[PATCH v4 0/4] Energy Micro efm32 support

[u.kleine-koenig@pengutronix.de (Uwe Kleine-König)]

Hello Arnd,

On Tue, Nov 19, 2013 at 02:11:41PM +0100, Arnd Bergmann wrote:
> On Tuesday 19 November 2013, Uwe Kleine-K?nig wrote:
> > I'd like to get patch 4 (ARM: new platform for Energy Micro's EFM32
> > Cortex-M3 SoCs) in. In the state as it was sent here it build depends on
> > patches 1 - 3.  Patches 2 and 3 are in Russell's patch tracker (7890/1
> > and 7889/1). What do you think about patch 1? Some of the
> > NEED_MACH_TIMEX_H are already fixed by patches that I sent out. The
> > options here are:
> > 
> >         - rework patch 4 to not depend on patch 1 (easy)
> >         - merge v5 of patch 1 (which is conservative, i.e. introduces
> >           more NEED_MACH_TIMEX_H as probably will be needed in 3.14-rc1
> >           and fix up later)
> 
> These both sound fine to me in retrospect, unless someone has objections.
> I would prefer the Kconfig solution I suggested (with the help text
> fixed to address Russell's objections, and the list of platforms changed
> to match your v4 patch), but I don't want to force you to go through
> more revisions for this.
I think with all my patches only 2 or 3 platforms are left that need
timex.h. (Note this is a related but still different problem to the
gettimeoffset stuff.) And I plan to get rid of it again for 3.15-rc1, so
not needing to rework it sounds right to me.

> >         - depend on all sent patches and coordinate accordingly (at
> >           least: watchdog, clocksource, rtc).
> > 
> > I'd prefer the 2nd option as I didn't get Acks on all patches needed for
> > the third. What do you think?
> 
> Makes sense. It would be nice to still follow up on those patches
> and merge them eventually.
Sure.

> > Russell, if you are happy with patches 2 and 3 and would apply them to
> > your tree I could prepare a branch for the arm-soc people to pull which
> > bases on your tree and has patch 1 (v5).
> 
> Did you get in touch with Jonathan about the patch set to make ARMv7-M
> support coexist with multiplatform? I think that would be the best
> solution in the long run, and IIRC there were no more objections at the
> ARM mini summit to the draft patch.
Yeah, the previous revision used multiarch. I dropped it because of
objections from Russell. It's only some Kconfig shuffling away to use it
again.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-K?nig            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |