On Wed, May 13, 2020 at 12:21:16PM +0100, Mark Brown wrote:
> On Wed, May 13, 2020 at 02:04:07PM +0300, Serge Semin wrote:
> > On Wed, May 13, 2020 at 11:23:24AM +0100, Mark Brown wrote:
> 
> > > The conversion to YAML format should be the very last thing in the patch
> > > series,
> 
> > Hm, haven't heard about this requirement. Could you point me out to a doc or
> > some discussion concerning this for future reference? It's not a first DT
> > conversion patch I've submitted and non of them were addressed with such
> > request. I do understand that the order of DT concerning patches can be
> > important and agree to fix it by updating the original legacy binding first,
> > then perform a conversion. But placing the conversion in a tail of the 
> > series
> > just seems unnecessary. The patch can be dropped from any place of the 
> > series
> > if for some reason Rob would be late with review.
> 
> This is a practical observation based on the fact that there is a huge
> backlog of reviews of DT binding conversions and that those conversions
> typically go through several review cycles and that not everyone who's
> sending patches to the kernel is fully up to speed on processes or has
> strong English.  By telling people (including other people who find
> instructions on the list) to put the conversion right at the end of the
> series I am avoiding any ambiguity or confusion about ordering with
> regard to any other patches to the DT, including any new patches that
> get added to the series.
> 
> > Personally I prefer placing all DT changes in the head of the series, so Rob
> > wouldn't need to search through the whole patchset looking for the 
> > DT-related
> > patches.
> 
> Ideally the YAML conversions would be done entirely separately to other
> development rather than as part of a bigger series, they're pretty much
> orthogonal anyway.  Sadly there's obvious content collisions with any
> new development that adds DT stuff so that's not always the most
> practical thing.

Ok. I see your point. I'll move the conversion patch to the tail of the series
after rebasing the patchset on top of the spi/for-next branch. Thanks for
clarification.

-Sergey

Reply via email to