LLVM Bytecode File Format
  1. Abstract
  2. Concepts
    1. Blocks
    2. Lists
    3. Fields
    4. Alignment
    5. Variable Bit-Rate Encoding
    6. Encoding Primitives
    7. Slots
  3. General Structure
  4. Block Definitions
    1. Signature Block
    2. Module Block
    3. Global Type Pool
    4. Module Info Block
    5. Global Constant Pool
    6. Function Definition
    7. Compaction Table
    8. Instruction List
    9. Symbol Table
  5. Version Differences
    1. Version 1.2 Differences From 1.3
    2. Version 1.1 Differences From 1.2
    3. Version 1.0 Differences From 1.1

Written by Reid Spencer

Abstract

This document describes the LLVM bytecode file format. It specifies the binary encoding rules of the bytecode file format so that equivalent systems can encode bytecode files correctly. The LLVM bytecode representation is used to store the intermediate representation on disk in compacted form.

The LLVM bytecode format may change in the future, but LLVM will always be backwards compatible with older formats. This document will only describe the most current version of the bytecode format. See Version Differences for the details on how the current version is different from previous versions.

Concepts

This section describes the general concepts of the bytecode file format without getting into specific layout details. It is recommended that you read this section thoroughly before interpreting the detailed descriptions.

Blocks

LLVM bytecode files consist simply of a sequence of blocks of bytes using a binary encoding Each block begins with an header of two unsigned integers. The first value identifies the type of block and the second value provides the size of the block in bytes. The block identifier is used because it is possible for entire blocks to be omitted from the file if they are empty. The block identifier helps the reader determine which kind of block is next in the file. Note that blocks can be nested within other blocks.

All blocks are variable length, and the block header specifies the size of the block. All blocks begin on a byte index that is aligned to an even 32-bit boundary. That is, the first block is 32-bit aligned because it starts at offset 0. Each block is padded with zero fill bytes to ensure that the next block also starts on a 32-bit boundary.

Lists

LLVM Bytecode blocks often contain lists of things of a similar type. For example, a function contains a list of instructions and a function type contains a list of argument types. There are two basic types of lists: length lists (llist), and null terminated lists (zlist), as described below in the Encoding Primitives.

Fields

Fields are units of information that LLVM knows how to write atomically. Most fields have a uniform length or some kind of length indication built into their encoding. For example, a constant string (array of bytes) is written simply as the length followed by the characters. Although this is similar to a list, constant strings are treated atomically and are thus fields.

Fields use a condensed bit format specific to the type of information they must contain. As few bits as possible are written for each field. The sections that follow will provide the details on how these fields are written and how the bits are to be interpreted.

Alignment

To support cross-platform differences, the bytecode file is aligned on certain boundaries. This means that a small amount of padding (at most 3 bytes) will be added to ensure that the next entry is aligned to a 32-bit boundary.

Variable Bit-Rate Encoding

Most of the values written to LLVM bytecode files are small integers. To minimize the number of bytes written for these quantities, an encoding scheme similar to UTF-8 is used to write integer data. The scheme is known as variable bit rate (vbr) encoding. In this encoding, the high bit of each byte is used to indicate if more bytes follow. If (byte & 0x80) is non-zero in any given byte, it means there is another byte immediately following that also contributes to the value. For the final byte (byte & 0x80) is false (the high bit is not set). In each byte only the low seven bits contribute to the value. Consequently 32-bit quantities can take from one to five bytes to encode. In general, smaller quantities will encode in fewer bytes, as follows:

Byte # Significant Bits Maximum Value
10-6127
27-1316,383
314-202,097,151
421-27268,435,455
528-3434,359,738,367
635-414,398,046,511,103
742-48562,949,953,421,311
849-5572,057,594,037,927,935
956-629,223,372,036,854,775,807
1063-691,180,591,620,717,411,303,423

Note that in practice, the tenth byte could only encode bit 63 since the maximum quantity to use this encoding is a 64-bit integer.

Signed VBR values are encoded with the standard vbr encoding, but with the sign bit as the low order bit instead of the high order bit. This allows small negative quantities to be encoded efficiently. For example, -3 is encoded as "((3 << 1) | 1)" and 3 is encoded as "(3 << 1) | 0)", emitted with the standard vbr encoding above.

Encoding Primitives

Each field in the bytecode format is encoded into the file using a small set of primitive formats. The table below defines the encoding rules for the various primitives used and gives them each a type name. The type names used in the descriptions of blocks and fields in the Detailed Layoutnext section. Any type name with the suffix _vbr indicates a quantity that is encoded using variable bit rate encoding as described above.

Type Rule
unsigned A 32-bit unsigned integer that always occupies four consecutive bytes. The unsigned integer is encoded using LSB first ordering. That is bits 20 through 27 are in the byte with the lowest file offset (little endian).
uint32_vbr A 32-bit unsigned integer that occupies from one to five bytes using variable bit rate encoding.
uint64_vbr A 64-bit unsigned integer that occupies from one to ten bytes using variable bit rate encoding.
int64_vbr A 64-bit signed integer that occupies from one to ten bytes using the signed variable bit rate encoding.
char A single unsigned character encoded into one byte
bit(n-m) A set of bit within some larger integer field. The values of n and m specify the inclusive range of bits that define the subfield. The value for m may be omitted if its the same as n.
string A uint32_vbr indicating the type of the constant string which also includes its length, immediately followed by the characters of the string. There is no terminating null byte in the string.
data An arbitrarily long segment of data to which no interpretation is implied. This is used for float, double, and constant initializers.
llist(x) A length list of x. This means the list is encoded as an uint32_vbr providing the length of the list, followed by a sequence of that many "x" items. This implies that the reader should iterate the number of times provided by the length.
zlist(x) A zero-terminated list of x. This means the list is encoded as a sequence of an indeterminate number of "x" items, followed by an uint32_vbr terminating value. This implies that none of the "x" items can have a zero value (or else the list terminates).
block A block of data that is logically related. A block begins with an unsigned that provides the block identifier (constant value) and an unsigned that provides the length of the block. Blocks may compose other blocks.
Field Notation

In the detailed block and field descriptions that follow, a regex like notation is used to describe optional and repeated fields. A very limited subset of regex is used to describe these, as given in the following table:

Character Meaning
? The question mark indicates 0 or 1 occurrences of the thing preceding it.
* The asterisk indicates 0 or more occurrences of the thing preceding it.
+ The plus sign indicates 1 or more occurrences of the thing preceding it.
() Parentheses are used for grouping.
, The comma separates sequential fields.

So, for example, consider the following specifications:

  1. string?
  2. (uint32_vbr,uin32_vbr)+
  3. (unsigned?,uint32_vbr)*
  4. (llist(unsigned))?

with the following interpretations:

  1. An optional string. Matches either nothing or a single string
  2. One or more pairs of uint32_vbr.
  3. Zero or more occurrences of either an unsigned followed by a uint32_vbr or just a uint32_vbr.
  4. An optional length list of unsigned values.
Slots

The bytecode format uses the notion of a "slot" to reference Types and Values. Since the bytecode file is a direct representation of LLVM's intermediate representation, there is a need to represent pointers in the file. Slots are used for this purpose. For example, if one has the following assembly:

%MyType = type { int, sbyte }
%MyVar = external global %MyType

there are two definitions. The definition of %MyVar uses %MyType. In the C++ IR this linkage between %MyVar and %MyType is explicit through the use of C++ pointers. In bytecode, however, there's no ability to store memory addresses. Instead, we compute and write out slot numbers for every Type and Value written to the file.

A slot number is simply an unsigned 32-bit integer encoded in the variable bit rate scheme (see encoding). This ensures that low slot numbers are encoded in one byte. Through various bits of magic LLVM attempts to always keep the slot numbers low. The first attempt is to associate slot numbers with their "type plane". That is, Values of the same type are written to the bytecode file in a list (sequentially). Their order in that list determines their slot number. This means that slot #1 doesn't mean anything unless you also specify for which type you want slot #1. Types are handled specially and are always written to the file first (in the Global Type Pool) and in such a way that both forward and backward references of the types can often be resolved with a single pass through the type pool.

Slot numbers are also kept small by rearranging their order. Because of the structure of LLVM, certain values are much more likely to be used frequently in the body of a function. For this reason, a compaction table is provided in the body of a function if its use would make the function body smaller. Suppose you have a function body that uses just the types "int*" and "{double}" but uses them thousands of time. Its worthwhile to ensure that the slot number for these types are low so they can be encoded in a single byte (via vbr). This is exactly what the compaction table does.

General Structure

This section provides the general structure of the LLVM bytecode file format. The bytecode file format requires blocks to be in a certain order and nested in a particular way so that an LLVM module can be constructed efficiently from the contents of the file. This ordering defines a general structure for bytecode files as shown below. The table below shows the order in which all block types may appear. Please note that some of the blocks are optional and some may be repeated. The structure is fairly loose because optional blocks, if empty, are completely omitted from the file.

ID Parent Optional? Repeated? Level Block Type Description
N/AFileNoNo0 Signature This contains the file signature (magic number) that identifies the file as LLVM bytecode.
0x01FileNoNo0 Module This is the top level block in a bytecode file. It contains all the other blocks.
0x15ModuleNoNo1    Global Type Pool This block contains all the global (module) level types.
0x14ModuleNoNo1    Module Globals Info This block contains the type, constness, and linkage for each of the global variables in the module. It also contains the type of the functions and the constant initializers.
0x12ModuleYesNo1    Module Constant Pool This block contains all the global constants except function arguments, global values and constant strings.
0x11ModuleYesYes1    Function Definitions* One function block is written for each function in the module. The function block contains the instructions, compaction table, type constant pool, and symbol table for the function.
0x12FunctionYesNo2       Function Constant Pool Any constants (including types) used solely within the function are emitted here in the function constant pool.
0x33FunctionYesNo2       Compaction Table This table reduces bytecode size by providing a funtion-local mapping of type and value slot numbers to their global slot numbers
0x32FunctionNoNo2       Instruction List This block contains all the instructions of the function. The basic blocks are inferred by terminating instructions.
0x13FunctionYesNo2       Function Symbol Table This symbol table provides the names for the function specific values used (basic block labels mostly).
0x13ModuleYesNo1    Module Symbol Table This symbol table provides the names for the various entries in the file that are not function specific (global vars, and functions mostly).

Use the links in the table for details about the contents of each of the block types.

Block Definitions

This section provides the detailed layout of the individual block types in the LLVM bytecode file format.

Signature Block

The signature occurs in every LLVM bytecode file and is always first. It simply provides a few bytes of data to identify the file as being an LLVM bytecode file. This block is always four bytes in length and differs from the other blocks because there is no identifier and no block length at the start of the block. Essentially, this block is just the "magic number" for the file.

Type Field Description
char Constant "l" (0x6C)
char Constant "l" (0x6C)
char Constant "v" (0x76)
char Constant "m" (0x6D)
Module Block

The module block contains a small pre-amble and all the other blocks in the file. The table below shows the structure of the module block. Note that it only provides the module identifier, size of the module block, and the format information. Everything else is contained in other blocks, described in other sections.

Type Field Description
unsigned Module Identifier (0x01)
unsigned Size of the module block in bytes
uint32_vbr Format Information
block Global Type Pool
block Module Globals Info
block Module Constant Pool
block* Function Definitions
block Module Symbol Table
Format Information

The format information field is encoded into a uint32_vbr as shown in the following table.

Type Description
bit(0) Target is big endian?
bit(1) On target pointers are 64-bit?
bit(2) Target has no endianess?
bit(3) Target has no pointer size?
bit(4-31) Bytecode format version

Of particular note, the bytecode format number is simply a 28-bit monotonically increase integer that identifies the version of the bytecode format (which is not directly related to the LLVM release number). The bytecode versions defined so far are (note that this document only describes the latest version, 1.3):

Note that we plan to eventually expand the target description capabilities of bytecode files to target triples.

Global Type Pool

The global type pool consists of type definitions. Their order of appearance in the file determines their slot number (0 based). Slot numbers are used to replace pointers in the intermediate representation. Each slot number uniquely identifies one entry in a type plane (a collection of values of the same type). Since all values have types and are associated with the order in which the type pool is written, the global type pool must be written as the first block of a module. If it is not, attempts to read the file will fail because both forward and backward type resolution will not be possible.

The type pool is simply a list of type definitions, as shown in the table below.

Type Field Description
unsigned Type Pool Identifier (0x15)
unsigned Size in bytes of the type pool block.
llist(type) A length list of type definitions.
Type Definitions

Types in the type pool are defined using a different format for each kind of type, as given in the following sections.

Primitive Types

The primitive types encompass the basic integer and floating point types

Type Description
uint32_vbr Type ID for the primitive types (values 1 to 11) 1
Notes:
  1. The values for the Type IDs for the primitive types are provided by the definition of the llvm::Type::TypeID enumeration in include/llvm/Type.h. The enumeration gives the following mapping:
    1. bool
    2. ubyte
    3. sbyte
    4. ushort
    5. short
    6. uint
    7. int
    8. ulong
    9. long
    10. float
    11. double

Function Types

Type Description
uint32_vbr Type ID for function types (13)
uint32_vbr Slot number of function's return type.
llist(uint32_vbr) Slot number of each argument's type.
uint32_vbr? Value 0 if this is a varargs function, missing otherwise.

Structure Types

Type Description
uint32_vbr Type ID for structure types (14)
zlist(uint32_vbr) Slot number of each of the element's fields.

Array Types

Type Description
uint32_vbr Type ID for Array Types (15)
uint32_vbr Slot number of array's element type.
uint32_vbr The number of elements in the array.

Pointer Types

Type Description
uint32_vbr Type ID For Pointer Types (16)
uint32_vbr Slot number of pointer's element type.

Opaque Types

Type Description
uint32_vbr Type ID For Opaque Types (17)
Module Global Info

The module global info block contains the definitions of all global variables including their initializers and the declaration of all functions. The format is shown in the table below:

Type Field Description
unsigned Module global info identifier (0x14)
unsigned Size in bytes of the module global info block.
zlist(globalvar) A zero terminated list of global var definitions occuring in the module.
zlist(uint32_vbr) A zero terminated list of function types occuring in the module.
Global Variable Field

Global variables are written using an uint32_vbr that encodes information about the global variable and a list of the constant initializers for the global var, if any.

The table below provides the bit layout of the first uint32_vbr that describes the global variable.

Type Description
bit(0) Is constant?
bit(1) Has initializer? Note that this bit determines whether the constant initializer field (described below) follows.
bit(2-4) Linkage type: 0=External, 1=Weak, 2=Appending, 3=Internal, 4=LinkOnce
bit(5-31) Slot number of type for the global variable.

The table below provides the format of the constant initializers for the global variable field, if it has one.

Type Description
(zlist(uint32_vbr))? An optional zero-terminated list of slot numbers of the global variable's constant initializer.
Constant Pool

A constant pool defines as set of constant values. There are actually two types of constant pool blocks: one for modules and one for functions. For modules, the block begins with the constant strings encountered anywhere in the module. For functions, the block begins with types only encountered in the function. In both cases the header is identical. The tables that follow, show the header, module constant pool preamble, function constant pool preamble, and the part common to both function and module constant pools.

Common Block Header

Type Field Description
unsigned Constant pool identifier (0x12)
unsigned Size in bytes of the constant pool block.

Module Constant Pool Preamble (constant strings)

Type Field Description
uint32_vbr The number of constant strings that follow.
uint32_vbr Zero. This identifies the following "plane" as containing the constant strings. This is needed to identify it uniquely from other constant planes that follow.
uint32_vbr+ Slot number of the constant string's type. Note that the constant string's type implicitly defines the length of the string.

Function Constant Pool Preamble (function types)

The structure of the types for functions is identical to the Global Type Pool. Please refer to that section for the details.

Common Part (other constants)

Type Field Description
uint32_vbr Number of entries in this type plane.
uint32_vbr Type slot number of this plane.
constant+ The definition of a constant (see below).
Constant Field

Constants come in many shapes and flavors. The sections that followe define the format for each of them. All constants start with a uint32_vbr encoded integer that provides the number of operands for the constant. For primitive, structure, and array constants, this will always be zero since those types of constants have no operands. In this case, we have the following field definitions:

When the number of operands to the constant is non-zero, we have a constant expression and its field format is provided in the table below.

Type Field Description
uint32_vbr Op code of the instruction for the constant expression.
uint32_vbr The slot number of the constant value for an operand.1
uint32_vbr The slot number for the type of the constant value for an operand.1
Notes:
  1. Both these fields are repeatable but only in pairs.
Function Definition

Function definitions contain the linkage, constant pool or compaction table, instruction list, and symbol table for a function. The following table shows the structure of a function definition.

