[PATCH 1/2] of: Add for_each_compatible_node_from iterator

[Sascha Hauer]

On Tue, Jan 05, 2016 at 09:05:35AM +0100, Yegor Yefremov wrote:
> On Tue, Jan 5, 2016 at 8:58 AM, Sascha Hauer <s.hauer at pengutronix.de> wrote:
> > On Mon, Jan 04, 2016 at 07:07:27PM +0000, Trent Piepho wrote:
> >> On Mon, 2016-01-04 at 09:32 +0100, Sascha Hauer wrote:
> >> > On Sat, Dec 19, 2015 at 12:13:59AM +0000, Trent Piepho wrote:
> >> > > The existing iterator for_each_compatible_node() searches for each
> >> > > compatible node starting from the root of the loaded device tree.
> >> > > This means it only works on the barebox device tree and not the tree
> >> > > to be passed to the Linux kernel, which is what an of_fixup would
> >> > > probably want to use.
> >> > >
> >> > > This adds for_each_compatible_node_from(), which takes an additional
> >> > > parameter of a root to search from.  This could be the device tree to
> >> > > be used for the kernel.  It could also be used to search just a
> >> > > subtree.
> >> > >
> >> > > Signed-off-by: Trent Piepho <tpiepho at kymetacorp.com>
> >> >
> >> > Applied, thanks
> >> >
> >> > > ---
> >> > >
> >> > > It's possible the fixups in cm_cogent_fixup() and hb_fixup() should
> >> > > be using this.  It's not clear to me if they want to modify the barebox
> >> > > device tree or the Linux device tree or both.
> >> >
> >> > It's always the Linux device tree that is fixed up in the OF fixups.
> >>
> >> Sent patch to fix them.
> >>
> >> Couldn't one also use the of fixup system to modify the barebox DT?  In
> >> order to support multiple board variants, I added DT nodes that specify
> >> what nodes should be enabled and/or disabled for different board
> >> versions.  An OF fixup applies this to the Linux DT.  I haven't had to
> >> modify the barebox DT for different boards but anticipate that happening
> >> for the next board and I was planning to use the same system.
> >
> > I think you don't need the fixup system to accomplish that. Just hook up
> > to an initcall early enough and modify the barebox device tree. It
> > shouldn't be necessary to register a callback first and then wait for
> > its execution.
> 
> What initcall can be used to change the device tree, that is already
> loaded into memory and before Linux is started?

You want to change the device tree that is passed to Linux, right? In
this case you can use of_register_fixup(). We were talking about
changing the device tree that barebox uses for itself. For this
everything after core_initcall will work. It should be early enough
though so that the device that is ought to be changed has not probed, so
I suggest doing it as early as possible, that would be
postcore_initcall.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |