core: provide main thread as optional default module

[miri64]

Contribution description

For a private project of mine I did not really need the main thread. At most I needed it for some device-setup (I will also provide a feature to do that without main() that), but the rest was basically happening in ISR (or for other applications the action might happen in other threads). While it isn't really necessary for my specific case, as the board has a large enough memory, to cope with that, but basically I am wasting a lot of memory, just for simple initialization. As such I am wondering if we could provide a mechanism to completely remove the main-thread and is stack. This is my proposal for that.

Testing procedure

All applications should still compile and work as before. When compiling with DISABLE_MODULE=core_main most applications should still compile, they might not work properly anymore, but they should be significantly smaller.

Issues/PRs references

Requires #10970

[jcarrano]

The way this is done, disabling the main thread also disables auto init. There are two options:

1. Accept it and note that dependency in the build system so that if both features are requested, an error is raised.
2. Change the kernel initialization so that auto-initialization is done even if the main thread is not launched.

The problem with (1) is that if there is no main thread and no auto_init, who initializes stuff (maybe I'm missing something?)

[miri64]

The way this is done, disabling the main thread also disables auto init. There are two options:

1. Accept it and note that dependency in the build system so that if both features are requested, an error is raised.
2. Change the kernel initialization so that auto-initialization is done even if the main thread is not launched.

The problem with (1) is that if there is no main thread and no auto_init, who initializes stuff (maybe I'm missing something?)

I was aware of that, but forgot to mention this in OP. For now I would prefer to keep (1) with a later move to (2) in a follow-up (which would be distinct from the application-specific setup routine I'm planning to propose, once the pseudo-module issue in #10970 is resolved)

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you want me to ignore this issue, please mark it with the "State: don't stale" label. Thank you for your contributions.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you want me to ignore this issue, please mark it with the "State: don't stale" label. Thank you for your contributions.