Scheduling enclaves

[erlingrj]

This PR implements the first version of scheduling enclaves in the C target. This PR is paired with lf-lang/lingua-franca#1872.

---

Things not supported yet:

1. Enclaved federations.
2. Physical connections between enclaves
3. Ability to specify a maximum lead between enclaves
4. Smart mutexes wich can catch violations of specified order of acquisition

[erlingrj]

@edwardalee and @lhstrh. I am tagging you for a review now. There is still a test failure on macOS on a test with coordinated shutdown of enclaves. I will look at this when I get my hands on a MacBook next week

[erlingrj]

The compiler errors now are due the fact that in reactor-c the preprocessor flags for picking the scheduler has changed, from NP -> SCHED_NP and so on. I think that my first step in November should be to craft a minimal lingua-franca PR which can me merged together with this PR. It should not introduce the AST transformation and enable enclaves. But just let us merge these refactorings of the RTI, which I think are good, into master.

[edwardalee]

This is just a review of rti_local.h/c. I don't understand the locking strategy here.

[edwardalee]

The compiler errors now are due the fact that in reactor-c the preprocessor flags for picking the scheduler has changed, from NP -> SCHED_NP and so on. I think that my first step in November should be to craft a minimal lingua-franca PR which can me merged together with this PR. It should not introduce the AST transformation and enable enclaves. But just let us merge these refactorings of the RTI, which I think are good, into master.

I have started doing this in lingua-franca/enclaves3, which is pointed to by reactor-c/enclaves3, which includes some small extensions of your reactor-c/enclaves2 branch. Feel free to push directly to either enclave3 branch or to your reactor-c/enclaves2 branch, which we can then merge into reactor-c/enclaves3.

[erlingrj]

Answering questions

[edwardalee]

I've finally gone through this whole PR and it looks good to me! Modulo small comments, I think we can merge this as soon as we can get tests to pass in lingua-franca/master with minimal changes to lingua-franca. This will not yet provide support for enclaves, but will get us much closer.

I'm proposing further changes in the enclaves3 branch, but we can review and merge those after merging this into main.

[erlingrj]

@edwardalee. I have addressed all comments. I removed the PROLOGUE/EPILOGUE macros which locked/unlocked mutex. I have made the documentation clearer with respect to which mutex are assumed held etc. Tests are passing here and in LF? Shall we merge?