Fix split tunneling packet loss

[dlon]

Fixes the following error:

Failed to forward to tun device: Message too long (os error 40)

There were a couple of issues:

• The MTU of the ST utun was potentially higher than that of the VPN tunnel. We now keep track of and sync the MTUs.
• Outbound packets weren't correctly filtered. Capture::direction seems to handle it correctly but doesn't work on macOS 13. Instead, we must discard packets with the wrong source IP.

Close DES-1115.

---

This change is 

[linear]

DES-1115 Fix "Message too long" issue

[MarkusPettersson98]

Reviewed 15 of 16 files at r1, all commit messages.
Reviewable status: 15 of 16 files reviewed, 3 unresolved discussions (waiting on @dlon)

---

talpid-core/src/split_tunnel/macos/tun.rs line 345 at r1 (raw file):

        if details.mtu == current_mtu {
            return;
        }

What sort of condition is this? And why would the mtu_updater completely return instead of continuing to listen for new mtu updates? 😊

Code quote:

        if details.mtu == current_mtu {
            return;
        }

---

talpid-core/src/split_tunnel/macos/tun.rs line 361 at r1 (raw file):

            log::error!("Failed to set MTU of {st_iface_name} to {mtu}: {error}");
            return;
        }

The return-statement seems excessive here - I assume it's a left over from before you refactored this piece of code into a function 😊

Code quote:

        Err(error) => {
            log::error!("Failed to set MTU of {st_iface_name} to {mtu}: {error}");
            return;
        }

---

talpid-net/src/unix.rs line 0 at r1 (raw file):
So much unsafe here without SAFETY sections :( Would you mind adding that? 😊

[dlon]

Reviewed 1 of 16 files at r1, 1 of 1 files at r2, all commit messages.
Reviewable status: 14 of 16 files reviewed, 3 unresolved discussions (waiting on @dlon and @MarkusPettersson98)

---

talpid-core/src/split_tunnel/macos/tun.rs line 345 at r1 (raw file):

Previously, MarkusPettersson98 (Markus Pettersson) wrote…

What sort of condition is this? And why would the mtu_updater completely return instead of continuing to listen for new mtu updates? 😊

Done. Nice catch!

---

talpid-core/src/split_tunnel/macos/tun.rs line 361 at r1 (raw file):

Previously, MarkusPettersson98 (Markus Pettersson) wrote…

The return-statement seems excessive here - I assume it's a left over from before you refactored this piece of code into a function 😊

Done. Nice catch!

---

talpid-net/src/unix.rs line at r1 (raw file):

Previously, MarkusPettersson98 (Markus Pettersson) wrote…

So much unsafe here without SAFETY sections :( Would you mind adding that? 😊

Done.

[MarkusPettersson98]

Reviewed 1 of 1 files at r2, 1 of 1 files at r3, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

[MarkusPettersson98]

Reviewed 4 of 4 files at r4, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved