Kernel Development Learning Pipeline

Midpoint Presentation Guidelines 🕴️

You will create a brief, pre-recorded presentation on research into a Linux kernel related topic of your choice

Presentation requirements

• 10 ± 1 minutes long

• Must include some sort of visual component

  • Slideshow

  • Diagrams

  • Etc.

• Must include some sort of live demo in the terminal, even if brief

• Must include discussion of some snippet(s) of code that is relevant to your topic

• Plenty of room to be creative 🙂

What and how to submit

• You will propose your choice of topic ahead of time

  • You may chose a topic from the list below or come up with your own

  • Each student must choose a unique topic, so check the #midpoint channel to see if your choice is available

  • Submit your proposal in the #midpoint channel on Matrix by midnight 12 July 2024

  • Your proposal must be approved by course staff

• You will submit your presentation by posting a link to your recording in Matrix

  • Suggestion: use OBS to record your presentation as a livestream to youtube and set the video visibility to "unlisted" so only those with the link can access

  • You may use any video hosting platform of your choice, but make sure anyone with the link can acces your video before you submit

  • There is no peer review and additional final submission for this assignment

  • Submit the link to your video in the #midpoint channel on Matrix by midnight 22 July 2024

Example topics

• What is the OOM killer?

• Why are there multiple versions of some syscalls with numbers at the end? What issue were they added to address?

• Why do some syscalls have a version ending in at? What issue were they added to address?

• What is the difference between Hurd and Linux?

• How does one leverage kernel features to do performance profiling?

• What role does the rust programming language play in the kernel?

• How does Read-Copy-Update (RCU) locking work?

• How do concurrent Linux kernel data structures mitigate race conditions?

• What is BPFML? (performance optimization in page fault prediction using ML) https://gitlab.com/aarcange/bpfml

• How does Facebook's sched_ext BPF extension re-implement the scheduler? https://github.com/sched-ext/sched_extt https://www.youtube.com/watch?v=8kAcnNVSAdI