[meta-freescale] No sound on nitrogen6x + linux-boundary 3.10.17

[Erik Botö]

On Mon, May 5, 2014 at 7:55 PM, Fabio Estevam <festevam at gmail.com> wrote:
> On Mon, May 5, 2014 at 2:40 PM, Fabio Estevam <festevam at gmail.com> wrote:
>> Erik,
>>
>> On Mon, May 5, 2014 at 3:51 AM, Erik Botö <erik.boto at pelagicore.com> wrote:
>>> Hi,
>>>
>>> I'm having trouble playing audio on this board using the sgtl5000. Can
>>> anyone verify that it works on their system? Also any ideas of what
>>> might be wrong are highly appreciated :)
>>>
>>> Using aplay it get stuck here:
>>
>> I remember that I have helped John Weber recently to get audio working
>> on 3.10.17 on wandboard.
>>
>> There were two patches that were needed:
>>
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/sound/soc/codecs/sgtl5000.c?id=9e13f345887c179068bbc1f7389b7177bf88f57e
>>
>> and this one as well:
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/sound/soc/codecs/sgtl5000.c?id=46a5905e1cd4a9d9d238ec7beece49ce49e2ad85
>
> For a more complete reference, please apply all the ASoC patches from:
> https://github.com/wandboard-org/linux/commits/wandboard_imx_3.10.17_1.0.0_ga
>
> and make sure the codec dts files also matches.

So I tried to apply the following commits from the wandboard kernel
git, they didn't apply cleanly but I think I was able to get their
function in:

ASoC-sgtl5000-Convert-to-use-regmap-directly
ASoC-sgtl5000-Read-SGTL5000_CHIP_ID-in-i2c_probe
ASoC-sgtl5000-Fix-driver-probe-after-reset
ASoC-sgtl5000-Let-the-codec-acquire-its-clock
ASoC-sgtl5000-defer-the-probe-if-clock-is-not-found

But unfortunately I still have the same problem.

I measured I2S_SCLK, I2S_LRCLK and SYS_MCLK on the sgtl5000 chip using
both the 3.10.17 and 3.0.35 (alsa playback works on 3.0.35 but not
3.10.17):
                       3.10.17           3.0.35
I2S_SCLK       ~1.4 Mhz      ~1.4 Mhz
I2S_LRCLK     44100 Hz      44100 Hz
SYS_MCLK    ~23 MHz     ~16,5 MHz

Looking at the datasheet [1] it looks like things are setup
differently depending on if SYS_MCLK is > 17 MHz which is the case
with 3.10.17. Currently I'm a bit confused about how this clock get
set up.

I'm a bit out of my regular comfort zone here so if you think I'm
wasting my time looking at SYS_MCLK please tell me :) If anyone using
sgtl5000 + 3.10.17 with alsa playback working has an oscilloscope
nearby I'd would love to know how your clocks look.

[1] - http://cache.freescale.com/files/analog/doc/data_sheet/SGTL5000.pdf

Cheers,
Erik