Date:   Wed, 27 Oct 2021 08:37:18 -0600
From:   David Ahern <dsahern@...il.com>
To:     Taehee Yoo <ap420073@...il.com>, davem@...emloft.net,
        kuba@...nel.org, dsahern@...nel.org, netdev@...r.kernel.org
Cc:     dkirjanov@...e.de
Subject: Re: [PATCH net-next 1/4 v4] amt: add control plane of amt interface

On 10/26/21 9:10 AM, Taehee Yoo wrote:
> diff --git a/drivers/net/amt.c b/drivers/net/amt.c
> new file mode 100644
> index 000000000000..8d4782c66cde
> --- /dev/null
> +++ b/drivers/net/amt.c
> @@ -0,0 +1,487 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/* Copyright (c) 2021 Taehee Yoo <ap420073@...il.com> */
> +
> +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
> +
> +#include <linux/module.h>
> +#include <linux/skbuff.h>
> +#include <linux/udp.h>
> +#include <linux/jhash.h>
> +#include <linux/if_tunnel.h>
> +#include <linux/net.h>
> +#include <linux/igmp.h>
> +#include <net/net_namespace.h>
> +#include <net/protocol.h>
> +#include <net/ip.h>
> +#include <net/udp.h>
> +#include <net/udp_tunnel.h>
> +#include <net/icmp.h>
> +#include <net/mld.h>
> +#include <net/amt.h>
> +#include <uapi/linux/amt.h>
> +#include <linux/security.h>
> +#include <net/gro_cells.h>
> +#include <net/ipv6.h>
> +#include <net/protocol.h>
> +#include <net/if_inet6.h>
> +#include <net/ndisc.h>
> +#include <net/addrconf.h>
> +#include <net/ip6_route.h>
> +#include <net/inet_common.h>
> +
> +static struct workqueue_struct *amt_wq;
> +
> +static struct socket *amt_create_sock(struct net *net, __be16 port)
> +{
> +	struct udp_port_cfg udp_conf;
> +	struct socket *sock;
> +	int err;
> +
> +	memset(&udp_conf, 0, sizeof(udp_conf));
> +	udp_conf.family = AF_INET;
> +	udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
> +
> +	udp_conf.local_udp_port = port;
> +
> +	err = udp_sock_create(net, &udp_conf, &sock);
> +	if (err < 0)
> +		return ERR_PTR(err);
> +
> +	return sock;
> +}
> +
> +static int amt_socket_create(struct amt_dev *amt)
> +{
> +	struct udp_tunnel_sock_cfg tunnel_cfg;
> +	struct socket *sock;
> +
> +	sock = amt_create_sock(amt->net, amt->relay_port);
> +	if (IS_ERR(sock))
> +		return PTR_ERR(sock);
> +
> +	/* Mark socket as an encapsulation socket */
> +	memset(&tunnel_cfg, 0, sizeof(tunnel_cfg));
> +	tunnel_cfg.sk_user_data = amt;
> +	tunnel_cfg.encap_type = 1;
> +	tunnel_cfg.encap_destroy = NULL;
> +	setup_udp_tunnel_sock(amt->net, sock, &tunnel_cfg);
> +
> +	rcu_assign_pointer(amt->sock, sock);
> +	return 0;
> +}
> +
> +static int amt_dev_open(struct net_device *dev)
> +{
> +	struct amt_dev *amt = netdev_priv(dev);
> +	int err;
> +
> +	amt->ready4 = false;
> +	amt->ready6 = false;
> +
> +	err = amt_socket_create(amt);
> +	if (err)
> +		return err;
> +
> +	spin_lock_bh(&amt->lock);
> +	amt->req_cnt = 0;
> +	get_random_bytes(&amt->key, sizeof(siphash_key_t));
> +	spin_unlock_bh(&amt->lock);

why the amt dev lock here? dev_open is called with rtnl lock held and
the device will not be receiving packets yet (the _bh).

> +
> +	amt->status = AMT_STATUS_INIT;
> +	return err;
> +}
> +

> +
> +static int amt_change_mtu(struct net_device *dev, int new_mtu)
> +{
> +	if (new_mtu > dev->max_mtu)
> +		new_mtu = dev->max_mtu;
> +	else if (new_mtu < dev->min_mtu)
> +		new_mtu = dev->min_mtu;

that is handled by dev_validate_mtu.

Since you are not doing anything special here, I believe you do not need
the ndo_change_mtu at all.

