07:57 <inode> how are the frequencies and voltages specified for cpu operating points usually calculated or derived?

07:59 <inode> my dmesg was being flooded often with a triplet of warnings beginning with "rockchip_cpuclk_pre_rate_change: Invalid rate : 1464000000 for cpuclk" until i commented out the operating point for that frequency (the maximum specified) from rk3229.dtsi

09:42 <mmind00> inode: the clk+volt pair in the dts is specified by the soc manufacturer, but needs a corresponding entry in rk3228_cpuclk_rates in drivers/clk/rockchip/clk-rk3228.c ... which seems to be missing

10:20 <inode> thanks mmind00

10:28 <inode> in a case like this where the dts selects something illogical, is it usually a better idea to adjust the dts file to include a modified duplication of an entry in an inherited dtsi, or to modify the dtsi file itself?

11:06 <mmind00> inode: I guess first step would be to compare against the vendor kernel ... rk3229 has been sort of a step-child in mainline so far

11:07 <mmind00> inode: aka I think you're the first one to actually try this in recent times ... (like I don't even have a rk3229-board in my boardfarm for testing)

11:13 <arturog> Hello everyone, I need some help with my nanopi M4 and the current output. Running the stock FriendlyElec ubuntu. lsusb -v says that MaxPower is around 250ma but under /sys the device has correctly reported around 900ma. (external hard drive)

11:13 <arturog> Does anybody know how to get the USB port to output the 900ma I need?

11:55 <inode> mmind00: i haven't been able to identify anything like operating point pairs in the vendor's dtb, intended for a build of 3.10.x - i guess i might be able to see some information about the scaling range from it's cpufreq directory in sysfs

13:14 <mmind00> inode: 3.10? ... I was refering to the develop-4.4 branch I pointed to yesterday

13:14 <mmind00> which is Rockchip's vendor kernel

13:16 <inode> ok got it, i assumed you had meant that which had shipped with the device

13:24 <inode> seems RK3228_CPUCLK_RATE(1464000000, 1, 5) is the only entry in rk3228_cpuclk_rates that is missing in drivers/clk/rockchip/clk-rk3228.c between develop-4.4 and linux-5.1

20:39 <mmind00> inode: great ... patches welcome by the way ;-) [aka if you like, make it a patch and send it to mainline]