Want abstraction over USDT/SDT split in probes

Many of the DTrace probes we have in OPTE are structured such that they automatically switch between USDT and SDT probes depending on whether the crate is compiled for userland (cargo test) or as part of the kernel module (xde):

impl Port {
    fn xyz_probe(&self, dir: Direction, pkt: &Packet<Parsed>, msg: String) {
        cfg_if::cfg_if! {
            if #[cfg(all(not(feature = "std"), not(test)))] {
                /* SDT-specific arg conversion: String->CString, ... */
                unsafe {
                    __dtrace_probe_xyz(
                        /* ... pointers to args ... */
                    );
                }
            } else if #[cfg(feature = "usdt")] {
                /* USDT-specific arg conversion */
                crate::opte_provider::xyz!(
                    || (/* ... Rust-friendly args (references, owned types) ... */)
                );
            } else {
                /* feature-flag-driven no-op */
                let (..) = (dir, pkt, msg);
            }
        }
    }
}

Many of the Port::*_probe methods have a structure more or less in line with this, which leads to a lot of duplication and noise. We should define a macro or similar abstraction to simplify the implementation of {current, future} DTrace probes we choose to insert.

Start by surveying the Port::*_probe methods and their repeated cfg_if! branches for SDT, USDT, and no-op behavior. Compare the argument-conversion paths and determine the intended macro or abstraction boundary. Done means current and future DTrace probes can use the shared abstraction without duplicating the conditional probe setup.