Re: [RFC PATCH] NET: Clone the sk_buff->iif field properly

[Jarek Poplawski <jarkao2@gmail.com>]

On Thu, Jan 03, 2008 at 04:20:06PM -0500, Paul Moore wrote:
...
> For me personally, I would argue the readability bit.  Whenever I see a 
> function/macro call I have to go find the function/macro definition 
> before I can understand what it is doing.  Granted, the macro is 
> defined "local" to the function but my point is that being able to look 
> at a line of code and understand it without having to look elsewhere is 
> a nice quality.  To loose that simply because someone wants to save a 
> few keystrokes is a mistake from my point of view.

When I first read this __skb_clone() I had mixed emotions about this
macro too. Later I didn't think about it, and only now, after your
question I've done a quick test, compared with __copy_skb_header() and
it seems there is really less rightwords eye moving. So, it was only
about this one very "local" macro.

I'm not macros fan in general: just yesterday I've cursed a bit at some
guy (I forgot the name...), who gave all these "meaningful" names to
macros in linux/pkt_cls.h. But, maybe after some time I'll start to
defend them as well... Especially when I try to imagine doing the same
without them.

Jarek P.