[deliverable/linux.git] / net / switchdev / switchdev.c

/*
 * net/switchdev/switchdev.c - Switch device API
 * Copyright (c) 2014 Jiri Pirko <jiri@resnulli.us>
 * Copyright (c) 2014-2015 Scott Feldman <sfeldma@gmail.com>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 */

#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/init.h>
#include <linux/mutex.h>
#include <linux/notifier.h>
#include <linux/netdevice.h>
#include <net/ip_fib.h>
#include <net/switchdev.h>

/**
 *	switchdev_port_attr_get - Get port attribute
 *
 *	@dev: port device
 *	@attr: attribute to get
 */
int switchdev_port_attr_get(struct net_device *dev, struct switchdev_attr *attr)
{
	const struct switchdev_ops *ops = dev->switchdev_ops;
	struct net_device *lower_dev;
	struct list_head *iter;
	struct switchdev_attr first = {
		.id = SWITCHDEV_ATTR_UNDEFINED
	};
	int err = -EOPNOTSUPP;

	if (ops && ops->switchdev_port_attr_get)
		return ops->switchdev_port_attr_get(dev, attr);

	if (attr->flags & SWITCHDEV_F_NO_RECURSE)
		return err;

	/* Switch device port(s) may be stacked under
	 * bond/team/vlan dev, so recurse down to get attr on
	 * each port.  Return -ENODATA if attr values don't
	 * compare across ports.
	 */

	netdev_for_each_lower_dev(dev, lower_dev, iter) {
		err = switchdev_port_attr_get(lower_dev, attr);
		if (err)
			break;
		if (first.id == SWITCHDEV_ATTR_UNDEFINED)
			first = *attr;
		else if (memcmp(&first, attr, sizeof(*attr)))
			return -ENODATA;
	}

	return err;
}
EXPORT_SYMBOL_GPL(switchdev_port_attr_get);

static int __switchdev_port_attr_set(struct net_device *dev,
				     struct switchdev_attr *attr)
{
	const struct switchdev_ops *ops = dev->switchdev_ops;
	struct net_device *lower_dev;
	struct list_head *iter;
	int err = -EOPNOTSUPP;

	if (ops && ops->switchdev_port_attr_set)
		return ops->switchdev_port_attr_set(dev, attr);

	if (attr->flags & SWITCHDEV_F_NO_RECURSE)
		return err;

	/* Switch device port(s) may be stacked under
	 * bond/team/vlan dev, so recurse down to set attr on
	 * each port.
	 */

	netdev_for_each_lower_dev(dev, lower_dev, iter) {
		err = __switchdev_port_attr_set(lower_dev, attr);
		if (err)
			break;
	}

	return err;
}

struct switchdev_attr_set_work {
	struct work_struct work;
	struct net_device *dev;
	struct switchdev_attr attr;
};

static void switchdev_port_attr_set_work(struct work_struct *work)
{
	struct switchdev_attr_set_work *asw =
		container_of(work, struct switchdev_attr_set_work, work);
	int err;

	rtnl_lock();
	err = switchdev_port_attr_set(asw->dev, &asw->attr);
	BUG_ON(err);
	rtnl_unlock();

	dev_put(asw->dev);
	kfree(work);
}

static int switchdev_port_attr_set_defer(struct net_device *dev,
					 struct switchdev_attr *attr)
{
	struct switchdev_attr_set_work *asw;

	asw = kmalloc(sizeof(*asw), GFP_ATOMIC);
	if (!asw)
		return -ENOMEM;

	INIT_WORK(&asw->work, switchdev_port_attr_set_work);

	dev_hold(dev);
	asw->dev = dev;
	memcpy(&asw->attr, attr, sizeof(asw->attr));

	schedule_work(&asw->work);

	return 0;
}

/**
 *	switchdev_port_attr_set - Set port attribute
 *
 *	@dev: port device
 *	@attr: attribute to set
 *
 *	Use a 2-phase prepare-commit transaction model to ensure
 *	system is not left in a partially updated state due to
 *	failure from driver/device.
 */
int switchdev_port_attr_set(struct net_device *dev, struct switchdev_attr *attr)
{
	int err;

	if (!rtnl_is_locked()) {
		/* Running prepare-commit transaction across stacked
		 * devices requires nothing moves, so if rtnl_lock is
		 * not held, schedule a worker thread to hold rtnl_lock
		 * while setting attr.
		 */

		return switchdev_port_attr_set_defer(dev, attr);
	}

	/* Phase I: prepare for attr set. Driver/device should fail
	 * here if there are going to be issues in the commit phase,
	 * such as lack of resources or support.  The driver/device
	 * should reserve resources needed for the commit phase here,
	 * but should not commit the attr.
	 */

	attr->trans = SWITCHDEV_TRANS_PREPARE;
	err = __switchdev_port_attr_set(dev, attr);
	if (err) {
		/* Prepare phase failed: abort the transaction.  Any
		 * resources reserved in the prepare phase are
		 * released.
		 */

