CodeView Symbol Records

Introduction

This document describes the usage and serialization format of the various CodeView symbol records that LLVM understands. Like CodeView Type Records, we describe only the important types which are generated by modern C++ toolchains.

Record Categories

Symbol records share one major similarity with type records: They start with the same record prefix, which we will not describe again (refer to the previous link for a description). As a result of this, a sequence of symbol records can be processed with largely the same code as that which processes type records. There are several important differences between symbol and type records:

  • Symbol records only appear in the The PDB Public Symbol Stream, The PDB Global Symbol Stream, and Module Info Streams.

  • Type records only appear in the TPI & IPI streams.

  • While types are referenced from other CodeView records via type indices, symbol records are referenced by the byte offset of the record in the stream that it appears in.

  • Types can reference types (via type indices), and symbols can reference both types (via type indices) and symbols (via offsets), but types can never reference symbols.

  • There is no notion of Leaf Records and Member Records as there are with types. Every symbol record describes is own length.

  • Certain special symbol records begin a “scope”. For these records, all following records up until the next S_END record are “children” of this symbol record. For example, given a symbol record which describes a certain function, all local variables of this function would appear following the function up until the corresponding S_END record.

Finally, there are three general categories of symbol record, grouped by where they are legal to appear in a PDB file. Public Symbols (which appear only in the publics stream), Global Symbols (which appear only in the globals stream) and module symbols (which appear in the module info stream).

Public Symbols

Public symbols are the CodeView equivalent of DWARF .debug_pubnames. There is one public symbol record for every function or variable in the program that has a mangled name. The Publics Stream, which contains these records, additionally contains a hash table that allows one to quickly locate a record by mangled name.

S_PUB32 (0x110e)

There is only type of public symbol, an S_PUB32 which describes a mangled name, a flag indicating what kind of symbol it is (e.g. function, variable), and the symbol’s address. The Section Map Substream of the DBI Stream can be consulted to determine what module this address corresponds to, and from there that module’s module debug stream can be consulted to locate full information for the symbol with the given address.

Global Symbols

While there is one public symbol for every symbol in the program with external linkage, there is one global symbol for every symbol in the program with linkage (including internal linkage). As a result, global symbols do not describe a mangled name or an address, since symbols with internal linkage need not have any mangling at all, and also may not have an address. Thus, all global symbols simply refer directly to the full symbol record via a module/offset combination.

Similarly to public symbols, all global symbols are contained in a single Globals Stream, which contains a hash table mapping fully qualified name to the corresponding record in the globals stream (which as mentioned, then contains information allowing one to locate the full record in the corresponding module symbol stream).

Note that a consequence and limitation of this design is that program-wide lookup by anything other than an exact textually matching fully-qualified name of whatever the compiler decided to emit is impractical. This differs from DWARF, where even though we don’t necessarily have O(1) lookup by basename within a given scope (including O(1) scope, we at least have O(n) access within a given scope).

Important

Program-wide lookup of names by anything other than an exact textually matching fully qualified name is not possible.

S_GDATA32

S_GTHREAD32 (0x1113)

S_PROCREF (0x1125)

S_LPROCREF (0x1127)

S_GMANDATA (0x111d)

Module Symbols

S_END (0x0006)

S_FRAMEPROC (0x1012)

S_OBJNAME (0x1101)

S_THUNK32 (0x1102)

S_BLOCK32 (0x1103)

S_LABEL32 (0x1105)

S_REGISTER (0x1106)

S_BPREL32 (0x110b)

S_LPROC32 (0x110f)

S_GPROC32 (0x1110)

S_REGREL32 (0x1111)

S_COMPILE2 (0x1116)

S_UNAMESPACE (0x1124)

S_TRAMPOLINE (0x112c)

S_SECTION (0x1136)

S_COFFGROUP (0x1137)

S_EXPORT (0x1138)

S_CALLSITEINFO (0x1139)

S_FRAMECOOKIE (0x113a)

S_COMPILE3 (0x113c)

S_ENVBLOCK (0x113d)

S_LOCAL (0x113e)

Defines a local variable. This record is followed by a series of S_DEFRANGE* records that define the live range of this variable.

struct LocalSym {
  /// The type of this variable (TPI type index).
  uint32_t Type;
  /// See enum below.
  uint16_t Flags;
  /// Name of the variable. A zero terminated string.
  char Name[];
};
enum class LocalSymFlags : uint16_t {
  None = 0,
  IsParameter = 1 << 0,
  IsAddressTaken = 1 << 1,
  IsCompilerGenerated = 1 << 2,
  IsAggregate = 1 << 3,
  IsAggregated = 1 << 4,
  IsAliased = 1 << 5,
  IsAlias = 1 << 6,
  IsReturnValue = 1 << 7,
  IsOptimizedOut = 1 << 8,
  IsEnregisteredGlobal = 1 << 9,
  IsEnregisteredStatic = 1 << 10,
};

All S_DEFRANGE* records consist of a header followed by LocalVariableAddrRange and a list of LocalVariableAddrGap (until the record length is reached) except for the S_DEFRANGE_FRAMEPOINTER_REL_FULL_SCOPE record.

