02:14 <set_> I have not done it.

02:15 <set_> I wanted to say, "Yes to me and I have done it!" But, truth tells all. I have not.

02:15 <set_> BBB...how are you and your persons of interest doing in the GSoC world?

02:33 <set_> Is this the mostly incorrect room to inquire an esquire about the GSoC program?

02:33 <set_> yes. it is the incorrect one. Blah.

02:33 <set_> Sorry.

09:24 <set_> It is basically this 120v sychronous motor, three RPM per minute, that is iffy. So, bypass!

14:08 <zmatt> 120v? try not to get yourself killed

14:14 <Konsgn> rcn-ee: Thank you for that, it does seem that it may be a difference in how the u-boot is configured. The yocto sources do not do anything with rtc in u-boot. Would it be possible to somehow enable the rtc in internal mode from the device tree on mainline 5.9?

14:15 <zmatt> that actually makes sense, it should probably be done by u-boot

14:16 <zmatt> though it's a bit weird, it looks like am33xx.dtsi only declares the internal clock?

14:20 <zmatt> oh I'm looking at old versions

14:21 <zmatt> konsgn: 5.9 ? you realize that that's an unmaintained kernel series?

14:21 <zmatt> (EOL december 2020)

14:22 <zmatt> 5.10 is LTS until december 2026

14:23 <Konsgn> Perhaps we should switch to that if it is available on the Dunfell branch of yocto

14:24 <zmatt> LTS releases are 5.4, 5.10, and 5.15

14:24 <Konsgn> got it. What makes the decision for LTS or not?

14:25 <zmatt> a declaration from the higher powers

14:25 <Konsgn> hhaha alright, thank you.

14:27 <zmatt> https://kernel.org/category/releases.html#what-is-the-next-longterm-release-going-to-be

14:29 <Konsgn> ahhh I see, 5.9 is the latest available from the meta-bbb layer

14:30 <zmatt> wouldn't you want to use a ti kernel when using yocto anyway?

14:31 <Konsgn> But the beaglebone project made such a nice configured kernel.

14:32 <zmatt> they did? I didn't know they did anything with yocto at all

14:32 <Konsgn> ... you mean meta-ti?

14:33 <Konsgn> nagh, but the meta-bbb tries to get close to the beagle provided setup in my understanding

14:33 <zmatt> I have no idea, I don't actually know what I'm talking about... I just know that TI's official SDKs are based on yocto, and they'll undoubtedly use TI kernels

14:35 <zmatt> (and TI kernels are all based on LTS mainline kernels)

15:05 <Konsgn> hmm, If i was looking to debug the rtc registers of the am33xx is there something I could look at in the /sys folder to check them?

15:07 <zmatt> dunno, maybe the clock tree?

15:08 <Konsgn> sys/kernel/debug/clk?

15:08 <zmatt> /sys/kernel/debug/clk/clk_summary ... but that's only if the relevant register is actually declared as clock mux to linux, which it might not be

15:12 <zmatt> it looks like the kernel driver, in kernel 5.10, never touches bit 3 of the RTC_OSC register

15:12 <zmatt> it doesn't even have a #define for that bit

15:13 <zmatt> so it should definitely be configured in u-boot

15:13 <zmatt> wait I'm blind

15:13 <zmatt> sorry

15:14 <zmatt> they just didn't put the #defines in the correct order

15:15 <zmatt> okay it's set depending on whether the rtc has a clock named "ext-clk"

15:30 <Konsgn> ... interesting.. the only reference I see to rtc is in the am33xx-l4.dtsi where the node is created, elsewhere I see the system-power-controller flag, but i don't see external parent clocks

15:30 <Konsgn> I am looking at 5.9 though...

20:40 <rcn-ee> konsgn: when i last looked at this (v4.4/v4.9) era, the register was u-boot only bit.. once in kernel, it was locked offf.

20:48 <Konsgn> Yea, current focus is update the uboot to get it set. I can see it verify that the external source is set, now to just get is adjusted to internal.

20:48 <zmatt> rcn-ee: dunno about 4.9 or older, but in 4.14 and later the kernel will use 32 kHz osc if the device has a clock named "ext-clk"

20:49 <zmatt> although if it doesn't then it doesn't seem to touch the bit

20:54 <zmatt> here's the relevant snippets from the driver: https://pastebin.com/nX5LRHsn

20:57 <zmatt> (this is from 4.14 but nothing of importance has changed in the driver since then)

20:58 <zmatt> (nothing of importance to clock source selection)

21:35 <Konsgn> I got it to work by adding in rtc32k_enable() to https://github.com/beagleboard/u-boot/blob/v2020.01/board/ti/am335x/board.c\

21:35 <Konsgn> Thank you!!

23:16 <set_> @zmatt: I know. Luckily, the motor in question just rotates at 3 RPM. Ha. It is a super slow motor for a doorway to control fans.

23:18 <set_> Fridge crud. The control board is closed source/proprietary and I the chips on it are even clean, i.e. no insignia.

23:18 <set_> So, I might switch some items out. Who knows?

23:19 <set_> I still have the older model GE greenbean thing.

23:22 <set_> I was thinking of putting the BBB to use but I think it gets too cold in the compartment where the electronics are located.