Currently when we detect an EOF on a read from the underlying host socket that we are proxying to and from we treat this as if the TCP socket has closed. Under these conditions we are sending a vsock packet that tells the guest we have shut down our read and write halves via VIRTIO_VSOCK_SHUTDOWN_F_SEND | VIRTIO_VSOCK_SHUTDOWN_F_RECEIVE.
It would be nice to be able to distinguish a close of the read and write halves independently.
Currently when we detect an
EOFon a read from the underlying host socket that we are proxying to and from we treat this as if the TCP socket has closed. Under these conditions we are sending a vsock packet that tells the guest we have shut down our read and write halves viaVIRTIO_VSOCK_SHUTDOWN_F_SEND | VIRTIO_VSOCK_SHUTDOWN_F_RECEIVE.It would be nice to be able to distinguish a close of the read and write halves independently.