[mirror_ubuntu-bionic-kernel.git] / net / mpls / mpls_iptunnel.c

/*
 * mpls tunnels	An implementation mpls tunnels using the light weight tunnel
 *		infrastructure
 *
 * Authors:	Roopa Prabhu, <roopa@cumulusnetworks.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/types.h>
#include <linux/skbuff.h>
#include <linux/net.h>
#include <linux/module.h>
#include <linux/mpls.h>
#include <linux/vmalloc.h>
#include <net/ip.h>
#include <net/dst.h>
#include <net/lwtunnel.h>
#include <net/netevent.h>
#include <net/netns/generic.h>
#include <net/ip6_fib.h>
#include <net/route.h>
#include <net/mpls_iptunnel.h>
#include <linux/mpls_iptunnel.h>
#include "internal.h"

static const struct nla_policy mpls_iptunnel_policy[MPLS_IPTUNNEL_MAX + 1] = {
	[MPLS_IPTUNNEL_DST]	= { .len = sizeof(u32) },
	[MPLS_IPTUNNEL_TTL]	= { .type = NLA_U8 },
};

static unsigned int mpls_encap_size(struct mpls_iptunnel_encap *en)
{
	/* The size of the layer 2.5 labels to be added for this route */
	return en->labels * sizeof(struct mpls_shim_hdr);
}

static int mpls_xmit(struct sk_buff *skb)
{
	struct mpls_iptunnel_encap *tun_encap_info;
	struct mpls_shim_hdr *hdr;
	struct net_device *out_dev;
	unsigned int hh_len;
	unsigned int new_header_size;
	unsigned int mtu;
	struct dst_entry *dst = skb_dst(skb);
	struct rtable *rt = NULL;
	struct rt6_info *rt6 = NULL;
	struct mpls_dev *out_mdev;
	struct net *net;
	int err = 0;
	bool bos;
	int i;
	unsigned int ttl;

	/* Find the output device */
	out_dev = dst->dev;
	net = dev_net(out_dev);

	skb_orphan(skb);

	if (!mpls_output_possible(out_dev) ||
	    !dst->lwtstate || skb_warn_if_lro(skb))
		goto drop;

	skb_forward_csum(skb);

	tun_encap_info = mpls_lwtunnel_encap(dst->lwtstate);

	/* Obtain the ttl using the following set of rules.
	 *
	 * LWT ttl propagation setting:
	 *  - disabled => use default TTL value from LWT
	 *  - enabled  => use TTL value from IPv4/IPv6 header
	 *  - default  =>
	 *   Global ttl propagation setting:
	 *    - disabled => use default TTL value from global setting
	 *    - enabled => use TTL value from IPv4/IPv6 header
	 */
	if (dst->ops->family == AF_INET) {
		if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DISABLED)
			ttl = tun_encap_info->default_ttl;
		else if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DEFAULT &&
			 !net->mpls.ip_ttl_propagate)
			ttl = net->mpls.default_ttl;
		else
			ttl = ip_hdr(skb)->ttl;
		rt = (struct rtable *)dst;
	} else if (dst->ops->family == AF_INET6) {
		if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DISABLED)
			ttl = tun_encap_info->default_ttl;
		else if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DEFAULT &&
			 !net->mpls.ip_ttl_propagate)
			ttl = net->mpls.default_ttl;
		else
			ttl = ipv6_hdr(skb)->hop_limit;
		rt6 = (struct rt6_info *)dst;
	} else {
		goto drop;
	}

	/* Verify the destination can hold the packet */
	new_header_size = mpls_encap_size(tun_encap_info);
	mtu = mpls_dev_mtu(out_dev);
	if (mpls_pkt_too_big(skb, mtu - new_header_size))
		goto drop;

	hh_len = LL_RESERVED_SPACE(out_dev);
	if (!out_dev->header_ops)
		hh_len = 0;

	/* Ensure there is enough space for the headers in the skb */
	if (skb_cow(skb, hh_len + new_header_size))
		goto drop;

	skb_set_inner_protocol(skb, skb->protocol);
	skb_reset_inner_network_header(skb);

	skb_push(skb, new_header_size);

	skb_reset_network_header(skb);

	skb->dev = out_dev;
	skb->protocol = htons(ETH_P_MPLS_UC);

