0%

blog-rcore-ctzsm

Stage 1 (Rustling)

I tried to learn Rust programming language many times, but it was from start to give up every time. Mostly because I never used it in my real job or other personal works. The recent one was trying to implement a vulkan game engine with Rust, but I do have other stuff to do, so that was never finished.

I spent about 3 nights to finish the rustling problem set, it was fun. For the first half, it was a refresh to my already rusty Rust memory. I did learn something different during the second part. It was indeed engaging. The compiler message is a very powerful tool in general, Many times I just need to follow what the compiler said to answer the question. I did have to Google a lot and consult The Book and the official Rust lang document.

With these practices, I started to get used to Rust, I wouldn’t say I am fully equipped comparing with my C++ and Java skill, but I think it is a good start.

Stage 2 (Lab)

Lab 1

After setting up the environment, I realized that I had read the tutorial book before, but just brief. It was probably because I followed an OS course from Nanjing University taught by Yanyan Jiang on Bilibili. I even have a qemu installed in my WSL2. In general, the guide book is very readable. I followed the book to understand the code structure and WeChat group helped some of the common questions on newer version of qemu, which saved my a lot of time.

However, 2 weeks are super short, I can’t finish reading the first two chapters to understand all the details and I have to start chapter 3 right away to meet the deadline. Fortunately the actual lab isn’t very hard. The first lab is just to implement sys_task_info, there is a big hint that the TaskControlBlock has a TaskStatus field, we also have that in TaskInfo struct, I immediately realized that we just need replace that field with the TaskInfo struct type. The rest of the work is fairly straightfoward.

Lab 2

This lab is indeed harder than the previous one, again I didn’t have time to read the book in a great detail, so I had to make a guess sometimes. But fortunately the code itself is very readable as well, I can find all the info I need in the code. The first hint I get is the translated_byte_buffer() function, which reminds me memcpy in C/C++, if I can write a function to convert everything to a byte array, I then can use translated_byte_buffer() with that to easily copy data from kernel space to user space. That’s exactly what I did in my code to solve this.

The second part is mmap and munmap, there are many restrictions on the params, which simplied the whole story. I am happy to see this so I can pass the test cases, but wondering how a real world OS implements them.

Lab 3

I spent more time than before to read chapter 5, which didn’ help me a ton in the lab, but I still think that is worth because it provides me some details about how this particular OS is designed.

I was planning to rewrite all of lab 2 and lab 1 requirements again, but I don’t have that time, so I spent about 1 hour to do cherry-pick from ch4. The spawn part was easy if we understood how fork and exec work.

The stride part is also not too hard since the problem statement is clear about what I need to do. I did have some bugs (and I probably still have some bugs even I passed the test cases), it was not too hard to debug them. I ended with using a binary heap to implement the algorithm.

Summary

That’s all I did so far, I wish I can finish the rest two labs soon, 2 weeks are super tight, not sure why this was arranged in this way, the curves in stage 1 and stage 2 are completely different. But maybe that was on purpose, so everyboday has to find their way out, I did have fun and finally had time to get my hands on code.

Big thanks to all the hosts and staff!