Re: [PATCH] irqchip/bcm2836: Move SMP startup code to arch/arm

[Stefan Wahren]

Am 10.05.2017 um 17:38 schrieb Marc Zyngier:
> On 10/05/17 16:32, Phil Elwell wrote:
>> On 10/05/2017 16:31, Marc Zyngier wrote:
>>> On 10/05/17 16:07, Phil Elwell wrote:
>>>> On 10/05/2017 14:32, Marc Zyngier wrote:
>>>>> On 10/05/17 14:26, Marc Zyngier wrote:
>>>>>> One of the RPi-2/3 irqchip's key features is that it contains some
>>>>>> SMP startup code for the 32bit ARM architecture version. The only
>>>>>> reason I can imagine for this is "RPi is special".
>>>>>>
>>>>>> Let's move this code where it belongs (in the platform support code),
>>>>>> creating a shared include file for this purpose.
>>>>>>
>>>>>> Signed-off-by: Marc Zyngier <marc.zyngier@xxxxxxx>
>>>>> Also:
>>>>> Fixes: 41f4988cc287 ("irqchip/bcm2836: Add SMP support for the 2836")
>>>>>
>>>>> 	M.
>>>>>
>>>> The patch has some checkpatch warnings, but otherwise:
>>> Bah...
>>>
>>> WARNING: line over 80 characters
>>> #51: FILE: arch/arm/mach-bcm/board_bcm2835.c:30:
>>> +static int bcm2836_smp_boot_secondary(unsigned int cpu, struct task_struct *idle)
>>>
>>> WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
>>> #194: 
>>> new file mode 100644
>>>
>>> Yup, the checkpatch stamp of approval. My screen has stopped being 
>>> limited to 80 chars at some point between 1989 and 1991. As for the
>>> MAINTAINERS file, that's probably for the maintainers to pick it up.
>>>
>>>> Reviewed-by: Phil Elwell <phil@xxxxxxxxxxxxxxx>
>>>> Tested-by: Phil Elwell <phil@xxxxxxxxxxxxxxx>
>>> Thanks. I assume you'll respin your DSB/SEV fix on top of this?
>> Yes - how does one manage the timing of interdependent patches?

Since Phil's patch is a critical bugfix for the near future. I prefer to
have it in 4.12.

Stefan

> You get the maintainers of this code to queue both patches at the same
> time. You can take them directly through the BCM tree if that makes
> things easier for you.
>
> Thanks,
>
> 	M.