HeapMemory — Custom Heap Memory Allocator

A low-level heap memory manager implemented in C/C++, exploring two distinct approaches to dynamic memory allocation from scratch — without relying on malloc or any standard allocator.

Overview

This project implements a custom heap allocator to deeply understand how memory management works at the OS boundary. Two allocator strategies are implemented, each with different trade-offs in complexity and control.

Implementations

1. sbrk/brk Allocator (/Sbrk) — ✅ Complete

A fully functional heap allocator using the sbrk syscall to grow the process data segment on demand.

How it works:

• Maintains a free list of memory blocks using a linked list structure
• Implements first-fit allocation strategy
• Coalesces adjacent free blocks to reduce fragmentation
• Uses sbrk() to request additional memory from the OS when no suitable free block exists
• Implements malloc, free, and realloc equivalents

Key concepts demonstrated:

• Process memory layout (heap segment)
• Free list management and block metadata
• Internal and external fragmentation handling
• Memory alignment

2. Virtual Memory Page Allocator (/virtualMemory) — 🔬 In Progress

An exploration of lower-level memory management using mmap to allocate memory directly at the virtual memory page level.

Goal: Manage memory at page granularity, bypassing the traditional heap entirely and interacting directly with the OS virtual memory subsystem.

Concepts being explored:

• mmap/munmap for anonymous memory mapping
• Page-aligned allocation
• Virtual memory layout and address space management
• Build specific structure for every type that is registered

Why This Project

Understanding how allocators work is foundational to systems programming and security research. Custom allocators will help me learn:

• How heap vulnerabilities like heap overflows, use-after-free, and double-free bugs arise at the implementation level
• The OS mechanisms (sbrk, mmap) that underpin all dynamic memory in user space

Future plans

I plan in future to finish vm version and to add version small buckets, large bucker in order to understand memory better in terms of security and to add intrusive lists to all projects

Tech Stack

• C / C++
• CMake
• Linux system calls (sbrk, mmap)
• Linked list block management