On Tue, Nov 19, 2019 at 12:01:19PM +0100, Stefano Garzarella wrote:

Ideas for long-term changes below.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

> diff --git a/MAINTAINERS b/MAINTAINERS
> index 760049454a23..c2a3dc3113ba 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -17239,6 +17239,7 @@ F:	net/vmw_vsock/diag.c
>  F:	net/vmw_vsock/af_vsock_tap.c
>  F:	net/vmw_vsock/virtio_transport_common.c
>  F:	net/vmw_vsock/virtio_transport.c
> +F:	net/vmw_vsock/vsock_loopback.c
>  F:	drivers/net/vsockmon.c
>  F:	drivers/vhost/vsock.c
>  F:	tools/testing/vsock/

At this point you are most active in virtio-vsock and I am reviewing
patches on a best-effort basis.  Feel free to add yourself as
maintainer.

> diff --git a/net/vmw_vsock/vsock_loopback.c b/net/vmw_vsock/vsock_loopback.c
> new file mode 100644
> index 000000000000..3d1c1a88305f
> --- /dev/null
> +++ b/net/vmw_vsock/vsock_loopback.c
> @@ -0,0 +1,217 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * loopback transport for vsock using virtio_transport_common APIs
> + *
> + * Copyright (C) 2013-2019 Red Hat, Inc.
> + * Author: Asias He <asias@redhat.com>
> + *         Stefan Hajnoczi <stefanha@redhat.com>
> + *         Stefano Garzarella <sgarzare@redhat.com>
> + *
> + */
> +#include <linux/spinlock.h>
> +#include <linux/module.h>
> +#include <linux/list.h>
> +#include <linux/virtio_vsock.h>

Is it time to rename the generic functionality in
virtio_transport_common.c?  This doesn't have anything to do with virtio
:).

> +
> +static struct workqueue_struct *vsock_loopback_workqueue;
> +static struct vsock_loopback *the_vsock_loopback;

the_vsock_loopback could be a static global variable (not a pointer) and
vsock_loopback_workqueue could also be included in the struct.

The RCU pointer is really a way to synchronize vsock_loopback_send_pkt()
and vsock_loopback_cancel_pkt() with module exit.  There is no other
reason for using a pointer.

It's cleaner to implement the synchronization once in af_vsock.c (or
virtio_transport_common.c) instead of making each transport do it.
Maybe try_module_get() and related APIs provide the necessary semantics
so that core vsock code can hold the transport module while it's being
used to send/cancel a packet.

> +MODULE_ALIAS_NETPROTO(PF_VSOCK);

Why does this module define the alias for PF_VSOCK?  Doesn't another
module already define this alias?