Debugging with Ozone fails on verification phase with nRF9161

[mmade]

This is same what same as in my previous ticket  Restarting debug session fails on firmware verification 

With nRF9160 Ozone solution worked just fine but when trying to debug nRF9161 with Ozone I receive error on Ozone Console and also get an popup telling me that Download failed.

Error on Ozone console.

Failed to verify @ address 0x00FF8108
Download failed: error during verification phase

If I have understood correctly the problem is in the UICR area memory verification. Is there workaround or solution for this. When I try to debugging directly with VS Code I get the same error as on previous ticket.

[Elfving]

Hello again,

In Ozone, are you using "Start debug session"->"Attach to running program" under the Debug tab?

Regards,

Elfving

[mmade]

When this error occurs I'm using "Download & Reset program".

If I change "Tools->J-Link Settings->Target Device" to "nRF9160_xxAA" I don't get an error message (when using nRF9161 as hardware) and I can debug software.

[Elfving]

mmade said:

When this error occurs I'm using "Download & Reset program".

Right. The reason why ozone should work better than VSC is that VSC is forcing you to re-program. And you would then have to erase the chip to write to UICR again.

So attaching to running program is the way to go. I'm surprised it worked differently on the 9160?

Regards,

Elfving

[mmade]

This means that I cannot no longer make small changes and use "Download & Reset program" to flash and debug. With nRF9160 that was quite convenient because we have settings on the flash also and we could keep those intact even after flashing new program.

Do you know what is the root cause for different behavior between nRF9160 and nRF9161 when using Ozone?

[Elfving]

mmade said:

This means that I cannot no longer make small changes and use "Download & Reset program" to flash and debug. With nRF9160 that was quite convenient because we have settings on the flash also and we could keep those intact even after flashing new program.

Once again I'm surprised it worked differently on the 9160. I think it should be the same.

Is there any difference in the FW that you use for the two?

Regards,

Elfving

[mmade]

Applications and bootloader settings are the same.

[Elfving]

I'm glad that you atleast have a workaround for the time-being.

Do you see the same issue with other samples?

Is your FW signed? And is that the case on both 60 and 61? Do you know what you are writing to UICR?

Regards,

Elfving

[mmade]

I haven't tried other samples.

My firmware is signed and code, keys, project settings, etc. are exactly same (only board related part of code is different based on Nordic sample board definitions).

I don't know what is written to UICR. I think it has same thing to do with bootloaders. My application doesn't currently use UICR for storing any data. The error address looks like it is related to UICR.

[Charlie]

I am not sure which sample you are working with. I tested v2.6.0\zephyr\samples\hello_world sample on a nRF9161DK. It works fine with Ozone.