Hi Pierre, how are you?
I also ported a machine driver code from a kernel Lenovo made available for a similar hardware (Lenovo yoga 2 1050 - uses the same baytrail Z3745 and wolfson wm5102 audio) [4]
There are quite a few references in the machine driver to concepts related to Clovertrail+ and early versions of Baytrail. I am fairly convinced the mentions of 'BYT_COMMS_BT' or 'modem mixing' are no longer supported with the DPCM platform driver/firmware you used.
Yes, the code I ported was originally wrote for kernel 3.10.x targeted for Android
In any case, I didn't enable this option on our kernel config
To enable this codec it would be simpler to start with a newer machine driver derived from the bytcr_rt5640 one, with just a basic SSP0-aif1 link for starters (+ MCLK).
Thanks for the tip. I can try it ( create a machine driver derived from bytcr_rt5640 ). Will just wait for Charles help on analyzing CODEC output I provided. [1] If he says this part seems ok, I will work on machine driver
I'm not sure this hardware uses SSP0 or SSP2, so I'm testing with both firmwares. Is there any method I can confirm it? (querying an ACPI entry listed on DSDT maybe?)
Better yet, do this work with a MinnowBoard and an WM5102 eval board and test all this with more control on how the hardware is setup.
This is probably the best approach (minnowboard + WM5102 eval board). But unfortunatelly my skills are limited...
If you stick to working directly on the Lenovo device, make sure you use all the DSP controls set in the UCM files used for baytrail/baytrail-CR (see github.com/plbossart/UCM).
In fact, currently I don't have this device. Some volunteers from Android-x86 google group are testing changes made on their hardware and provide the info I ask for (logcat / dmesg / mixer output, etc)
Thanks Pstglia
[1] - https://drive.google.com/file/d/0BxO6THtB865fS2NxUkdOa18zUlE/view?usp=sharin...