Pass warp sync target block to parachain node's SyncingEngine during initialization

[dmitry-markin]

Currently, warp sync target block header is passed to parachain node syncing via oneshot channel, which is later polled in SyncingEngine:

polkadot-sdk/substrate/client/network/sync/src/engine.rs

Lines 659 to 667 in 6508866

	warp_target_block_header = &mut self.warp_sync_target_block_header_rx_fused => {
	if let Err(_) = self.pass_warp_sync_target_block_header(warp_target_block_header) {
	error!(
	target: LOG_TARGET,
	"Failed to set warp sync target block header, terminating `SyncingEngine`.",
	);
	return
	}
	},

and passed all the way down to WarpSync:

polkadot-sdk/substrate/client/network/sync/src/strategy/warp.rs

Lines 290 to 301 in 6508866

	pub fn set_target_block(&mut self, header: B::Header) {
	let Phase::PendingTargetBlock = self.phase else {
	error!(
	target: LOG_TARGET,
	"Attempt to set warp sync target block in invalid phase.",
	);
	debug_assert!(false);
	return
	};
	self.phase = Phase::TargetBlock(header);
	}

This is cumbersome, requires checks for the correct phase of syncing, leads to bugs like #3496, and should be reworked to pass the target block header as a parameter during the initialization of SyncingEngine.

[nazar-pc]

This PR should resolve it I think: #5431