Lanes used to be separate HostInfos, each with its own handshake, its own
half-established states, its own lifetime and its own slot bookkeeping. That
bought nothing: a lane is the same tunnel over a different underlay 5-tuple.
Derive lane sessions instead. Noise leaves us with A.eKey == B.dKey, so both
sides HKDF-expand the same two keys with the same per-lane label and land on a
matched pair without exchanging anything. Lanes now cost no handshake, have no
half-established state, and die exactly when their base tunnel does. Which lane
a packet belongs to rides the low byte of the nebula header's Reserved field,
inside the AEAD's associated data.
Receiving on a lane needs no permission, since the session exists the moment the
base handshake completes. Sending on one needs proof the new 5-tuple works, so a
lane stays down until a probe on it is acked and falls back to the base tunnel
the moment it stops being acked. Probing is demand-driven off the connection
manager's per-tunnel traffic tick: a peer we exchange a trickle with never costs
more than its base tunnel, however many lanes are configured. The ack rides the
base session on purpose, so a broken reverse lane can't fail a working one.
Because the data now rides lane counters, the rehandshake, exhaustion and
swap-primary checks take the max counter across the base session and its lanes;
otherwise the base counter would sit near zero while a lane ran its keys past
the nonce ceiling.
Removes OutboundLaneTimer, EnsureLanes, startLaneHandshake, completeLane,
completeLaneResponder, makeLaneTrafficDecision and the lane fields on HostInfo.
Handshake payload field 3 (the per-lane handshake index) is permanently
reserved; peers advertise a TxLanes count instead.
Add support for the "fips140" mode of Go:
- https://go.dev/doc/security/fips140
- https://csrc.nist.gov/projects/cryptographic-module-validation-program/certificate/5247
You can build with `make fips140`, see the README changes for more info.
Some differences from the boringcrypto builds:
- We switch to using `go:linkname crypto/tls.aeadAESGCMTLS13`, which gives us the fips implementation for both `boringcrypto` and `fips140` modes. This means we also no longer need `-checklinkname=0`
- Go native `fips140` doesn't need CGO_ENABLED=1
- We decide if we should use the fips140 GCM at runtime, if `fips140.Enabled()` is true. If you use the `make release-fips140`, we build with build tag `fips140enforce` which ensures the binary is running with fips140 enabled and that only P256 / AES-GCM is being used. If you don't want this enforce mode, you can build without the build tag.
* add PKCS11 support
* add pkcs11 build option to the makefile, add a stub pkclient to avoid forcing CGO onto people
* don't print the pkcs11 option on nebula-cert keygen if not compiled in
* remove linux-arm64-pkcs11 from the all target to fix CI
* correctly serialize ec keys
* nebula-cert: support PKCS#11 for sign and ca
* fix gofmt lint
* clean up some logic with regard to closing sessions
* pkclient: handle empty correctly for TPM2
* Update Makefile and Actions
---------
Co-authored-by: Morgan Jones <me@numin.it>
Co-authored-by: John Maguire <contact@johnmaguire.me>
* add test for GOEXPERIMENT=boringcrypto
* fix NebulaCertificate.Sign
Set the PublicKey field in a more compatible way for the tests. The
current method grabs the public key from the certificate, but the
correct thing to do is to derive it from the private key. Either way
doesn't really matter as I don't think the Sign method actually even
uses the PublicKey field.
* assert boring
* cleanup tests
* Support NIST curve P256
This change adds support for NIST curve P256. When you use `nebula-cert ca`
or `nebula-cert keygen`, you can specify `-curve P256` to enable it. The
curve to use is based on the curve defined in your CA certificate.
Internally, we use ECDSA P256 to sign certificates, and ECDH P256 to do
Noise handshakes. P256 is not supported natively in Noise Protocol, so
we define `DHP256` in the `noiseutil` package to implement support for
it.
You cannot have a mixed network of Curve25519 and P256 certificates,
since the Noise protocol will only attempt to parse using the Curve
defined in the host's certificate.
* verify the curves match in VerifyPrivateKey
This would have failed anyways once we tried to actually use the bytes
in the private key, but its better to detect the issue up front with
a better error message.
* add cert.Curve argument to Sign method
* fix mismerge
* use crypto/ecdh
This is the preferred method for doing ECDH functions now, and also has
a boringcrypto specific codepath.
* remove other ecdh uses of crypto/elliptic
use crypto/ecdh instead
Wade Simmons