On Fri, Apr 08, 2022 at 10:32:28PM +0200, Peter Zijlstra wrote: > > > This weird option is having us upgrade quite a few 'inline' to > > > '__always_inline'. > > > > As is, the assumption that __init functions only call other __init > > functions or __always_inline is a brittle house of cards that leads to > > a "what color is your function" [0] scenario, and leads to code that > > happens to not emit warnings for compiler X (or compiler X version Y). > > There's also curious exceptions in modpost that look like memory leaks > > to me. So I don't see __always_inline that way (also I'm in the 'inline' should be '__always_inline' camp). To me inline is more like: 'instantiate that pattern *here*'. It's like CPP macros, only less horrible. You get the code generated according to the local rules (instrumentation yes/no, section, and whatever other function attributes we have that affect code-gen). So with inline we can get the same pattern instantiated a number of different times, leading to different actual code, without having to type the whole thing multiple times (which would be terrible for maintenance) etc.. Combine __always_inline with constant propagation of inline function 'pointers' and you get do beautiful things ;-) /me runs