		attr->trans = SWITCHDEV_TRANS_ABORT;
		__switchdev_port_attr_set(dev, attr);

		return err;
	}

	/* Phase II: commit attr set.  This cannot fail as a fault
	 * of driver/device.  If it does, it's a bug in the driver/device
	 * because the driver said everythings was OK in phase I.
	 */

	attr->trans = SWITCHDEV_TRANS_COMMIT;
	err = __switchdev_port_attr_set(dev, attr);
	BUG_ON(err);

	return err;
}
EXPORT_SYMBOL_GPL(switchdev_port_attr_set);

static DEFINE_MUTEX(switchdev_mutex);
static RAW_NOTIFIER_HEAD(switchdev_notif_chain);

/**
 *	register_switchdev_notifier - Register notifier
 *	@nb: notifier_block
 *
 *	Register switch device notifier. This should be used by code
 *	which needs to monitor events happening in particular device.
 *	Return values are same as for atomic_notifier_chain_register().
 */
int register_switchdev_notifier(struct notifier_block *nb)
{
	int err;

	mutex_lock(&switchdev_mutex);
	err = raw_notifier_chain_register(&switchdev_notif_chain, nb);
	mutex_unlock(&switchdev_mutex);
	return err;
}
EXPORT_SYMBOL_GPL(register_switchdev_notifier);

/**
 *	unregister_switchdev_notifier - Unregister notifier
 *	@nb: notifier_block
 *
 *	Unregister switch device notifier.
 *	Return values are same as for atomic_notifier_chain_unregister().
 */
int unregister_switchdev_notifier(struct notifier_block *nb)
{
	int err;

	mutex_lock(&switchdev_mutex);
	err = raw_notifier_chain_unregister(&switchdev_notif_chain, nb);
	mutex_unlock(&switchdev_mutex);
	return err;
}
EXPORT_SYMBOL_GPL(unregister_switchdev_notifier);

/**
 *	call_switchdev_notifiers - Call notifiers
 *	@val: value passed unmodified to notifier function
 *	@dev: port device
 *	@info: notifier information data
 *
 *	Call all network notifier blocks. This should be called by driver
 *	when it needs to propagate hardware event.
 *	Return values are same as for atomic_notifier_call_chain().
 */
int call_switchdev_notifiers(unsigned long val, struct net_device *dev,
			     struct switchdev_notifier_info *info)
{
	int err;

	info->dev = dev;
	mutex_lock(&switchdev_mutex);
	err = raw_notifier_call_chain(&switchdev_notif_chain, val, info);
	mutex_unlock(&switchdev_mutex);
	return err;
}
EXPORT_SYMBOL_GPL(call_switchdev_notifiers);

/**
 *	switchdev_port_bridge_setlink - Notify switch device port of bridge
 *	port attributes
 *
 *	@dev: port device
 *	@nlh: netlink msg with bridge port attributes
 *	@flags: bridge setlink flags
 *
 *	Notify switch device port of bridge port attributes
 */
int switchdev_port_bridge_setlink(struct net_device *dev,
				  struct nlmsghdr *nlh, u16 flags)
{
	const struct net_device_ops *ops = dev->netdev_ops;

	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
		return 0;

	if (!ops->ndo_bridge_setlink)
		return -EOPNOTSUPP;

	return ops->ndo_bridge_setlink(dev, nlh, flags);
}
EXPORT_SYMBOL_GPL(switchdev_port_bridge_setlink);

/**
 *	switchdev_port_bridge_dellink - Notify switch device port of bridge
 *	port attribute delete
 *
 *	@dev: port device
 *	@nlh: netlink msg with bridge port attributes
 *	@flags: bridge setlink flags
 *
 *	Notify switch device port of bridge port attribute delete
 */
int switchdev_port_bridge_dellink(struct net_device *dev,
				  struct nlmsghdr *nlh, u16 flags)
{
	const struct net_device_ops *ops = dev->netdev_ops;

	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
		return 0;

	if (!ops->ndo_bridge_dellink)
		return -EOPNOTSUPP;

	return ops->ndo_bridge_dellink(dev, nlh, flags);
}
EXPORT_SYMBOL_GPL(switchdev_port_bridge_dellink);

/**
 *	ndo_dflt_switchdev_port_bridge_setlink - default ndo bridge setlink
 *						 op for master devices
 *
 *	@dev: port device
 *	@nlh: netlink msg with bridge port attributes
 *	@flags: bridge setlink flags
 *
 *	Notify master device slaves of bridge port attributes
 */
int ndo_dflt_switchdev_port_bridge_setlink(struct net_device *dev,
					   struct nlmsghdr *nlh, u16 flags)
{
	struct net_device *lower_dev;
	struct list_head *iter;
	int ret = 0, err = 0;

	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
		return ret;

	netdev_for_each_lower_dev(dev, lower_dev, iter) {
		err = switchdev_port_bridge_setlink(lower_dev, nlh, flags);
		if (err && err != -EOPNOTSUPP)
			ret = err;
	}

	return ret;
}
EXPORT_SYMBOL_GPL(ndo_dflt_switchdev_port_bridge_setlink);

/**
 *	ndo_dflt_switchdev_port_bridge_dellink - default ndo bridge dellink
 *						 op for master devices
 *
 *	@dev: port device
 *	@nlh: netlink msg with bridge port attributes
 *	@flags: bridge dellink flags
 *
 *	Notify master device slaves of bridge port attribute deletes
 */
int ndo_dflt_switchdev_port_bridge_dellink(struct net_device *dev,
					   struct nlmsghdr *nlh, u16 flags)
{
	struct net_device *lower_dev;
	struct list_head *iter;
	int ret = 0, err = 0;

	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
		return ret;

	netdev_for_each_lower_dev(dev, lower_dev, iter) {
		err = switchdev_port_bridge_dellink(lower_dev, nlh, flags);
		if (err && err != -EOPNOTSUPP)
			ret = err;
	}

	return ret;
}
EXPORT_SYMBOL_GPL(ndo_dflt_switchdev_port_bridge_dellink);

static struct net_device *switchdev_get_lowest_dev(struct net_device *dev)
{
	const struct switchdev_ops *ops = dev->switchdev_ops;
	struct net_device *lower_dev;
	struct net_device *port_dev;
	struct list_head *iter;

	/* Recusively search down until we find a sw port dev.
	 * (A sw port dev supports switchdev_port_attr_get).
	 */

	if (ops && ops->switchdev_port_attr_get)
		return dev;

	netdev_for_each_lower_dev(dev, lower_dev, iter) {
		port_dev = switchdev_get_lowest_dev(lower_dev);
		if (port_dev)
			return port_dev;
	}

	return NULL;
}

static struct net_device *switchdev_get_dev_by_nhs(struct fib_info *fi)
{
	struct switchdev_attr attr = {
		.id = SWITCHDEV_ATTR_PORT_PARENT_ID,
	};
	struct switchdev_attr prev_attr;
	struct net_device *dev = NULL;
	int nhsel;

	/* For this route, all nexthop devs must be on the same switch. */

	for (nhsel = 0; nhsel < fi->fib_nhs; nhsel++) {
		const struct fib_nh *nh = &fi->fib_nh[nhsel];

		if (!nh->nh_dev)
			return NULL;

		dev = switchdev_get_lowest_dev(nh->nh_dev);
		if (!dev)
			return NULL;

		if (switchdev_port_attr_get(dev, &attr))
			return NULL;

		if (nhsel > 0) {
			if (prev_attr.ppid.id_len != attr.ppid.id_len)
				return NULL;
			if (memcmp(prev_attr.ppid.id, attr.ppid.id,
				   attr.ppid.id_len))
				return NULL;
		}

		prev_attr = attr;
	}

	return dev;
}

/**
 *	switchdev_fib_ipv4_add - Add IPv4 route entry to switch
 *
 *	@dst: route's IPv4 destination address
 *	@dst_len: destination address length (prefix length)
 *	@fi: route FIB info structure
 *	@tos: route TOS
 *	@type: route type
 *	@nlflags: netlink flags passed in (NLM_F_*)
 *	@tb_id: route table ID
 *
 *	Add IPv4 route entry to switch device.
 */
int switchdev_fib_ipv4_add(u32 dst, int dst_len, struct fib_info *fi,
			   u8 tos, u8 type, u32 nlflags, u32 tb_id)
{
	struct net_device *dev;
	const struct switchdev_ops *ops;
	int err = 0;

	/* Don't offload route if using custom ip rules or if
	 * IPv4 FIB offloading has been disabled completely.
	 */

#ifdef CONFIG_IP_MULTIPLE_TABLES
	if (fi->fib_net->ipv4.fib_has_custom_rules)
		return 0;
#endif

	if (fi->fib_net->ipv4.fib_offload_disabled)
		return 0;

	dev = switchdev_get_dev_by_nhs(fi);
	if (!dev)
		return 0;
	ops = dev->switchdev_ops;

	if (ops->switchdev_fib_ipv4_add) {
		err = ops->switchdev_fib_ipv4_add(dev, htonl(dst), dst_len,
						  fi, tos, type, nlflags,
						  tb_id);
		if (!err)
			fi->fib_flags |= RTNH_F_EXTERNAL;
	}

	return err;
}
EXPORT_SYMBOL_GPL(switchdev_fib_ipv4_add);

/**
 *	switchdev_fib_ipv4_del - Delete IPv4 route entry from switch
 *
 *	@dst: route's IPv4 destination address
 *	@dst_len: destination address length (prefix length)
 *	@fi: route FIB info structure
 *	@tos: route TOS
 *	@type: route type
 *	@tb_id: route table ID
 *
 *	Delete IPv4 route entry from switch device.
 */
int switchdev_fib_ipv4_del(u32 dst, int dst_len, struct fib_info *fi,
			   u8 tos, u8 type, u32 tb_id)
{
	struct net_device *dev;
	const struct switchdev_ops *ops;
	int err = 0;

	if (!(fi->fib_flags & RTNH_F_EXTERNAL))
		return 0;

	dev = switchdev_get_dev_by_nhs(fi);
	if (!dev)
		return 0;
	ops = dev->switchdev_ops;

	if (ops->switchdev_fib_ipv4_del) {
		err = ops->switchdev_fib_ipv4_del(dev, htonl(dst), dst_len,
						  fi, tos, type, tb_id);
		if (!err)
			fi->fib_flags &= ~RTNH_F_EXTERNAL;
	}

	return err;
}
EXPORT_SYMBOL_GPL(switchdev_fib_ipv4_del);

/**
 *	switchdev_fib_ipv4_abort - Abort an IPv4 FIB operation
 *
 *	@fi: route FIB info structure
 */
void switchdev_fib_ipv4_abort(struct fib_info *fi)
{
	/* There was a problem installing this route to the offload
	 * device.  For now, until we come up with more refined
	 * policy handling, abruptly end IPv4 fib offloading for
	 * for entire net by flushing offload device(s) of all
	 * IPv4 routes, and mark IPv4 fib offloading broken from
	 * this point forward.
	 */

	fib_flush_external(fi->fib_net);
	fi->fib_net->ipv4.fib_offload_disabled = true;
}
EXPORT_SYMBOL_GPL(switchdev_fib_ipv4_abort);
Commit	Line	Data
007f790c JP	1	/*
	2	* net/switchdev/switchdev.c - Switch device API
	3	* Copyright (c) 2014 Jiri Pirko <jiri@resnulli.us>
f8f21471	4	* Copyright (c) 2014-2015 Scott Feldman <sfeldma@gmail.com>
007f790c JP	5	*
	6	* This program is free software; you can redistribute it and/or modify
	7	* it under the terms of the GNU General Public License as published by
	8	* the Free Software Foundation; either version 2 of the License, or
	9	* (at your option) any later version.
	10	*/
	11
	12	#include <linux/kernel.h>
	13	#include <linux/types.h>
	14	#include <linux/init.h>
03bf0c28 JP	15	#include <linux/mutex.h>
03bf0c28 JP	16	#include <linux/notifier.h>
007f790c	17	#include <linux/netdevice.h>
5e8d9049	18	#include <net/ip_fib.h>
007f790c JP	19	#include <net/switchdev.h>
007f790c JP	20
3094333d SF	21	/**
	22	* switchdev_port_attr_get - Get port attribute
	23	*
	24	* @dev: port device
	25	* @attr: attribute to get
	26	*/
	27	int switchdev_port_attr_get(struct net_device dev, struct switchdev_attr attr)
	28	{
	29	const struct switchdev_ops *ops = dev->switchdev_ops;
	30	struct net_device *lower_dev;
	31	struct list_head *iter;
	32	struct switchdev_attr first = {
	33	.id = SWITCHDEV_ATTR_UNDEFINED
	34	};
	35	int err = -EOPNOTSUPP;
	36
	37	if (ops && ops->switchdev_port_attr_get)
	38	return ops->switchdev_port_attr_get(dev, attr);
	39
	40	if (attr->flags & SWITCHDEV_F_NO_RECURSE)
	41	return err;
	42
	43	/* Switch device port(s) may be stacked under
	44	* bond/team/vlan dev, so recurse down to get attr on
	45	* each port. Return -ENODATA if attr values don't
	46	* compare across ports.
	47	*/
	48
	49	netdev_for_each_lower_dev(dev, lower_dev, iter) {
	50	err = switchdev_port_attr_get(lower_dev, attr);
	51	if (err)
	52	break;
	53	if (first.id == SWITCHDEV_ATTR_UNDEFINED)
	54	first = *attr;
	55	else if (memcmp(&first, attr, sizeof(*attr)))
	56	return -ENODATA;
	57	}
	58
	59	return err;
	60	}
	61	EXPORT_SYMBOL_GPL(switchdev_port_attr_get);
	62
	63	static int __switchdev_port_attr_set(struct net_device *dev,
	64	struct switchdev_attr *attr)
	65	{
	66	const struct switchdev_ops *ops = dev->switchdev_ops;
	67	struct net_device *lower_dev;
	68	struct list_head *iter;
	69	int err = -EOPNOTSUPP;
	70
	71	if (ops && ops->switchdev_port_attr_set)
	72	return ops->switchdev_port_attr_set(dev, attr);
	73
	74	if (attr->flags & SWITCHDEV_F_NO_RECURSE)
	75	return err;
	76
	77	/* Switch device port(s) may be stacked under
	78	* bond/team/vlan dev, so recurse down to set attr on
	79	* each port.
	80	*/
	81
	82	netdev_for_each_lower_dev(dev, lower_dev, iter) {
	83	err = __switchdev_port_attr_set(lower_dev, attr);
	84	if (err)
85	break;
86	}
87
88	return err;
89	}
90
91	struct switchdev_attr_set_work {
92	struct work_struct work;
93	struct net_device *dev;
94	struct switchdev_attr attr;
95	};
96
97	static void switchdev_port_attr_set_work(struct work_struct *work)
98	{
99	struct switchdev_attr_set_work *asw =
100	container_of(work, struct switchdev_attr_set_work, work);
101	int err;
102
103	rtnl_lock();
104	err = switchdev_port_attr_set(asw->dev, &asw->attr);
105	BUG_ON(err);
106	rtnl_unlock();
107
108	dev_put(asw->dev);
109	kfree(work);
110	}
111
112	static int switchdev_port_attr_set_defer(struct net_device *dev,
113	struct switchdev_attr *attr)
114	{
115	struct switchdev_attr_set_work *asw;
116
117	asw = kmalloc(sizeof(*asw), GFP_ATOMIC);
118	if (!asw)
119	return -ENOMEM;
120
121	INIT_WORK(&asw->work, switchdev_port_attr_set_work);
122
123	dev_hold(dev);
124	asw->dev = dev;
125	memcpy(&asw->attr, attr, sizeof(asw->attr));
126
127	schedule_work(&asw->work);
128
129	return 0;
130	}
131
132	/**
133	* switchdev_port_attr_set - Set port attribute
134	*
135	* @dev: port device
136	* @attr: attribute to set
137	*
138	* Use a 2-phase prepare-commit transaction model to ensure
139	* system is not left in a partially updated state due to
140	* failure from driver/device.
141	*/
142	int switchdev_port_attr_set(struct net_device dev, struct switchdev_attr attr)
143	{
144	int err;
145
146	if (!rtnl_is_locked()) {
147	/* Running prepare-commit transaction across stacked
