Allow running and testing network samples in automatic way

[jukkar]

Is your enhancement proposal related to a problem? Please describe.
The network samples can bit rot easily as they are only compile tested by sanitycheck script.
We need to run the network samples in some semi-automatic way so that we can be sure that there is no regressions in network stack.

Describe the solution you'd like
Use Docker to to run scripts and programs from net-tools, then have qemu / native_posix to connect to that Docker instance. Have also IP addresses setup properly so that connectivity between these two works.

The task can can be split up as follows:

• Create a script to run sample tests against the Docker host tools from 'net-tools'
• Add sample tests where the Docker 'net-tools' image can serve as the client end as the result is easy to have available when the client stops executing in the Docker container
• Investigate how to utilize e.g. native_posix Zephyr target to exit with a return value after a completed test where Zephyr acts as a client and the Docker image is the server end
• Create sample tests where the native_posix Zephyr target can serve as the test client, if the above point works out properly
  - [ ] Write the test script in some other language than bash, python has been proposed (future development, if needed)
• Run samples against Docker net-tools container #19677 Is tracking items

[dleach02]

@jukkar, I note that all the subtasks are checked off but it is unclear to me if there are actual PRs or items merged into main?

[pfl]

note that all the subtasks are checked off but it is unclear to me if there are actual PRs or items merged into main?

We have #19677 in the pipeline implementing the ticked off items. The support in net-tools was handled by zephyrproject-rtos/net-tools#37 which was already merged.