[PATCH] Some patches around profiling

[David Disseldorp]

On Thu, 2 Oct 2014 08:31:27 -0700, Richard Sharpe wrote:

> On Thu, Oct 2, 2014 at 8:19 AM, David Disseldorp <ddiss at suse.de> wrote:
> > On Thu, 2 Oct 2014 17:00:12 +0200, Volker Lendecke wrote:
> >
> >> Review&push would be appreciated.
> >
> > These look really good at first glance. I'm all for moving ugly ifdef
> > logic out of the source and into the makefiles where possible :-).
> 
> While I don't have strong feelings about this, it does seem that it
> complicates the lives of people who are learning their way around the
> code. This means they need to read the wscript to understand what is
> going on.
> 
> However, I do not like the conditional compilation style that mixes in
> #ifdefs all over the place, and prefer that there be only on #ifdef
> and two blocks of code, one with the code that is used when the define
> is set and one that is used when the define is not set.

I'm happy with anything that stops them from being littered through the
.c sources. The kernel style also tries to limit #ifdefs to header
files:
http://www.kroah.com/linux/talks/ols_2002_kernel_codingstyle_talk/html/mgp00029.html

Cheers, David