[PATCH v2 0/4] Move plat-mxc gpio driver into drivers/gpio

Grant Likely grant.likely at secretlab.ca
Fri Jun 3 11:47:06 EDT 2011


On Fri, Jun 3, 2011 at 9:26 AM, Shawn Guo <shawn.guo at freescale.com> wrote:
> Hi Grant,
>
> On Fri, Jun 03, 2011 at 08:55:58AM -0600, Grant Likely wrote:
>> On Fri, Jun 3, 2011 at 8:34 AM, Shawn Guo <shawn.guo at freescale.com> wrote:
>> > Hi Russell,
>> >
>> > On Fri, Jun 03, 2011 at 08:52:01AM +0100, Russell King - ARM Linux wrote:
>> >> On Thu, Jun 02, 2011 at 11:33:48AM +0800, Shawn Guo wrote:
>> >> >  arch/arm/plat-mxc/gpio.c                        |  361 -------------------
>> >> >  drivers/gpio/gpio-mxc.c                         |  433 +++++++++++++++++++++++
>> >>
>> > My bad here.  I should have used 'git diff --stat -M' to show the
>> > the following.
>> >
>> > .../arm/plat-mxc/gpio.c => drivers/gpio/gpio-mxc.c |  216 +++++++++++++-------
>> >
>> >> I'm wondering why just moving this driver into drivers/gpio has
>> >> resulted in it growing by 72 lines - and it's not clear from the
>> >> diffs why that is because of the way they're broken up.
>> >>
>> > Yes, I agree.  But when I did something like that to ease the review,
>> > people think it's not necessary :)
>> >
>> > http://permalink.gmane.org/gmane.linux.kernel/1143257
>>
>> The issue was bisectability: it looked like the build would break
>> after applying the first patch.  The first patch should move the
>
> Yes, the build would break only if you change Kconfig/Makefile to
> actually build it.  The patch does not enable the build of the driver
> in the patch.
>
>> driver without breaking the build, and then you can follow up with
>> driver fixes.  I don't want to see functional changes mixed in with
>> the file move change.
>>
> Understood.  Do you want me to resend the gpio-mxs and gpio-mxc patch
> sets for that?  Or can I follow the practice you and Russell
> suggested in the future posts?  I have learnt the lesson.

Please repost.

>
>> > +
>> > +static struct platform_driver mxc_gpio_driver = {
>> > +       .driver         = {
>> > +               .name   = "gpio-mxc",
>>
>> .owner = THIS_MODULE,
>>
> Do I need to re-spin the patch set to fix it, or maintainer (you or
> Sascha) can help to fix it up?
>
> BTW, do you and Sascha get the agreement on which tree the gpio-mxs
> and gpio-mxc should go through?

It will probably go through my tree because there will be a lot of
potentially conflicting GPIO changes going in this cycle.

g.



More information about the linux-arm-kernel mailing list