device: fix TOCTOU race during session state update (#77)

API introduced in a927a66e has two cases of state determination
happening out of critical section for the state value:

(1) expiredSession loads sessionExpiresNano, then releases all locks
and calls noteSessionState(Expired). So a concurrent refresh that
lands in that gap gets clobbered by a stale Expired -- and sticks
until the next re-key.

(2) Likewise in noteSessionHandshakeStopped, hasKeyMaterial check
happens out of the session state lock and races with ZeroAndFlushAll.

Both lead to a wrong state emitted via the device.sessionState.fn,
but are otherwise benign.

This moves the expiry timestamp under a lock to address the former,
and provides a noteSessionStateLocked helper for the latter.
Also changes API semantics to serialize events per-peer, to avoid
sharing a single lock for all timestamps.

Updates tailscale/corp#42874

Signed-off-by: Alex Valiushko <alexvaliushko@tailscale.com>
Change-Id: Iee2cdf135375519e58a8e84362349d966a6a6964
This commit is contained in:
Alex Valiushko 2026-06-24 14:09:25 -07:00 committed by 世界
parent 35a60acb84
commit 15b912c1c0
No known key found for this signature in database
GPG key ID: CD109927C34A63C4
3 changed files with 61 additions and 47 deletions

View file

@ -141,12 +141,13 @@ func expiredZeroKeyMaterial(peer *Peer) {
}
func expiredSession(peer *Peer) {
expires := peer.sessionExpiresNano.Load()
if expires == 0 || time.Now().UnixNano() < expires {
peer.sessionState.Lock()
defer peer.sessionState.Unlock()
if peer.sessionState.sessionExpires.IsZero() || time.Now().Before(peer.sessionState.sessionExpires) {
return
}
peer.device.log.Verbosef("%s - Session expired after %d seconds", peer, int(RejectAfterTime.Seconds()))
peer.noteSessionState(PeerSessionExpired)
peer.noteSessionStateLocked(PeerSessionExpired)
}
func expiredPersistentKeepalive(peer *Peer) {
@ -208,11 +209,15 @@ func (peer *Peer) timersHandshakeComplete() {
/* Should be called after an ephemeral key is created, which is before sending a handshake response or after receiving a handshake response. */
func (peer *Peer) timersSessionDerived() {
if peer.timersActive() {
peer.sessionExpiresNano.Store(time.Now().Add(RejectAfterTime).UnixNano())
peer.sessionState.Lock()
peer.sessionState.sessionExpires = time.Now().Add(RejectAfterTime)
peer.noteSessionStateLocked(PeerSessionEstablished)
peer.sessionState.Unlock()
peer.timers.sessionExpired.Mod(RejectAfterTime)
peer.timers.zeroKeyMaterial.Mod(RejectAfterTime * 3)
} else {
peer.noteSessionState(PeerSessionEstablished)
}
peer.noteSessionState(PeerSessionEstablished)
}
/* Should be called before a packet with authentication -- keepalive, data, or handshake -- is sent, or after one is received. */