diff mbox series

[nf-next,2/3] netfilter: Add support for IPv6 segment routing 'SEG6' target

Message ID 20180816145143.12963-2-amsalam20@gmail.com
State Changes Requested
Delegated to: Pablo Neira
Headers show
Series [nf-next,1/3] netfilter: export SRH processing functions from seg6local | expand

Commit Message

Ahmed Abdelsalam Aug. 16, 2018, 2:51 p.m. UTC
Service Function Chaining (SFC) is one of the main use-cases of
IPv6 Segment Routing (SRv6) [1].

The Segment Routing Header (SRH) allows including a list of
segments in the IPv6 packet. This segment list can be used
to steer the packetthough a set of Virtual Network Functions
(VNFS) e.g., Firewall, DPI, IDS.

The edge node (SR ingress node) classifies the traffic and
consequently includes the segment lists in the IPv6 packet.

The matching of SRH previous, next and last SIDs has been
 added to ip6tables in kernel 4.16.  Once a packet matches
one of these options, we can perform some SR specific actions.

The IPv6 segment routing 'SEG6' target supports a set of
SR-specific actions.

The first set of actions include those actions necessary to
allow branching within a pre-defined SRv6 SFC policy instead
of the usual linear exploration of the VNF chain.

This patch implements the following SR-specific actions:
(1) go-next: sends packets towards the next SID from SRH.
(2) skip-next: skips the next SID in the SRH.
(3) go-last: skips the remaining part of the segment list
and send  the packet towards the last segment.

[1] https://tools.ietf.org/html/draft-xuclad-spring-sr-service-programming-00

Signed-off-by: Ahmed Abdelsalam <amsalam20@gmail.com>
---
 include/uapi/linux/netfilter_ipv6/ip6t_SEG6.h |  22 +++++
 net/ipv6/netfilter/Kconfig                    |  15 ++++
 net/ipv6/netfilter/Makefile                   |   1 +
 net/ipv6/netfilter/ip6t_SEG6.c                | 116 ++++++++++++++++++++++++++
 4 files changed, 154 insertions(+)
 create mode 100644 include/uapi/linux/netfilter_ipv6/ip6t_SEG6.h
 create mode 100644 net/ipv6/netfilter/ip6t_SEG6.c

Comments

Pablo Neira Ayuso Aug. 16, 2018, 6:25 p.m. UTC | #1
On Thu, Aug 16, 2018 at 04:51:42PM +0200, Ahmed Abdelsalam wrote:
[...]
> diff --git a/net/ipv6/netfilter/Kconfig b/net/ipv6/netfilter/Kconfig
> index 339d0762b027..a2502c54a837 100644
> --- a/net/ipv6/netfilter/Kconfig
> +++ b/net/ipv6/netfilter/Kconfig
> @@ -344,6 +344,21 @@ config IP6_NF_TARGET_NPT
>  
>  endif # IP6_NF_NAT
>  
> +if IPV6_SEG6_LWTUNNEL

Given this depends on the lwtunnel infra, could you instead extend
net/netfilter/nft_tunnel.c to support seg6 from there?

Please, see:

https://marc.info/?l=netfilter-devel&m=153323641010277&w=2
https://marc.info/?l=netfilter-devel&m=153323641410281&w=2
https://marc.info/?l=netfilter-devel&m=153323641210280&w=2

Thanks.
Ahmed Abdelsalam Aug. 16, 2018, 8:16 p.m. UTC | #2
We depend on IPV6_SEG6_LWTUNNEL only to be sure that three sympols 
seg6_get_srh(), seg6_advance_nextseg(), and seg6_lookup_nexthop exist. 

On the other hand, the actions go-next, skip-next, and go-last are
based the ip6tables SRH match of previous, next and last sid. 

https://patchwork.ozlabs.org/patch/910693/
https://patchwork.ozlabs.org/patch/904107/
https://patchwork.ozlabs.org/patch/904108/

With the combination of the srh matching options 
and the seg6 target actions it's possible to branching
within a SFC. 