148	* devices requires nothing moves, so if rtnl_lock is
149	* not held, schedule a worker thread to hold rtnl_lock
150	* while setting attr.
151	*/
152
153	return switchdev_port_attr_set_defer(dev, attr);
154	}
155
156	/* Phase I: prepare for attr set. Driver/device should fail
157	* here if there are going to be issues in the commit phase,
158	* such as lack of resources or support. The driver/device
159	* should reserve resources needed for the commit phase here,
160	* but should not commit the attr.
161	*/
162
163	attr->trans = SWITCHDEV_TRANS_PREPARE;
164	err = __switchdev_port_attr_set(dev, attr);
165	if (err) {
166	/* Prepare phase failed: abort the transaction. Any
167	* resources reserved in the prepare phase are
168	* released.
169	*/
170
171	attr->trans = SWITCHDEV_TRANS_ABORT;
172	__switchdev_port_attr_set(dev, attr);
173
174	return err;
175	}
176
177	/* Phase II: commit attr set. This cannot fail as a fault
178	* of driver/device. If it does, it's a bug in the driver/device
179	* because the driver said everythings was OK in phase I.
180	*/
181
182	attr->trans = SWITCHDEV_TRANS_COMMIT;
183	err = __switchdev_port_attr_set(dev, attr);
184	BUG_ON(err);
185
186	return err;
187	}
188	EXPORT_SYMBOL_GPL(switchdev_port_attr_set);
189
ebb9a03a JP	190	static DEFINE_MUTEX(switchdev_mutex);
ebb9a03a JP	191	static RAW_NOTIFIER_HEAD(switchdev_notif_chain);
03bf0c28 JP	192
03bf0c28 JP	193	/**
ebb9a03a	194	* register_switchdev_notifier - Register notifier
03bf0c28 JP	195	* @nb: notifier_block
	196	*
	197	* Register switch device notifier. This should be used by code
	198	* which needs to monitor events happening in particular device.
	199	* Return values are same as for atomic_notifier_chain_register().
	200	*/
ebb9a03a	201	int register_switchdev_notifier(struct notifier_block *nb)
03bf0c28 JP	202	{
	203	int err;
	204
ebb9a03a JP	205	mutex_lock(&switchdev_mutex);
	206	err = raw_notifier_chain_register(&switchdev_notif_chain, nb);
	207	mutex_unlock(&switchdev_mutex);
03bf0c28 JP	208	return err;
03bf0c28 JP	209	}
ebb9a03a	210	EXPORT_SYMBOL_GPL(register_switchdev_notifier);
03bf0c28 JP	211
03bf0c28 JP	212	/**
ebb9a03a	213	* unregister_switchdev_notifier - Unregister notifier
03bf0c28 JP	214	* @nb: notifier_block
	215	*
	216	* Unregister switch device notifier.
	217	* Return values are same as for atomic_notifier_chain_unregister().
	218	*/
ebb9a03a	219	int unregister_switchdev_notifier(struct notifier_block *nb)
03bf0c28 JP	220	{
	221	int err;
	222
ebb9a03a JP	223	mutex_lock(&switchdev_mutex);
	224	err = raw_notifier_chain_unregister(&switchdev_notif_chain, nb);
	225	mutex_unlock(&switchdev_mutex);
03bf0c28 JP	226	return err;
03bf0c28 JP	227	}
ebb9a03a	228	EXPORT_SYMBOL_GPL(unregister_switchdev_notifier);
03bf0c28 JP	229
03bf0c28 JP	230	/**
ebb9a03a	231	* call_switchdev_notifiers - Call notifiers
03bf0c28 JP	232	* @val: value passed unmodified to notifier function
	233	* @dev: port device
	234	* @info: notifier information data
	235	*
	236	* Call all network notifier blocks. This should be called by driver
	237	* when it needs to propagate hardware event.
	238	* Return values are same as for atomic_notifier_call_chain().
	239	*/
ebb9a03a JP	240	int call_switchdev_notifiers(unsigned long val, struct net_device *dev,
ebb9a03a JP	241	struct switchdev_notifier_info *info)
03bf0c28 JP	242	{
	243	int err;
	244
	245	info->dev = dev;
ebb9a03a JP	246	mutex_lock(&switchdev_mutex);
	247	err = raw_notifier_call_chain(&switchdev_notif_chain, val, info);
	248	mutex_unlock(&switchdev_mutex);
03bf0c28 JP	249	return err;
03bf0c28 JP	250	}
ebb9a03a	251	EXPORT_SYMBOL_GPL(call_switchdev_notifiers);
8a44dbb2 RP	252
8a44dbb2 RP	253	/**
ebb9a03a	254	* switchdev_port_bridge_setlink - Notify switch device port of bridge
8a44dbb2 RP	255	* port attributes
	256	*
	257	* @dev: port device
	258	* @nlh: netlink msg with bridge port attributes
	259	* @flags: bridge setlink flags
	260	*
	261	* Notify switch device port of bridge port attributes
	262	*/
ebb9a03a JP	263	int switchdev_port_bridge_setlink(struct net_device *dev,
ebb9a03a JP	264	struct nlmsghdr *nlh, u16 flags)
8a44dbb2 RP	265	{
	266	const struct net_device_ops *ops = dev->netdev_ops;
	267
