> Yah, churning code for no apparent gain is really just kinda dumb in my
> book.  Removing __P() gains us ABSOLUTELY NOTHING.

Well, I had to retract my type argument, as soon as I sent the mail I
realized that nothing happens if the whole call is replaced. (Duhhh, *slap*.)

But __P still preserves the prototype syntax, which is good, and it
does no harm.

And as the developer count explodes, I've got another argument against
source code churns: they set a bad example ... when the established
developers do it (and to the kernel!) it implies that it's a good thing
for every developer to do somewhere.