[PATCH v8 0/9] initial suport for Alphascale ASM9260

[Jason Cooper]

On Sun, Nov 02, 2014 at 08:56:46PM +0100, Oleksij Rempel wrote:
> Am 02.11.2014 um 19:31 schrieb Jason Cooper:
> > On Sun, Nov 02, 2014 at 07:51:42AM +0100, Oleksij Rempel wrote:
> >> Am 02.11.2014 um 03:11 schrieb Jason Cooper:
> >>> On Sun, Oct 26, 2014 at 03:39:02PM +0100, Oleksij Rempel wrote:
> >>>> Will it be better to split this patch set?
> >>>>
> >>>> For example:
> >>>> part 1
> >>>>>   ARM: add mach-asm9260
> >>>>>   ARM: add lolevel debug support for asm9260
> >>>> part2
> >>>>>   ARM: irqchip: mxs: prepare driver for HW with different offsets
> >>>>>   ARM: irqchip: mxs: add Alpascale ASM9260 support
> >>>> and all other patches separately.
> >>>>
> >>>> this will reduce review time for you and give some hope for me :D
> >>>
> >>> I honestly prefer to keep the series together.  The subject lines make
> > 
> > To be clear, I meant the emails being in the same thread.
> 
> can be split in to parts, but should be within this tread?

Sure, no need to over-analyze it.  I was just expressing a preference
for seeing the big picture.  Some maintainers don't care, as long as
they are sent the part they are responsible for, and one or two -only-
want what they are responsible for.

The most important thing, which I mentioned before, is letting us know if
the changes in one subsystem have a compile-time dependency on the
changes in another subsystem.  Then we need to be careful how we merge
the patches.

> >>> it clear which parts I need to worry about merging.  The big thing is
> >>> if there are compile-time dependencies between the different subsystems.
> >>> It doesn't look like there are, but if so, just bring it to our
> >>> attention.
> >>
> >> Ok, i'll resend updated version against latest arm-soc/for-next.git, or
> >> should i take other branch?
> > 
> > In general it's best to base against one of Linus' tags (eg v3.18-rc1)
> > and then rebase against something different only if asked.  That'll be
> > per subsystem, though.
> 
> just notice, arch/arm related patches are ok on arm-soc/master but fail
> fail on arm-soc/for-next.git

arm-soc/master is still against v3.17-rc3.  arm-soc/for-next is against
v3.18-rc1.  How does this series fare against v3.18-rc1?

I've heard it frequently in the past that the goal isn't to remove all
conflicts.  Maintainers, like arm-soc, prefer to see conflicts because
it lets them know where two developers were working on the same code.
As a courtesy, we try to let them know of the conflict ahead of time,
but that's mainly the job of the sub-arch maintainers.

thx,

Jason.