Attention is currently required from: fixeria.
1 comment:
Patchset:
I think we can simply try to set a large UDP send buffer and ask the kernel to do any buffering by switcing to non-blocking mode and discarding anything when we hit the send buffer limit (just like we would with the write-queue limit). Using osmo_wqueue with datagram sockets doesn't really make all that much sense to me, as it just adds overhead to the program for something the kernel could do just as well (and kernel overhead can be done on other CPU cores).
To view, visit change 41685. To unsubscribe, or for help writing mail filters, visit settings.