	/* Push the new labels */
	hdr = mpls_hdr(skb);
	bos = true;
	for (i = tun_encap_info->labels - 1; i >= 0; i--) {
		hdr[i] = mpls_entry_encode(tun_encap_info->label[i],
					   ttl, 0, bos);
		bos = false;
	}

	mpls_stats_inc_outucastpkts(out_dev, skb);

	if (rt)
		err = neigh_xmit(NEIGH_ARP_TABLE, out_dev, &rt->rt_gateway,
				 skb);
	else if (rt6)
		err = neigh_xmit(NEIGH_ND_TABLE, out_dev, &rt6->rt6i_gateway,
				 skb);
	if (err)
		net_dbg_ratelimited("%s: packet transmission failed: %d\n",
				    __func__, err);

	return LWTUNNEL_XMIT_DONE;

drop:
	out_mdev = out_dev ? mpls_dev_get(out_dev) : NULL;
	if (out_mdev)
		MPLS_INC_STATS(out_mdev, tx_errors);
	kfree_skb(skb);
	return -EINVAL;
}

static int mpls_build_state(struct nlattr *nla,
			    unsigned int family, const void *cfg,
			    struct lwtunnel_state **ts,
			    struct netlink_ext_ack *extack)
{
	struct mpls_iptunnel_encap *tun_encap_info;
	struct nlattr *tb[MPLS_IPTUNNEL_MAX + 1];
	struct lwtunnel_state *newts;
	u8 n_labels;
	int ret;

	ret = nla_parse_nested(tb, MPLS_IPTUNNEL_MAX, nla,
			       mpls_iptunnel_policy, extack);
	if (ret < 0)
		return ret;

	if (!tb[MPLS_IPTUNNEL_DST]) {
		NL_SET_ERR_MSG(extack, "MPLS_IPTUNNEL_DST attribute is missing");
		return -EINVAL;
	}

	/* determine number of labels */
	if (nla_get_labels(tb[MPLS_IPTUNNEL_DST], MAX_NEW_LABELS,
			   &n_labels, NULL, extack))
		return -EINVAL;

	newts = lwtunnel_state_alloc(sizeof(*tun_encap_info) +
				     n_labels * sizeof(u32));
	if (!newts)
		return -ENOMEM;

	tun_encap_info = mpls_lwtunnel_encap(newts);
	ret = nla_get_labels(tb[MPLS_IPTUNNEL_DST], n_labels,
			     &tun_encap_info->labels, tun_encap_info->label,
			     extack);
	if (ret)
		goto errout;

	tun_encap_info->ttl_propagate = MPLS_TTL_PROP_DEFAULT;

	if (tb[MPLS_IPTUNNEL_TTL]) {
		tun_encap_info->default_ttl = nla_get_u8(tb[MPLS_IPTUNNEL_TTL]);
		/* TTL 0 implies propagate from IP header */
		tun_encap_info->ttl_propagate = tun_encap_info->default_ttl ?
			MPLS_TTL_PROP_DISABLED :
			MPLS_TTL_PROP_ENABLED;
	}

	newts->type = LWTUNNEL_ENCAP_MPLS;
	newts->flags |= LWTUNNEL_STATE_XMIT_REDIRECT;
	newts->headroom = mpls_encap_size(tun_encap_info);

	*ts = newts;

	return 0;

errout:
	kfree(newts);
	*ts = NULL;

	return ret;
}

static int mpls_fill_encap_info(struct sk_buff *skb,
				struct lwtunnel_state *lwtstate)
{
	struct mpls_iptunnel_encap *tun_encap_info;
	
	tun_encap_info = mpls_lwtunnel_encap(lwtstate);

	if (nla_put_labels(skb, MPLS_IPTUNNEL_DST, tun_encap_info->labels,
			   tun_encap_info->label))
		goto nla_put_failure;

	if (tun_encap_info->ttl_propagate != MPLS_TTL_PROP_DEFAULT &&
	    nla_put_u8(skb, MPLS_IPTUNNEL_TTL, tun_encap_info->default_ttl))
		goto nla_put_failure;

	return 0;

nla_put_failure:
	return -EMSGSIZE;
}

static int mpls_encap_nlsize(struct lwtunnel_state *lwtstate)
{
	struct mpls_iptunnel_encap *tun_encap_info;
	int nlsize;

	tun_encap_info = mpls_lwtunnel_encap(lwtstate);

	nlsize = nla_total_size(tun_encap_info->labels * 4);

	if (tun_encap_info->ttl_propagate != MPLS_TTL_PROP_DEFAULT)
		nlsize += nla_total_size(1);

	return nlsize;
}

static int mpls_encap_cmp(struct lwtunnel_state *a, struct lwtunnel_state *b)
{
	struct mpls_iptunnel_encap *a_hdr = mpls_lwtunnel_encap(a);
	struct mpls_iptunnel_encap *b_hdr = mpls_lwtunnel_encap(b);
	int l;

	if (a_hdr->labels != b_hdr->labels ||
	    a_hdr->ttl_propagate != b_hdr->ttl_propagate ||
	    a_hdr->default_ttl != b_hdr->default_ttl)
		return 1;

	for (l = 0; l < a_hdr->labels; l++)
		if (a_hdr->label[l] != b_hdr->label[l])
			return 1;
	return 0;
}

static const struct lwtunnel_encap_ops mpls_iptun_ops = {
	.build_state = mpls_build_state,
	.xmit = mpls_xmit,
	.fill_encap = mpls_fill_encap_info,
	.get_encap_size = mpls_encap_nlsize,
	.cmp_encap = mpls_encap_cmp,
	.owner = THIS_MODULE,
};

static int __init mpls_iptunnel_init(void)
{
	return lwtunnel_encap_add_ops(&mpls_iptun_ops, LWTUNNEL_ENCAP_MPLS);
}
module_init(mpls_iptunnel_init);

static void __exit mpls_iptunnel_exit(void)
{
	lwtunnel_encap_del_ops(&mpls_iptun_ops, LWTUNNEL_ENCAP_MPLS);
}
module_exit(mpls_iptunnel_exit);

