The School of Computing Science is introducing a new course, CMPT 201 Systems Programming, that will satisfy the curriculum requirements currently met by CMPT 300 Operating Systems for both CS and SoSy students. In other words, any students who take this course will be able to use it toward satisfying CMPT 300 requirements.
By taking this course, students will learn:
- User-level services and abstractions provided by operating systems, and
- How to write low-level programs that utilize those services and abstractions correctly, efficiently, and reliably, using Linux.
Although this course is known to be a challenging course, it is not designed for an advanced programmer. The training that this course aims to provide is to get students out of the beginning stage and move them to an early intermediate stage. Students have to make that jump at some point in their college career. This course forces students to make that jump, so they can advance much further with upper-level courses.
Prerequisites are:
- (CMPT 125 or CMPT 135) and MACM 101, all with a minimum grade of C-
- This course assumes that students thoroughly understand the intro programming materials and they do not struggle with them at all. This includes, but not limited to, basic programming constructs (e.g., loops, conditionals, functions/methods/modules, etc.) as well as basic data structures (e.g., linked lists, hashing and hash tables, etc.).
- Proficiency in C programming
- This course uses C but does not teach C.
- This course assumes that students thoroughly know C. This includes, but not limited to, the
syntax of C, C
structs, C strings and how to manipulate them, arrays and pointers, the use ofmalloc()andfree(), etc. - Knowing C++ is better but not enough. C is still different from C++.
- A very basic test is asking yourself, "Can I write a doubly-linked list in C without getting any help (with all the features, creating/appending/deleting a node and traversing the list)?"
- The following seem to be good resources to learn C.
- C for Python programmers: I think this can serve as a general intro to C, not just for Python programmers but any programmers.
- C for Java programmers
- Here's another one that talks about pointers and strings, which might be helpful.
- Experience in using the command-line interface on Linux
- We do not expect students to have proficiency but some experience.
Upon completion of this course, students are expected to achieve proficiency in the following areas:
- Interacting with, and programming in, a Linux environment using the command-line interface.
- Utilizing programming tools such as build systems and debuggers.
- Programming with processes and threads.
- Understanding concurrency concepts such as semaphores, locks, critical sections, deadlocks, common concurrency patterns, etc.
- Manipulating and managing data in memory correctly.
- Understanding file system abstractions and programming with them.
- Understanding communication abstractions such as IPC, sockets, and RPC and programming with them.
- Understanding cryptographic functions and programming with them.
- Tu 12:30PM - 2:20PM (K9500)
- Th 12:30PM - 2:20PM (Images Theatre in RCB Hall)
- Th 2:30PM - 3:20PM (ASB9838)
- Th 2:30PM - 3:20PM (ASB9838)
- Th 3:30PM - 4:20PM (ASB9838)
- Th 3:30PM - 4:20PM (ASB9838)
- Steve Ko <steveyko@sfu.ca>
- Kimia Khabiri <kimia_khabiri@sfu.ca>
- Parsa Hoseininejad <parsa_hosseininejad@sfu.ca>
- Mohammad Omidvar <m_omidvart@sfu.ca>
- Shishir Gopinath <shishir_gopinath@sfu.ca>
- Steve Ko: 10 am - 12 pm Mondays at TASC1 8019
- Shishir Gopinath: 2:30 pm - 4:30 pm Tuesdays at ASB 9812
- Kimia Khabiri: 2:30 pm to 4:30 pm Wednesdays at ASB 9810
- Mohammad Omidvar: 9:00 am - 11:00 am Thursdays at ASB 9812
- Parsa Hoseininejad: 2:30 pm to 4:30 pm Fridays at ASB 9810
The Linux Programming Interface: A Linux and UNIX System Programming Handbook, Michael Kerrisk, 2010.
- This book is arguably the best book on systems programming using Linux.
- If someone is looking to purchase one book on systems programming, this book is that book.
- A lot of the course materials are based on this book.
- This book is almost required. There are two reasons why it is almost.
- The book covers much more than what this course covers.
- The book is not inexpensive.
- If there is a student struggling with the materials of this course, it is recommended to purchase this book and read along.
Operating Systems: Three Easy Pieces, Remzi H. Arpaci-Dusseau and Andrea C. Arpaci-Dusseau
- This is a free online book on operating systems.
- This course teaches some of the concepts described in the book. However, the book goes much more in detail than what this course teaches, hence it is not a required book.
- Nevertheless, the course lecture notes refer to some chapters of the book and students are expected to read those chapters.
Computer Systems: A Programmer's Perspective, 3rd Edition, Randal E. Bryant and David R. O'Hallaron, 2015.
