> > > But such a hack is unlikely to ever get incorporated in the official
> > > kernel sources.
> >
> > Why? Because the authors find such a concept distasteful? (For similar
> > logic look at the recent raw/buffered device thread).
>
> I do sometimes wonder how much Linux growth/development is stunted
> because of personal biases...
Dont just look at these corner cases where it results in open
confrontation. Look at all the other cases where delaying a feature has
resulted in the right code.
OTOH, you are free to implement it, just dont request broken design from
the authors. AFAIR, turning off the routing cache is a _onliner_ patch...
-- mingo