Re: [PATCH 2/2] Driver core: reduce duplicated code

["Uwe Kleine-König" <u.kleine-koenig@pengutronix.de>]

Hi Greg,

On Wed, Jun 16, 2010 at 01:53:32PM -0700, Greg KH wrote:
> On Tue, Jun 15, 2010 at 11:05:00AM +0200, Uwe Kleine-König wrote:
> > On Tue, Jun 15, 2010 at 10:47:56AM +0200, Uwe Kleine-König wrote:
> > > This makes the two similar functions platform_device_register_simple
> > > and platform_device_register_data one line inline functions using a new
> > > generic function platform_device_register_resndata.
> > I forgot to add some comments to this mail, ... sorry.
> > 
> >  - I'm not completely happy with the name of the new function.  If
> >    someone has a better name please tell me.
> 
> I don't like it either, what is "resndata" supposed to stand for?
resources and data -> res'n'data -> resndata.

> >  - can platform_device_register_resndata be moved to __init_or_module?
> 
> I doubt it, but try it and see if a build warns about it.
for x86_64_defconfig + MODULES=n there are two section mismatches:
	dock_add
	regulatory_init

regulatory_init is only called from

	static int __init cfg80211_init(void)

.  (I just sent a patch[1] that moves regulatory_init to .init.text.)

dock_add (in drivers/acpi/dock.c) is a bit harder.  I will take a look
later if it can go to .init.text (together with a few more functions),
too.

> >  - I moved the kernel docs to the header but didn't test if they are
> >    picked up when generating docs.  Even if not, there is no better
> >    place, is there?
> 
> No, that's the proper place, but make sure the docbook source is also
> picking up the .h file, I don't know if it currently does.
It does not, will fix in a v2.

Thanks
Uwe

[1] http://mid.gmane.org/1276846735-12105-2-git-send-email-u.kleine-koenig@pengutronix.de
-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |