20:10 <gbisson> hi all, playing around more with a311d brought a couple of questions: 1) isn't there any cpufreq driver available? can't seem to find any scaling governors for the CPU. 2) is the internal PHY (10/100M) supposed to work out of the box? doesn't seem to probe on my end

20:47 <xdarklight> gbisson: both should work. are you using the arm64 defconfig or a custom one?

20:48 <xdarklight> gbisson: for cpufreq you want CONFIG_CPUFREQ_DT + CONFIG_REGULATOR_PWM + meson PWM driver ...

20:55 <gbisson> xdarklight: I based my defconfig on khadas one (their 5.14 branch) I'll check the PWM driver

20:57 <gbisson> xdarklight: looks like i have all the config already enabled :(

20:58 <gbisson> xdarklight: you confirm that in your case it appears under /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors?

20:59 <xdarklight> gbisson: I don't have any A311D based boards but it should list all frequencies which are defined in arch/arm64/boot/dts/amlogic/meson-g12b-a311d.dtsi

21:00 <xdarklight> gbisson: can you please share your kernel log (dmesg)?

21:04 <gbisson> xdarklight: here: https://gist.github.com/gibsson/218503150c874551df7fa1064e85e7d3

21:08 <xdarklight> gbisson: just checked on an GXL based board: /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governor has the following in it: ondemand userspace powersave performance

21:09 <xdarklight> gbisson: do you see the frequencies which you expect in /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies ?

21:09 <gbisson> xdarklight: no I don't see any cpufreq folder which is why I was surprised

21:10 <xdarklight> gbisson: can you please share the output of "cat /sys/kernel/debug/regulator/regulator_summary" and "cat /sys/kernel/debug/clk/clk_summary" ?

21:12 <xdarklight> gbisson: also in your kernel log I see that it detected an AR8035 RGMII PHY. are you expecting this to work at the same time as the internal 100M PHY?

21:14 <gbisson> xdarklight: for the ethernet I expected to be able to have 2 ports: 1 external PHY + 1 internal PHY, but I now realize that there's only 1 MAC

21:15 <xdarklight> gbisson: yep, one MAC means the choice between external or internal PHY is exclusive

21:16 <gbisson> xdarklight: thanks for the confirmation. I edited the gist to add the output you requested.

21:19 <xdarklight> gbisson: thanks. regulator output is suspicious since it's not showing any VDDCPU* regulator. can you please check your .dts and make sure that 1) each CPU node has a cpu-supply property 2) this cpu-supply refers a regulator (typically one which has compatible = "pwm-regulator") and 3) that the PWM controller which is used in the "pwms" property of the regulator has status = "okay";

21:19 <xdarklight> for step 3) you can also check cat /sys/kernel/debug/pwm

21:21 <gbisson> xdarklight: oh actually we're using FAN parts for the regulators, maybe i'm simply missing their conf, lemme check

21:22 <gbisson> xdarklight: indeed the fan5355 were missing their drivers...

21:24 <xdarklight> gbisson: ah, great that it works now :-)

21:26 <gbisson> xdarklight: thanks for the support!

21:28 <xdarklight> gbisson: you're welcome!