	268	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
	269	return 0;
	270
	271	if (!ops->ndo_bridge_setlink)
	272	return -EOPNOTSUPP;
	273
	274	return ops->ndo_bridge_setlink(dev, nlh, flags);
	275	}
ebb9a03a	276	EXPORT_SYMBOL_GPL(switchdev_port_bridge_setlink);
8a44dbb2 RP	277
8a44dbb2 RP	278	/**
ebb9a03a	279	* switchdev_port_bridge_dellink - Notify switch device port of bridge
8a44dbb2 RP	280	* port attribute delete
	281	*
	282	* @dev: port device
	283	* @nlh: netlink msg with bridge port attributes
	284	* @flags: bridge setlink flags
	285	*
	286	* Notify switch device port of bridge port attribute delete
	287	*/
ebb9a03a JP	288	int switchdev_port_bridge_dellink(struct net_device *dev,
ebb9a03a JP	289	struct nlmsghdr *nlh, u16 flags)
8a44dbb2 RP	290	{
	291	const struct net_device_ops *ops = dev->netdev_ops;
	292
	293	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
	294	return 0;
	295
	296	if (!ops->ndo_bridge_dellink)
	297	return -EOPNOTSUPP;
	298
	299	return ops->ndo_bridge_dellink(dev, nlh, flags);
	300	}
ebb9a03a	301	EXPORT_SYMBOL_GPL(switchdev_port_bridge_dellink);
8a44dbb2 RP	302
8a44dbb2 RP	303	/**
ebb9a03a JP	304	* ndo_dflt_switchdev_port_bridge_setlink - default ndo bridge setlink
ebb9a03a JP	305	* op for master devices
8a44dbb2 RP	306	*
	307	* @dev: port device
	308	* @nlh: netlink msg with bridge port attributes
	309	* @flags: bridge setlink flags
	310	*
	311	* Notify master device slaves of bridge port attributes
	312	*/
ebb9a03a JP	313	int ndo_dflt_switchdev_port_bridge_setlink(struct net_device *dev,
ebb9a03a JP	314	struct nlmsghdr *nlh, u16 flags)
8a44dbb2 RP	315	{
	316	struct net_device *lower_dev;
	317	struct list_head *iter;
	318	int ret = 0, err = 0;
	319
	320	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
	321	return ret;
	322
	323	netdev_for_each_lower_dev(dev, lower_dev, iter) {
ebb9a03a	324	err = switchdev_port_bridge_setlink(lower_dev, nlh, flags);
8a44dbb2 RP	325	if (err && err != -EOPNOTSUPP)
	326	ret = err;
	327	}
	328
	329	return ret;
	330	}
ebb9a03a	331	EXPORT_SYMBOL_GPL(ndo_dflt_switchdev_port_bridge_setlink);
8a44dbb2 RP	332
8a44dbb2 RP	333	/**
ebb9a03a JP	334	* ndo_dflt_switchdev_port_bridge_dellink - default ndo bridge dellink
ebb9a03a JP	335	* op for master devices
8a44dbb2 RP	336	*
	337	* @dev: port device
	338	* @nlh: netlink msg with bridge port attributes
	339	* @flags: bridge dellink flags
	340	*
	341	* Notify master device slaves of bridge port attribute deletes
	342	*/
ebb9a03a JP	343	int ndo_dflt_switchdev_port_bridge_dellink(struct net_device *dev,
ebb9a03a JP	344	struct nlmsghdr *nlh, u16 flags)
8a44dbb2 RP	345	{
	346	struct net_device *lower_dev;
	347	struct list_head *iter;
	348	int ret = 0, err = 0;
	349
	350	if (!(dev->features & NETIF_F_HW_SWITCH_OFFLOAD))
	351	return ret;
	352
	353	netdev_for_each_lower_dev(dev, lower_dev, iter) {
ebb9a03a	354	err = switchdev_port_bridge_dellink(lower_dev, nlh, flags);
8a44dbb2 RP	355	if (err && err != -EOPNOTSUPP)
	356	ret = err;
	357	}
	358
	359	return ret;
	360	}
ebb9a03a	361	EXPORT_SYMBOL_GPL(ndo_dflt_switchdev_port_bridge_dellink);
5e8d9049	362
ebb9a03a	363	static struct net_device switchdev_get_lowest_dev(struct net_device dev)
b5d6fbde	364	{
9d47c0a2	365	const struct switchdev_ops *ops = dev->switchdev_ops;
b5d6fbde SF	366	struct net_device *lower_dev;
	367	struct net_device *port_dev;
	368	struct list_head *iter;
	369
	370	/* Recusively search down until we find a sw port dev.
f8e20a9f	371	* (A sw port dev supports switchdev_port_attr_get).
b5d6fbde SF	372	*/
b5d6fbde SF	373
f8e20a9f	374	if (ops && ops->switchdev_port_attr_get)
b5d6fbde SF	375	return dev;
	376
	377	netdev_for_each_lower_dev(dev, lower_dev, iter) {
ebb9a03a	378	port_dev = switchdev_get_lowest_dev(lower_dev);
b5d6fbde SF	379	if (port_dev)
	380	return port_dev;
	381	}
	382
	383	return NULL;
	384	}
	385
ebb9a03a	386	static struct net_device switchdev_get_dev_by_nhs(struct fib_info fi)
b5d6fbde	387	{
f8e20a9f SF	388	struct switchdev_attr attr = {
	389	.id = SWITCHDEV_ATTR_PORT_PARENT_ID,
	390	};
	391	struct switchdev_attr prev_attr;
b5d6fbde SF	392	struct net_device *dev = NULL;
	393	int nhsel;
	394
	395	/* For this route, all nexthop devs must be on the same switch. */
