A complete guide to Linux process scheduling (2015) [pdf]

trepo.tuni.fi

192 points·kklisura··8 comments

How does io/uring pollepoll asyncio fit into 2023 Linux user process scheduling? Does this or any other material cover this for rust async await, Python async await?

For your second question, you could write a basic async/await Python or Rust application, then use strace and ftrace to figure out what system calls are happening and what the kernel is doing with them.

https://strace.io/

https://www.kernel.org/doc/html/v5.0/trace/ftrace.html

People shouldn’t give their pdfs names like “GRADU-1428493916.pdf.”

It's a thesis from a university publication catalogue, and the file name might be just conforming to some kind of a guideline from their university library.

"Gradu" is a common shorthand for "pro gradu" which is the term used in Finland for a master's thesis. The numbers might just be some kind of an id. Many other theses from the same archive also seem to follow some kind of a similar (though not quite identical) pattern in their file names.

It's not the most descriptive file name but it might not be the author's choice.

I’m not blaming the author at all. But surely the server administrator could map the file path onto a nicer name?

Should append [2015] to the submission title. This is practically ancient news due to the rate of changes in the Linux kernel in this domain.

Is there a 2023 version? What parts are obsolete now?

The CFS scheduler went into Linux in 2007 and is still the main scheduler, as reflected in this paper, so nothing major has changed in that regard. Most changes are minor.

The paper talks about process priority and niceness - a lot of how that works on Unix has not changed in almost fifty years.

Reply on news.ycombinator.com