[Bulk] [PATCH] timer: vt8500: Move timer code to drivers/clocksource

[Olof Johansson]

On Mon, Jan 14, 2013 at 8:12 PM, Tony Prisk <linux at prisktech.co.nz> wrote:
> On Mon, 2013-01-14 at 12:07 -0800, Olof Johansson wrote:
>> On Mon, Jan 14, 2013 at 06:47:35PM +1300, Tony Prisk wrote:
>> > On Mon, 2013-01-14 at 18:13 +1300, Tony Prisk wrote:
>> > > On Mon, 2013-01-14 at 18:09 +1300, Tony Prisk wrote:
>> > > > This patch moves arch-vt8500/timer.c into drivers/clocksource and
>> > > > updates the necessary Kconfig/Makefile options.
>> > > >
>> > > > Signed-off-by: Tony Prisk <linux at prisktech.co.nz>
>> > > > ---
>> > > >  arch/arm/mach-vt8500/Kconfig       |    1 +
>> > > >  arch/arm/mach-vt8500/Makefile      |    2 +-
>> > > >  arch/arm/mach-vt8500/common.h      |    1 -
>> > > >  arch/arm/mach-vt8500/timer.c       |  184 ------------------------------------
>> > > >  arch/arm/mach-vt8500/vt8500.c      |    1 +
>> > > >  drivers/clocksource/Kconfig        |    3 +
>> > > >  drivers/clocksource/Makefile       |    1 +
>> > > >  drivers/clocksource/vt8500_timer.c |  184 ++++++++++++++++++++++++++++++++++++
>> > > >  include/linux/vt8500_timer.h       |   22 +++++
>> > > >  9 files changed, 213 insertions(+), 186 deletions(-)
>> > > >  delete mode 100644 arch/arm/mach-vt8500/timer.c
>> > > >  create mode 100644 drivers/clocksource/vt8500_timer.c
>> > > >  create mode 100644 include/linux/vt8500_timer.h
>> > >
>> > > Darn.. forgot the -m again. I'll await your feedback regarding the
>> > > basing of the patch first (and any other feedback), then I'll redo it
>> > > with the correct stats.
>> > >
>> > > Regards
>> > > Tony P
>> >
>> > Oh grr.. forget this completely. It doesn't take into account the
>> > patches I already sent for WM8850.
>> >
>> > I guess it needs to be based on timer/cleanup + vt8500/wm8x50.
>> >
>> > Need a little advise on how to handle this one please :)
>>
>> The normal way to handle these kind of dependencies is to base them on merges
>> of the needed branches. Based on the later email, you only seem to need
>> timer/cleanup, but if you would have needed the other one, then you'd merge
>> that on top of timer/cleanup, and then add your patches.
>>
>> Of course, ideally you would do the cleanup, then add the wm8x50 features,
>> but in reality work doesn't always pan out that way, so you end up with
>> cleanups that depend on including new features in the same (sweeping)
>> cleanup since they have already been merged. That's when things sometimes
>> get hairy, and we need to start a second cleanup branch that's "after"
>> the feature branch in the sequence of topics. But it should be rare,
>> and in your case it seems like it wasn't needed.
>>
>>
>> -Olof
>>
>
> Just to clarify what I did (and to make sure it was as you understood
> it):
>
> #1) I wrote the patch on top of timer/cleanup. This is the branch the
> patch was written for.
>
> #2) I then pulled timer/cleanup and merged vt8500/wm8x50 on top, then
> reapplied the patch from #1 - it applied cleanly.
>
> What I have just realised is that you might?? get a conflict when you
> merge vt8500/wm8x50 on top of wherever this patch ends up due to the few
> lines at the top of arch-vt8500/Kconfig having changed (the addition of
> SELECT VT8500_TIMER). This should be trivial to fix (I assume).

Yeah, a couple of trivial conflicts are fine, but when they start to
stack up we normally have to push back since some of these are exposed
all the way up to Linus. We can also be a little more strategic in how
we chose to pick bases for the topic branches to avoid some of this,
which is what I intend to do this release cycle if it pans out.

In this case, there are no manual conflict resolution to be done,
everything panned out just fine as it was.


-Olof