On Thu, Mar 25, 2010 at 05:46:27PM +0200, Jukka Ruohonen wrote: > Well wouldn't it then be better to document that "they are not of any use to > the general kernel"? From a quick look around, I see also quite a few > instances of stathz and profhz being used/defined. Sure, because each platform has a specific implementation of the clock interrupts. That doesn't change anything about stathz being an implementation detail.
there's a big difference to something been an implementation detail, and something being part of the API between MI and MD code for some implementation. i think all these, and more -- hardclock() and friends -- could do with a nice document, and i think MI/MD apis should have man pages. .mrg.