Mathew Heard

@iczero Have you thought about aproaching this in a way that could be used for general purpose routing? You might be able to kill a few birds with one PR...

@gsliepen what would be too much traffic in your opinion? Perhaps: Each node sends a packet to each of it's directly connected nodes with the `[node, rtt]` of each subnet...

> We want to limit background traffic as much as possible. We need some keepalive packets for the meta-connections, and we have on-demand UDP probe packets (ie, probe packets are...

Can also confirm segfaults on Windows too.

The main thing is multihop routing e.g `Node A -> Relay 1 -> Node B` with wireguard without going into userspace. This is difficult because a wireguard tunnel is done...

@nh2 we too are facing this issue on our DigitalOcean cloud VMs. We are running the older 1.0 branch (for stability reasons) and currently seeing only rx: 23.56 Mbit/s (3257...

@nh2 isn't IFF_MULTI_READ for the read side of tun not the write. Am I misunderstanding that patch?

@dechamps I can confirm that on machines with more complicated routing rules that outgoing performance on tinc does decrease. I haven't got any benchmarks currently but it was something our...

Is this recent, perhaps related to #302?

We don't use the USB interface. I'd say check your configuration though. Likely you have a missconfiguration (preprocessor / global).