[scheduler] Another task-stealing optimization

[RobinMorisset]

The main win here is the third commit.
It is a followup to #9594, and was tested in prod on multiple kinds of machines and multiple services.
We observed some small but significant wins in scheduler utilisation, including at peak load which was the weak point of the previous batch of optimisations.
cc @tomas-abrahamsson since you saw large wins from the previous batch of scheduler enhancements.

[github-actions]

CT Test Results

    3 files    135 suites   49m 11s ⏱️
1 649 tests 1 592 ✅ 57 💤 0 ❌
2 285 runs  2 209 ✅ 76 💤 0 ❌

Results for commit 09c6a67.

♻️ This comment has been updated with latest results.

To speed up review, make sure that you have read Contributing to Erlang/OTP and that all checks pass.

See the TESTING and DEVELOPMENT HowTo guides for details about how to run test locally.

Artifacts

• Complete CT logs (Download Logs)
• HTML Documentation (Download HTML Docs)
• Windows Installer

// Erlang/OTP Github Action Bot

[RobinMorisset]

I appear to have missed some failed assert in debug builds, I'm looking into it.

[rickard-green]

The ASSERT(!is_normal_sched || (flags & ERTS_RUNQ_FLG_NONEMPTY)); hit since the ERTS_RUNQ_FLG_NONEMPTY flag more or less has been used as "the scheduler is not waiting" as well as "the run queue isn't empty", in the accompanying no_empty_run_queues variable, which are not the same thing. Since the fix in your first commit, the assert hits.

I added a new commit "Track waiting schedulers" removing the assert that hit and added another way of keeping track of waiting schedulers. It is not used everywhere where it could be used, but I think that the legacy configurations that this is the case for will behave more as before by not changing them.

Please remove the meta internal stuff like the following from the commit messages:

    Test Plan: CI + canary
    
    Reviewers:
    
    Subscribers:
    
    Tasks:
    
    Tags:
    
    
    Differential Revision: https://phabricator.intern.facebook.com/D72957513

[RobinMorisset]

Thanks for fixing the assertion, and sorry for the messed-up commit messages.
I've fixed them and pushed the new version of the PR.