Re: [PATCH v5 00/21] Move Hisilicon 6421v600 SPMI driver set out of staging

[Mauro Carvalho Chehab <mchehab+huawei@kernel.org>]

Em Wed, 27 Jan 2021 11:19:36 +0100
Greg Kroah-Hartman <gregkh@linuxfoundation.org> escreveu:

> > This patch series finish addressing support for Hikey 970
> > SPMI controller, PMIC and regulators.
> > 
> > This version was generated with -M, in order to make easier
> > to merge upstream.  Also, rebased on the top of v5.10,
> > without any dependencies from the other patch series
> > I'm submitting for this board.
> > 
> > Yet,  patch 18 to 20 modifies drivers/staging/hikey9xx/Kconfig
> > and drivers/staging/hikey9xx/Makefile. So, trivial conflicts
> > will rise if they're applied via different trees, as they all
> > remove some lines from such files.   
> 
> I've applied the first 13 patches, except for patch 3, as that did not
> apply, to my tree, thanks.

Ok. I'll rebase the remaining patches on the top of staging-testing branch.

> On Wed, Jan 27, 2021 at 10:08:16AM +0000, Lee Jones wrote:
> > On Wed, 27 Jan 2021, Greg Kroah-Hartman wrote:
> >   
> > > On Tue, Jan 26, 2021 at 06:11:24PM +0000, Mark Brown wrote:  
> > > > On Tue, Jan 26, 2021 at 07:02:39PM +0100, Greg Kroah-Hartman wrote:  
> > > > > On Tue, Jan 26, 2021 at 05:57:52PM +0000, Mark Brown wrote:  
> > > >   
> > > > > > Is there a branch we can pull from?  
> > > >   
> > > > > Once 0-day passes, you can pull from my staging-testing branch from
> > > > > staging.git on git.kernel.org if you want.  Give it 24 hours to pass
> > > > > before it hits that location.  
> > > > 
> > > > Thanks.  
> > > 
> > > Should be out there now if you want to pull.
> > >   
> > > > > Do you need a tag to pull from?  
> > > > 
> > > > It'd be nice but not essential.  
> > > 
> > > Why do you want/need this?  Having these changes in your tree is good,
> > > but what about other coding style cleanups that I will end up applying
> > > over time before the 5.12-rc1 merge window opens?  Are you wanting to
> > > take the moved driver in your tree, or something else?
> > > 
> > > Traditionally moving drivers out of staging can be done 2 ways:
> > > 	- all happens in the staging tree, I take an ack from the
> > > 	  subsystem maintainer that this is ok to do.
> > > 	- A new driver enters the "real" subsystem tree, and then I
> > > 	  delete the driver in the staging tree.  This doesn't preserve
> > > 	  history as well (not at all), but can be easier for trees that
> > > 	  move quickly (like networking.)
> > > 
> > > Which ever works for you is fine with me, but relying on the code to
> > > stay "not touched" in my tree after you pull it almost never happens due
> > > to the number of drive-by coding style cleanups that end up in the
> > > staging tree every week.  
> > 
> > I would have expected the whole set to be merged as a set into a
> > single tree, placed on an immutable branch and a pull-request to be
> > sent out for the other maintainers to pull from (if they so wished).
> > 
> > This would ensure development could continue on any/all of the
> > affected drivers/files.
> > 
> > If it's not too late, I'd be more than happy to facilitate.  
> 
> Given these patches are already in my public tree, that might be a bit
> harder, why the huge work for this?  Worst case, I just keep all of the
> patches that do not actually move the code in my tree, and then things
> can move after 5.12-rc1.

Whatever works best for Lee/Mark.

From my side, I can re-submit the move patches and the DTS ones to
be applied after 5.12-rc1, if this would be the preferred way.

Thanks,
Mauro