Advanced Programming in the UNIX Environment, 3rd Edition, W. Stevens and Stephen Rago, 2013
Unix Network Programming, Volume 1: The Sockets Networking API, 3rd Edition, W. Stevens, Bill Fenner, and Andrew Rudoff, 2003
UNIX Network Programming, Volume 2: Interprocess Communications, 2nd Edition, W. Richard Stevens, 1998
The Art of Unix Programming, Eric Steven Raymond, 2003
C Programming Language, 2nd Edition, Dennis Ritchie and Brian Kernighan, 1988
- Midterm: 15%
- Final: 15%
- Programming assignments: 66%
- 8 short assignments: 2% each
- 5 long assignment: 10% each
- In-class activities: 4%
The Lecture Notes repo contains the current lecture notes. The schedule generally looks as follows.
- Week 1 - Week 2: Tour of Computer Systems
- Week 2 - Week 3:
fork(),exec(),wait(), anderrno& Signals - Week 4 - Week 5: Scheduling Algorithms & Memory Management
- Week 5 - Week 6: Virtual Memory
- Week 6: Midterm
- Week 7 - Week 8: Threads and Synchronization
- Week 9 - Week 10: File I/O
- Week 10 - Week 11: Network Programming
- Week 12: IPC
- Week 13: Basic Cryptography
- Week 14: Final
There are two types of assignments in this course---short assignments and long assignments. Short assignments are designed to be straightforward and tutorial-like, while long assignments require a significant time investment. The difficulty level of long assignments is substantially higher than that of short assignments.
Students should consider these assignments take-home exams rather than traditional homework. This is because we require students to complete them in a highly-controlled environment using our Docker setup. Modifying the environment is not allowed and considered a violation of academic integrity.
The first type of assignments in this course is short assignments.
- In the first few weeks of the course, students are expected to finish two short assignments per week.
- Each short assignment is in the style of a tutorial and takes about 2-3 hours to complete.
- These assignments are designed to teach students basic tools and concepts that students use for in-class activities and long assignments.
- Unlike long assignments, short assignments are designed to be straightforward.
Below are the short assignments, provided for your reference. Each offering of CMPT 201 has its own private repos for these assignments and instructions may vary slightly. Therefore, if you are taking CMPT 201, please make sure you use the correct repos corresponding to your term instead of the repos below.
- A0: Basics of the Command-Line Interface
- A1: Neovim
- A2: Bash Scripting
- A3: Basics of Compilation
- A4: Make and CMake
- A5: Debugging Tools
- A6: Memory Layout
- A7: More on Memory Layout, Data Types, and Alignment
The second type of assignments in this course is long assignments.
- Long assignments are exponentially more challenging than short ones, specifically designed to push students to test and enhance their programming skills.
- The expectation is that each assignment will take 15-20 hours to complete. However, students sometimes report spending 30-40 hours on each long assignment. This discrepancy is likely due to varying levels of preparedness with the prerequisites.
- In other words, the level of challenge posed by the assignments can vary significantly, depending on the student's stage of proficiency in general programming and their understanding of the prerequisite materials (e.g., beginner-beginner, intermediate-beginner, advanced-beginner).
- Long assignments typically do not provide a detailed step-by-step guide. Students are expected to figure things out independently. The course is designed to train and challenge students, fostering self-reliance and problem-solving skills.
Below are the long assignments, provided for your reference. Each offering of CMPT 201 has its own private repos for these assignments and instructions may vary slightly. Therefore, if you are taking CMPT 201, please make sure you use the correct repos corresponding to your term instead of the repos below.
- A8: Simple Shell
- A9: Memory Management
- A10: Simple MapReduce
- A11: A Simple Group Chat Server with Fuzzing Clients
- A12: Simple Blockchain
Exams from Fall 2023 are available for reference.
- You will use a Docker container for everything in this course.
- You are not allowed to modify this setup in any way. Doing so is considered an academic integrity violation.
- Setting up the Docker container will take some time, e.g., a few hours, and it is necessary for the first assignment, which will take another few hours. So make sure you start this right away. Downloading and running the container (as described below) is not the end of the setup process. There are more things to do after you log in.
- First, install Docker Desktop. Choose the correct version for
your platform. If you already have Docker installed, make sure you have the latest version. Once
you install it, start it. You can run it in the background as you don't need the GUI.
- On Linux, it may be simpler to just install Docker Engine
instead of Docker Desktop. It is lighter weight than Docker Desktop. If you decide to do it,
after installation, you should also follow the instructions on Manage Docker as a non-root
user.
