Generate quantum resistant keypairs ahead of connecting

[Serock3]

---

This change is 

[linear]

DES-1614 Generate quantum-resistant keypairs ahead of connection attempt

[hulthe]

Reviewable status: 0 of 3 files reviewed, 1 unresolved discussion (waiting on @Serock3)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 32 at r1 (raw file):

    // capacity of the channel by one
    let bufsize = bufsize.checked_sub(1).expect("bufsize must be at least 1");
    let (tx, rx) = mpsc::channel(bufsize);

mpsc::channel panics if the buffer size is 0, so spawn_keypair_worker(1) would also panic. Sadly, unlike std, tokio does not have rendezvous channels afaik :(

[hulthe]

Reviewable status: 0 of 3 files reviewed, 2 unresolved discussions (waiting on @Serock3)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 21 at r1 (raw file):

type KeyPair = (PublicKey<'static>, SecretKey<'static>);

pub static KEYPAIR_RX: OnceLock<Mutex<mpsc::Receiver<KeyPair>>> = OnceLock::new();

⛏️ Definitely bikeshedding here, but I'd personally keep KEYPAIR_RX private and expose a public funcion to initialize it.

[hulthe]

Reviewed 2 of 3 files at r1.
Reviewable status: 1 of 3 files reviewed, 3 unresolved discussions (waiting on @Serock3)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 57 at r1 (raw file):

        .recv()
        .await
        .expect("Failed to receive key pair, generating working expectedly closed.")

MORE BIKESHEDDING 🚲 🏚️

According to the Rust docs, the message in expect should tell us why we expect it not to panic; the assumptions you make. E.g. "keypair_worker should never exit while receiver exists" or something.

That said, these are more like guidelines than actual rules, yarr. 🏴‍☠️

[dlon]

Reviewed 1 of 3 files at r1, 1 of 1 files at r2, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @hulthe and @Serock3)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 21 at r1 (raw file):

Previously, hulthe (Joakim Hulthe) wrote…

⛏️ Definitely bikeshedding here, but I'd personally keep KEYPAIR_RX private and expose a public funcion to initialize it.

+1

[Serock3]

Reviewable status: 1 of 3 files reviewed, 2 unresolved discussions (waiting on @dlon and @hulthe)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 21 at r1 (raw file):

Previously, dlon (David Lönnhager) wrote…

+1

Done

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 32 at r1 (raw file):

Previously, hulthe (Joakim Hulthe) wrote…

mpsc::channel panics if the buffer size is 0, so spawn_keypair_worker(1) would also panic. Sadly, unlike std, tokio does not have rendezvous channels afaik :(

Damn it. I think I've figured out a fix, but it became more verbose than I would have wanted. I added a fairly large note about this to motivate the choice, but maybe that isn't necessary?

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 57 at r1 (raw file):

Previously, hulthe (Joakim Hulthe) wrote…

MORE BIKESHEDDING 🚲 🏚️

According to the Rust docs, the message in expect should tell us why we expect it not to panic; the assumptions you make. E.g. "keypair_worker should never exit while receiver exists" or something.

That said, these are more like guidelines than actual rules, yarr. 🏴‍☠️

Sure, I'll reword it. My impression was that we didn't bother following that convention, but maybe it doesn't hurt.

[hulthe]

Reviewed 1 of 2 files at r3, all commit messages.
Reviewable status: 2 of 3 files reviewed, 2 unresolved discussions (waiting on @dlon and @Serock3)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 32 at r1 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

Damn it. I think I've figured out a fix, but it became more verbose than I would have wanted. I added a fairly large note about this to motivate the choice, but maybe that isn't necessary?

tx.reserve() 💯

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 57 at r1 (raw file):

Previously, Serock3 (Sebastian Holmin) wrote…

Sure, I'll reword it. My impression was that we didn't bother following that convention, but maybe it doesn't hurt.

It's up to you. Just though I'd mention the convention in case you wasn't aware of it :)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 29 at r4 (raw file):

/// It can take upwards of 200 ms to generate McEliece key pairs so it needs to be done before we
/// start connecting to the tunnel.
pub fn spawn_keypair_worker(bufsize: usize) -> mpsc::Receiver<KeyPair> {

Could you note the possible panic when bufsize=0 here in the docstring?

[hulthe]

Reviewed 1 of 1 files at r5, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 29 at r4 (raw file):

Previously, hulthe (Joakim Hulthe) wrote…

Could you note the possible panic when bufsize=0 here in the docstring?

Nice, thanks :D

[dlon]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @Serock3)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 74 at r4 (raw file):

/// The first call will spawn the worker which immedietly starts to compute and buffer [`BUFSIZE`]
/// of key pairs.
pub fn get_or_init_keypair_receiver<'a>() -> &'a Mutex<mpsc::Receiver<KeyPair>> {

More bikeshedding: The receiver/return value does not have to be public

[Serock3]

Reviewable status: 1 of 3 files reviewed, all discussions resolved (waiting on @hulthe)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 74 at r4 (raw file):

Previously, dlon (David Lönnhager) wrote…

More bikeshedding: The receiver/return value does not have to be public

Done

[dlon]

Reviewed 2 of 2 files at r7, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 35 at r7 (raw file):

///
/// Panics if the buffer capacity is 0.
pub fn spawn_keypair_worker(bufsize: usize) -> mpsc::Receiver<KeyPair> {

Does this need to be pub?

[Serock3]

Reviewable status: 2 of 3 files reviewed, all discussions resolved (waiting on @dlon)

---

talpid-tunnel-config-client/src/classic_mceliece.rs line 35 at r7 (raw file):

Previously, dlon (David Lönnhager) wrote…

Does this need to be pub?

Suppose not

[dlon]

Reviewed 1 of 1 files at r8, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved