That worked. Thanks so much.

@Lazar-Demin @György-Farkas thanks, I was starting to think I was stupid. This explains some inconsistent outcomes I was experiencing.

readValue needs to be passed as a reference, since the contents will be modified by the i2c_read function. That's likely causing the segmentation fault.

Or change readValue to a point and allocate two bytes before passing it to i2c_read

See the documentation for the i2c_read function for more details: http://docs.onion.io/omega2-docs/i2c-c-library.html#i2c-c-read-multiple-bytes

@Don-Elwell I've wondered about the same thing when thinking about an Omega2(S) with a PCIe graphics card 😅 Let me know if you find out something..

I now stumbled upon similar case, if not the same with another board. I was able to connect access point of Omega2S and open putty to run commands. After I have run command wifisetup and select option 1, link with omega broke, and I think it also rebooted. I'm not sure about all circumstances, as I didn't connect via serial port.

Then I made conection with O2S again and check ifconfig.

I noticed that there is suspecious eth0.1 interface. So I decided to run command ifconfig eth0.1 down to disable that interface.
Then I tried to run wifisetup and this time it was succesful! I managed to finish instalation of all needed software and it works like a charm.

I can't check if it's the same with one that I had problems with, as it is not on it's board anymore, but for this one it worked.

@Jeff-Seese

Does anybody else have ... a real hard reset solution?

Here you are a "brute force" solution by Onion for the HW_RST button problem of Omega2S+.
It should work theoretically if you put a resistor between the base (1) and emitter (2) of the Q2 PNP transistor. ;-)
I wonder what about in practice because I do not have such devices.
Omega2S-Dev-Board.pdf

@Zheng-Han Thank you for your reply. Sorry to be a pain but can I specify my previous question?
According to @Larry-Pinney's photo Omega2+ has a Winbond 25Q256FV flash.
Does Omega2S+ also have a Winbond 25Q256FV or some fully equivalent flash chip?

@Zheng-Han said in Newly minted Omega2S:

yes, Omega2S+ has the same spec as Omega2+ in terms of CPU, RAM and Flash

Because if yes, that would mean it has the same 3B / 4B address mode issue than the Omega2+ has and its HW Reset is broken too.

Please correct me if I'm wrong.

@Richard-Chabot
kernel patch for correct omega2+ reboot

Caveat: It always reboots, even with the poweroff command.

@wdu , hi! Can you share with us the details? Did you just write the binary to the raw flash memory? Starting at which address? I imagine at least some part of the 32MB of flash must be partitioned with a FS (jffs2?), so we must be careful not to overwrite it, right?

Can someone point me to or share the flash's map / partitioning scheme?

@Pavel-Metrokhin , can you please confirm if building the source @ https://github.com/OnionIoT/omega2-bootloader will generate the same bootloader provided @ http://repo.onion.io/omega2/bootloader/ ? In other words, is the GitHub repo complete and up to date or you build the provided bootloader from a private repo?

Thank you!

First of all, there is no PCI (or PCIe) bus on Omega2, thus no go on that route.

Regarding USB-VGA output dongle:
It needs a driver on the MT7688 side (Omega2 side) and not only that,
it needs a large frame buffer on the Omega2. That USB-VGA dongle is just a data pump/VGA output formatter to continuously refresh/display the content of the frame buffer.
In such case, Omega2 does not have large enough memory, nor enough horse power to isochronously send the FB data out over USB.
In another word: dead-end.

Use the device as what its is most suitable for.
Your case needs a device such as Raspberry Pi zero.

Here you are, unofficial (my own project) but usable Kicad library, footprint and 3D shape:
Onion_Omega_2.7z