This scanning thread either polls libusb or checks every 500ms for a
change depending on host capabilities. The GC Adapter can now be plugged
and unplugged at any time when dolphin is open, it will be used if the
direct connect option is set.
With some specific, STUN-hostile routers, the netplay client can get
stuck forever while trying to connect to the stun server. This adds a
5 seconds (much more than should be necessary if it works) timer until
a failure is registered and the attempt stops.
Otherwise, it would work but any async sending would be delayed by 4ms or
wait until the next packet was received.
Also increase the client timeout to 250ms, since enet_host_service is now
really interrupted.
Add std::unique_ptr<sf::Packet> objects to a queue instead of functions,
makes things easier to read, and avoids headaches while checking the
lifetime of the concerned objects.
If we successfully detach the kernel driver from the interface, we
should continue instead of aborting the setup.
And we should not use libusb_handle_events(), as the API says it is only
for backwards compatibility. Additionally, if the adapter thread is not
active, dolphin will take 60 seconds to close because the
libusb_handle_events() timeout is hardcoded to 60 seconds. Instead, use
libusb_handle_events_timeout_completed() with a timeout of 1 second.
Also, cancel the libusb transfers before the join(), to be able to close
the usb device without libusb screaming in the background (and
potentially crashing).
And finally, split the Init() and Shutdown() functions to avoid having
to init and exit libusb every time we neeed to detect the adapter.
mathieui