This way, you can avoid using
sudoto use Docker.
- On Linux, it may be simpler to just install Docker Engine
instead of Docker Desktop. It is lighter weight than Docker Desktop. If you decide to do it,
after installation, you should also follow the instructions on Manage Docker as a non-root
user.
This way, you can avoid using
- Next, make sure you have a good terminal emulator installed. For Windows, install Windows Terminal. For Mac, install iTerm2. Do not use the default Terminal app since it's not going to display certain things correctly. For Linux, the default terminal is fine at least for GNOME. We have not tested the default terminal for KDE. There are many other options such as foot, Alacritty, etc., but you need to be comfortable with editing configuration files with these. You can use those if you know what you're doing or if you are more adventurous.
- You will spend a lot of time on the terminal in this course. Therefore, it is important to use a
terminal setting that is comfortable. Change the terminal setting as follows.
- Change the width to more than 100, e.g., 110.
- Change the height to a length that is sufficiently long.
- Change the color scheme to
solarized dark. Windows Terminal, iTerm2, and GNOME Terminal should support it out of the box. Google how to change the color scheme if you are not sure. For other terminal emulators, it's likely that you need to install the theme separately.
- Open the terminal and enter the following commands.
docker create -it --name cmpt201 ghcr.io/sfu-cmpt-201/base- This command downloads our container image (
ghcr.io/sfu-cmpt-201/base) and creates an interactive (-it) container namedcmpt201. - Note that you only need to execute this once as you don't need to create an image multiple times.
- This command downloads our container image (
docker start -ai cmpt201- This command starts the container interactively (
-ai) and lets you enter the container.
- This command starts the container interactively (
- Once again, from now on, you do not need to create a container image. You only need to start
the created container. In other words, you only need to execute
docker start -ai cmpt201.
- At this point, you should see a greeting message and a command-line prompt (
~❯). - If you enter
ls, you should see two files (start_here.shandunits). - Enter
./start_here.shand read through. It will show you what to do. - When you want to stop using the container, enter
exit. - When you want to use the container again, enter the start command (
docker start -ai cmpt201). - Make sure you do not delete the container (e.g.,
docker rm). You will lose all the files, i.e., your work.
- If you see that you have a pull request, please merge it. Pull requests contain updates to the assignments.
- In order to do it, go to your repo's webpage, check if your
Pull requeststab shows a number. If it doesn't show a number, you don't need to do anything as there is no pull requests. - If it does show a number, click the tab, click a pull request, scroll all the way down, and click
Merge pull request(green button). If there are more than one pull request, please merge each and every one.
- All submissions should be done on GitHub Classroom and/or CourSys. Email submissions are not accepted.
- Submissions that do not follow the instructions down to the letter receive a zero.
- Late submissions are not accepted.
- All members of the same group receive the same grade (if there is a group).
- Students must attain an overall passing grade on the weighted average of exams (quizzes/midterms/final) in the course in order to obtain a clear pass (C- or better).
- A private Piazza post is the only way to request a re-grading. Make sure you add all
instructors (i.e., post to
Instructors). If you do not post to all instructors, we will not accept the request. - We will not accept email requests either.
- Before requesting a re-grading for an assignment, make sure you run the checker/grader/test cases and see the output. Request a re-grading only when the grade is different from your expectation based on the output.
- In your re-grading request, clearly tell us the assignment name, your SFU ID, your GitHub username, and how your grade is different from what you expect from the checker/grader/test cases output.
- If you have either chosen a wrong ID or not chosen an ID at all for GitHub Classroom, you get a 0 and there is no re-grading.
- Lastly, you have one week to request a re-grading from the time your grade is released.
- In this course, Piazza is mainly for you to help each other. If you know how to answer a question, please provide your answer and help out others.
- If no sensitive information is involved, always consider posting a question as a public question, not as a private question. It could be beneficial for other people in class.
- Depending on the volume of questions, the teaching staff may not be able to answer all questions in a timely fashion.
- The teaching staff monitors Piazza only during regular work days (Mon-Fri, excluding holidays) and regular work hours (9 am - 5 pm).
- For coding questions, it is most likely necessary to examine your code carefully and perhaps run it as well. This means that most of the times, it is much better to use the office hours than Piazza.
- For a followup discussion, always come back and mark it as resolved if the answer has been provided.
- All submitted work should be your own. For example, if you copy code from the Internet or from your peers, it is not your own work.
- You should not allow others to copy your code either. For example, your peer should not copy your code and you should not post your code on the Internet.
- Any violations of academic integrity receive an F for the course.
- We generally follow the academic integrity policies from the university.