08:22 <konradybcio> aka parallel to work

08:57 <srinik> aka [m]: we can query the remote service version info by calling q6core_get_svc_api_info().

09:00 <srinik> aka [m]: q6afe already has this info, this will give the version for afe rather than clk in particular. Hopefully that should help.

09:36 <aka_[m]> <srinik> "aka [m]: q6afe already has..." <- then we match for example v2.6 to afe_clk_v1?

09:36 <aka_[m]> i was thinking on adding query to apq8016 driver then doing clocks based on that

09:37 <aka_[m]> any idea how we can use v1 clks via q6lpass-clock?

09:37 <aka_[m]> i think we could match IDs with preset packets

09:38 <aka_[m]> i have digital codec in apq8016_sbc card but we have digital codec for wcd which want to consume mclk

09:40 <aka_[m]> i mean msm8916-wcd-digital.c wants to manage mclk and i cannot pass it from qdsp6 card driver

10:21 <srinik> aka [m]: Am guessing the clock ids are different on v1, Normally for such cases we add proper clk support in qcom,q6dsp-lpass-ports.h q6afe-clocks.c and then corresponding clk ids based on device compatible or some logic in soundcard

10:40 <aka_[m]> <srinik> "aka [m]: Am guessing the clock..." <- my limited understanding is that on v2 you have ID for each type of clock like primary ebit/ibit on v1 you set type of clock like BIT , its source (external or internal) and rate, and all of this has to be done in context of port

10:40 <aka_[m]> so for setting quinary ibit you send bit clock with internal source and rate in packet with port id of quinary

10:41 <aka_[m]> on v2 you might not need to send "source" anymroe

10:44 <aka_[m]> for now v1 afe clk is used on 8916/8909/8976

10:44 <aka_[m]> 8909/8976 no longer have digital codec clocks in GCC space

10:48 <minecrell> aka_[m]: it's in GCC anyway even on 8909, you're just not supposed to access it

10:48 <minecrell> but same applies to 8916

10:51 <aka_[m]> minecrell: on 8976 i see it under -- MODULE ELDARION.LPASS.LPASS_LPASS_CC_REG

10:51 <aka_[m]> DIGCODEC_CMD_RCGR ADDRESS 0x2C000

10:52 <aka_[m]> and its all under lpass space

10:52 <aka_[m]> which is like 0x0c000000

10:53 <minecrell> ah

10:53 <minecrell> guess they changed it on 8976

10:53 <minecrell> it's still in gcc on 8909

10:53 <minecrell> I guess 8976 has separate adsp again?

10:55 <aka_[m]> yea

10:55 <aka_[m]> no combined adsp/modem like on 8916

10:55 <aka_[m]> or you mean something else?

10:56 <minecrell> yep, guess the clock is part of the adsp then

10:56 <minecrell> I mean, lpasscc is owned by the adsp

12:50 <Danct12> gm, i have qcom,gcc setup on my device tree (this is a completely new platform) and when the driver initializes, the screen (IPS) fades to black

12:50 <Danct12> does anyone know what i'm missing? i have clk_ignore_unused

12:59 <konradybcio> pd_ignore_unused?

13:04 <Danct12> konradybcio: there's no power domain drivers now

13:04 <konradybcio> don't you have GDSCs in your GCC?

13:04 <Danct12> uh, yes of course

13:04 <konradybcio> well guess what they are registered as :D

13:06 <Danct12> oh damn :D

13:06 <Danct12> so that was it

13:07 <konradybcio> "power domains" or "genpds" (generic power domains) is a generic concept, what you had in mind was probably the RPM(h)PD driver which is used to represent some voltage resources that RPM(h) exposes to us as genpds