Re: [PATCH] mfd: fix dependency warning for CHROME_PLATFORMS on !X86, !ARM

[Javier Martinez Canillas <javier.martinez@collabora.co.uk>]

On 06/24/2015 04:21 PM, Heiko Stübner wrote:
> Am Mittwoch, 24. Juni 2015, 15:17:50 schrieb Lee Jones:
>> On Sun, 21 Jun 2015, Javier Martinez Canillas wrote:
>> > Hello Paul,
>> > 
>> > On 06/21/2015 02:01 AM, Paul Gortmaker wrote:
>> > > In commit 062476f24aa7cf714169342cc50626fd9bbb93da ("mfd: cros_ec:
>> > > Move protocol helpers out of the MFD driver") there was a select
>> > > added on CHROME_PLATFORMS, which leads to the following for ppc/mips:
>> > > 
>> > > warning: (MFD_CROS_EC) selects CHROME_PLATFORMS which has unmet
>> > > direct dependencies (X86 || ARM)
>> > 
>> > Right, sorry for missing that and thank a lot for taking care of it.
>> > 
>> > > Presumably the above means MFD_CROS_EC is only currently useful
>> > > on x86 and ARM, so lets limit it to there.
>> > 
>> > That is correct, although I wonder if CHROME_PLATFORMS should not
>> > also depend on || COMPILE_TEST and same for MFD_CROS_EC so those
>> > have at least build coverage on other architectures.
>> > 
>> > I can't think for a reason to not be buildable on other archs...
>> 
>> Right.  Will you fix that Javier?
>

Yes, I was waiting to see if Paul agreed and wanted to re-spin his patches
or if anyone else had other opinions.
 
> looking at the ChromeOS kernel tree (3.18 / oak) it looks like ARM64 might 
> also be a nice addition to this list, but I'm not sure if there are more ec 
> changes involved.
> 

Right, on a second thought I wonder if having CHROME_PLATFORMS depends on
depends on X86 || ARM even makes sense. I see that the individual drivers
already depend on a certain architecture. i.e: CROS_EC_LPC on x86 because
LPC is only found on x86, etc.

So maybe we can just remove the constrain from CHROME_PLATFORMS and let
it be selected on all architectures if someone choose to.

I'll double check and post a patch on top of Paul's tomorrow.

Best regards,
Javier