eth: fix transaction sender cache miss before broadcast (#31657)

BroadcastTransactions needs the Sender address to route message flows
from the same Sender address consistently to the same random subset of
peers. It however spent considerable time calculating the Sender
addresses, even if the Sender address was already calculated and cached
in other parts of the code.

Since we only need the mapping, we can use any signer, and the one that
had already been used is a better choice because of cache reuse.

eth/handler.go

@@ -484,7 +484,7 @@ func (h *handler) BroadcastTransactions(txs types.Transactions) {
 	total := new(big.Int).Exp(direct, big.NewInt(2), nil) // Stabilise total peer count a bit based on sqrt peers
 
 	var (
-		signer = types.LatestSignerForChainID(h.chain.Config().ChainID) // Don't care about chain status, we just need *a* sender
+		signer = types.LatestSigner(h.chain.Config()) // Don't care about chain status, we just need *a* sender
 		hasher = crypto.NewKeccakState()
 		hash   = make([]byte, 32)
 	)