But to be direct, I'm lacking also a bit the free time to spend on it.
That's my main issue as well :(
Regarding the peer and locks. Seems like this is the only situation where it might be invalidated under normal operation? I guess that is the reason calling Rcv_Process after the big switch without a lock (code in master) works so well. I assume most users use a static setup.
if(p->is_dynamic && config->drop_unknown_peers) {
remove_peer(p);
free_peer(p, 1);
break;
}
As a quick fix/work around, would you prefer that approach for Snd_Message too over the proposed change in this PR?
To await doing that until after out of the critical zone described earlier.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.
mtryfoss left a comment