Type Field Description
unsigned Function definition block identifier (0x11)
unsigned Size in bytes of the function definition block.
uint32_vbr The linkage type of the function: 0=External, 1=Weak, 2=Appending, 3=Internal, 4=LinkOnce1
block The constant pool block for this function.2
block The compaction table block for the function.2
block The instruction list for the function.
block The function's symbol table containing only those symbols pertinent to the function (mostly block labels).
Notes:
  1. Note that if the linkage type is "External" then none of the other fields will be present as the function is defined elsewhere.
  2. Note that only one of the constant pool or compaction table will be written. Compaction tables are only written if they will actually save bytecode space. If not, then a regular constant pool is written.
Compaction Table

Compaction tables are part of a function definition. They are merely a device for reducing the size of bytecode files. The size of a bytecode file is dependent on the value of the slot numbers used because larger values use more bytes in the variable bit rate encoding scheme. Furthermore, the compressed instruction format reserves only six bits for the type of the instruction. In large modules, declaring hundreds or thousands of types, the values of the slot numbers can be quite large. However, functions may use only a small fraction of the global types. In such cases a compaction table is created that maps the global type and value slot numbers to smaller values used by a function. Functions will contain either a function-specific constant pool or a compaction table but not both. Compaction tables have the format shown in the table below.

Type Field Description
uint32_vbr The number of types that follow
uint32_vbr+ The slot number in the global type plane of the type that will be referenced in the function with the index of this entry in the compaction table.
type_len An encoding of the type and number of values that follow. This field's encoding varies depending on the size of the type plane. See Type and Length for further details.
uint32_vbr+ The slot number in the globals of the value that will be referenced in the function with the index of this entry in the compaction table
Type and Length

The type and length of a compaction table type plane is encoded differently depending on the length of the plane. For planes of length 1 or 2, the length is encoded into bits 0 and 1 of a uint32_vbr and the type is encoded into bits 2-31. Because type numbers are often small, this often saves an extra byte per plane. If the length of the plane is greater than 2 then the encoding uses a uint32_vbr for each of the length and type, in that order.

Instruction List

The instructions in a function are written as a simple list. Basic blocks are inferred by the terminating instruction types. The format of the block is given in the following table.

Type Field Description
unsigned Instruction list identifier (0x33).
unsigned Size in bytes of the instruction list.
instruction+ An instruction. Instructions have a variety of formats. See Instructions for details.
Instructions

For brevity, instructions are written in one of four formats, depending on the number of operands to the instruction. Each instruction begins with a uint32_vbr that encodes the type of the instruction as well as other things. The tables that follow describe the format of this first word of each instruction.

Instruction Format 0

This format is used for a few instructions that can't easily be optimized because they have large numbers of operands (e.g. PHI Node or getelementptr). Each of the opcode, type, and operand fields is as successive fields.

Type Field Description
uint32_vbr Specifies the opcode of the instruction. Note that for compatibility with the other instruction formats, the opcode is shifted left by 2 bits. Bits 0 and 1 must have value zero for this format.
uint32_vbr Provides the slot number of the result type of the instruction
uint32_vbr The number of operands that follow.
uint32_vbr+ The slot number of the value(s) for the operand(s). 1
Notes:
  1. Note that if the instruction is a getelementptr and the type of the operand is a sequential type (array or pointer) then the slot number is shifted up two bits and the low order bits will encode the type of index used, as follows: 0=uint, 1=int, 2=ulong, 3=long.

Instruction Format 1

This format encodes the opcode, type and a single operand into a single uint32_vbr as follows:

Bits Type Field Description
0-1constant "1" These two bits must be the value 1 which identifies this as an instruction of format 1.
2-7opcode Specifies the opcode of the instruction. Note that the maximum opcode value is 63.
8-19unsigned Specifies the slot number of the type for this instruction. Maximum slot number is 212-1=4095.
20-31unsigned Specifies the slot number of the value for the first operand. Maximum slot number is 212-1=4095. Note that the value 212-1 denotes zero operands.

