The Magic Ring Buffer (2012)

Can you prove they are always safe? (at least on the level of linux kernel LOCKDEP) I bet you cannot, because 1.1.0 changelog mentions "fixing" a deadlock by tossing functionality.

Likewise, I see a lot of places that do reference counting (those _ref/_unref calls). Those are probable deallocations on heap from RT side, which have unbounded runtime unless you have a realtime memory allocator. I haven't noticed anything like that in the code.

This coming from a guy who supposedly "has carefully read the documentation for every audio backend and understands the purpose of every line of code."

For people who want to get a primer on such issues, I recommend this book: http://www.amazon.com/Real-Time-Systems-Programming-Language... It's a bit old despite updates (e.g. no C++11, no C99), but still highly relevant. The techniques themselves and the analysis parts are well written.

Uggh. I know how you feel. For a while now, I've wanted to create some sort of audio processing engine (either a very versatile LADSPA/VST/AU/etc plugin or a minimal DAW) wherein all the different components that usually run their own code to generate/process audio are described as data instead (e.g. transfer functions), and then the host decides how to actually compute the audio. This would give lots of room to try new optimizations that aren't usually possible, and you only have to optimize one area of critical code, rather than optimizing each individual plugin. This is especially beneficial when trying to make conditional use of processor extensions like SSE.

On recent kernels the fast implementation of remap_file_pages has been removed because keeping non linear mapping from regressing was too much of a burden; instead a slow emulation has been added. The emulation should still be faster than doing the mmap dance by hand.

The documentation is pretty terrible, though, so some experimentation may be necessary. It's also even less like the POSIX parts than the NS bits are...

(Apple doesn't seem to tell you much; OK references are https://www.gnu.org/software/hurd/gnumach-doc/Virtual-Memory... http://www.amazon.co.uk/Mac-OS-Internals-Systems-Approach/dp..., and, of course, http://www.opensource.apple.com/source/xnu/. The source has some reference-style documentation in it.