Re: [PATCH] if (foo) kfree(foo) /fs cleanup + reverted JBD code path

[Zwane Mwaikambo]

You present some good arguments against the patch, i did it initially to
remove the double check and i would think the most common path would be a
kfree in any case (I can't back this up with proof) so its not that
expensive, so instead of;

if (foo) {
	kfree(foo);
	foo = NULL;
}
frob();

we can /* as whoever wrote kfree intended */

kfree(foo);
foo = NULL;
frob();

I would think in any case if we were looking at optimisation, the
assignment is roughly on par with the compare and saves you a bit on the
other case.

Alan Cox also suggested the BUG() trigger in there, with the
resultant breakage that would bring about. But i think its ok as long as
you don't call kfree with uninitialised variables or some such or forget
to set them to NULL at some point.

Thanks,
	Zwane Mwaikambo