Instruction Format 2

This format encodes the opcode, type and two operands into a single uint32_vbr as follows:

Bits Type Field Description
0-1constant "2" These two bits must be the value 2 which identifies this as an instruction of format 2.
2-7opcode Specifies the opcode of the instruction. Note that the maximum opcode value is 63.
8-15unsigned Specifies the slot number of the type for this instruction. Maximum slot number is 28-1=255.
16-23unsigned Specifies the slot number of the value for the first operand. Maximum slot number is 28-1=255.
24-31unsigned Specifies the slot number of the value for the second operand. Maximum slot number is 28-1=255.

Instruction Format 3

This format encodes the opcode, type and three operands into a single uint32_vbr as follows:

Bits Type Field Description
0-1constant "3" These two bits must be the value 3 which identifies this as an instruction of format 3.
2-7opcode Specifies the opcode of the instruction. Note that the maximum opcode value is 63.
8-13unsigned Specifies the slot number of the type for this instruction. Maximum slot number is 26-1=63.
14-19unsigned Specifies the slot number of the value for the first operand. Maximum slot number is 26-1=63.
20-25unsigned Specifies the slot number of the value for the second operand. Maximum slot number is 26-1=63.
26-31unsigned Specifies the slot number of the value for the third operand. Maximum slot number is 26-1=63.
Symbol Table

A symbol table can be put out in conjunction with a module or a function. A symbol table is a list of type planes. Each type plane starts with the number of entries in the plane and the type plane's slot number (so the type can be looked up in the global type pool). For each entry in a type plane, the slot number of the value and the name associated with that value are written. The format is given in the table below.

Type Field Description
unsigned Symbol Table Identifier (0x13)
unsigned Size in bytes of the symbol table block.
uint32_vbr Number of entries in type plane
symtab_entry* Provides the slot number of the type and its name.
symtab_plane* A type plane containing value slot number and name for all values of the same type.
Symbol Table Plane

A symbol table plane provides the symbol table entries for all values of a common type. The encoding is given in the following table:

Type Field Description
uint32_vbr Number of entries in this plane.
uint32_vbr Slot number of type for this plane.
symtab_entry+ The symbol table entries for this plane.
Symbol Table Entry

A symbol table entry provides the assocation between a type or value's slot number and the name given to that type or value. The format is given in the following table:

Type Field Description
uint32_vbr Slot number of the type or value being given a name.
uint32_vbr Length of the character array that follows.
char+ The characters of the name.
Version Differences

This section describes the differences in the Bytecode Format across LLVM versions. The versions are listed in reverse order because it assumes the current version is as documented in the previous sections. Each section here describes the differences between that version and the one that follows.

Version 1.2 Differences From 1.3
Type Derives From Value

In version 1.2, the Type class in the LLVM IR derives from the Value class. This is not the case in version 1.3. Consequently, in version 1.2 the notion of a "Type Type" was used to write out values that were Types. The types always occuped plane 12 (corresponding to the TypeTyID) of any type planed set of values. In 1.3 this representation is not convenient because the TypeTyID (12) is not present and its value is now used for LabelTyID. Consequently, the data structures written that involve types do so by writing all the types first and then each of the value planes according to those types. In version 1.2, the types would have been written intermingled with the values.

Restricted getelementptr Types

In version 1.2, the getelementptr instruction required a ubyte type index for accessing a structure field and a long type index for accessing an array element. Consequently, it was only possible to access structures of 255 or fewer elements. Starting in version 1.3, this restriction was lifted. Structures must now be indexed with uint constants. Arrays may now be indexed with int, uint, long, or ulong typed values. The consequence of this was that the bytecode format had to change in order to accommodate the larger range of structure indices.

Version 1.1 Differences From 1.2
Explicit Primitive Zeros

In version 1.1, the zero value for primitives was explicitly encoded into the bytecode format. Since these zero values are constant values in the LLVM IR and never change, there is no reason to explicitly encode them. This explicit encoding was removed in version 1.2.

Inconsistent Module Global Info

In version 1.1, the Module Global Info block was not aligned causing the next block to be read in on an unaligned boundary. This problem was corrected in version 1.2.

Version 1.0 Differences From 1.1

None. Version 1.0 and 1.1 bytecode formats are identical.


Valid CSS! Valid HTML 4.01! Reid Spencer and Chris Lattner
The LLVM Compiler Infrastructure
Last modified: $Date$