device: change Peer.endpoint locking to reduce contention

Access to Peer.endpoint was previously synchronized by Peer.RWMutex.
This has now moved to Peer.endpoint.Mutex. Peer.SendBuffers() is now the
sole caller of Endpoint.ClearSrc(), which is signaled via a new bool,
Peer.endpoint.clearSrcOnTx. Previous Callers of Endpoint.ClearSrc() now
set this bool, primarily via peer.markEndpointSrcForClearing().
Peer.SetEndpointFromPacket() clears Peer.endpoint.clearSrcOnTx when an
updated conn.Endpoint is stored. This maintains the same event order as
before, i.e. a conn.Endpoint received after peer.endpoint.clearSrcOnTx
is set, but before the next Peer.SendBuffers() call results in the
latest conn.Endpoint source being used for the next packet transmission.

These changes result in throughput improvements for single flow,
parallel (-P n) flow, and bidirectional (--bidir) flow iperf3 TCP/UDP
tests as measured on both Linux and Windows. Latency under load improves
especially for high throughput Linux scenarios. These improvements are
likely realized on all platforms to some degree, as the changes are not
platform-specific.

Co-authored-by: James Tucker <james@tailscale.com>
Signed-off-by: James Tucker <james@tailscale.com>
Signed-off-by: Jordan Whited <jordan@tailscale.com>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
This commit is contained in:
Jordan Whited
2023-11-20 16:49:06 -08:00
committed by Jason A. Donenfeld
parent d0bc03c707
commit 4ffa9c2032
6 changed files with 83 additions and 81 deletions

View File

@@ -17,17 +17,20 @@ import (
type Peer struct {
isRunning atomic.Bool
sync.RWMutex // Mostly protects endpoint, but is generally taken whenever we modify peer
keypairs Keypairs
handshake Handshake
device *Device
endpoint conn.Endpoint
stopping sync.WaitGroup // routines pending stop
txBytes atomic.Uint64 // bytes send to peer (endpoint)
rxBytes atomic.Uint64 // bytes received from peer
lastHandshakeNano atomic.Int64 // nano seconds since epoch
disableRoaming bool
endpoint struct {
sync.Mutex
val conn.Endpoint
clearSrcOnTx bool // signal to val.ClearSrc() prior to next packet transmission
disableRoaming bool
}
timers struct {
retransmitHandshake *Timer
@@ -74,8 +77,6 @@ func (device *Device) NewPeer(pk NoisePublicKey) (*Peer, error) {
// create peer
peer := new(Peer)
peer.Lock()
defer peer.Unlock()
peer.cookieGenerator.Init(pk)
peer.device = device
@@ -97,7 +98,11 @@ func (device *Device) NewPeer(pk NoisePublicKey) (*Peer, error) {
handshake.mutex.Unlock()
// reset endpoint
peer.endpoint = nil
peer.endpoint.Lock()
peer.endpoint.val = nil
peer.endpoint.disableRoaming = false
peer.endpoint.clearSrcOnTx = false
peer.endpoint.Unlock()
// init timers
peer.timersInit()
@@ -116,14 +121,19 @@ func (peer *Peer) SendBuffers(buffers [][]byte) error {
return nil
}
peer.RLock()
defer peer.RUnlock()
if peer.endpoint == nil {
peer.endpoint.Lock()
endpoint := peer.endpoint.val
if endpoint == nil {
peer.endpoint.Unlock()
return errors.New("no known endpoint for peer")
}
if peer.endpoint.clearSrcOnTx {
endpoint.ClearSrc()
peer.endpoint.clearSrcOnTx = false
}
peer.endpoint.Unlock()
err := peer.device.net.bind.Send(buffers, peer.endpoint)
err := peer.device.net.bind.Send(buffers, endpoint)
if err == nil {
var totalLen uint64
for _, b := range buffers {
@@ -267,10 +277,20 @@ func (peer *Peer) Stop() {
}
func (peer *Peer) SetEndpointFromPacket(endpoint conn.Endpoint) {
if peer.disableRoaming {
peer.endpoint.Lock()
defer peer.endpoint.Unlock()
if peer.endpoint.disableRoaming {
return
}
peer.Lock()
peer.endpoint = endpoint
peer.Unlock()
peer.endpoint.clearSrcOnTx = false
peer.endpoint.val = endpoint
}
func (peer *Peer) markEndpointSrcForClearing() {
peer.endpoint.Lock()
defer peer.endpoint.Unlock()
if peer.endpoint.val == nil {
return
}
peer.endpoint.clearSrcOnTx = true
}