[alsa-devel] [PATCH] ALSA: jack: fix a randconfig build issue

[Takashi Iwai]

At Sat, 2 May 2015 07:09:50 +0000,
Jie, Yang wrote:
> 
> > -----Original Message-----
> > From: Takashi Iwai [mailto:tiwai at suse.de]
> > Sent: Saturday, May 02, 2015 3:01 PM
> > To: Jie, Yang
> > Cc: broonie at kernel.org; alsa-devel at alsa-project.org; Girdwood, Liam R
> > Subject: Re: [PATCH] ALSA: jack: fix a randconfig build issue
> > 
> > At Sat, 2 May 2015 04:51:34 +0000,
> > Jie, Yang wrote:
> > >
> > > > -----Original Message-----
> > > > From: Takashi Iwai [mailto:tiwai at suse.de]
> > > > Sent: Friday, May 01, 2015 3:11 PM
> > > > To: Jie, Yang
> > > > Cc: broonie at kernel.org; alsa-devel at alsa-project.org; Girdwood, Liam
> > > > R
> > > > Subject: Re: [PATCH] ALSA: jack: fix a randconfig build issue
> > > >
> > > > At Fri,  1 May 2015 11:10:29 +0800,
> > > > Jie Yang wrote:
> > > > >
> > > > > There may be no input_dev for jack(e.g. phantom jack), when
> > > > > CONFIG_INPUT is not selected, building errors such as:
> > > > > undefined reference to `input_xxx'...
> > > > >
> > > > > Here add #ifdef judgement to fix the issue.
> > > >
> > > > Well, the standard idiom so far is to enable SND_JACK selectively.
> > > > Take a look at the comment in sound/core/Kconfig.
> > > >
> > > > # To be effective this also requires INPUT - users should say:
> > > > #    select SND_JACK if INPUT=y || INPUT=SND
> > > > # to avoid having to force INPUT on.
> > > > config SND_JACK
> > > > 	bool
> > > >
> > > > And, looking at sound/pci/hda/Kconfig, there is still
> > SND_HDA_INPUT_JACK.
> > > >
> > > > So, what you need to fix is:
> > > > - Remove SND_HDA_INPUT_JACK from sound/pci/hda/Kconfig and the
> > last
> > > >   ifdef in hda_codec.h
> > > >
> > > > - Replace the line "select SND_JACK" with
> > > > 	select SND_JACK if INPUT=y || INPUT=SND_HDA
> > >
> > > OK, will follow this. What I only concern is that if user want
> > > SND_JACk when INPUT=N(e.g. phantom jack?), it becomes impossible.
> > 
> > Right, it's impossible now.  But this is the case really no one would care much
> > (who wants the jack detection on a system without any input capability?).
>  
> Got it. Another question here, I found that this change in sound/pci/had/Kconfig
> doesn't fix the building issue:
> 	select SND_JACK if INPUT=y || INPUT=SND_HDA
> but this do:
> 	select SND_JACK if INPUT=y || INPUT=SND
> 
> suppose we should use the latter one, right?

Yes, correct.

> Do you know why the former one
> doesn't fix?(seems here "INPUT=SND..." added for both CONFIG_INPUT=m and
> SND...=m case?)

SND may be y while SND_HDA=m and INPUT=m.


Takashi