Re: [PATCH alsa-ucm-conf v3 1/2] sof-soundwire: Add basic support for cs42l43

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Dec 08, 2023 at 12:00:26PM +0000, Charles Keepax wrote:
> On Thu, Dec 07, 2023 at 02:56:17PM +0100, Jaroslav Kysela wrote:
> > On 07. 12. 23 10:55, Charles Keepax wrote:
> > >On Wed, Dec 06, 2023 at 06:26:17PM +0100, Jaroslav Kysela wrote:
> > >>On 06. 12. 23 17:46, Charles Keepax wrote:
> > >Oh, I was not aware we could do that. I would yes much rather
> > >handle the switches and volumes in this way. I will see if I can
> > >figure it out, but if you had any good examples that already
> > >exist that would really be handy?
> > 
> > You may look for "LibraryConfig.remap.Config" and
> > "Include.ctl-remap.File" strings in .conf files.
> 
> Apologies still struggling to get this working. I think there must
> some important boiler plate or limitation I am missing. Would really
> appreciate if you could have a look at this and let me know if it
> looks sane. I am starting out with just the simplest thing I can
> think of, just trying to rename a control:
> 

Ok it seems starting with the simplest thing was not the best
idea :-)

If you only have a remap, no map then it looks like one needs to
do something like this:

+++ b/src/control/control_remap.c
@@ -1192,7 +1192,7 @@ int snd_ctl_remap_open(snd_ctl_t **handlep, const char *name, snd_config_t *rema
                goto _err;
        }
 
-   priv->numid_remap_active = priv->map_items > 0;
+ priv->numid_remap_active = priv->map_items > 0 || priv->remap_items > 0;

Otherwise the check at the start of remap_find_numid_app will
keep failing and you get stuck in an infinite loop in
remap_numid_child_new. I will likely send a patch soon (might
slip into the new year, getting close to the end of the year and I
am rather unfamiliar with the user side of the alsa code), but I
want double check things a little more first.

Any input on this change would be greatly appreciated?

> LibraryConfig.remap.Config {
> 	ctl.default.remap {
> 		"name='cs42l43 PDM2 Switch'" "name='cs42l43 Decimator 3 Switch'"
> 	}
> }

To answer part of my own question, and in the hope that if anyone
else is having similar difficulties they will find this thread,
the remapping works the other way around, it should go:

"name='cs42l43 Decimator 3 Switch'" "name='cs42l43 PDM2 Switch'"

With the newly mapped control second, kinda confusing as the .map
sections do it the other way around, but fair enough.

> 
> SectionDevice."Mic"
> {
> 	Comment "Microphones"
> 
> 	EnableSequence
> 	[
> 		cset "name='cs42l43 PDM2 Switch' 1"
> 	]
> 
> 	DisableSequence
> 	[
> 		cset "name='cs42l43 PDM2 Switch' 0"
> 	]
> 
> 	Value
> 	{
> 		CapturePriority 100
> 		CapturePCM "hw:${CardId},4"
> 	}
> }

It would seem the primary issue is here, one needs to add:

CaptureCTL "default:${CardId}"
PlaybackCTL "default:${CardId}"

This is because the control device will normally be the hw: one,
which doesn't have the renaming applied to it, you need to switch
to the device with the remapping, which was added as the one
named default in lib/ctl-remap.conf.

Seems weird you have to do the PlaybackCTL in a capture only use
case, but otherwise it complains about them being mismatched. Also
weirdly you need to add this into some other devices in the same
use-case, otherwise it appears it just ignores these lines and
uses the hw device anyway. Still looking at what is going on with
that but I would appreciate any thoughts on that as well?

> Also if there are any docs I should read happy to go there first?

Updating this to any suggestions on where to add some docs would
also be appreciated?

Hopefully I can find sometime to document some of this a little
and save someone else spending the quite large amount of time I
have sunk into working this lot out so far.

> 

Finally, does anyone have any idea what is going on with the
current users of the remap. It looks like rt5660, rt5677, rt5651,
rt5645, rt5640, rt5682 all currently have remap sections in their
config. However almost the remapped controls are never used, which
might not be surprising given the likely bug at the start of this
email. But curious if anyone has any ideas that the remapping is
actually being used for something non-obvious on those devices?
Kinda wonder if we should just remove some of the unused
remapping as I found it quite confusing whilst trying to figure
this out.

Anyway any extra thoughts/corrections on all this would be really
really appreciated. I am still very new to the ALSA userspace
side of things, so sure I am still making lots of mistakes.

Thanks,
Charles



[Index of Archives]     [ALSA User]     [Linux Audio Users]     [Pulse Audio]     [Kernel Archive]     [Asterisk PBX]     [Photo Sharing]     [Linux Sound]     [Video 4 Linux]     [Gimp]     [Yosemite News]

  Powered by Linux