Jędrzej Stuczyński
d9d549fd0f
* Changed identity keypair to use ed25519 * Encryption key is now x25519 based + compatibiltiy with sphinx * Pathing and import fixes * Moved all asymmetric keys to sub-module in crypto * Extracted aes to separate module * kdf module in crypto * Ability to perform diffie hellman on encryption keys * ecdsa on identity keys * Extremely rough and incomplete registration handshake * Authentication primitives * Creating new random authenticationIV * Wrapper type for the derived shared key * Removed AuthToken in favour of using SharedKey for authentication * Gateway identity keys * Registration handshake without error mapping * Gateway address in client config * Added extra key for gateway presence * Updated pemstore to work on borrows instead * Gateway client trying to perform the handshake * Gateway changes to allow for handshake and shared key * Debug trait on sharedkey * native client using updated gateway client * Slightly updated gateway API * Minor cleanup * Fixed pemstore to correctly save multiple keypairs * Gateway actually deriving shared key during handshake * Gateway sending correct mid-handshake message * Missing quotation mark in client config template * Fixed template for correct shared key serialization * Fixed gateway authentication * Fixed tests * Using correct gateway key when converting to sphinx node * "get_all_clients" takes them from gateways as opposed to providers now * cargo fmt * Renamed pemstore methods * Unused import * Encryption of forward requests between client and gateway * Updated sphinx dependency to use public revision * Sending 'error' on handshake processing error * Removed some dead code * Dead code I forgot to remove before * Extracted AckAes128Key into a struct * Slight pemstore revamp allowing for symmetric key store * ibid. * PemStorableKey for SharedKey * Introduced single location responsible for key management for client * WIP * Sphinx version update * Stop using NodeAddressBytes for two distinct and confusing purposes * Abstracting away SocketAddr from sphinx forwarding * Passing the bool for reply surbs * Attack plan for replies + encryption * Comment + removed variable binding * ReplySURB usage * Topology import in nymsphinx * Sphinx version update * Changed 'Recipient' to contain client's encryption key * Message preparation taking shape! * reply surb also containing the encryption key * Very initial message receiver * Sphinx version update * A possibly working way of receiving surbs * Fixed incorrect field name in client config template * camel casing all request arguments * Renamed and moved `MessageMode` to more appropriate file * Restored reconstruction tests * Removed dead code from chunking * Made rust examples compilable * reply SURB key storage * Replies as an InputMessage * Forgotten commented code * No retransmission processing for cover or replies * Received reply processing * Renamed client pathfinder to something more appropriate * Made HasherOutputSize public * Added key store path to config * Reply surb attaching key digest when used * Changes due to previous renaming * Removed comment * Fixed insert_encryption_key * Assigning initial value of key store path * Computing key digest with correct algorithm * Initial and presumably temporary request serialization * hacky way of introducing 'FragmentIdentifier' for replies * Moved responsibility of reply encryption, padding, etc, to message preparer * Optional recipient in try_get_valid_topology_ref * Handling new reply surbs with acks and padding * Updated go and python examples to include replies in text and binary cases * Updated rust examples + binaryserverresponse * Helpers in rust examples * And updated JS example * Moved shared key generation function to crypto crate * Cover traffic encryption! * hmac computation in crypto * Updated aes imports due to new dependencies * hkdf made more generic * crypto cleanup + algorithms in params * Clippy cleanup pass * Generating encryption+mac shared keys between client and gateway * MACs attached to forward requests to gateway * Gateway messages encrypted and mac'd * Lowered logging level * compiler warning cleanup * Some minor cleanup * Generic stream cipher * Generic shared key derivation + algorithm definitions * Project-wide AES clean-up * Comment fix * Removed commented imports * Updated comments * Fixed topology test fixture
69 lines
2.7 KiB
Rust
69 lines
2.7 KiB
Rust
// Copyright 2020 Nym Technologies SA
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
|
|
use crate::asymmetric::encryption;
|
|
use crate::hkdf;
|
|
use digest::{BlockInput, FixedOutput, Reset, Update};
|
|
use generic_array::{typenum::Unsigned, ArrayLength};
|
|
use rand::{CryptoRng, RngCore};
|
|
use stream_cipher::{Key, NewStreamCipher, SyncStreamCipher};
|
|
|
|
/// Generate an ephemeral encryption keypair and perform diffie-hellman to establish
|
|
/// shared key with the remote.
|
|
pub fn new_ephemeral_shared_key<C, D, R>(
|
|
rng: &mut R,
|
|
remote_key: &encryption::PublicKey,
|
|
) -> (encryption::KeyPair, Key<C>)
|
|
where
|
|
C: SyncStreamCipher + NewStreamCipher,
|
|
D: Update + BlockInput + FixedOutput + Reset + Default + Clone,
|
|
D::BlockSize: ArrayLength<u8>,
|
|
D::OutputSize: ArrayLength<u8>,
|
|
R: RngCore + CryptoRng,
|
|
{
|
|
let ephemeral_keypair = encryption::KeyPair::new_with_rng(rng);
|
|
|
|
// after performing diffie-hellman we don't care about the private component anymore
|
|
let dh_result = ephemeral_keypair.private_key().diffie_hellman(remote_key);
|
|
|
|
// there is no reason for this to fail as our okm is expected to be only C::KeySize bytes
|
|
let okm = hkdf::extract_then_expand::<D>(None, &dh_result, None, C::KeySize::to_usize())
|
|
.expect("somehow too long okm was provided");
|
|
|
|
let derived_shared_key =
|
|
Key::<C>::from_exact_iter(okm).expect("okm was expanded to incorrect length!");
|
|
|
|
(ephemeral_keypair, derived_shared_key)
|
|
}
|
|
|
|
/// Recompute shared key using remote public key and local private key.
|
|
pub fn recompute_shared_key<C, D>(
|
|
remote_key: &encryption::PublicKey,
|
|
local_key: &encryption::PrivateKey,
|
|
) -> Key<C>
|
|
where
|
|
C: SyncStreamCipher + NewStreamCipher,
|
|
D: Update + BlockInput + FixedOutput + Reset + Default + Clone,
|
|
D::BlockSize: ArrayLength<u8>,
|
|
D::OutputSize: ArrayLength<u8>,
|
|
{
|
|
let dh_result = local_key.diffie_hellman(remote_key);
|
|
|
|
// there is no reason for this to fail as our okm is expected to be only C::KeySize bytes
|
|
let okm = hkdf::extract_then_expand::<D>(None, &dh_result, None, C::KeySize::to_usize())
|
|
.expect("somehow too long okm was provided");
|
|
|
|
Key::<C>::from_exact_iter(okm).expect("okm was expanded to incorrect length!")
|
|
}
|