/// A live range of a variable.
struct LocalVariableAddrRange {
  /// Starting offset in the section
  uint32_t OffsetStart;
  /// Index of the section
  uint16_t ISectStart;
  /// Size of the range
  uint16_t Range;
};
/// A subrange of a `LocalVariableAddrRange` where a variable is _not_ live.
struct LocalVariableAddrGap {
  /// Start of the range, relative to `LocalVariableAddrRange::OffsetStart`
  uint16_t GapStartOffset;
  /// Size of the range
  uint16_t Range;
};

The following records only describe the header.

S_DEFRANGE (0x113f)

A live range expressed as a DIA program.

struct DefrangeSymHeader {
  /// DIA program to evaluate the value of the symbol
  uint32_t Program;
};

S_DEFRANGE_SUBFIELD (0x1140)

A live range of sub field of variable (e.g. local.i).

struct DefrangeSubfieldSymHeader {
  /// DIA program to evaluate the value of the symbol
  uint32_t Program;
  /// Offset in parent variable.
  uint32_t OffsetInParent;
};

S_DEFRANGE_REGISTER (0x1141)

A live range of a variable living in a register.

struct DefrangeRegisterSymHeader {
  /// Register to hold the value of the symbol
  uint16_t Register;
  /// May have no user name on one control flow path
  uint16_t MayHaveNoName : 1;
  uint16_t Padding : 15;
};

S_DEFRANGE_FRAMEPOINTER_REL (0x1142)

A live range of frame variable. The register for the frame pointer is specified in S_FRAMEPROC.

struct DefrangeFramepointerSymHeader {
  /// Offset from the frame pointer
  int32_t Offset;
};

S_DEFRANGE_SUBFIELD_REGISTER (0x1143)

A live range of sub field of variable (e.g. local.i).

struct DefrangeSubfieldRegisterSymHeader {
  /// Register to hold the value of the symbol
  uint16_t Register;
  /// May have no user name on one of control flow path
  uint16_t MayHaveNoName : 1;
  uint16_t Padding1 : 15;
  /// Offset in parent variable
  uint32_t OffsetInParent : 12;
  uint32_t Padding2 : 20;
};

S_DEFRANGE_FRAMEPOINTER_REL_FULL_SCOPE (0x1144)

A frame variable valid in all function scope.

struct DefrangeFramepointerRelFullScopeSymHeader {
  /// Offset from the frame pointer
  int32_t Offset;
};

S_DEFRANGE_REGISTER_REL (0x1145)

A live range of variable relative to a register (range version of S_REGREL32).

struct DefrangeRegisterRelSymHeader {
  /// Register to hold the base pointer of the symbol
  uint16_t Register;
  /// Spilled member for s.i
  uint16_t SpilledUdtMember : 1;
  uint16_t Padding : 3;
  /// Offset in parent variable.
  uint16_t OffsetInParent : 12;
  /// Offset to register
  int32_t BasePointerOffset;
};

S_DEFRANGE_REGISTER_REL_INDIR (0x1177)

A live range of variable indirectly relative to a register (range version of S_REGREL32_INDIR).

struct DefrangeRegisterRelIndirSymHeader {
  /// Register to hold the base pointer of the symbol
  uint16_t Register;
  /// Spilled member for s.i
  uint16_t SpilledUdtMember : 1;
  uint16_t Padding : 3;
  /// Offset in parent variable.
  uint16_t OffsetInParent : 12;
  /// Offset to register
  int32_t BasePointerOffset;
  /// Offset to add after dereferencing `Register + BasePointerOffset`
  int32_t OffsetInUdt;
};

S_LPROC32_ID (0x1146)

S_GPROC32_ID (0x1147)

S_BUILDINFO (0x114c)

S_INLINESITE (0x114d)

S_INLINESITE_END (0x114e)

S_PROC_ID_END (0x114f)

S_FILESTATIC (0x1153)

S_LPROC32_DPC (0x1155)

S_LPROC32_DPC_ID (0x1156)

S_CALLEES (0x115a)

S_CALLERS (0x115b)

S_HEAPALLOCSITE (0x115e)

S_INLINEES (0x1168)

S_REGREL32_INDIR (0x1171)

This encodes a variable at the location *($Register + Offset) + OffsetInUdt. It’s equivalent to the following DWARF location expression:

DW_OP_breg{corresponding DWARF register} {Offset}
DW_OP_deref
DW_OP_plus_uconst {OffsetInUdt}

It’s used in C++ 17 structured bindings for example:

struct Foo { int a, b; };

void fn() {
  Foo f = {1, 2};
  //  ╰─ S_REGREL32{ reg = rsp, offset = 0 }
  auto &[x, y] = f;
  //     │  ╰─ S_REGREL32_INDIR{ reg = rsp, offset = 8, offset-in-udt = 4, type = int }
  //     ╰─ S_REGREL32_INDIR{ reg = rsp, offset = 8, offset-in-udt = 0, type = int }
}

The S_REGREL32_INDIR symbol for y from above looks like this:

Offset

Type

OffsetInUdt

Register

Name

08000000

74000000

04000000

4F01

7900

8

int

4

RSP

“a”

Symbols which can go in either/both of the module info stream & global stream

S_CONSTANT (0x1107)

S_UDT (0x1108)

S_LDATA32 (0x110c)

S_LTHREAD32 (0x1112)

S_LMANDATA (0x111c)

S_MANCONSTANT (0x112d)