Skip to content

Latest commit

 

History

History
48 lines (29 loc) · 1.79 KB

revisiting-plt-and-got.md

File metadata and controls

48 lines (29 loc) · 1.79 KB
description
06/09/2024

Revisiting: PLT & GOT

Introduction

Why revisit?

I needed a little refresher and wanted to share some quick/easy ways to remember both the PLT and GOT.

🚨 These are a MUST understand when working with dynamically-linked executables in Linux.

Procedure Linkage Table (PLT)

This can be looked at as a table of "stubs" (trampolines) that are used to call functions defined within shared libraries, like libc.

"PLT is a 'jump table' for function calls"

It simply helps a program jump to "external" functions defined within shared libraries.

Global Offset Table (GOT)

This acts as a "storage area" for globals and functions for their corresponding virtual addresses.

When the program needs to access a global variable or call a function, it looks up the address in the GOT. These addresses are all resolved at runtime.

How They Work Together

PLT -> GOT -> Linker (ld)
  1. When the program first calls a dynamically-linked function, it will store it in the PLT.
  • PLT Entry for the function will then point to the linker
  • The linker finds the address of the function, stores it in the GOT, and updates the PLT to point at the GOT for future calls
    • This is for fast acquisition and accessibility to these functions for future use
    • Meaning, the first call will always take slightly longer due to the lookup
    • Subsequent calls will be blazingly faster
  1. Subsequent calls PLT -> GOT
  • The PLT accesses the GOT for quick access
  • Example: foo@plt

Ultimately, remember this, the PLT is a jump table, possessing stubs (trampoline code) for corresponding function calls stored in dynamic libraries. Meanwhile, the GOT is a "storage area" for all resolved addresses.