Re: [PATCH v2] mac80211: better rate control algorithm selection

[Sam Ravnborg <sam@ravnborg.org>]

On Wed, Jan 02, 2008 at 06:45:01PM +0100, Johannes Berg wrote:
> 
> > > +# objects for PID algorithm
> > > +rc80211_pid-y := rc80211_pid_algo.o
> > > +rc80211_pid-$(CONFIG_MAC80211_DEBUGFS) += rc80211_pid_debugfs.o
> 
> > > +# build helper for PID algorithm
> > > +rc-pid-y := $(rc80211_pid-y)
> > 
> > Is this extra indirection ( rc80211_pid-y ) a preparation
> > for future rate algorithms or was it just coped from my proposal?
> > It is not needed and it does not help readability.
> 
> As far as I can tell rc80211_pid-y is needed to build rc80211_pid.o from
> rc80211_pid_algo.o and rc80211_pid_debugfs.o for the modular case, no?

Correct - I just missed that when reading your patch.
And I guess this was one of the things I got wrong in my first proposal too.

Looks good then.

> Hence, I just wanted to keep it symmetric for the non-modular case.
> 
> > A second note - but more as a personal taste..
> > I prefer += assignment in preference of continued lines.
> > So 
> > +mac80211-y := \
> >  	ieee80211.o \
> >   	ieee80211_ioctl.o \
> >  	sta_info.o \
> > 
> > would become:
> > +mac80211-y := ieee80211.o ieee80211_ioctl.o
> > +mac80211-y += sta_info.o
> 
> Yeah, I was hoping the diff would be smaller this way because it was
> already this way before.

Understood.

	Sam