can: add Virtual CAN Tunnel driver (vxcan)
authorOliver Hartkopp <socketcan@hartkopp.net>
Tue, 25 Apr 2017 06:19:44 +0000 (08:19 +0200)
committerMarc Kleine-Budde <mkl@pengutronix.de>
Tue, 25 Apr 2017 07:04:30 +0000 (09:04 +0200)
Similar to the virtual ethernet driver veth, vxcan implements a
local CAN traffic tunnel between two virtual CAN network devices.
See Kconfig entry for details.

Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
drivers/net/can/Kconfig
drivers/net/can/Makefile
drivers/net/can/vxcan.c [new file with mode: 0644]
include/uapi/linux/can/vxcan.h [new file with mode: 0644]

index aa20b69d2a1ad319b0ccafb2fc4d18289ec014f7..ac4ff394bc569273403cd7219dd2d06c62f2f371 100644 (file)
@@ -9,6 +9,24 @@ config CAN_VCAN
          This driver can also be built as a module.  If so, the module
          will be called vcan.
 
+config CAN_VXCAN
+       tristate "Virtual CAN Tunnel (vxcan)"
+       ---help---
+         Similar to the virtual ethernet driver veth, vxcan implements a
+         local CAN traffic tunnel between two virtual CAN network devices.
+         When creating a vxcan, two vxcan devices are created as pair.
+         When one end receives the packet it appears on its pair and vice
+         versa. The vxcan can be used for cross namespace communication.
+
+         In opposite to vcan loopback devices the vxcan only forwards CAN
+         frames to its pair and does *not* provide a local echo of sent
+         CAN frames. To disable a potential echo in af_can.c the vxcan driver
+         announces IFF_ECHO in the interface flags. To have a clean start
+         in each namespace the CAN GW hop counter is set to zero.
+
+         This driver can also be built as a module.  If so, the module
+         will be called vxcan.
+
 config CAN_SLCAN
        tristate "Serial / USB serial CAN Adaptors (slcan)"
        depends on TTY
index 8581e2b3e87f3a66f0d341593d8f76f9052ed73d..4aabbee133b898a6f30aec7626f6580c12e24054 100644 (file)
@@ -3,6 +3,7 @@
 #
 
 obj-$(CONFIG_CAN_VCAN)         += vcan.o
+obj-$(CONFIG_CAN_VXCAN)                += vxcan.o
 obj-$(CONFIG_CAN_SLCAN)                += slcan.o
 
