-0x00000004)
+# CHECK-NEXT:       id = {0x0000003f}, ranges = [0x00000001-0x00000002)[0x00000003-0x00000004)
 
         .text
         .p2align 12
 ranges:
         nop
 .Lblock1_begin:
 lookup_ranges:
         nop

lldb/test/Shell/SymbolFile/DWARF/debug_rnglists-dwo.s

 # REQUIRES: x86
 
 # RUN: cd %T
 # RUN: llvm-mc -triple=x86_64-pc-linux -filetype=obj %s >debug_rnglists-dwo.o
 # RUN: %lldb debug_rnglists-dwo.o -o "image lookup -v -s lookup_rnglists" \
 # RUN:   -o exit | FileCheck %s
 
 # CHECK-LABEL: image lookup -v -s lookup_rnglists
-# CHECK:  Function: id = {0x00000028}, name = "rnglists", range = [0x0000000000000000-0x0000000000000003)
-# CHECK:    Blocks: id = {0x00000028}, range = [0x00000000-0x00000003)
-# CHECK-NEXT:       id = {0x00000037}, range = [0x00000001-0x00000002)
+# CHECK:  Function: id = {0x4000000000000028}, name = "rnglists", range = [0x0000000000000000-0x0000000000000003)
+# CHECK:    Blocks: id = {0x4000000000000028}, range = [0x00000000-0x00000003)
+# CHECK-NEXT:       id = {0x4000000000000037}, range = [0x00000001-0x00000002)
 
         .text
 rnglists:
         nop
 .Lblock1_begin:
 lookup_rnglists:
         nop
 .Lblock1_end:

lldb/test/Shell/SymbolFile/DWARF/debug_rnglists.s

 # REQUIRES: x86
 
 # RUN: llvm-mc -triple=x86_64-pc-linux -filetype=obj %s > %t
 # RUN: %lldb %t -o "image lookup -v -s lookup_rnglists" \
 # RUN:   -o "image lookup -v -s lookup_rnglists2" -o exit | FileCheck %s
 
 # CHECK-LABEL: image lookup -v -s lookup_rnglists
-# CHECK:  Function: id = {0x7fffffff00000030}, name = "rnglists", range = [0x0000000000000000-0x0000000000000004)
-# CHECK:    Blocks: id = {0x7fffffff00000030}, range = [0x00000000-0x00000004)
-# CHECK-NEXT:       id = {0x7fffffff00000046}, ranges = [0x00000001-0x00000002)[0x00000003-0x00000004)
+# CHECK:  Function: id = {0x00000030}, name = "rnglists", range = [0x0000000000000000-0x0000000000000004)
+# CHECK:    Blocks: id = {0x00000030}, range = [0x00000000-0x00000004)
+# CHECK-NEXT:       id = {0x00000046}, ranges = [0x00000001-0x00000002)[0x00000003-0x00000004)
 
 # CHECK-LABEL: image lookup -v -s lookup_rnglists2
-# CHECK:  Function: id = {0x7fffffff0000007a}, name = "rnglists2", range = [0x0000000000000004-0x0000000000000007)
-# CHECK:    Blocks: id = {0x7fffffff0000007a}, range = [0x00000004-0x00000007)
-# CHECK-NEXT:       id = {0x7fffffff00000091}, range = [0x00000005-0x00000007)
+# CHECK:  Function: id = {0x0000007a}, name = "rnglists2", range = [0x0000000000000004-0x0000000000000007)
+# CHECK:    Blocks: id = {0x0000007a}, range = [0x00000004-0x00000007)
+# CHECK-NEXT:       id = {0x00000091}, range = [0x00000005-0x00000007)
 
         .text
         .p2align 12
 rnglists:
         nop
 .Lblock1_begin:
 lookup_rnglists:
         nop

lldb/test/Shell/SymbolFile/DWARF/dwarf5_locations.s

 # This tests that lldb is able to process DW_OP_addrx tags introduced in dwarf5.
 
 # REQUIRES: lld, x86
 
 # RUN: llvm-mc -g -dwarf-version=5 -triple x86_64-unknown-linux-gnu %s -filetype=obj > %t.o
 # RUN: ld.lld -m elf_x86_64 %t.o -o %t 
 # RUN: lldb-test symbols %t | FileCheck %s
 
-# CHECK: Variable{0x7fffffff00000011}, name = "color"
+# CHECK: Variable{0x00000011}, name = "color"
 # CHECK-SAME: location = DW_OP_addrx 0x0
 
         .text
         .section        .debug_str,"MS",@progbits,1
 .Lstr_offsets_base0:
         .asciz  "color"
 
         .section        .debug_str_offsets,"",@progbits

lldb/test/Shell/SymbolFile/DWARF/dwp-debug-types.s

 # CHECK-NEXT: }
 
 # CHECK-LABEL: target variable A
 # CHECK: (ENUM0) A = case0
 # CHECK: (ENUM1) A = case0
 
 # Make sure each entity is present in the index only once.
 # SYMBOLS:      Globals and statics:
-# SYMBOLS-DAG: 3fffffff/INFO/00000023 "A"
-# SYMBOLS-DAG: 3fffffff/INFO/0000005a "A"
+# SYMBOLS-DAG: INFO/00000023 "A"
+# SYMBOLS-DAG: INFO/0000005a "A"
 # SYMBOLS-EMPTY:
 
 # SYMBOLS: Types:
-# SYMBOLS-DAG: 3fffffff/TYPE/00000018 "ENUM0"
-# SYMBOLS-DAG: 3fffffff/TYPE/0000004d "ENUM1"
-# SYMBOLS-DAG: 3fffffff/TYPE/0000002d "int"
-# SYMBOLS-DAG: 3fffffff/TYPE/00000062 "int"
+# SYMBOLS-DAG: TYPE/00000018 "ENUM0"
+# SYMBOLS-DAG: TYPE/0000004d "ENUM1"
+# SYMBOLS-DAG: TYPE/0000002d "int"
+# SYMBOLS-DAG: TYPE/00000062 "int"
 # SYMBOLS-EMPTY:
 
 .ifdef MAIN
         .section        .debug_abbrev,"",@progbits
         .byte   1                       # Abbreviation Code
         .byte   17                      # DW_TAG_compile_unit
         .byte   0                       # DW_CHILDREN_no
         .ascii  "\260B"                 # DW_AT_GNU_dwo_name