webnet/tailscale
1
0
Fork 0
Code Issues Pull Requests 1 Packages Projects Releases Wiki Activity

control/controlclient,ipn/ipnlocal,wgengine: avoid restarting wireguard when key is learned via tsmp (#19142)

Browse Source 
When disco keys are learned on a node that is connected to control and
has a mapSession, wgengine will see the key as having changed, and
assume that any existing connections will need to be reset.

For keys learned via TSMP, the connection should not be reset as that
key is learned via an active wireguard connection. If wgengine resets
that connetion, a 15s timeout will occur.

This change adds a map to track new keys coming in via TSMP, and removes
them from the list of keys that needs to trigger wireguard resets. This
is done with an interface chain from controlclient down via localBackend
to userspaceEngine via the watchdog.

Once a key has been actively used for preventing a wireguard reset, the
key is removed from the map.

If mapSession becomes a long lived process instead of being dependent on
having a connection to control. This interface chain can be removed, and
the event sequence from wrap->controlClient->userspaceEngine, can be
changed to wrap->userspaceEngine->controlClient as we know the map will
not be gunked up with stale TSMP entries.

Updates #12639

Signed-off-by: Claus Lensbøl <claus@tailscale.com>
This commit is contained in:
Claus Lensbøl
2026-03-30 14:26:08 -04:00
committed by GitHub
parent 99f8039101
commit bf467727fc
8 changed files with 323 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files
control/controlclient/direct.go
+20 -7
View File
@@ -228,6 +228,15 @@ type NetmapDeltaUpdater interface {
UpdateNetmapDelta([]netmap.NodeMutation) (ok bool)
}
// patchDiscoKeyer is an optional interface that can be implemented by an [Observer] to be
// notified about node disco keys received out-of-band from control, via
// existing connection state.
type patchDiscoKeyer interface {
// PatchDiscoKey reports to the receiver that the specified disco key
// for node was obtained out-of-band from control.
PatchDiscoKey(key.NodePublic, key.DiscoPublic)
}
var nextControlClientID atomic.Int64
// NewDirect returns a new Direct client.
@@ -367,7 +376,7 @@ func NewDirect(opts Options) (*Direct, error) {
// mapSession has gone away, we want to fall back to pushing the key
// further down the chain.
if err := c.streamingMapSession.updateDiscoForNode(
peer.ID(), update.Key, time.Now(), false); err == nil ||
peer.ID(), peer.Key(), update.Key, time.Now(), false); err == nil ||
!errors.Is(err, ErrChangeQueueClosed) {
return
}
@@ -377,10 +386,7 @@ func NewDirect(opts Options) (*Direct, error) {
// not have a mapSession (we are not connected to control) or because the
// mapSession queue has closed.
c.logf("controlclient direct: updating discoKey for %v via magicsock", update.Src)
discoKeyPub.Publish(events.PeerDiscoKeyUpdate{
Src: update.Src,
Key: update.Key,
})
discoKeyPub.Publish(events.PeerDiscoKeyUpdate(update))
})
return c, nil
@@ -859,8 +865,10 @@ func (c *Direct) PollNetMap(ctx context.Context, nu NetmapUpdater) error {
// update it observed. It is used by tests and [NetmapFromMapResponseForDebug].
// It will report only the first netmap seen.
type rememberLastNetmapUpdater struct {
last *netmap.NetworkMap
done chan any
last *netmap.NetworkMap
lastTSMPKey key.NodePublic
lastTSMPDisco key.DiscoPublic
done chan any
}
func (nu *rememberLastNetmapUpdater) UpdateFullNetmap(nm *netmap.NetworkMap) {
@@ -871,6 +879,11 @@ func (nu *rememberLastNetmapUpdater) UpdateFullNetmap(nm *netmap.NetworkMap) {
}
}
func (nu *rememberLastNetmapUpdater) PatchDiscoKey(key key.NodePublic, disco key.DiscoPublic) {
nu.lastTSMPKey = key
nu.lastTSMPDisco = disco
}
// FetchNetMapForTest fetches the netmap once.
func (c *Direct) FetchNetMapForTest(ctx context.Context) (*netmap.NetworkMap, error) {
var nu rememberLastNetmapUpdater