Re: [PATCH] irqchip: nps: add 64BIT dependency

From: Thomas Gleixner
Date: Tue May 17 2016 - 15:41:51 EST


On Fri, 13 May 2016, Vineet Gupta wrote:
> On Friday 13 May 2016 03:55 PM, Marc Zyngier wrote:
> > On 13/05/16 10:51, Arnd Bergmann wrote:
> >> On Friday 13 May 2016 14:05:41 Vineet Gupta wrote:
> >>> On Friday 13 May 2016 01:54 PM, Marc Zyngier wrote:
> >>>> On 12/05/16 22:03, Arnd Bergmann wrote:
> >>> ...
> >>>>>
> >>>>> config EZNPS_GIC
> >>>>> bool "NPS400 Global Interrupt Manager (GIM)"
> >>>>> + depends on ARC || (COMPILE_TEST && !64BIT)
> >>>>> select IRQ_DOMAIN
> >>>>> help
> >>>>> Support the EZchip NPS400 global interrupt controller
> >>>>>
> >>>>
> >>>> Acked-by: Marc Zyngier <narc.zyngier@xxxxxxx>
> >>>>
> >>>> As I've already started collecting fixes that are aimed at -rc1 (mostly
> >>>> to avoid dependencies), I can queue that as well.
> >>>
> >>> There is a slight logistics issue here - as agreed the driver will go in 4.7-rc1
> >>> via ARC tree. So either I pick the fix for rc1 or you apply it post rc1 - or
> >>> towards the end of rc1 ?
> >>>
> >>
> >> I'd say the best option is to have you pick up the fix for the ARC tree,
> >> but either way works.
> >
> > That'd work for me too (I've acked it anyway). Just let me know what you
> > decide to do.
>
> I'd prefer Marc takes it post rc1. The reason being chances of merge conflicts
> between ARC and tip trees increase with ARC tree changing drivers/irqchip/*. We've
> seen two of those already which Stephen fixed up in linux-next. Although
> admittedly the conflicts are trivial and given the location of this diff hunk it
> might not happen at all....

If the driver is new and in ARC then the fix should go into ARC and shipped
with the pull request.

Thanks,

tglx