Metadata and the PE File Structure
Metadata is stored in one section of a .NET Framework portable executable (PE) file, while Microsoft intermediate language (MSIL) is stored in another section of the PE file. The metadata portion of the file contains a series of table and heap data structures. The MSIL portion contains MSIL and metadata tokens that reference the metadata portion of the PE file. You might encounter metadata tokens when you use tools such as the MSIL Disassembler (Ildasm.exe) to view your code's MSIL or the Runtime Debugger (Cordbg.exe) to perform a memory dump.
Metadata Tables and Heaps
Each metadata table holds information about the elements of your program. For example, one metadata table describes the classes in your code, another table describes the fields, and so on. If you have ten classes in your code, the class table will have tens rows, one for each class. Metadata tables reference other tables and heaps. For example, the metadata table for classes references the table for methods.
Metadata also stores information in four heap structures: string, blob, user string, and GUID. All the strings used to name types and members are stored in the string heap. For example, a method table does not directly store the name of a particular method, but points to the method's name stored in the string heap.
Metadata Tokens
Each row of each metadata table is uniquely identified in the MSIL portion of the PE file by a metadata token. Metadata tokens are conceptually similar to pointers, persisted in MSIL, that reference a particular metadata table.
A metadata token is a four-byte number. The top byte denotes the metadata table to which a particular token refers (method, type, and so on). The remaining three bytes specify the row in the metadata table that corresponds to the programming element being described. If you define a method in C# and compile it into a PE file, the following metadata token might exist in the MSIL portion of the PE file:
0x06000004
The top byte (0x06
) indicates that this is a MethodDef token. The lower three bytes (000004
) tells the common language runtime to look in the fourth row of the MethodDef table for the information that describes this method definition.
Metadata within a PE File
When a program is compiled for the common language runtime, it is converted to a PE file that consists of three parts. The following table describes the contents of each part.
PE section | Contents of PE section |
---|---|
PE header |
The index of the PE file's main sections and the address of the entry point. The runtime uses this information to identify the file as a PE file and to determine where execution starts when loading the program into memory. |
MSIL instructions |
The Microsoft intermediate language instructions (MSIL) that make up your code. Many MSIL instructions are accompanied by metadata tokens. |
Metadata |
Metadata tables and heaps. The runtime uses this section to record information about every type and member in your code. This section also includes custom attributes and security information. |