[alsa-devel] idt blue jack patch

[Wu Fengguang]

On Wed, Feb 18, 2009 at 04:24:58PM +0100, Takashi Iwai wrote:
> At Wed, 18 Feb 2009 10:44:26 +0100,
> Vedran Miletić wrote:
> > 
> > On Wed, Feb 18, 2009 at 10:04 AM, Paulo Cavalcanti <promac at gmail.com> wrote:
> > > Takashi,
> > >
> > > Fengguang got the same board I have (DG45ID)
> > > and he confirmed that the blue jack was output.
> > > The only difference between his computer and mine
> > > was that his had an older bios.
> > >
> > > This is what he said to me before sending me the patch:
> > >
> > > "OK, so it's in fact a general issue. I'll look into it. But I'm afraid
> > > I have difficulty in allocating time for it in the near future..."
> > >
> > > Therefore, I think we need another quirk.
> > >
> > > Thanks.
> > >
> > >
> > > --
> > > Paulo Roma Cavalcanti
> > > LCG - UFRJ
> > > _______________________________________________
> > > Alsa-devel mailing list
> > > Alsa-devel at alsa-project.org
> > > http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
> > >

[sorry for being late]

> > 
> > Can you find someone who has different board with same codec to verify
> > how general issue this really is?
>
> In general, the pin config value can't be generic.
> As port-C is a multi-purpose jack, I don't think this can be applied
> to all cases.

I made that hacking patch based on IDT's spec. The value I wrote is in
fact the default value listed in the spec. So it's the Intel BIOS that
changed the default value and make it an output pin.

> So, now the question is when to apply -- under which condition.
> Apparently, the BIOS of Paulo is broken.  And, Wu Fengguang's case
> is unclear, whether it comes from BIOS or from a static pin cfg
> table in patch_sigmatel.c.

After rebooting and force using the generic codec, the 0x0c node still
shows "Pin Default 0x01113014: [Jack] Speaker at Ext Rear". So the
value is from the BIOS.

The only problem is that the pin color is Blue both logically and
physically, which should be input instead of output according to the
convention of color codes.

If you think it's OK, I can refine that patch for submitting.  The
possible regression could be that users connecting speakers to that
pin will find it no longer producing sound after upgrading kernel.

Thanks,
Fengguang