[PATCH 2/2] staging: bcm2835-audio: Disable extra virtual cards by default
Michael Zoran
mzoran at crowfest.net
Tue Mar 14 04:16:28 PDT 2017
On Tue, 2017-03-14 at 14:05 +0300, Dan Carpenter wrote:
> On Tue, Mar 14, 2017 at 01:55:01AM -0700, Michael Zoran wrote:
> > On Tue, 2017-03-14 at 09:17 +0100, Stefan Wahren wrote:
> > > Am 14.03.2017 um 07:45 schrieb Michael Zoran:
> > > > The extra virtual cards created for HDMI and Headphone output
> > > > may not be compatible with operating systems that expose
> > > > very simple audio controls.
> > > >
> > > > By default only enable the traditional mode that exposes a
> > > > single
> > > > card for maximum compatibility.
> > > >
> > > > Add options to the build configuration for enabling the extra
> > > > virtual cards.
> > >
> > > This looks like a workaround for broken applications. Could
> > > please
> > > you
> > > name them?
> > >
> > > Btw avoid those many #if defined(). They are bad for readibility
> > > and
> > > compile test. Sometimes a single IS_ENABLED() at the right place
> > > is
> > > better.
> >
> > I did all my testing on debian sid which has rather complex and
> > detailed controls with pulseaudio. But Raspbian has a custom
> > simple
> > mixer and I'm concerned about breaking that. I didn't want to name
> > names.
>
> Heh... Feel free to name names. Everyone accepts that bugs are part
> of
> being human and that UI disagreements are the other half of being
> human.
>
> This patch doesn't sound like the right thing at all. What you're
> saying is that we'd like to have 3 volume knobs, one for HDMI, one
> for
> headphones and one for both, but some user space applications can
> only
> handle one? Is that right?
>
> Maybe we should make this a module option instead of a compile time
> option?
>
> regards,
> dan carpenter
>
That's exactly correct.
A module option sounds fine, I just couldn't decide on what it should
be. Perhaps the default should be to just have HDMI controls +
Headphone controls with no common contols but the module option to
change things.
Raspbian ships a customized version of this code anyway, so if they
can't have the loader pass a module option then they are always free
to make a small change in the github tree that makes best sense for
Raspbian.
I'll send out V2 with your suggestion in awhile.
More information about the linux-rpi-kernel
mailing list