* groom poh_recorder
* fixup
* nits
* slot() from the outside means "the slot the recorder is working on"
* remove redundant check
* review comments, put next_tick back in the "is reset" check
* remove redundant check
* Some counters for real poh perf analysis
* more metrics
* Comment on CPU affinity change, and reduce hash batch size based on TPS perf
* review comments
* Remove unused PohServiceConfig::Step
* Clarify variable name
* Poh::hash() now takes an iteration counter
* man -> max
* Inline functions with single call site
* Move PohServiceConfig into GenesisBlock
* Add plumbing to enable real PoH on testnets
* Batch hashes to improve PoH hash rate
* Ensure a constant hashes_per_tick
* Remove PohEntry mixin field
* Poh/PohEntry no longer maintains tick_height
* Ensure a constant hashes_per_tick
* ci/localnet-sanity.sh: Use real PoH
* Rework Poh/PohService to keep PohRecorder unlocked as much as possible while hashing
* Restart node test (#3459)
* Add test to local_cluster for restarting a node
* fix so that we don't hit end of epoch - leader not found before trying to transfer
* Do not look for confirmations, b/c nobody is voting on empty transmissions in this single node test
* Don't vote for empty leader transmissions
* Add is_delta flag to bank to detect empty leader transmissions
* Plumb new is_votable flag through replay stage
* Fix PohRecorder tests
* Change is_delta to AtomicBool to avoid making Bank references mutable
* Reset start slot in poh_recorder when working bank is cleared, so that connsecutive TPU's will start from the correct place
* Use proper max tick height calculation
* Test for not voting on empty transmission
* tests for is_votable
* wake up replay stage when the poh bank is cleared
* bump ticks per second
* Increase ticks per slot to match faster tick rate
* Remove check that working bank must be the bank for the greatest slot
* Make start_leader() skip starting TPU for slots we've already been leader for