Re: [PATCH 2/9] ASoC: Intel: Kconfig: Simplify-clarify ACPI/PCI dependencies

[Andy Shevchenko]

On Fri, 2017-12-15 at 06:34 -0600, Pierre-Louis Bossart wrote:
> On 12/15/17 4:58 AM, Andy Shevchenko wrote:
> > On Thu, 2017-12-14 at 18:44 -0600, Pierre-Louis Bossart wrote:
> > > PCI/ACPI selections should not happen in Kconfig for machine
> > > drivers,
> > > move to SOC selections.
> > > 
> > > Add distinction between PCI and ACPI HiFi2 platforms and help
> > > text.
> > > 
> > > There should be no functionality change.
> > > 
> > > The PCI-based platforms may be removed at some point since
> > > Medfield
> > > is not really supported by anyone, and with Edison now defunct
> > > support
> > > for
> > > Merrifield/Edison is to be determined.
> > > 
> > > -	depends on SND_DMA_SGBUF
> > > +	depends on SND_DMA_SGBUF && ACPI
> > >   	depends on DMADEVICES
> > 
> > Perhaps
> > 
> > depends on DMADEVICES && ACPI
> > 
> > to be in align with below?
> > 
> > >   config SND_SOC_INTEL_BAYTRAIL
> > >   	tristate "Intel ASoC SST driver for Baytrail (legacy)"
> > > -	depends on DMADEVICES
> > > +	depends on DMADEVICES && ACPI
> 
> I must admit I don't know why it's different, I just cleaned-up what
> I 
> could.

What I'm talking about is to use same pattern, so, in the first case 

	depends on SND_DMA_SGBUF
   	depends on DMADEVICES && ACPI

The question WRT SGBUF is indeed another one to Liam or whoever knows
better.

>  Maybe Liam can comment since I believe he contributed to both 
> Haswell and the initial Baytrail driver?

-- 
Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Intel Finland Oy