Normalise subnet withdraw functions on L1 and L2

The L1 subnet contract has the public function[ `withdraw-stx` ](https://github.com/hirosystems/stacks-subnets/blob/b47bc51e5718c409d99054290147f3d1cc8c67b2/core-contracts/contracts/subnet.clar#L686)and the L2 subnet contract has [`stx-withdraw?`](https://github.com/hirosystems/stacks-subnets/blob/b47bc51e5718c409d99054290147f3d1cc8c67b2/src/chainstate/stacks/boot/subnet.clar#L148).

It can be confusing, and the methods don't actually do the same thing.

L2 could have `initiate-stx-withdrawal?` (being explicit on the fact that the withdrawal won't be complete with only this call).
L1 could have `claim-stx-withdrawal?`

Some for ft and nft.