Add WireGuard over Shadowsocks

[dlon]

---

This change is 

[dlon]

Reviewable status: 0 of 33 files reviewed, 6 unresolved discussions (waiting on @Serock3)

---

docs/relay-selector.md line 58 at r2 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

Or something to clarify that shadowsocks is used while connecting and not for communication

Done.

---

mullvad-relay-selector/src/relay_selector/helpers.rs line 145 at r2 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

Is this correct? Should we really compare the extra addresses to the main address, not to the users IP setting?

Good question. I guess the IP setting determines the family used by wg_in_addr, so will it not be the same?

---

mullvad-relay-selector/src/relay_selector/matcher.rs line 43 at r2 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

For consistentcy with the above formatting

Done.

[dlon]

Reviewable status: 0 of 34 files reviewed, 6 unresolved discussions (waiting on @Serock3)

---

mullvad-relay-selector/src/relay_selector/matcher.rs line 166 at r2 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

Nightly fmt

        // If Shadowsocks is specifically asked for, we must check if the specific relay supports
        // our port. If there are extra addresses, then all ports are available, so we do
        // not need to do this.

Done.

---

mullvad-relay-selector/src/relay_selector/matcher.rs line 176 at r2 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

This line does not compile for me for some reason, even though it should be fine (and rust-analyzer doesn't complain). Cargo doesn't seem to find the from impl.

Is this still the case now?

[Serock3]

[Serock3]

Reviewed 10 of 32 files at r1, 1 of 10 files at r2, 7 of 18 files at r4, 2 of 4 files at r5, 3 of 7 files at r6, 6 of 9 files at r9, 4 of 9 files at r12, 1 of 1 files at r13, 1 of 2 files at r14, 2 of 3 files at r15, 1 of 1 files at r16, 1 of 1 files at r17, 7 of 9 files at r18, 7 of 7 files at r19, all commit messages.
Reviewable status: all files reviewed, 13 unresolved discussions (waiting on @dlon)

---

mullvad-relay-selector/src/relay_selector/helpers.rs line 145 at r2 (raw file):

Previously, dlon (David Lönnhager) wrote…

Good question. I guess the IP setting determines the family used by wg_in_addr, so will it not be the same?

After digging a bit, it seems that the address type of wg_in_addr will be determined by the user settings according to the resolve_ip_version fn. I guess we miss the opportunity to switch to IPv6, in case the user IP version is "any" and the extra address is IPv6, but that is probably so unlikely that we don't have to worry about it. I suggest adding a here comment clarifying this behavior though, in case we add many IPv6 extra addresses and want to make it smarter in the future.

Code snippet:

pub fn resolve_ip_version(ip_version: Constraint<IpVersion>) -> IpVersion {
    match ip_version {
        Constraint::Any | Constraint::Only(IpVersion::V4) => IpVersion::V4,
        Constraint::Only(IpVersion::V6) => IpVersion::V6,
    }
}

---

En gammal draft-kommentar från reviewable som följde med när jag godkände