Re: [PATCH 5/8] iio/counter: add FlexTimer Module Quadrature decoder counter driver

[Patrick Havelange]

Hi Jonathan,

Thanks for your comments, I'll make a new version of the patch based
on your input.

William, I'll rebase the next version on top of your branch.

I'm glad the counter subsystem effort is progressing :)


Patrick Havelange.

On Thu, Feb 21, 2019 at 9:27 AM William Breathitt Gray
<vilhelm.gray@xxxxxxxxx> wrote:
>
> On Thu, Feb 21, 2019 at 10:09:54AM +0900, William Breathitt Gray wrote:
> > On Wed, Feb 20, 2019 at 04:41:54PM +0000, Jonathan Cameron wrote:
> > > On Mon, 18 Feb 2019 15:03:18 +0100
> > > Patrick Havelange <patrick.havelange@xxxxxxxxxxxxx> wrote:
> > >
> > > > This driver exposes the counter for the quadrature decoder of the
> > > > FlexTimer Module, present in the LS1021A soc.
> > > >
> > > > Signed-off-by: Patrick Havelange <patrick.havelange@xxxxxxxxxxxxx>
> > > > Reviewed-by: Esben Haabendal <esben@xxxxxxxxxxxx>
> > > Given you cc'd William, I'm guessing you know about the counter
> > > subsystem effort.  I would really rather not take any drivers
> > > into IIO if we have any hope of getting that upstreamed soon
> > > (which I personally think we do and should!).  The reason is
> > > we end up having to maintain old ABI just because someone might be using
> > > it and it makes the drivers very messy.
> > >
> > > I'll review as is though as may be there are some elements that will
> > > cross over.
> > >
> > > Comments inline.  William: Looks like a straight forward conversion if
> > > it makes sense to get this lined up as part of your initial submission?
> > > You have quite a few drivers so I wouldn't have said it needs to be there
> > > at the start, but good to have it soon after.
> > >
> > > Jonathan
> >
> > I agree, we should try to merge this as part of Counter subsystem
> > introduction rather than as another IIO Counter driver. As we determined
> > when adding support for the STM32 timers, the existing IIO Counter API
> > is fundamentally unsuitable for representing counter devices. So
> > regardless of how a new Counter API is merged, the existing IIO Counter
> > API must be deprecated.
> >
> > Patrick, I apologize for the confusion this has caused. Would you be
> > able to convert this driver to use the proposed Counter subsystem API
> > from this patchset that I believe you encountered before:
> > https://marc.info/?l=linux-arm-kernel&m=153229982404051
> >
> > Although it was last updated in October, I believe you should be able to
> > rebase that Counter subsystem introduction patchset cleanly on top of
> > the IIO tree (if there are any merge conflicts send me an email). Take a
> > look at the generic-counter.rst file under the Documentation/driver-api/
> > directory for an overview of the API; the counter drivers under the
> > drivers/counter/ directory also make good references.
> >
> > If you have any difficulties understanding the API, or any other
> > troubles, don't hesitate to ask. Hopefully, I've made the documentation
> > clear enough to make the conversion of this driver quick and easy -- and
> > if not, then it's something I need to fix, so let me know. :-)
> >
> > William Breathitt Gray
>
> Patrick,
>
> It looks like there were some minor conflicts with the v9 patchset, so
> I've rebased it on top of the latest iio tree testing branch and
> resolved the conflicts in my personal repository. Please pull from my
> personal repository at https://gitlab.com/vilhelmgray/iio.git and base
> your patches on top of the generic_counter_v10 branch.
>
> William Breathitt Gray