MODULE_ALIAS_RTNL_LWT(MPLS);
MODULE_DESCRIPTION("MultiProtocol Label Switching IP Tunnels");
MODULE_LICENSE("GPL v2");
Commit	Line	Data
e3e4712e RP	1	/*
	2	* mpls tunnels An implementation mpls tunnels using the light weight tunnel
	3	* infrastructure
	4	*
	5	* Authors: Roopa Prabhu, <roopa@cumulusnetworks.com>
	6	*
	7	* This program is free software; you can redistribute it and/or
	8	* modify it under the terms of the GNU General Public License
	9	* as published by the Free Software Foundation; either version
	10	* 2 of the License, or (at your option) any later version.
	11	*
	12	*/
	13	#include <linux/types.h>
	14	#include <linux/skbuff.h>
	15	#include <linux/net.h>
	16	#include <linux/module.h>
	17	#include <linux/mpls.h>
	18	#include <linux/vmalloc.h>
	19	#include <net/ip.h>
	20	#include <net/dst.h>
	21	#include <net/lwtunnel.h>
	22	#include <net/netevent.h>
	23	#include <net/netns/generic.h>
	24	#include <net/ip6_fib.h>
	25	#include <net/route.h>
	26	#include <net/mpls_iptunnel.h>
	27	#include <linux/mpls_iptunnel.h>
	28	#include "internal.h"
	29
	30	static const struct nla_policy mpls_iptunnel_policy[MPLS_IPTUNNEL_MAX + 1] = {
7d282163	31	[MPLS_IPTUNNEL_DST] = { .len = sizeof(u32) },
a59166e4	32	[MPLS_IPTUNNEL_TTL] = { .type = NLA_U8 },
e3e4712e RP	33	};
	34
	35	static unsigned int mpls_encap_size(struct mpls_iptunnel_encap *en)
	36	{
	37	/* The size of the layer 2.5 labels to be added for this route */
	38	return en->labels * sizeof(struct mpls_shim_hdr);
	39	}
	40
14972cbd	41	static int mpls_xmit(struct sk_buff *skb)
e3e4712e RP	42	{
	43	struct mpls_iptunnel_encap *tun_encap_info;
	44	struct mpls_shim_hdr *hdr;
	45	struct net_device *out_dev;
	46	unsigned int hh_len;
	47	unsigned int new_header_size;
	48	unsigned int mtu;
	49	struct dst_entry *dst = skb_dst(skb);
	50	struct rtable *rt = NULL;
	51	struct rt6_info *rt6 = NULL;
27d69105	52	struct mpls_dev *out_mdev;
a59166e4	53	struct net *net;
e3e4712e RP	54	int err = 0;
	55	bool bos;
	56	int i;
	57	unsigned int ttl;
	58
27d69105 RS	59	/* Find the output device */
27d69105 RS	60	out_dev = dst->dev;
a59166e4	61	net = dev_net(out_dev);
e3e4712e RP	62
	63	skb_orphan(skb);
	64
e3e4712e	65	if (!mpls_output_possible(out_dev) \|\|
61adedf3	66	!dst->lwtstate \|\| skb_warn_if_lro(skb))
e3e4712e RP	67	goto drop;
	68
	69	skb_forward_csum(skb);
	70
61adedf3	71	tun_encap_info = mpls_lwtunnel_encap(dst->lwtstate);
e3e4712e	72
a59166e4 RS	73	/* Obtain the ttl using the following set of rules.
	74	*
	75	* LWT ttl propagation setting:
	76	* - disabled => use default TTL value from LWT
	77	* - enabled => use TTL value from IPv4/IPv6 header
	78	* - default =>
	79	* Global ttl propagation setting:
	80	* - disabled => use default TTL value from global setting
	81	* - enabled => use TTL value from IPv4/IPv6 header
	82	*/
	83	if (dst->ops->family == AF_INET) {
	84	if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DISABLED)
	85	ttl = tun_encap_info->default_ttl;
	86	else if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DEFAULT &&
	87	!net->mpls.ip_ttl_propagate)
	88	ttl = net->mpls.default_ttl;
	89	else
	90	ttl = ip_hdr(skb)->ttl;
	91	rt = (struct rtable *)dst;
	92	} else if (dst->ops->family == AF_INET6) {
	93	if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DISABLED)
