Tcache Exploitation Skill

A comprehensive guide for identifying and exploiting Tcache vulnerabilities in binary exploitation challenges.

When to Use This Skill

Use this skill when working with:

• Heap exploitation challenges (CTF, security research, vulnerability analysis)
• Binaries compiled with Glibc 2.26 or newer (Tcache introduced in 2.26)
• Vulnerabilities involving

  malloc

  ,

  free

  , double free, use-after-free, or heap overflow
• Challenges requiring libc address leaks or arbitrary writes
• Any mention of

  tcache

  ,

  malloc_hook

  ,

  free_hook

  , or heap bin manipulation

Core Concepts

What is Tcache?

Tcache (thread cache) was introduced in Glibc 2.26 to improve memory allocation performance. It maintains per-thread caches of freed chunks before returning them to the main heap.

Key properties:

• Per-thread cache (not shared across threads)
• Stores up to 7 chunks per size class by default
• Uses a singly-linked list (FD pointer only)
• No size validation on allocation (unlike fast bins)
• No alignment checks in older versions (error:

  malloc(): unaligned tcache chunk detected

  in newer versions)

Tcache Poisoning Attack

The Tcache attack (also called Tcache poisoning) is similar to fast bin attacks:

Goal: Overwrite the

next

Basic flow:

1. Allocate a chunk and free it (adds to tcache)
2. Overflow/modify the freed chunk's FD pointer to point to target address
3. Allocate a chunk of the same size → returns the target address
4. Use the returned chunk to overwrite critical data (e.g.,

   __malloc_hook

   )

Important: The target address must be aligned (8-byte aligned on 64-bit systems). If not aligned, you may need to run the binary multiple times or find an aligned target.

Tcache Index Attack

The tcache structure itself contains metadata that can be exploited:

Tcache structure layout:

• Array of chunk counts per size index
• Array of head pointers for each tcache bin

Attack pattern:

1. Modify the tcache metadata (chunk count or head pointer)
2. Make a tcache bin's head pointer point to a desired address (e.g.,

   __malloc_hook

   )
3. Allocate a chunk of that size → returns the target address
4. Overwrite the target (e.g., write one-gadget to

   __malloc_hook

   )

Common Attack Patterns

Pattern 1: Double Free Tcache Poisoning

Vulnerability: Double free of the same chunk

Exploitation:

1. Allocate chunk A
2. Free chunk A (A → NULL in tcache)
3. Free chunk A again (A → A in tcache, self-referential)
4. Allocate chunk A (returns A, tcache head now points to A's FD)
5. Modify A's FD to point to

   __free_hook

6. Allocate again → returns chunk at

   __free_hook

7. Write

   system

   address to

   __free_hook

8. Free a chunk containing "/bin/sh" → shell

Pattern 2: Heap Overflow + Size Manipulation

Vulnerability: 1-byte heap overflow allowing size header modification

Exploitation:

1. Allocate chunk A (small size)
2. Overflow to increase A's size header
3. Free chunk A (goes to tcache of fake larger size)
4. Allocate chunk of fake size → returns A but with larger apparent size
5. Use the extra space to overwrite next chunk's FD pointer
6. Point FD to

   __malloc_hook

7. Allocate twice: first gets legit chunk, second gets

   __malloc_hook

8. Write one-gadget to

   __malloc_hook

9. Trigger malloc → shell

Pattern 3: Tcache Index Corruption

Vulnerability: Ability to free arbitrary heap addresses

Exploitation:

1. Allocate and free a chunk of specific size that creates value 0x100 in tcache metadata
2. The tcache stores chunk counts in specific bytes, creating exploitable values
3. Free the address 0x100 (looks like a valid chunk)
4. This adds 0x100 to the tcache index for size 0x100
5. Allocate size 0x100 → returns address 0x100
6. Use this to overwrite other tcache indexes
7. Point an index to

   __malloc_hook

   , allocate that size, write one-gadget

Pattern 4: Libc Leak via Unsorted Bin

Vulnerability: Use-after-free or double free

Exploitation:

1. Fill all tcaches (7 chunks per size class)
2. Add one more chunk → goes to unsorted bin
3. Empty the tcache (allocate all cached chunks)
4. Re-allocate the unsorted bin chunk
5. Overwrite only first 8 bytes, leaving second address intact
6. The second address is a libc pointer → leak libc base

Step-by-Step Exploitation Workflow

Step 1: Identify the Vulnerability

Look for:

• Double free vulnerabilities
• Use-after-free conditions
• Heap overflow (especially 1-byte overflows)
• Arbitrary free capabilities
• Uninitialized pointers in heap chunks

Step 2: Determine Glibc Version

Check if Tcache is available:

