[nas] [PATCH] changed method of setting the input gain

[Erik Auerswald]

Hi,

On Thu, Jul 27, 2006 at 08:46:53AM -0400, Paul Fox wrote:
> erik wrote:
> documentation thought:  i think we need to start stating, in both
> the man page and in the config file comments, what the default
> behavior is for each of our booleans.  there's a tendency to edit
> the originally-installed file to make changes, and then any
> indication of the server default is lost.

I agree.

>  > The patch is attached to this mail. It is not without issues, though:
> 
> one change:  when reading the gain, you have to do inverse gain scaling:
> 
> --- server/dda/voxware/auvoxware.c.erik	Thu Jul 27 08:18:38 2006
> +++ server/dda/voxware/auvoxware.c	Thu Jul 27 08:21:29 2006
> @@ -361,6 +361,11 @@
>                       sndStatOut.mixer, mixerfd, strerror(errno));
>          }
>          pcm_level = pcm_level >> 8;
> +        if (sndStatOut.gainScale ) {
> +	    pcm_level *= 100;
> +	    pcm_level /= sndStatOut.gainScale;
> +	    if (pcm_level > 100) pcm_level = 100;
> +	}
>      } else {
>          pcm_level = sndStatOut.gain;
>      }

I'll do this in the next version of the patch, thanks.

> the tracking seems to work for me, after applying this.  thanks!
> 
> another question though:  i haven't tried it, but since you've
> eliminated the last* variables, if the mixer is closed, you
> always report the default sndStatOut.mixer.

Exactly. Without a mixer device no true value can be reported. This way
the user tries to change it but nothing happens, which is at least
consistent.

> but if the user sets
> keepmixer to false, doesn't this mean that no nas-based mixer can
> report a true value?  i'm confused.  (in fact, i did just try
> setting keepmixer to "no", and sure enough, my "audiooss aumix"
> is now locked at 50%.  oh.  wait.  i didn't have a stream
> playing.  if i play a stream, then the nas mixer works correctly. 
> is the total lockout when the pcm device isn't open the expected
> behavior?  i thought it would work like it did before.

Before (in the development version) changes made while there is no mixer
are kept and applied when the mixer is re-opened. In the released
version the server tells it changed the value but it is lost as soon as
the mixer is re-opened.

>  > When reading the IGain value from the mixer the ioctl returns -1 with
>  > errno == 0 (success) most of the time. The value was read correctly
>  > every time I tested this regardless of error reported. I have no idea
>  > why this happens. The patch works for me, it justs spams the logfiles
>  > with lots of error: success messages. I was using aupanel to control the
>  > NAS mixer, I did not check if auctl behaves the same (it should).
> 
> my logs look like this.  i got one "success" message, but a bunch of
> EINTR. calls.  not sure what those are from.

I've found my error: I checked ioctl() != -1 and thought I would check
ioctl() == -1... It will be corrected in the next version of the patch
along with error handling besides reporting (I took this out after the
errors seemed to be bogus).

Erik