Faster op-node startup on replicas of OP Stack blockchains

[qbzzt]

Is your feature request related to a problem? Please describe.

When I have an existing replica of an OP Stack chain, and I restart op-node after a break, it takes a very long time to start advancing again. While waiting I get a lot of op-node log entries like this:

t=2024-05-11T12:37:23-0500 lvl=info msg="Walking back L1Block by hash" curr=0xcc342fa25773b876a0df0b1061fb30d95929ba31a2470e12197fd363512c6be1:19602253 next=0x2077592483faaa035a1f7aeb4a39fc66d149e8efc4a8d0cb4a56bec06e061b8a:19602252 l2block=0xe68165a993b32282d4fece1e7c079e1664226580caea4098514a5356604044bf:144487

While it is looking for the L1 block from which to start retrieving state.

Describe the solution you'd like

op-node can ask op-geth what is the latest L1 block and start looking from there, all it takes is doing the equivalent of

cast call 0x4200000000000000000000000000000000000015 "number()" | cast to-dec

The next input commitment is (I use Redstone) should be after that point, but not very far from it.