Add process control plug

[MonicaisHer]

This PR will resolve #8. As the issue described, there is a seccomp violation on a sched_setattr system call, and snappy-debug suggests adding the process-control plug.

chip-tool toggle command success rates :

process-control plug	without	with
chip-tool toggle command success rates	8/10 (2 Delay) 9/10 (1 Timeout) 8/10 (2 Timeout) 9/10(1 Delay) 9/10 (1 Timeout)	10/10 7/10 (3 Timeout) 10/10 10/10 9/10 (1 Timeout)

Note: Timeout is a fatal error; Delay is not a fatal error.

Overall, using a process-control plug provides chip-tool snap with slightly higher success rates when controlling a device.

[farshidtz]

The PR addresses two unrelated issues, one for documenting connections on UC, and another for changing the packaging to allow granting access to a new interface (which itself isn't being documented).

[farshidtz]

Why is process-control interface being added? As issue #8 shows, snappy-debug suggests this interface because there is a seccomp violation on a sched_setattr system call. Is this affecting the functionality?

See also https://snapcraft.io/docs/debug-snaps#heading--seccomp

[MonicaisHer]

Yes, it does affect the functionality. Please see the PR description has been updated to include an explanation.

The solution, which involves running the scmp_sys_resolver command to solve the Seccomp violation, appears to be less user-friendly compareing to simply adding the process-control plug.

[farshidtz]

Yes, it does affect the functionality. Please see the PR description has been updated to include an explanation.

The solution, which involves running the scmp_sys_resolver command to solve the Seccomp violation, appears to be less user-friendly compareing to simply adding the process-control plug.

The data you shared shows the same success rate (4/50 timeouts) with and without connecting the interfaces. The delays may be caused by external networking and computing factors; it appears that the tests have been done in batches without randomization.

I'm not sure what solution are you referring to, involving scmp_sys_resolver. The shared document uses that command to resolve the syscall name from its ID.