CompilersLinkers & LoadersMedium

Parse and Display Symbol Tables

Master the fundamental concepts of linkers & loaders through this focused micro-challenge.

~45 minutes Linkers & Loaders Ready to learn

Topic Overview

Symbols Bridge Compilation Units

Each .o file exports defined symbols and imports undefined ones. The linker's job is to match undefined call foo relocations with global foo definitions. nm, objdump -t, and LLD's verbose mode show the same tables.

Symbol Attributes

Name, value (address), size, type (FUNC, OBJECT), binding (LOCAL, GLOBAL, WEAK), and section index. Undefined symbols have section index 0 and value 0 until resolved.

c
Loading…

T/t: text (code), defined/local
U: undefined, needs linker resolution
Weak symbols allow override by strong definitions
Static locals get local binding

Connecting to real compilers

Production compilers embed this step inside a longer pipeline. GCC flows through cpp, cc1, assembly, and ld; Clang uses the driver, Sema, LLVM IR passes, and a target backend. LLVM bitcode, JVM bytecode, and WASM are other familiar IRs at the same layer. The exercise isolates one pass so you can test it alone before chaining it to the next stage.

Why for this exercise

You will parse and display symbol tables from object files. This exercise requires reading symbol entries and reporting name, address, type, and binding for each symbol.

Implementation Task

Your Task

Read an ELF object's symbol table (the list of functions, variables and external references a linker resolves) and present it the way readelf -s does, plus a summary grouped by kind.

Input

The raw bytes of one ELF64 or ELF32 little-endian file as hexadecimal pairs separated by whitespace (the format of Parse and Understand Object File Formats (ELF); the header and section-header fields are decoded the same way).

Finding the symbols

Find the first section header whose sh_type is 2 (SYMTAB).
Its sh_link (offset 40 in ELF64 / 24 in ELF32) is the index of the string table holding symbol names; its sh_entsize (56 / 36) is the size of one entry; sh_size / sh_entsize is the number of entries.
Each entry (ELF64 / ELF32 layout):

Field	ELF64 offset, size	ELF32 offset, size
`st_name`	0, 4	0, 4
`st_info`	4, 1	12, 1
`st_shndx`	6, 2	14, 2
`st_value`	8, 8	4, 4
`st_size`	16, 8	8, 4

st_info holds the type in its low 4 bits (0 NOTYPE, 1 OBJECT, 2 FUNC, 3 SECTION, 4 FILE, 5 COMMON, 6 TLS) and the binding in its high 4 bits (0 LOCAL, 1 GLOBAL, 2 WEAK).

Output

c
Loading…

NDX is UND for 0, ABS for 0xfff1, COM for 0xfff2, otherwise the name of section st_shndx. An empty NAME prints as -.
The summary lines skip entry 0 and list names in table order (none if empty): functions = FUNC symbols that are defined (st_shndx ≠ 0); objects = defined OBJECT symbols; undefined = named symbols with st_shndx 0; weak = WEAK-bound symbols.
locals/globals count entries 1..N-1 by binding (WEAK counts as global).

Example

Input:

c
Loading…

Output:

c
Loading…

Requirements

Reuse your ELF header / section-header reader; add a symbol-entry reader that switches layout on the class.
Resolve section indices to names through the section-name string table.

Hidden tests include an AArch64 object with a SECTION symbol and weak undefined references, and an object with a COMMON symbol, a TLS symbol and a defined weak function.