Headphone volume adjustment

[Mark Brown]

On Fri, Aug 16, 2013 at 08:45:42AM -0600, Daniel Drake wrote:

> The last time we tried PulseAudio, it segfaulted on startup, and we
> didn't get into the diagnosis. It was years ago though, and we haven't
> tried since. We do want to give it another go but it has never taken a
> high enough priority for us to actually do it.

It'd probably be helpful for the people running Fedora at least.

> We also agree that UCM looks like it will solve some of our problems,
> and we'd like to fix/use the dynamic routing in the upstream codec
> driver. Just need to find time to work on it.

Do you know if a mainline kernel will actually run sensibly on your
system?  I have a XO-1.75 (this was part of the reason I was looking)
but I'm mostly interested in advancing the state of the art.

> 3 of us have already spent a considerable amount of time on the
> routing issue with little progress :/  Manually validating the links
> between the components against the spec seems to be a very time
> consuming process and prone to human error. Maybe you have some
> suggestions or tools to help on that front. The issue is that upon
> playback, nothing is reproduced, and the dapm files in sysfs show that
> almost everything on the codec is powered down.

Hrm, no issues reported upstream with this :(

There's a couple of scripts written by Dimitris Papastamos in the git
repository at git://opensource.wolfsonmicro.com/asoc-tools.git that help
with visualisation of the graph.  

Essentially always it's a case of starting at one of the ends of the
audio path you're trying to create and looking for the point at which a
widget is missing an input or output (depending on which direction
you're working through) - the tools liked above can help do that
visually, or you can do it by looking at the widget files in debugfs
since all the inputs and outputs for each widget are listed.  As you say
the fact that there's a problem with the path not being connected tends
to be fairly obvious since nothing gets powered up so it should be clear
when to look at the graph.

Otherwise it's mostly just a question of looking for mutes and low
gains in the path, but that only applies once things are powered up.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.laptop.org/pipermail/devel/attachments/20130816/a7f9a61a/attachment.sig>