chore(Test): Revert #20308 precautionary serial testing

[bartelink]

Right now, main has precautionary serial running of some Async tests applied in #20308 in response to #20306

The original increase of parallelization that was suspected of causing the instability arose in #19785 and #20258

Sample size 1, but: Build WindowsNoRealsig_testDesktop => 97m vs 100m for c251d06

My contention is that:

• the key instability that was present was that a failsafe timeout on a wait in a new test was 5s, as opposed to the 30s as we learned makes sense for heavily loaded CI boxes running parallelized tests
• whether or not those Async/Task suites are run concurrently or not is not a factor (xunit conservative mode doesn't have SynchronizationContext-sensitive behavior, and we there's no reason to think we'd reintroduce that mode at a future point; we're more likely to move to ParallelizationMode.All)

resolves #20306

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[github-actions]

✅ No release notes required

[T-Gro]

The reasoning holds up under inspection, so this is a clean revert to make.

The only shared mutable state in either file is Async.CancelDefaultToken, and every call to it lives inside the AsyncType module and the AsyncType class — both of which keep their [<Collection(nameof NotThreadSafeResourceCollection)>]. The four suites you release back to parallel execution (AsyncModuleFunctionsTests, AsyncAwaitTaskLikeTests, AsyncStartTaskImmediateTaskLikeTests, AsyncAwaitStackTraceTests) touch no process-global state, so there was never a correctness reason for them to be serialized — the 30s failsafe in verifyTaskCompletion is the part that actually addressed the loaded-CI flakiness, and that stays.

A full green fsharp-ci run with no host timeouts is the empirical half of the argument.

One tiny thing, not blocking: the AsyncTaskLikeAwaitTests → AsyncAwaitTaskLikeTests rename is a nice readability tidy-up but is unrelated to the revert, so a one-line note in the PR body would save the next archaeologist a git blame.