17:16 <ad__> hi, is it correct H2 (bananapi m2 zero) has only 1 PWM pin available ?

17:53 <apritzel> ad__: the H3 SoC itself has one PWM pin on PA5, but a second one (controlled by S_PWM) on PL10. Don't know which of those the BPi M2 Zero exposes ...

19:44 <apritzel> ad__: according to the schematics, PA5 is UART0-RX, and PL10 is connected to the power LED. So you can make the LED change brightness, but it's probably not easily available if you need a second external PWM signal

22:23 <IlikeTech> Hi, I'm trying to get the F1C100s USB working correctly, and I have it successfully acting as a host (enabling it's own regulator to provide vbus and being able to attach devices) if I set it's role specifically to "host" in the device tree. However, I'm having no luck getting it to work in otg mode. This is with mainline Linux 6.10.5. Does

22:23 <IlikeTech> anyone have any experience as to what's needed to be configured in the kernel and / or in the device tree to make this work?

22:43 <IlikeTech> I can't seem to get peripheral / gadget mode to work at all, even

23:01 <apritzel> IlikeTech: so OTG (as in: determined at runtime via the ID pin) is known to have issue (on AW in mainline Linux in general), but dedicating to peripheral via the DT should work

23:02 <IlikeTech> okay, at the moment neither seem to be working, but the latter may be an issue on my part. I know there's an older, working tree for this chip so I'm going to try building that and see if it works

23:02 <IlikeTech> good to know that it has issues though

23:02 <IlikeTech> I assigned the pins as noted here: https://www.kernel.org/doc/Documentation/devicetree/bindings/phy/allwinner%2Csuniv-f1c100s-usb-phy.yaml, but no such luck

23:04 <apritzel> IlikeTech: look at the suniv-f1c200s-lctech-pi.dts in mainline, that should have peripheral mode working

23:04 <IlikeTech> interestingly with the mode set to OTG, it seems to default to host mode but also doesn't turn on the supply, so it doesn't work

23:05 <IlikeTech> should the matching defconfig actually present as a peripheral, or do I need to enable a gadget option?

23:05 <IlikeTech> (other then just gadget support)

23:06 <apritzel> for a mainline kernel?

23:06 <IlikeTech> yeah

23:06 <apritzel> I'd go with the modern way of dynamic configuation via configfs

23:07 <IlikeTech> alright

23:07 <apritzel> that looks somewhat more complicated at first, but gives you much more flexibility, and you can put the commands in a script once and be done

23:09 <IlikeTech> makes sense, thanks for the help

23:09 <IlikeTech> I'll try that and report back

23:09 <apritzel> and that would be generic, for Linux in general, so follow anything that Google turns up out there

23:09 <IlikeTech> yeah

23:10 <apritzel> the F1C100s is a bit different, since there seems to be no HCI controller, which means OTG might actually work, but I forgot all the details about this SoC

23:12 <IlikeTech> I designed a PCB around one and jumped in head first, the hardware was the easy part lol :/

23:12 <IlikeTech> giving myself a crash course in working on u-boot and the kernel

23:18 <apritzel> IlikeTech: since the SoC is somewhat supported in both mainlines, it should really just be a matter of providing a .dts file and the U-Boot defconfig

23:19 <apritzel> both of which you can copy from some existing files

23:19 <IlikeTech> yeah, that's what I've been doing

23:19 <IlikeTech> it's very possible this is a misunderstanding on my part about how device trees / config options lol

23:19 <apritzel> and just to be clear: the .dts files for both U-Boot and the kernel are identical

23:20 <IlikeTech> oh, that could definitely be the issue

23:20 <IlikeTech> u-boot is using the in-tree config for the lichee pi nano

23:20 <IlikeTech> the kernel is using the config I based off of that one that I'm providing via the out-of-tree dts option in buildroot

23:21 <IlikeTech> at least, afaik? I'm assuming u-boot integrates a dts into it's "u-boot-with-spl.bin"

23:22 <apritzel> config & DT are independent: for the kernel just use multi_v5_defconfig, and the DT describes the hardware, so there is nothing to really *configure* there

23:22 <apritzel> similar with the U-Boot defconfig: just copy one, and adjust at least the DT name

23:23 <apritzel> I should actually check whether we can switch the F1C100s in U-Boot to OF_UPSTREAM, which means it literally uses the very same DT files, imported from the kernel repo

23:25 <IlikeTech> hmm, alright, that makes sense

23:29 <IlikeTech> is there a reason to use multi_v5_defconfig over sunxi_defconfig?

23:32 <apritzel> sunxi_defconfig configures a ARMv6/ARMv7 kernel, which does not run on the ARM9 core in the F1C100s - ARM9 is ARMv5

23:33 <apritzel> most of sunxi, starting with the venerable A10, is ARMv7, or ARMv8 for the "newer" Cortex-A53 based SoCs

23:35 <IlikeTech> interesting, I've been using the sunxi_defconfig and I haven't had any issues with it booting

23:35 <IlikeTech> sounds like that's surprising lol

23:36 <apritzel> interesting, haven't checked, but I was under the impression the kernel would use ARMv6+ instructions pretty quickly

23:37 <IlikeTech> I'll try the other config

23:39 <IlikeTech> gonna go eat something, back in a little bit, thanks for the help!