On 04/25/2018 07:56 PM, Tony Lindgren wrote:
- Peter Ujfalusi peter.ujfalusi@ti.com [180425 16:06]:
On 04/25/2018 04:50 PM, Tony Lindgren wrote:
- Jarkko Nikula jarkko.nikula@bitmer.com [180425 03:56]:
What I noticed some time ago on N810 but not time to fix it that OMAP mux change dropped muxing audio McBSP pins. Those pins are muxed by reset or bootloader to another function.
Probably a few lines to arch/arm/boot/dts/omap2420-n810.dts etc.
Hmm do you have a link or commit id to a patch that does the muxing to see which pins need to be muxed?
FWIW I have this at the moment: https://pastebin.com/WNT9wnD5
I see that the card is there, but I'm fighting buildroot to have nfs filesystem with aplay :o
There are few things going to be needed to get things in a shape, like converting the sound/soc/omap/n810.c to platform driver to be probed via DT. Proper bindings for the card, support for mclk handling in aic3x codec driver (n810 is hackinig that in the machine driver)
In theory audio _should_ work with the DT changes and the fixup in the n810.c...
Should...
Cool. The pinctrl fixes at least make sense to me, that could be a separate patch maybe with Fixes: 5a87cde490e1 ("ARM: OMAP2+: Remove legacy booting support for n8x0") as pointed out by Jarkko.
The two GPIO for the headphone and speaker amp might also need pinctrl entry.
The clock sys_clkout2 reparenting and rate setting is done with the assigned-clock stuff.
As I mentioned the only thing that we will need is the mclk handling in the codec driver.
Most likely we could have a series dealing with DT only, then rewrite the n810.c.
The only thing might cause issue is that if the aic3x driver starts to handle the sys_clkout2 (mclk) and the n810.c also wants to grab it and control it... That might break along the way.
Regards,
Tony