I can't help myself digging into the referenced source code. The membarrier syscall can fail to allocate, returning ENOMEM. The way Folly calls it, the program would abort. Which I guess is a fair strategy but it's good to know when your synchronization primitive is actually SynchronizeOrCrash.

This is a great article but it goes into a lot of detail that can be intimidating at first.

For me, the reading that made asymmetric fences "click" is this: https://pvk.ca/Blog/2019/01/09/preemption-is-gc-for-memory-r...

It might be easier to read that first, as it also goes into practical applications, and then this one.

[flagged]