> +
> +	dev->mtu = new_mtu;
> +	return 0;
> +}
> +
> +static const struct net_device_ops amt_netdev_ops = {
> +	.ndo_init               = amt_dev_init,
> +	.ndo_uninit             = amt_dev_uninit,
> +	.ndo_open		= amt_dev_open,
> +	.ndo_stop		= amt_dev_stop,
> +	.ndo_get_stats64        = dev_get_tstats64,
> +	.ndo_change_mtu         = amt_change_mtu,
> +};
> +
> +static void amt_link_setup(struct net_device *dev)
> +{
> +	dev->netdev_ops         = &amt_netdev_ops;
> +	dev->needs_free_netdev  = true;
> +	SET_NETDEV_DEVTYPE(dev, &amt_type);
> +	dev->min_mtu		= ETH_MIN_MTU;
> +	dev->max_mtu		= ETH_MAX_MTU;
> +	dev->type		= ARPHRD_NONE;
> +	dev->flags		= IFF_POINTOPOINT | IFF_NOARP | IFF_MULTICAST;
> +	dev->hard_header_len	= 0;
> +	dev->addr_len		= 0;
> +	dev->priv_flags		|= IFF_NO_QUEUE;
> +	dev->features		|= NETIF_F_LLTX;
> +	dev->features		|= NETIF_F_GSO_SOFTWARE;
> +	dev->features		|= NETIF_F_NETNS_LOCAL;
> +	dev->hw_features	|= NETIF_F_SG | NETIF_F_HW_CSUM;
> +	dev->hw_features	|= NETIF_F_FRAGLIST | NETIF_F_RXCSUM;
> +	dev->hw_features	|= NETIF_F_GSO_SOFTWARE;
> +	eth_hw_addr_random(dev);
> +	eth_zero_addr(dev->broadcast);
> +	ether_setup(dev);
> +}
> +
> +static const struct nla_policy amt_policy[IFLA_AMT_MAX + 1] = {
> +	[IFLA_AMT_MODE]		= { .type = NLA_U32 },
> +	[IFLA_AMT_RELAY_PORT]	= { .type = NLA_U16 },
> +	[IFLA_AMT_GATEWAY_PORT]	= { .type = NLA_U16 },
> +	[IFLA_AMT_LINK]		= { .type = NLA_U32 },
> +	[IFLA_AMT_LOCAL_IP]	= { .len = sizeof_field(struct iphdr, daddr) },
> +	[IFLA_AMT_REMOTE_IP]	= { .len = sizeof_field(struct iphdr, daddr) },
> +	[IFLA_AMT_DISCOVERY_IP]	= { .len = sizeof_field(struct iphdr, daddr) },
> +	[IFLA_AMT_MAX_TUNNELS]	= { .type = NLA_U32 },
> +};
> +
> +static int amt_validate(struct nlattr *tb[], struct nlattr *data[],
> +			struct netlink_ext_ack *extack)
> +{
> +	if (!data)
> +		return -EINVAL;
> +
> +	if (!data[IFLA_AMT_LINK]) {
> +		NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_LINK],
> +				    "link interface should not be empty");

How about: "Link attribute is required".

Similar for the checks below.

> +		return -EINVAL;
> +	}
> +
> +	if (!data[IFLA_AMT_MODE] ||
> +	    nla_get_u32(data[IFLA_AMT_MODE]) > AMT_MODE_MAX) {
> +		NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_MODE],
> +				    "mode should not be empty");

For the extack message to make sense, you need separate checks here: one
that the attribute is set and one that its value is valid. I believe the
latter can be managed through the policy and netlink_range_validation.

> +		return -EINVAL;
> +	}
> +
> +	if (!data[IFLA_AMT_LOCAL_IP]) {
> +		NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_DISCOVERY_IP],
> +				    "local should not be empty");
> +		return -EINVAL;
> +	}
> +
> +	if (!data[IFLA_AMT_DISCOVERY_IP] &&
> +	    nla_get_u32(data[IFLA_AMT_MODE]) == AMT_MODE_GATEWAY) {
> +		NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_LOCAL_IP],
> +				    "discovery should not be empty");
> +		return -EINVAL;
> +	}
> +
> +	return 0;
> +}
> +
> +static int amt_newlink(struct net *net, struct net_device *dev,
> +		       struct nlattr *tb[], struct nlattr *data[],
> +		       struct netlink_ext_ack *extack)
> +{
> +	struct amt_dev *amt = netdev_priv(dev);
> +	int err;
> +
> +	amt->net = net;
> +	amt->mode = nla_get_u32(data[IFLA_AMT_MODE]);
> +
> +	if (data[IFLA_AMT_MAX_TUNNELS])
> +		amt->max_tunnels = nla_get_u32(data[IFLA_AMT_MAX_TUNNELS]);
> +	else
> +		amt->max_tunnels = AMT_MAX_TUNNELS;
> +
> +	spin_lock_init(&amt->lock);
> +	amt->max_groups = AMT_MAX_GROUP;
> +	amt->max_sources = AMT_MAX_SOURCE;
> +	amt->hash_buckets = AMT_HSIZE;
> +	amt->nr_tunnels = 0;
> +	get_random_bytes(&amt->hash_seed, sizeof(amt->hash_seed));
> +	amt->stream_dev = dev_get_by_index(net,
> +					   nla_get_u32(data[IFLA_AMT_LINK]));
> +	if (!amt->stream_dev) {
> +		NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_LINK],
> +				    "Can't find stream device");
> +		return -ENODEV;
> +	}
> +
> +	if (amt->stream_dev->type != ARPHRD_ETHER) {
> +		NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_LINK],
> +				    "Invalid stream device type");
> +		dev_put(amt->stream_dev);
> +		return -EINVAL;
> +	}
> +
> +	amt->local_ip = nla_get_in_addr(data[IFLA_AMT_LOCAL_IP]);

Any sanity checks needed for the local_ip? broadcast, multicast, local
ip is assigned locally.

