* lock state mutex while  computing sync actions

* rename batcher.state to batcher.channelManager

Closes https://github.com/ethereum-optimism/optimism/issues/13280

* move channel manager mutex up into driver

encapsulate imperative code and mutex locking/unlocking into new method executeSyncActions

* don't use pointer to mutex

* rename

* throttlingLoop uses channelMgrMutex

* unblock throttling loop by sending pendingBytes on a channel

* disable proposer in 4844 system test

it mostly adds noise to the logs

* remove unused ticker

* Revert "remove unused ticker"

This reverts commit b2e9762c5e6e92e8b62771439d864d9985dcc943.

* reinstate throttle on tick

used cached value

* make the main loop release the chMgr lock more often

* simplify

* add some more mutex calls

* simplify some more

* push mutex locking down into l.publishTxToL1

* do not signal the throttling loop inside publishTxToL1

This is a change in behaviour that is outside the scope of this PR.

* prefer deferring mutex unlock