A few questions about an audio CODEC

Thread Starter

Robert Murch

Joined Nov 2, 2015
43
@nsaspook
I realize its been awhile since I last posted on this topic but this project got sidelined. So since I last posted I have managed to get sound working for the most part. There was a lot of hoops to jump through but I am close. Basically, if I can figure out how to properly customize this asound.play file (known as asound.state for most systems not sure why mine is .play but it is) then the sound should be working as expected. The default configuration allows me to play a tone but the sound for the actual applications doesn't work.
 
Last edited:

Thread Starter

Robert Murch

Joined Nov 2, 2015
43
Alright well, after a lot of battling I have final managed to get sound working! In summary:

Installed new codec.
Changed all references from wm8753 to wm8750 in machine driver and machine startup code.
Installed a new timing crystal because the codecs required different clock rates.
Generated a new default asound.play file (normally called asound.state but this system is old) and
hand altered it until I was able to generate a tone.
Created a shell script to handle muting function on amplifier.
 

nsaspook

Joined Aug 27, 2009
13,086
That was fun, right? You're now an expert at obsolete Linux software technology.:D Now you understand why the kernel dev's restructured the static code complexity to be modular with device-tree.

If this is product that will continue to be supported for a while I would update that ancient Linux kernel 2.6.35.14 to something in the 4.x range so you can have experience with modern Linux embedded system kernels.
 

Thread Starter

Robert Murch

Joined Nov 2, 2015
43
That was fun, right? You're now an expert at obsolete Linux software technology.:D Now you understand why the kernel dev's restructured the static code complexity to be modular with device-tree.

If this is product that will continue to be supported for a while I would update that ancient Linux kernel 2.6.35.14 to something in the 4.x range so you can have experience with modern Linux embedded system kernels.

lol I know right! I think I may try to update the kernel. I get the feeling that it would be tough considering the DIGI SOM overlay. I would have to port the startup code and I don't know how easy that would be. That being said the product will most definitely continue being supported for a few more years.
 

bogosort

Joined Sep 24, 2011
696
lol I know right! I think I may try to update the kernel. I get the feeling that it would be tough considering the DIGI SOM overlay. I would have to port the startup code and I don't know how easy that would be. That being said the product will most definitely continue being supported for a few more years.
While I agree that updating to a modern kernel is worth it, be forewarned that lots has changed from the 2.6 days. In particular, if you have any custom device drivers that use hardware interrupts, much of the interrupt system has changed and the old "big kernel lock" is completely gone.

For work, I had to port a custom 2.6 kernel module to 4.x; so much had changed that I ended up rewriting the thing.
 
Top