 obj-$(CONFIG_CAN_DEV)          += can-dev.o
diff --git a/drivers/net/can/vxcan.c b/drivers/net/can/vxcan.c
new file mode 100644 (file)
index 0000000..7fbb247
--- /dev/null
@@ -0,0 +1,316 @@
+/*
+ * vxcan.c - Virtual CAN Tunnel for cross namespace communication
+ *
+ * This code is derived from drivers/net/can/vcan.c for the virtual CAN
+ * specific parts and from drivers/net/veth.c to implement the netlink API
+ * for network interface pairs in a common and established way.
+ *
+ * Copyright (c) 2017 Oliver Hartkopp <socketcan@hartkopp.net>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the version 2 of the GNU General Public License
+ * as published by the Free Software Foundation
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/if_ether.h>
+#include <linux/can.h>
+#include <linux/can/dev.h>
+#include <linux/can/skb.h>
+#include <linux/can/vxcan.h>
+#include <linux/slab.h>
+#include <net/rtnetlink.h>
+
+#define DRV_NAME "vxcan"
+
+MODULE_DESCRIPTION("Virtual CAN Tunnel");
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Oliver Hartkopp <socketcan@hartkopp.net>");
+MODULE_ALIAS_RTNL_LINK(DRV_NAME);
+
+struct vxcan_priv {
+       struct net_device __rcu *peer;
+};
+
+static netdev_tx_t vxcan_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+       struct vxcan_priv *priv = netdev_priv(dev);
+       struct net_device *peer;
+       struct canfd_frame *cfd = (struct canfd_frame *)skb->data;
+       struct net_device_stats *peerstats, *srcstats = &dev->stats;
+
+       if (can_dropped_invalid_skb(dev, skb))
+               return NETDEV_TX_OK;
+
+       rcu_read_lock();
+       peer = rcu_dereference(priv->peer);
+       if (unlikely(!peer)) {
+               kfree_skb(skb);
+               dev->stats.tx_dropped++;
+               goto out_unlock;
+       }
+
+       skb = can_create_echo_skb(skb);
+       if (!skb)
+               goto out_unlock;
+
+       /* reset CAN GW hop counter */
+       skb->csum_start = 0;
+       skb->pkt_type   = PACKET_BROADCAST;
+       skb->dev        = peer;
+       skb->ip_summed  = CHECKSUM_UNNECESSARY;
+
+       if (netif_rx_ni(skb) == NET_RX_SUCCESS) {
+               srcstats->tx_packets++;
+               srcstats->tx_bytes += cfd->len;
+               peerstats = &peer->stats;
+               peerstats->rx_packets++;
+               peerstats->rx_bytes += cfd->len;
+       }
+
+out_unlock:
+       rcu_read_unlock();
+       return NETDEV_TX_OK;
+}
+
+
+static int vxcan_open(struct net_device *dev)
+{
+       struct vxcan_priv *priv = netdev_priv(dev);
+       struct net_device *peer = rtnl_dereference(priv->peer);
+
+       if (!peer)
+               return -ENOTCONN;
+
+       if (peer->flags & IFF_UP) {
+               netif_carrier_on(dev);
+               netif_carrier_on(peer);
+       }
+       return 0;
+}
+
+static int vxcan_close(struct net_device *dev)
+{
+       struct vxcan_priv *priv = netdev_priv(dev);
+       struct net_device *peer = rtnl_dereference(priv->peer);
+
+       netif_carrier_off(dev);
+       if (peer)
+               netif_carrier_off(peer);
+
+       return 0;
+}
+
+static int vxcan_get_iflink(const struct net_device *dev)
+{
+       struct vxcan_priv *priv = netdev_priv(dev);
+       struct net_device *peer;
+       int iflink;
+
+       rcu_read_lock();
+       peer = rcu_dereference(priv->peer);
+       iflink = peer ? peer->ifindex : 0;
+       rcu_read_unlock();
+
+       return iflink;
+}
+
+static int vxcan_change_mtu(struct net_device *dev, int new_mtu)
+{
+       /* Do not allow changing the MTU while running */
+       if (dev->flags & IFF_UP)
+               return -EBUSY;
+
+       if (new_mtu != CAN_MTU && new_mtu != CANFD_MTU)
+               return -EINVAL;
+
+       dev->mtu = new_mtu;
+       return 0;
+}
+
+static const struct net_device_ops vxcan_netdev_ops = {
+       .ndo_open       = vxcan_open,
+       .ndo_stop       = vxcan_close,
+       .ndo_start_xmit = vxcan_xmit,
+       .ndo_get_iflink = vxcan_get_iflink,
+       .ndo_change_mtu = vxcan_change_mtu,
+};
+
+static void vxcan_setup(struct net_device *dev)
+{
+       dev->type               = ARPHRD_CAN;
+       dev->mtu                = CAN_MTU;
+       dev->hard_header_len    = 0;
+       dev->addr_len           = 0;
+       dev->tx_queue_len       = 0;
+       dev->flags              = (IFF_NOARP|IFF_ECHO);
+       dev->netdev_ops         = &vxcan_netdev_ops;
+       dev->destructor         = free_netdev;
+}
+
+/* forward declaration for rtnl_create_link() */
+static struct rtnl_link_ops vxcan_link_ops;
+
+static int vxcan_newlink(struct net *net, struct net_device *dev,
+                        struct nlattr *tb[], struct nlattr *data[])
+{
+       struct vxcan_priv *priv;
+       struct net_device *peer;
