Rethinking race-free process signaling

Purely timestamp-based PIDs will never fly because they'll never be able to perform well enough. An x86 HPET, shared between all cores in the system, is only guaranteed to provide 10MHz/100ns granularity and is fairly expensive to access on top of that. TSCs are basically free to read but nigh impossible to read _consistently_ in a frequency-hopping multicore system.

A scheme that encodes the CPU ID and TSC might be pretty efficient and would work, if we were to abandon the guarantee that IDs have any correlation to time. But, this being Linux, someone out there almost certainly depends on that implementation detail.