#1829 BLOC Trial-3: Control of VREF_OUT and HPF is coupled to Analog Input control

[Zarro Boogs per Child]

#1829: Control of VREF_OUT and HPF is coupled to Analog Input control
----------------------+-----------------------------------------------------
  Reporter:  arjs     |       Owner:  dilinger
      Type:  defect   |      Status:  new     
  Priority:  blocker  |   Milestone:  Trial-3 
 Component:  kernel   |     Version:          
Resolution:           |    Keywords:  relnote 
  Verified:  0        |  
----------------------+-----------------------------------------------------
Comment (by arjs):

 > I wrote up a patch to drop the generic AD1888 HPF control, and instead
 replace it with OLPC's >HPF (that also toggles AnalogInput).  Are you
 saying that we should expose it to the user as >AnalogInput, or HPF?

 IMHO as 'Analog Input'


 Another idea, how about we change the name from 'Analog Input' to 'DC
 Mode' ?


 >
 > Noted.  HPF should certainly be set to a default state.  Does VREF need
 to be set?  VREF is part of the AD1888/AC97 stuff, and the driver (none of
 the drivers, afaik) have bothered to reset VREF.
 >

 VREF needs to be enabled whenever capture device is opened. Enabling VREF
 is like switching on the mic, without which mic wouldn't be able to
 record. In the current kernel whenever we open capture device (for example
 record audio using arecord), VREF is getting enabled (as is indicated by
 the LED turning ON) - this is infact the desired behavior.



 > >
 > > 3. Requirements regarding capture_close: If an application enables
 VREF and doesn't disable it on quit, the LED will remain ON all the time.
 (the LED turns on whenever VREF is enabled). So on capture_close we need
 to disable VREF. We should also disable AnalogInput on capture_close.
 >
 > Yep, I verified that VREF controls the LED last night; I have a patch
 for this..
 >

 So on capture_close we plan to be disabling VREF and Analog Input right ?


 > > 4. Name of VREF in Alsamixer: The name V_REFOUT doesn't make sense to
 a non-technical user. Thinking back a little, why would a user want to
 toggle this control at all ? IMHO in a case when he/she wants to simply
 say something into the mic/external-mic  and hear it in real time (not
 capture). So why not rename the control to something like MIC_ENABLE ?
 >
 > This is part of the AC97 code.  If V_REF acts the same for OLPC as it
 does for other AD1888 users, then I can send a patch upstream to change
 the name to MIC_ENABLE; if it does something different, we need to
 consider whether it's worth changing.
 >

 I look forward to comments from jg,wmb on this.


 >
 > If you're around today, I can give you a kernel to test..

 On a 256kbps connection I think it'd take me way lot of time to download
 something ~200mb :) I'd try to get access to a faster connection on
 Saturday and will test it then and let you know.

-- 
Ticket URL: <https://dev.laptop.org/ticket/1829#comment:19>
One Laptop Per Child <http://laptop.org/>