Thanks, 
Ahmed 

On Thu, 16 Aug 2018 20:25:13 +0200
Pablo Neira Ayuso <pablo@netfilter.org> wrote:

> On Thu, Aug 16, 2018 at 04:51:42PM +0200, Ahmed Abdelsalam wrote:
> [...]
> > diff --git a/net/ipv6/netfilter/Kconfig b/net/ipv6/netfilter/Kconfig
> > index 339d0762b027..a2502c54a837 100644
> > --- a/net/ipv6/netfilter/Kconfig
> > +++ b/net/ipv6/netfilter/Kconfig
> > @@ -344,6 +344,21 @@ config IP6_NF_TARGET_NPT
> >  
> >  endif # IP6_NF_NAT
> >  
> > +if IPV6_SEG6_LWTUNNEL
> 
> Given this depends on the lwtunnel infra, could you instead extend
> net/netfilter/nft_tunnel.c to support seg6 from there?
> 
> Please, see:
> 
> https://marc.info/?l=netfilter-devel&m=153323641010277&w=2
> https://marc.info/?l=netfilter-devel&m=153323641410281&w=2
> https://marc.info/?l=netfilter-devel&m=153323641210280&w=2
> 
> Thanks.
Florian Westphal Aug. 21, 2018, 4:13 p.m. UTC | #3
Ahmed Abdelsalam <amsalam20@gmail.com> wrote:
> +static int seg6_check(const struct xt_tgchk_param *par)
> +{
> +	/**
> +	 * In the future, some new action may require using
> +	 * this function for doing some checks
> +	 */

Please add add a check on seg6->action and return -EOPNOSTUPP
in case it cannot be handled.

If possible, also sanity check other target parameters.
Otherwise newer userspace version can't be sure kernel
will actually know how to act on whatever action they've requested.

This will also allow you to remove the pr_err() call from the
evaluation function.
Ahmed Abdelsalam Aug. 21, 2018, 11 p.m. UTC | #4
Thanks Florian!
I'll send a v2 addressing the comments. 
Ahmed 

On Tue, 21 Aug 2018 18:13:08 +0200
Florian Westphal <fw@strlen.de> wrote:

> Ahmed Abdelsalam <amsalam20@gmail.com> wrote:
> > +static int seg6_check(const struct xt_tgchk_param *par)
> > +{
> > +	/**
> > +	 * In the future, some new action may require using
> > +	 * this function for doing some checks
> > +	 */
> 
> Please add add a check on seg6->action and return -EOPNOSTUPP
> in case it cannot be handled.
> 
> If possible, also sanity check other target parameters.
> Otherwise newer userspace version can't be sure kernel
> will actually know how to act on whatever action they've requested.
> 
> This will also allow you to remove the pr_err() call from the
> evaluation function.
diff mbox series

Patch

diff --git a/include/uapi/linux/netfilter_ipv6/ip6t_SEG6.h b/include/uapi/linux/netfilter_ipv6/ip6t_SEG6.h
new file mode 100644
index 000000000000..443805d0d1b2
--- /dev/null
+++ b/include/uapi/linux/netfilter_ipv6/ip6t_SEG6.h
@@ -0,0 +1,22 @@ 
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+
+#ifndef _IP6T_SEG6_H
+#define _IP6T_SEG6_H
+
+#include <linux/types.h>
+
+/* SEG6 action options */
+enum ip6t_seg6_action {
+	IP6T_SEG6_GO_NEXT,
+	IP6T_SEG6_SKIP_NEXT,
+	IP6T_SEG6_GO_LAST,
+	IP6T_SEG6_BSID,
+};
+
+struct ip6t_seg6_info {
+	__u32			action; /* SEG6 action */
+	struct	in6_addr	bsid;	/* SRv6 Bind SID */
+	unsigned int		tbl;	/* Routing table of bsid */
+};
+
+#endif /*_IP6T_SEG6_H*/
diff --git a/net/ipv6/netfilter/Kconfig b/net/ipv6/netfilter/Kconfig
index 339d0762b027..a2502c54a837 100644
--- a/net/ipv6/netfilter/Kconfig
+++ b/net/ipv6/netfilter/Kconfig
@@ -344,6 +344,21 @@  config IP6_NF_TARGET_NPT
 
 endif # IP6_NF_NAT
 
+if IPV6_SEG6_LWTUNNEL
+
+config IP6_NF_TARGET_SEG6
+	tristate 'IPv6 Segment Routing "SEG6" target support'
+	depends on NETFILTER_ADVANCED
+	help
+	 SEG6 is an ip6tables target for IPv6 Segment Routing encapsualted
+	 packets. It supports a set of Segment Routing specific actions
+	 that can be applied based on SRH information. It is useful for
+	 SRv6 Service Function chaining use-cases.
+
+	 To compile it as a module, choose M here.  If unsure, say N.
+
+endif # IPV6_SEG6_LWTUNNEL
+
 endif # IP6_NF_IPTABLES
 endmenu
 