+       struct net *peer_net;
+
+       struct nlattr *peer_tb[IFLA_MAX + 1], **tbp = tb;
+       char ifname[IFNAMSIZ];
+       unsigned char name_assign_type;
+       struct ifinfomsg *ifmp = NULL;
+       int err;
+
+       /* register peer device */
+       if (data && data[VXCAN_INFO_PEER]) {
+               struct nlattr *nla_peer;
+
+               nla_peer = data[VXCAN_INFO_PEER];
+               ifmp = nla_data(nla_peer);
+               err = rtnl_nla_parse_ifla(peer_tb,
+                                         nla_data(nla_peer) +
+                                         sizeof(struct ifinfomsg),
+                                         nla_len(nla_peer) -
+                                         sizeof(struct ifinfomsg),
+                                         NULL);
+               if (err < 0)
+                       return err;
+
+               tbp = peer_tb;
+       }
+
+       if (tbp[IFLA_IFNAME]) {
+               nla_strlcpy(ifname, tbp[IFLA_IFNAME], IFNAMSIZ);
+               name_assign_type = NET_NAME_USER;
+       } else {
+               snprintf(ifname, IFNAMSIZ, DRV_NAME "%%d");
+               name_assign_type = NET_NAME_ENUM;
+       }
+
+       peer_net = rtnl_link_get_net(net, tbp);
+       if (IS_ERR(peer_net))
+               return PTR_ERR(peer_net);
+
+       peer = rtnl_create_link(peer_net, ifname, name_assign_type,
+                               &vxcan_link_ops, tbp);
+       if (IS_ERR(peer)) {
+               put_net(peer_net);
+               return PTR_ERR(peer);
+       }
+
+       if (ifmp && dev->ifindex)
+               peer->ifindex = ifmp->ifi_index;
+
+       err = register_netdevice(peer);
+       put_net(peer_net);
+       peer_net = NULL;
+       if (err < 0) {
+               free_netdev(peer);
+               return err;
+       }
+
+       netif_carrier_off(peer);
+
+       err = rtnl_configure_link(peer, ifmp);
+       if (err < 0) {
+               unregister_netdevice(peer);
+               return err;
+       }
+
+       /* register first device */
+       if (tb[IFLA_IFNAME])
+               nla_strlcpy(dev->name, tb[IFLA_IFNAME], IFNAMSIZ);
+       else
+               snprintf(dev->name, IFNAMSIZ, DRV_NAME "%%d");
+
+       err = register_netdevice(dev);
+       if (err < 0) {
+               unregister_netdevice(peer);
+               return err;
+       }
+
+       netif_carrier_off(dev);
+
+       /* cross link the device pair */
+       priv = netdev_priv(dev);
+       rcu_assign_pointer(priv->peer, peer);
+
+       priv = netdev_priv(peer);
+       rcu_assign_pointer(priv->peer, dev);
+
+       return 0;
+}
+
+static void vxcan_dellink(struct net_device *dev, struct list_head *head)
+{
+       struct vxcan_priv *priv;
+       struct net_device *peer;
+
+       priv = netdev_priv(dev);
+       peer = rtnl_dereference(priv->peer);
+
+       /* Note : dellink() is called from default_device_exit_batch(),
+        * before a rcu_synchronize() point. The devices are guaranteed
+        * not being freed before one RCU grace period.
+        */
+       RCU_INIT_POINTER(priv->peer, NULL);
+       unregister_netdevice_queue(dev, head);
+
+       if (peer) {
+               priv = netdev_priv(peer);
+               RCU_INIT_POINTER(priv->peer, NULL);
+               unregister_netdevice_queue(peer, head);
+       }
+}
+
+static const struct nla_policy vxcan_policy[VXCAN_INFO_MAX + 1] = {
+       [VXCAN_INFO_PEER] = { .len = sizeof(struct ifinfomsg) },
+};
+
+static struct net *vxcan_get_link_net(const struct net_device *dev)
+{
+       struct vxcan_priv *priv = netdev_priv(dev);
+       struct net_device *peer = rtnl_dereference(priv->peer);
+
+       return peer ? dev_net(peer) : dev_net(dev);
+}
+
+static struct rtnl_link_ops vxcan_link_ops = {
+       .kind           = DRV_NAME,
+       .priv_size      = sizeof(struct vxcan_priv),
+       .setup          = vxcan_setup,
+       .newlink        = vxcan_newlink,
+       .dellink        = vxcan_dellink,
+       .policy         = vxcan_policy,
+       .maxtype        = VXCAN_INFO_MAX,
+       .get_link_net   = vxcan_get_link_net,
+};
+
+static __init int vxcan_init(void)
+{
+       pr_info("vxcan: Virtual CAN Tunnel driver\n");
+
+       return rtnl_link_register(&vxcan_link_ops);
+}
+
+static __exit void vxcan_exit(void)
+{
+       rtnl_link_unregister(&vxcan_link_ops);
+}
+
+module_init(vxcan_init);
+module_exit(vxcan_exit);
diff --git a/include/uapi/linux/can/vxcan.h b/include/uapi/linux/can/vxcan.h
new file mode 100644 (file)
index 0000000..ffb0b71
--- /dev/null
@@ -0,0 +1,12 @@
+#ifndef _UAPI_CAN_VXCAN_H
+#define _UAPI_CAN_VXCAN_H
+
+enum {
+       VXCAN_INFO_UNSPEC,
+       VXCAN_INFO_PEER,
+
+       __VXCAN_INFO_MAX
+#define VXCAN_INFO_MAX (__VXCAN_INFO_MAX - 1)
+};
+
+#endif