Consider using kqueue in hypervisor socket implementation

The core of hyperkit uses kqueue via mevent.c but pci_virtio_sock.c uses plain old select.

If hyperkit is used together with vpnkit in Docker for Mac and large numbers (> 1024) of connections are port forwarded then hyperkit becomes unable to process any more AF_VSOCK connections due to the accepted file descriptor being greater than FD_SETSIZE, see for example https://github.com/moby/hyperkit/blob/3ace9850121a2ef270e0309a3ff6c2f991357842/src/lib/pci_virtio_sock.c#L1364

This manifests as errors under load, for example

$ docker ps
Error response from daemon: Bad response from Docker engine

from https://github.com/docker/for-mac/issues/2841

This scalability limit could be removed by switching from select to kqueue (or poll) in pci_virtio_sock.c.

Start with src/lib/pci_virtio_sock.c around the select and accepted file-descriptor handling, then compare the event-management approach in mevent.c. Determine whether kqueue or poll is appropriate, preserve AF_VSOCK connection processing, and verify that descriptors above FD_SETSIZE can be handled under the reported connection load.