On Thu, Oct 26, 2017 at 09:00:03AM +0200, Greg KH wrote: > On Thu, Oct 26, 2017 at 12:59:08AM +0200, Jason A. Donenfeld wrote: > > On Thu, Oct 26, 2017 at 12:27 AM, Tobin C. Harding <m...@tobin.cc> wrote: > > > How good is unlikely()? > > > > It places that branch way at the bottom of the function so that it's > > less likely to pollute the icache. > > But always measure it. Lots of times (old numbers were 90% or so), we > get the marking wrong, so please, always benchmark the thing to verify > it actually is doing what you think it should be doing, otherwise it > could make the code worse.
Does this come under 'premature optimization is the root of all evil'? Should we be leaving out things like unlikely() and __read_only until the code has been profiled? thanks, Tobin.
