This is a repository for Pintos projects, done for the CSED312 course at POSTECH (Fall 2025). Coding is done up to project 3. For each project, the submission code is organized in branches. (1_threads, 2_userprog, 3_vm)

There are a few things to note about our implementation:

• Project 2 is built on top of ONLY the alarm clock functionality of Project 1.
• Project 3 is built on top of the entirety of Project 2.
• In project 2, our implementation for zombie child processes is a bit wonky. In particular, when a child process dies before the parent, then instead of the thread being deleted, the child process's state stalls in the sys_exit() function by a semaphore zombie_sema. So, the child process's thread_status (in struct thread) would be in THREAD_BLOCKED. (This is our implementation of the "zombie" state.)
  • Only when the parent dies will it reap all the zombie children, i.e., increase every children's zombie_sema. This will let the stalled zombie children to start the dying sequence. Also, children that will die in the future will not be stalled by the zombie_sema anymore and die safely.
  • A better design would be, if the child dies before the parent, to delete the struct thread of the children while keeping the info of the children process in another (more lightweight) struct.
• In project 3, we don't correctly utilize the "used bits". In particular, In our implementation, every frame that was not allocated via mmap is sent directly to swap as soon as it is selected for eviction, without any additional checks.
  • For frames originally backed by files, such as executable pages, if the dirty bit is not set, it would be much more efficient from the perspective of swap-space management to simply discard the contents and change the SPTE state to FILE_LAZY instead of writing them out to swap.
  • We expect this could be fixed by adding a flag such as bool file_backed to struct spt_entry and checking that flag right before eviction.

Also, note about the reports: the design reports are very rushed and may contain many errors. The final reports are much more polished. I tried to store all the core ideas and mechanisms for my code.

Good luck! Also, start early and thank yourself later. I never started early, and it damaged my mental health a bit..