• Glibc 2.26+ has Tcache
• Use

  ldd --version

  or check binary's linked libc
• Tcache errors:

  malloc(): unaligned tcache chunk detected

Step 3: Map the Heap Layout

Understand:

• Where chunks are allocated
• Where tcache metadata is stored
• Distance between chunks and targets (

  __malloc_hook

  ,

  __free_hook

  )
• Alignment requirements (8-byte on 64-bit)

Step 4: Choose Attack Pattern

Based on vulnerability type:

• Double free → Pattern 1 (self-referential tcache)
• Heap overflow → Pattern 2 (size manipulation)
• Arbitrary free → Pattern 3 (tcache index corruption)
• Need libc leak → Pattern 4 (unsorted bin leak)

Step 5: Craft the Exploit

1. Fill tcaches if needed (7 chunks per size class)
2. Trigger the vulnerability (double free, overflow, etc.)
3. Manipulate FD pointers to point to targets
4. Allocate chunks to get control over target addresses
5. Write payloads (one-gadget, system address, etc.)
6. Trigger the payload (malloc, free, etc.)

Step 6: Handle Alignment

If getting

unaligned tcache chunk detected

• Find aligned target addresses
• Run binary multiple times until alignment works
• Use heap feng-shui to control alignment

Key Addresses to Target

__malloc_hook

__free_hook

stdout

tcache

Common One-Gadgets

Find one-gadgets in libc:

one_gadget /path/to/libc.so.6

Common patterns:

• ret; pop rdi; ret

  +

  /bin/sh

  address

• execve

  gadgets

• system

  gadgets

Debugging Tips

Check Tcache State

Use GDB with

heap

heap analyze
heap chunks
heap tcache

Common Errors

• malloc(): unaligned tcache chunk detected

  → Address not 8-byte aligned

• double free or corruption

  → Tcache not being used, or older Glibc

• corrupted size vs. prev_size

  → Size header corruption

Verify Exploit

1. Check if libc leak works (compare with

   ldd

   output)
2. Verify FD pointer overwrites (GDB

   x/gx

   command)
3. Confirm hook addresses are writable
4. Test one-gadget execution

Learning Resources

CTF Writeups (Study These Patterns)

1. DCQuals19 Babyheap - Libc leak + Tcache attack with 1-byte overflow

   • https://guyinatuxedo.github.io/29-tcache/dcquals19_babyheap/index.html

2. Plaid19 Cpp - Double free Tcache poisoning

   • https://guyinatuxedo.github.io/29-tcache/plaid19_cpp/index.html

3. CSAW19 Popping Caps0 - Tcache index attack

   • https://guyinatuxedo.github.io/44-more_tcache/csaw19_popping_caps0/index.html

4. CSAW19 Popping Caps1 - Optimized tcache index attack

   • https://guyinatuxedo.github.io/44-more_tcache/csaw19_popping_caps1/index.html

5. HTB Math Door - Write-after-free + heap feng-shui

   • https://7rocky.github.io/en/ctf/other/htb-cyber-apocalypse/math-door/

Reference Material

• Guy In A Tuxedo Tcache Explanation - Comprehensive Tcache guide

  • https://guyinatuxedo.github.io/29-tcache/tcache_explanation/index.html

• Bins and Memory Allocations - Heap bin fundamentals

  • See:

    bins-and-memory-allocations.md

Quick Reference

Tcache Size Classes

Tcache has 7 chunks per size class. Common sizes:

• 0x20, 0x30, 0x40, 0x50, 0x60, 0x70, 0x80, 0x90, 0xa0, 0xb0, 0xc0, 0xd0, 0xe0, 0xf0, 0x100

Chunk Structure (64-bit)

[prev_size: 8 bytes]
[size: 8 bytes] (includes prev_inuse flag)
[FD pointer: 8 bytes] (tcache only, no BK)
[payload: size - 16 bytes]

Exploit Checklist

• Identify vulnerability type
• Confirm Glibc 2.26+
• Map heap layout
• Choose attack pattern
• Fill tcaches if needed
• Trigger vulnerability
• Overwrite FD pointers
• Allocate target chunks
• Write payload
• Trigger payload
• Handle alignment issues

When This Skill Applies

Use this skill for:

• CTF heap exploitation challenges
• Security research on heap vulnerabilities
• Binary analysis involving malloc/free
• Learning heap exploitation techniques
• Writing exploit code for Tcache vulnerabilities

Do NOT use this skill for:

• Non-heap vulnerabilities (stack, format string, etc.)
• Glibc versions before 2.26 (no Tcache)
• Non-Glibc allocators (jemalloc, tcmalloc, etc.)
• High-level application bugs without heap primitives