09:27 <aka_[m]> Marijn: sorry for bothering, anything done on 8976 front?

15:26 <z3ntu> Is there a way to configure mi2s interface on newer SoCs like sm8250 with regards to data width (16, 20, 24, 32), format (standards i2s, left-justified mode, right-justified mode), sampling frequency (8 kHz, 11.025 kHz, etc), slot length (16, 24, 32) etc? Done in devicetree, alsa config or something?

15:34 <z3ntu> Having troubles getting speaker to work on sm6350, as far as I can tell lpi-pinctrl, q6afedai, mm1-dai-link, i2s-dai-link and the other stuff is configured correctly, but the smart amplifier is complaining that the reference clock (which uses the i2s bit clock as source) is not running and I'm kinda stuck

15:35 <z3ntu> alsa config is basically only: amixer -c0 cset name='QUIN_MI2S_RX Audio Mixer MultiMedia1' 1 . Maybe I'm missing here something too to actually enable the i2s interface or something

15:35 <srinik> z3ntu: for I2s from dsp side there is nothing much w.r.t mixers.

15:36 <srinik> z3ntu: have you set thee sysclk associcated with that dai in machine driver

15:37 <z3ntu> ah, I'm re-using sound/soc/qcom/sm8250.c and that only seems to set the clk for tertiary i2s (Q6AFE_LPASS_CLK_ID_TER_MI2S_IBIT)

15:37 <z3ntu> maybe that's the magic solution

15:39 <srinik> z3ntu: if the speakers are connected to ter Mi2S then it would clock it but do you know which interface is the speaker connected to in your setup?

15:39 <z3ntu> quinary

15:40 <z3ntu> datasheet also likes calling it LPI_MI2S

15:40 <z3ntu> data line 0 for playback (for some reason called RX) and data line 1 for echo reference I believe (that one's called TX for some reason)

15:42 <z3ntu> oh wow, I think my amp is happy now

15:43 <z3ntu> srinik: You think that copy-pasting the tertiary case for quinary and using Q6AFE_LPASS_CLK_ID_QUI_MI2S_IBIT in the sm8250.c is ok for my use case?

15:44 <srinik> z3ntu: yes, that is what is required..

15:45 <z3ntu> but e.g. no clue what "codec_dai_fmt |= SND_SOC_DAIFMT_NB_NF | SND_SOC_DAIFMT_I2S;" does

15:46 <srinik> z3ntu: these are bits to indicate clk polarity to sample the data

15:47 <srinik> z3ntu: there is some good description in the headers in case you are interested.

15:50 <z3ntu> srinik: ok thanks. Maybe maybe it's just left now to please this amplifier driver I copy-pasted from downstream, because it originally wants to make some custom afe calls to adsp firmware, but I mostly patched that out and return errors, but it's possible that the amp driver is now unhappy

15:50 <z3ntu> But at least putting this clk in sm8250.c seems to have solved something, so thanks a lot :D

15:52 <z3ntu> at least amp hardware seems happier :)

22:11 <Marijn[m]> aka_: Thanks for reaching out, I've been at work all day and am only looking at IRC just now. There are a few more DT (bindings checker) issues to fix, and general improvements to apply in light of recent patches that were sent to clean up existing DT. Then I have to shuffle some nodes into separate patches or postpone them altogether, and finally have to make sure the patches actually work in isolation as I've only been using them with a

22:11 <Marijn[m]> full tree on top.