diff --git a/net/ipv6/netfilter/Makefile b/net/ipv6/netfilter/Makefile
index 200c0c235565..68e86e18f15f 100644
--- a/net/ipv6/netfilter/Makefile
+++ b/net/ipv6/netfilter/Makefile
@@ -58,3 +58,4 @@  obj-$(CONFIG_IP6_NF_TARGET_MASQUERADE) += ip6t_MASQUERADE.o
 obj-$(CONFIG_IP6_NF_TARGET_NPT) += ip6t_NPT.o
 obj-$(CONFIG_IP6_NF_TARGET_REJECT) += ip6t_REJECT.o
 obj-$(CONFIG_IP6_NF_TARGET_SYNPROXY) += ip6t_SYNPROXY.o
+obj-$(CONFIG_IP6_NF_TARGET_SEG6) += ip6t_SEG6.o
diff --git a/net/ipv6/netfilter/ip6t_SEG6.c b/net/ipv6/netfilter/ip6t_SEG6.c
new file mode 100644
index 000000000000..03b5f1a1c8bf
--- /dev/null
+++ b/net/ipv6/netfilter/ip6t_SEG6.c
@@ -0,0 +1,116 @@ 
+/**
+ * IPv6 Segment Routing target module (SEG6).
+ *
+ * Author:
+ * Ahmed Abdelsalam <amsalam20@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.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/gfp.h>
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/icmpv6.h>
+#include <linux/netdevice.h>
+#include <linux/netfilter/x_tables.h>
+#include <linux/netfilter_ipv6/ip6_tables.h>
+#include <linux/netfilter_ipv6/ip6t_SEG6.h>
+
+#include <net/flow.h>
+#include <net/seg6.h>
+#include <net/ip6_route.h>
+
+static int seg6_go_next(struct sk_buff *skb, struct ipv6_sr_hdr *srh)
+{
+	if (srh->segments_left == 0)
+		return NF_DROP;
+	seg6_advance_nextseg(srh, &ipv6_hdr(skb)->daddr);
+	seg6_lookup_nexthop(skb, NULL, 0);
+	dst_input(skb);
+	return NF_STOLEN;
+}
+
+static int seg6_skip_next(struct sk_buff *skb, struct ipv6_sr_hdr *srh)
+{
+	if (srh->segments_left < 2)
+		return NF_DROP;
+	seg6_advance_nextseg(srh, &ipv6_hdr(skb)->daddr);
+	seg6_advance_nextseg(srh, &ipv6_hdr(skb)->daddr);
+	seg6_lookup_nexthop(skb, NULL, 0);
+	dst_input(skb);
+	return NF_STOLEN;
+}
+
+static int seg6_go_last(struct sk_buff *skb, struct ipv6_sr_hdr *srh)
+{
+	if (srh->segments_left == 0)
+		return NF_DROP;
+	srh->segments_left = 1;
+	seg6_advance_nextseg(srh, &ipv6_hdr(skb)->daddr);
+	seg6_lookup_nexthop(skb, NULL, 0);
+	dst_input(skb);
+	return NF_STOLEN;
+}
+
+static unsigned int
+seg6_tg6(struct sk_buff *skb, const struct xt_action_param *par)
+{
+	struct ipv6_sr_hdr *srh;
+	const struct ip6t_seg6_info *seg6 = par->targinfo;
+
+	srh = seg6_get_srh(skb);
+	if (!srh)
+		return NF_DROP;
+
+	switch (seg6->action) {
+	case IP6T_SEG6_GO_NEXT:
+		return seg6_go_next(skb, srh);
+	case IP6T_SEG6_SKIP_NEXT:
+		return seg6_skip_next(skb, srh);
+	case IP6T_SEG6_GO_LAST:
+		return seg6_go_last(skb, srh);
+	}
+
+	pr_err("Unknown SEG6 action\n");
+	return NF_DROP;
+}
+
+static int seg6_check(const struct xt_tgchk_param *par)
+{
+	/**
+	 * In the future, some new action may require using
+	 * this function for doing some checks
+	 */
+	return 0;
+}
+
+static struct xt_target seg6_tg6_reg __read_mostly = {
+	.name		= "SEG6",
+	.family		= NFPROTO_IPV6,
+	.target		= seg6_tg6,
+	.targetsize	= sizeof(struct ip6t_seg6_info),
+	.checkentry	= seg6_check,
+	.me		= THIS_MODULE
+};
+
+static int __init seg6_tg6_init(void)
+{
+	return xt_register_target(&seg6_tg6_reg);
+}
+
+static void __exit seg6_tg6_exit(void)
+{
+	xt_unregister_target(&seg6_tg6_reg);
+}
+
+module_init(seg6_tg6_init);
+module_exit(seg6_tg6_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Xtables: IPv6 Segment Routing Target (SEG6)");
+MODULE_AUTHOR("Ahmed Abdelsalam <amsalam20@gmail.com>");