	94	ttl = tun_encap_info->default_ttl;
	95	else if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DEFAULT &&
	96	!net->mpls.ip_ttl_propagate)
	97	ttl = net->mpls.default_ttl;
	98	else
	99	ttl = ipv6_hdr(skb)->hop_limit;
	100	rt6 = (struct rt6_info *)dst;
	101	} else {
	102	goto drop;
	103	}
	104
e3e4712e RP	105	/* Verify the destination can hold the packet */
	106	new_header_size = mpls_encap_size(tun_encap_info);
	107	mtu = mpls_dev_mtu(out_dev);
	108	if (mpls_pkt_too_big(skb, mtu - new_header_size))
	109	goto drop;
	110
	111	hh_len = LL_RESERVED_SPACE(out_dev);
	112	if (!out_dev->header_ops)
	113	hh_len = 0;
	114
	115	/* Ensure there is enough space for the headers in the skb */
	116	if (skb_cow(skb, hh_len + new_header_size))
	117	goto drop;
	118
48d2ab60 DA	119	skb_set_inner_protocol(skb, skb->protocol);
	120	skb_reset_inner_network_header(skb);
	121
e3e4712e	122	skb_push(skb, new_header_size);
48d2ab60	123
e3e4712e RP	124	skb_reset_network_header(skb);
	125
	126	skb->dev = out_dev;
	127	skb->protocol = htons(ETH_P_MPLS_UC);
	128
	129	/* Push the new labels */
	130	hdr = mpls_hdr(skb);
	131	bos = true;
	132	for (i = tun_encap_info->labels - 1; i >= 0; i--) {
	133	hdr[i] = mpls_entry_encode(tun_encap_info->label[i],
	134	ttl, 0, bos);
	135	bos = false;
	136	}
	137
27d69105 RS	138	mpls_stats_inc_outucastpkts(out_dev, skb);
27d69105 RS	139
e3e4712e RP	140	if (rt)
	141	err = neigh_xmit(NEIGH_ARP_TABLE, out_dev, &rt->rt_gateway,
	142	skb);
	143	else if (rt6)
	144	err = neigh_xmit(NEIGH_ND_TABLE, out_dev, &rt6->rt6i_gateway,
	145	skb);
	146	if (err)
	147	net_dbg_ratelimited("%s: packet transmission failed: %d\n",
	148	__func__, err);
	149
14972cbd	150	return LWTUNNEL_XMIT_DONE;
e3e4712e RP	151
e3e4712e RP	152	drop:
27d69105 RS	153	out_mdev = out_dev ? mpls_dev_get(out_dev) : NULL;
	154	if (out_mdev)
	155	MPLS_INC_STATS(out_mdev, tx_errors);
e3e4712e RP	156	kfree_skb(skb);
	157	return -EINVAL;
	158	}
	159
30357d7d	160	static int mpls_build_state(struct nlattr *nla,
127eb7cd	161	unsigned int family, const void *cfg,
9ae28727 DA	162	struct lwtunnel_state **ts,
9ae28727 DA	163	struct netlink_ext_ack *extack)
e3e4712e RP	164	{
	165	struct mpls_iptunnel_encap *tun_encap_info;
	166	struct nlattr *tb[MPLS_IPTUNNEL_MAX + 1];
	167	struct lwtunnel_state *newts;
1511009c	168	u8 n_labels;
e3e4712e RP	169	int ret;
	170
	171	ret = nla_parse_nested(tb, MPLS_IPTUNNEL_MAX, nla,
9ae28727	172	mpls_iptunnel_policy, extack);
e3e4712e RP	173	if (ret < 0)
	174	return ret;
	175
a1f10abe DA	176	if (!tb[MPLS_IPTUNNEL_DST]) {
a1f10abe DA	177	NL_SET_ERR_MSG(extack, "MPLS_IPTUNNEL_DST attribute is missing");
e3e4712e	178	return -EINVAL;
a1f10abe	179	}
e3e4712e	180
1511009c	181	/* determine number of labels */
a1f10abe DA	182	if (nla_get_labels(tb[MPLS_IPTUNNEL_DST], MAX_NEW_LABELS,
a1f10abe DA	183	&n_labels, NULL, extack))
1511009c DA	184	return -EINVAL;
	185
	186	newts = lwtunnel_state_alloc(sizeof(*tun_encap_info) +
	187	n_labels * sizeof(u32));
e3e4712e RP	188	if (!newts)
	189	return -ENOMEM;
	190
e3e4712e	191	tun_encap_info = mpls_lwtunnel_encap(newts);
1511009c	192	ret = nla_get_labels(tb[MPLS_IPTUNNEL_DST], n_labels,
a1f10abe DA	193	&tun_encap_info->labels, tun_encap_info->label,
