I like(*) to connect to the micro-USB port of my OrangePi Zero (H2+)
a USB peripheral device, so the port at the OPi must act as host.
I have activated the port in armbian-config. Then, following
forum posts such as [1], I switched the dr_mode in the
port's device tree section to "host". (Decompile, edit, compile, then
placed the resulting .dtb file to /boot/dtb-.../, then rebooted.)
Nevertheless, the port does not seem to work as host, at least
no device is detected in dmesg when plugging the stick in
at the port. (same with lsusb)

 

I have tested the stick itself in another hardware setting, and the
cabling near the OPi's micro-USB port in so far as I can get g_serial to work.
(I.e. OPi in "peripheral" mode works perfectly, so the cable is good.)

 

Initially I thought it was just an overlay issue, but trying both "host" and
"otg" as dr_mode values did not deliver any progress.

 

Any ideas what could be the cause?

 

My "uname -a" is:
Linux orangepizero 4.14.18-sunxi #24 SMP Fri Feb 9 16:24:32 CET 2018 armv7l GNU/Linux

NB: I had problems to "rmmod" the g_serial module; the command hangs indefinitely
when run at console. I guess the g_serial module needs eventually to be
unloaded at least, to get the device work in host mode.
(I read an Igor's post somewhere saying to that effect.)

[1]

 

(*) I will try the other three ports as well as workaround as soon I have the hardware parts,

but if a simple solution for the micro-USB exists I'll go with it.

 

15 hours ago, lionz said:

Linux orangepizero 4.14.18-sunxi #24 SMP Fri Feb 9 16:24:32 CET 2018

Your firmware is quite old. Nobody will waste time to dig into unsupported/outdated software.

That's acknowledged. When writing the previous, I was nearly about to add the attribute "quite old" myself.

 

I had not upgraded it for long, following the "never change a running system".

 

Hope to come to this in course of the evening and tomorrow early morning. I guess however that it is still a risky process.

 

 

 

 

 

 

 

 

 

You can always do a backup beforehand by simply creating an img of your microSD card.

On 8/2/2021 at 6:56 PM, lionz said:

I guess however that it is still a risky process.

 

Overall I think we are providing better quality Linux than anyone else, but still its not perfect, it has problems and there is nothing we can do about - we are already spending more private resources than we should to keep this level.

 

We are trying to implement professional quality control and testings (can be easily full time position, senior in any case), but since this is very expensive process nobody is willing to cover on the needed degree, things are as they are.

 

Even we are trying hard to fix problems, we can't guarantee it will not break and manual testing on all boards + functions + combinations is simply too insane operation / out of question.