CS:APP Labs

Computer Systems: A Programmer's Perspective (3rd ed.) · CMU 15-213

Lab solutions with source-backed design walkthroughs.

📖   Read the notes → leexsean.github.io/CSAPP-Lab-Solutions-lx

---

Every completed solution below is checked against the lab's actual grader; the result is shown in the table and at the top of each writeup.

#	Lab	Focus	Result	Writeup
01	Data Lab	Bit manipulation, integer & float encoding	36/36	Notes →
02	Bomb Lab	Reverse engineering · GDB · x86-64	6 + secret	Notes →
03	Attack Lab	Buffer overflow · code injection · ROP	5/5	Notes →
04	Cache Lab	LRU cache simulator · blocked transpose	53/61	Notes →
05	Shell Lab	Processes · signals · job control	16/16	Notes →
06	Malloc Lab	Allocator · segregated free lists	97/100	Notes →
07	Proxy Lab	Concurrent, caching HTTP proxy	70/70	Notes →
08	SFS Lab (AI-assisted)	On-disk file system · fine-grained locking	22/22	Notes →

A few design highlights

• Data Lab — every integer puzzle solved within its operator budget; floats manipulated purely at the bit level.
• Bomb Lab — all six phases plus the hidden phase (a recursive binary-tree search).
• Attack Lab — code injection for ctarget; a ROP chain that defeats NX and ASLR for rtarget.
• Cache Lab — an O(E) LRU simulator; the 64×64 transpose splits each 8×8 block into four 4×4 sub-blocks, staging data in B to dodge conflict misses.
• Shell Lab — signal masking erases the fork/addjob race; sigsuspend replaces the busy-wait in waitfg.
• Malloc Lab — segregated explicit free lists, footer-less allocated blocks via a prev_alloc bit, and a realloc that grows in place before copying.
• Proxy Lab — one detached worker per connection; binary-safe response relay and a reader–writer-locked approximate-LRU cache.
• SFS Lab — explicitly AI-assisted; atomic rename and per-file locking reach 3.40× the calibrated global-mutex baseline.

The notes site

The writeups live in docs/ as an MkDocs Material site and are published to GitHub Pages automatically on every push (see .github/workflows/deploy.yml). To run it locally:

python -m venv .venv && source .venv/bin/activate
pip install mkdocs-material
mkdocs serve            # then open http://127.0.0.1:8000

Acknowledgment

Thanks to virgiling for the helpful blog that guided me through these labs.