a1f10abe DA	194	extack);
e3e4712e RP	195	if (ret)
e3e4712e RP	196	goto errout;
a59166e4 RS	197
	198	tun_encap_info->ttl_propagate = MPLS_TTL_PROP_DEFAULT;
	199
	200	if (tb[MPLS_IPTUNNEL_TTL]) {
	201	tun_encap_info->default_ttl = nla_get_u8(tb[MPLS_IPTUNNEL_TTL]);
	202	/* TTL 0 implies propagate from IP header */
	203	tun_encap_info->ttl_propagate = tun_encap_info->default_ttl ?
	204	MPLS_TTL_PROP_DISABLED :
	205	MPLS_TTL_PROP_ENABLED;
	206	}
	207
e3e4712e	208	newts->type = LWTUNNEL_ENCAP_MPLS;
14972cbd RP	209	newts->flags \|= LWTUNNEL_STATE_XMIT_REDIRECT;
14972cbd RP	210	newts->headroom = mpls_encap_size(tun_encap_info);
e3e4712e RP	211
	212	*ts = newts;
	213
	214	return 0;
	215
	216	errout:
	217	kfree(newts);
	218	*ts = NULL;
	219
	220	return ret;
	221	}
	222
	223	static int mpls_fill_encap_info(struct sk_buff *skb,
	224	struct lwtunnel_state *lwtstate)
	225	{
	226	struct mpls_iptunnel_encap *tun_encap_info;
	227
	228	tun_encap_info = mpls_lwtunnel_encap(lwtstate);
	229
	230	if (nla_put_labels(skb, MPLS_IPTUNNEL_DST, tun_encap_info->labels,
	231	tun_encap_info->label))
	232	goto nla_put_failure;
	233
a59166e4 RS	234	if (tun_encap_info->ttl_propagate != MPLS_TTL_PROP_DEFAULT &&
	235	nla_put_u8(skb, MPLS_IPTUNNEL_TTL, tun_encap_info->default_ttl))
	236	goto nla_put_failure;
	237
e3e4712e RP	238	return 0;
	239
	240	nla_put_failure:
	241	return -EMSGSIZE;
	242	}
	243
	244	static int mpls_encap_nlsize(struct lwtunnel_state *lwtstate)
	245	{
	246	struct mpls_iptunnel_encap *tun_encap_info;
a59166e4	247	int nlsize;
e3e4712e RP	248
	249	tun_encap_info = mpls_lwtunnel_encap(lwtstate);
	250
a59166e4 RS	251	nlsize = nla_total_size(tun_encap_info->labels * 4);
	252
	253	if (tun_encap_info->ttl_propagate != MPLS_TTL_PROP_DEFAULT)
	254	nlsize += nla_total_size(1);
	255
	256	return nlsize;
e3e4712e RP	257	}
	258
	259	static int mpls_encap_cmp(struct lwtunnel_state a, struct lwtunnel_state b)
	260	{
	261	struct mpls_iptunnel_encap *a_hdr = mpls_lwtunnel_encap(a);
	262	struct mpls_iptunnel_encap *b_hdr = mpls_lwtunnel_encap(b);
	263	int l;
	264
a59166e4 RS	265	if (a_hdr->labels != b_hdr->labels \|\|
	266	a_hdr->ttl_propagate != b_hdr->ttl_propagate \|\|
	267	a_hdr->default_ttl != b_hdr->default_ttl)
e3e4712e RP	268	return 1;
e3e4712e RP	269
1511009c	270	for (l = 0; l < a_hdr->labels; l++)
e3e4712e RP	271	if (a_hdr->label[l] != b_hdr->label[l])
	272	return 1;
	273	return 0;
	274	}
	275
	276	static const struct lwtunnel_encap_ops mpls_iptun_ops = {
	277	.build_state = mpls_build_state,
14972cbd	278	.xmit = mpls_xmit,
e3e4712e RP	279	.fill_encap = mpls_fill_encap_info,
	280	.get_encap_size = mpls_encap_nlsize,
	281	.cmp_encap = mpls_encap_cmp,
88ff7334	282	.owner = THIS_MODULE,
e3e4712e RP	283	};
	284
	285	static int __init mpls_iptunnel_init(void)
	286	{
	287	return lwtunnel_encap_add_ops(&mpls_iptun_ops, LWTUNNEL_ENCAP_MPLS);
	288	}
	289	module_init(mpls_iptunnel_init);
	290
	291	static void __exit mpls_iptunnel_exit(void)
	292	{
	293	lwtunnel_encap_del_ops(&mpls_iptun_ops, LWTUNNEL_ENCAP_MPLS);
	294	}
	295	module_exit(mpls_iptunnel_exit);
	296
b2b04edc	297	MODULE_ALIAS_RTNL_LWT(MPLS);
e3e4712e RP	298	MODULE_DESCRIPTION("MultiProtocol Label Switching IP Tunnels");
e3e4712e RP	299	MODULE_LICENSE("GPL v2");