layered: actually get slice_ns and max_exec_ns from CO:RE

[JakeHillion]

Layered previously hardcoded the defaults for slice_ns and max_exec_us. This
was changed in a recent PR to get these from CO:RE in the running kernel. This
change doesn't work because the old values are still set a few lines below.

Check for the presence of the CLI options and default to the CO:RE values
otherwise. Remove setting the layer struct's slice_ns as we check if it's equal
to 0 anyway and default to the global in the BPF code.

Test plan:

• CI

[etsal]

I think ideally we should have --slice-us and --max-exec-us to not have a default and only override the system default if set. The current code definitely does not use the builtin values because there is both a default and a possibly user-defined value provided from the scheduler, and it makes sense for both to override the system default. I think it makes sense to removes the scheduler-provided defaults and use the system ones instead, but the current code AFAICT breaks the --slice-us and --max-exec-us switches.

[JakeHillion]

I think ideally we should have --slice-us and --max-exec-us to not have a default and only override the system default if set. The current code definitely does not use the builtin values because there is both a default and a possibly user-defined value provided from the scheduler, and it makes sense for both to override the system default. I think it makes sense to removes the scheduler-provided defaults and use the system ones instead, but the current code AFAICT breaks the --slice-us and --max-exec-us switches.

Ah sure, you're right! Will update and test this properly. Spotted going through that the current behaviour (setting it twice) clearly wasn't what was intended and thought it was an easy fix.

[JakeHillion]

Closing for now as it would need rebasing and isn't my priority at the moment. Would still be good to fix at some point in the future.