	396
	397	for (nhsel = 0; nhsel < fi->fib_nhs; nhsel++) {
	398	const struct fib_nh *nh = &fi->fib_nh[nhsel];
	399
	400	if (!nh->nh_dev)
	401	return NULL;
	402
ebb9a03a	403	dev = switchdev_get_lowest_dev(nh->nh_dev);
b5d6fbde SF	404	if (!dev)
	405	return NULL;
	406
f8e20a9f	407	if (switchdev_port_attr_get(dev, &attr))
b5d6fbde SF	408	return NULL;
	409
	410	if (nhsel > 0) {
f8e20a9f	411	if (prev_attr.ppid.id_len != attr.ppid.id_len)
b5d6fbde	412	return NULL;
f8e20a9f SF	413	if (memcmp(prev_attr.ppid.id, attr.ppid.id,
f8e20a9f SF	414	attr.ppid.id_len))
b5d6fbde SF	415	return NULL;
	416	}
	417
f8e20a9f	418	prev_attr = attr;
b5d6fbde SF	419	}
	420
	421	return dev;
	422	}
	423
5e8d9049	424	/**
ebb9a03a	425	* switchdev_fib_ipv4_add - Add IPv4 route entry to switch
5e8d9049 SF	426	*
	427	* @dst: route's IPv4 destination address
	428	* @dst_len: destination address length (prefix length)
	429	* @fi: route FIB info structure
	430	* @tos: route TOS
	431	* @type: route type
f8f21471	432	* @nlflags: netlink flags passed in (NLM_F_*)
5e8d9049 SF	433	* @tb_id: route table ID
	434	*
	435	* Add IPv4 route entry to switch device.
	436	*/
ebb9a03a JP	437	int switchdev_fib_ipv4_add(u32 dst, int dst_len, struct fib_info *fi,
ebb9a03a JP	438	u8 tos, u8 type, u32 nlflags, u32 tb_id)
5e8d9049	439	{
b5d6fbde	440	struct net_device *dev;
9d47c0a2	441	const struct switchdev_ops *ops;
b5d6fbde SF	442	int err = 0;
b5d6fbde SF	443
8e05fd71 SF	444	/* Don't offload route if using custom ip rules or if
	445	* IPv4 FIB offloading has been disabled completely.
	446	*/
	447
e1315db1 SF	448	#ifdef CONFIG_IP_MULTIPLE_TABLES
	449	if (fi->fib_net->ipv4.fib_has_custom_rules)
	450	return 0;
	451	#endif
	452
	453	if (fi->fib_net->ipv4.fib_offload_disabled)
104616e7 SF	454	return 0;
104616e7 SF	455
ebb9a03a	456	dev = switchdev_get_dev_by_nhs(fi);
b5d6fbde SF	457	if (!dev)
b5d6fbde SF	458	return 0;
9d47c0a2	459	ops = dev->switchdev_ops;
b5d6fbde	460
9d47c0a2 JP	461	if (ops->switchdev_fib_ipv4_add) {
	462	err = ops->switchdev_fib_ipv4_add(dev, htonl(dst), dst_len,
	463	fi, tos, type, nlflags,
	464	tb_id);
b5d6fbde SF	465	if (!err)
	466	fi->fib_flags \|= RTNH_F_EXTERNAL;
	467	}
	468
	469	return err;
5e8d9049	470	}
ebb9a03a	471	EXPORT_SYMBOL_GPL(switchdev_fib_ipv4_add);
5e8d9049 SF	472
5e8d9049 SF	473	/**
ebb9a03a	474	* switchdev_fib_ipv4_del - Delete IPv4 route entry from switch
5e8d9049 SF	475	*
	476	* @dst: route's IPv4 destination address
	477	* @dst_len: destination address length (prefix length)
	478	* @fi: route FIB info structure
	479	* @tos: route TOS
	480	* @type: route type
	481	* @tb_id: route table ID
	482	*
	483	* Delete IPv4 route entry from switch device.
	484	*/
ebb9a03a JP	485	int switchdev_fib_ipv4_del(u32 dst, int dst_len, struct fib_info *fi,
ebb9a03a JP	486	u8 tos, u8 type, u32 tb_id)
5e8d9049	487	{
b5d6fbde	488	struct net_device *dev;
9d47c0a2	489	const struct switchdev_ops *ops;
b5d6fbde SF	490	int err = 0;
	491
	492	if (!(fi->fib_flags & RTNH_F_EXTERNAL))
	493	return 0;
	494
ebb9a03a	495	dev = switchdev_get_dev_by_nhs(fi);
b5d6fbde SF	496	if (!dev)
b5d6fbde SF	497	return 0;
9d47c0a2	498	ops = dev->switchdev_ops;
b5d6fbde	499
9d47c0a2 JP	500	if (ops->switchdev_fib_ipv4_del) {
	501	err = ops->switchdev_fib_ipv4_del(dev, htonl(dst), dst_len,
	502	fi, tos, type, tb_id);
b5d6fbde SF	503	if (!err)
	504	fi->fib_flags &= ~RTNH_F_EXTERNAL;
	505	}
	506
	507	return err;
5e8d9049	508	}
ebb9a03a	509	EXPORT_SYMBOL_GPL(switchdev_fib_ipv4_del);
8e05fd71 SF	510
8e05fd71 SF	511	/**
ebb9a03a	512	* switchdev_fib_ipv4_abort - Abort an IPv4 FIB operation
8e05fd71 SF	513	*
	514	* @fi: route FIB info structure
	515	*/
ebb9a03a	516	void switchdev_fib_ipv4_abort(struct fib_info *fi)
8e05fd71 SF	517	{
	518	/* There was a problem installing this route to the offload
	519	* device. For now, until we come up with more refined
	520	* policy handling, abruptly end IPv4 fib offloading for
	521	* for entire net by flushing offload device(s) of all
	522	* IPv4 routes, and mark IPv4 fib offloading broken from
	523	* this point forward.
	524	*/
	525
	526	fib_flush_external(fi->fib_net);
	527	fi->fib_net->ipv4.fib_offload_disabled = true;
	528	}
ebb9a03a	529	EXPORT_SYMBOL_GPL(switchdev_fib_ipv4_abort);