> +	if (data[IFLA_AMT_RELAY_PORT])
> +		amt->relay_port = nla_get_be16(data[IFLA_AMT_RELAY_PORT]);
> +	else
> +		amt->relay_port = htons(IANA_AMT_UDP_PORT);
> +
> +	if (data[IFLA_AMT_GATEWAY_PORT])
> +		amt->gw_port = nla_get_be16(data[IFLA_AMT_GATEWAY_PORT]);
> +	else
> +		amt->gw_port = htons(IANA_AMT_UDP_PORT);
> +
> +	if (!amt->relay_port) {
> +		NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
> +				    "relay port must not be 0");
> +		return -EINVAL;
> +	}
> +	if (amt->mode == AMT_MODE_RELAY) {
> +		amt->qrv = amt->net->ipv4.sysctl_igmp_qrv;
> +		amt->qri = 10;
> +		dev->needed_headroom = amt->stream_dev->needed_headroom +
> +				       AMT_RELAY_HLEN;
> +		dev->mtu = amt->stream_dev->mtu - AMT_RELAY_HLEN;
> +		dev->max_mtu = dev->mtu;
> +		dev->min_mtu = ETH_MIN_MTU + AMT_RELAY_HLEN;
> +	} else {
> +		if (!data[IFLA_AMT_DISCOVERY_IP]) {
> +			NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
> +					    "discovery must be set in gateway mode");
> +			return -EINVAL;
> +		}
> +		if (!amt->gw_port) {
> +			NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
> +					    "gateway port must not be 0");
> +			return -EINVAL;
> +		}
> +		amt->remote_ip = 0;
> +		amt->discovery_ip = nla_get_in_addr(data[IFLA_AMT_DISCOVERY_IP]);
> +		if (ipv4_is_loopback(amt->discovery_ip) ||
> +		    ipv4_is_multicast(amt->discovery_ip)) {
> +			NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
> +					    "discovery must be unicast");
> +			return -EINVAL;
> +		}
> +
> +		dev->needed_headroom = amt->stream_dev->needed_headroom +
> +				       AMT_GW_HLEN;
> +		dev->mtu = amt->stream_dev->mtu - AMT_GW_HLEN;
> +		dev->max_mtu = dev->mtu;
> +		dev->min_mtu = ETH_MIN_MTU + AMT_GW_HLEN;
> +	}
> +	amt->qi = AMT_INIT_QUERY_INTERVAL;
> +
> +	err = register_netdevice(dev);
> +	if (err < 0) {
> +		netdev_dbg(dev, "failed to register new netdev %d\n", err);
> +		dev_put(amt->stream_dev);
> +		return err;
> +	}
> +
> +	err = netdev_upper_dev_link(amt->stream_dev, dev, extack);
> +	if (err < 0) {
> +		dev_put(amt->stream_dev);
> +		unregister_netdevice(dev);
> +		return err;
> +	}
> +
> +	return 0;
> +}
> +

> diff --git a/include/uapi/linux/amt.h b/include/uapi/linux/amt.h
> new file mode 100644
> index 000000000000..641ef7f51253
> --- /dev/null
> +++ b/include/uapi/linux/amt.h
> @@ -0,0 +1,31 @@
> +/* SPDX-License-Identifier: GPL-2.0-only WITH Linux-syscall-note */
> +/*
> + * Copyright (c) 2021 Taehee Yoo <ap420073@...il.com>
> + */
> +#ifndef _UAPI_AMT_H_
> +#define _UAPI_AMT_H_
> +
> +enum ifla_amt_mode {
> +	AMT_MODE_GATEWAY = 0,
> +	AMT_MODE_RELAY,
> +	__AMT_MODE_MAX,
> +};
> +
> +#define AMT_MODE_MAX (__AMT_MODE_MAX - 1)
> +
> +enum {
> +	IFLA_AMT_UNSPEC,
> +	IFLA_AMT_MODE,
> +	IFLA_AMT_RELAY_PORT,
> +	IFLA_AMT_GATEWAY_PORT,
> +	IFLA_AMT_LINK,
> +	IFLA_AMT_LOCAL_IP,
> +	IFLA_AMT_REMOTE_IP,
> +	IFLA_AMT_DISCOVERY_IP,
> +	IFLA_AMT_MAX_TUNNELS,
> +	__IFLA_AMT_MAX,
> +};
> +
> +#define IFLA_AMT_MAX (__IFLA_AMT_MAX - 1)
> +
> +#endif /* _UAPI_AMT_H_ */
> 

Document each attribute type. Application developer should be able to
read this file and properly use the API.

Powered by blists - more mailing lists