resilien
/
nebula
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

fix regression with LightHouseHandler and punchBack (#346) 

The change introduced by #320 incorrectly re-uses the output buffer for
sending punchBack packets. Since we are currently spawning a new
goroutine for each send here, we need to allocate a new buffer each
time. We can come back and optimize this in the future, but for now we
should fix the regression.
This commit is contained in:
Wade Simmons 2020-11-25 17:49:26 -05:00 committed by GitHub
parent ee7c27093c
commit ce9ad37431
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
lighthouse.go
View File

@ -447,7 +447,10 @@ func (lhh *LightHouseHandler) HandleRequest(rAddr *udpAddr, vpnIp uint32, p []by
go func() {
time.Sleep(time.Second * 5)
l.Debugf("Sending a nebula test packet to vpn ip %s", IntIp(n.Details.VpnIp))
f.SendMessageToVpnIp(test, testRequest, n.Details.VpnIp, []byte(""), lhh.nb, lhh.out[:0])
// TODO we have to allocate a new output buffer here since we are spawning a new goroutine
// for each punchBack packet. We should move this into a timerwheel or a single goroutine
// managed by a channel.
f.SendMessageToVpnIp(test, testRequest, n.Details.VpnIp, []byte(""), make([]byte, 12, 12), make([]byte, mtu))
}()
}
}