fix(cosmic-swingset): Provide blockTime to bootstrap

[mhofman]

closes: #7124

Description

call timer.poll() before running the bootstrap code.

The wiring of the device is performed synchronously when the devices are built, so there the implementation won't throw.
The implementation of poll should only access its own state since no callbacks would be registered yet, which seems safe to do before any bootstrap code has actually run. I have not fully audited the device code however to check this, just a cursory look.

Security Considerations

None

Scaling Considerations

None

Documentation Considerations

None

Testing Considerations

Tested locally with the loadgen to make sure this change actually solves #7124 and that no timer related errors occur.

[datadog-full-agoric]

Datadog Report

Branch report: mhofman/7124-bootstrap-blocktime
Commit report: 3e413c2

✅ agoric-sdk: 0 Failed, 0 New Flaky, 3660 Passed, 57 Skipped, 15m 15.39s Wall Time

[warner]

That looks fine for start-from-scratch chains. Will this also work for the bulldozer upgrade, or will blockHeight be wrong? Should we be copying blockHeight out of action or something?

If this works for both cases, go ahead and land. If it won't work for the bulldozer upgrade, please land it and then file an issue to remind us to fix the second case before the release.

[mhofman]

If this works for both cases, go ahead and land. If it won't work for the bulldozer upgrade, please land it and then file an issue to remind us to fix the second case before the release.

We're not currently using the blockHeight anywhere for bootstrap, and I'm actually not sure what the blockHeight should be in the bulldozer case. 0 is treated somewhat special in hangover checks, but bootstrap doesn't actually set that value itself, so it seems a bit arbitrary. @michaelfig any ideas? Should we attempt to use a different more accurate value for bootstrap blockHeight, and in which case, what would the value be?