[mirror_ubuntu-bionic-kernel.git] / include / net / inet_ecn.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _INET_ECN_H_
#define _INET_ECN_H_

#include <linux/ip.h>
#include <linux/skbuff.h>

#include <net/inet_sock.h>
#include <net/dsfield.h>

enum {
	INET_ECN_NOT_ECT = 0,
	INET_ECN_ECT_1 = 1,
	INET_ECN_ECT_0 = 2,
	INET_ECN_CE = 3,
	INET_ECN_MASK = 3,
};

extern int sysctl_tunnel_ecn_log;

static inline int INET_ECN_is_ce(__u8 dsfield)
{
	return (dsfield & INET_ECN_MASK) == INET_ECN_CE;
}

static inline int INET_ECN_is_not_ect(__u8 dsfield)
{
	return (dsfield & INET_ECN_MASK) == INET_ECN_NOT_ECT;
}

static inline int INET_ECN_is_capable(__u8 dsfield)
{
	return dsfield & INET_ECN_ECT_0;
}

/*
 * RFC 3168 9.1.1
 *  The full-functionality option for ECN encapsulation is to copy the
 *  ECN codepoint of the inside header to the outside header on
 *  encapsulation if the inside header is not-ECT or ECT, and to set the
 *  ECN codepoint of the outside header to ECT(0) if the ECN codepoint of
 *  the inside header is CE.
 */
static inline __u8 INET_ECN_encapsulate(__u8 outer, __u8 inner)
{
	outer &= ~INET_ECN_MASK;
	outer |= !INET_ECN_is_ce(inner) ? (inner & INET_ECN_MASK) :
					  INET_ECN_ECT_0;
	return outer;
}

static inline void INET_ECN_xmit(struct sock *sk)
{
	inet_sk(sk)->tos |= INET_ECN_ECT_0;
	if (inet6_sk(sk) != NULL)
		inet6_sk(sk)->tclass |= INET_ECN_ECT_0;
}

static inline void INET_ECN_dontxmit(struct sock *sk)
{
	inet_sk(sk)->tos &= ~INET_ECN_MASK;
	if (inet6_sk(sk) != NULL)
		inet6_sk(sk)->tclass &= ~INET_ECN_MASK;
}

#define IP6_ECN_flow_init(label) do {		\
      (label) &= ~htonl(INET_ECN_MASK << 20);	\
    } while (0)

#define	IP6_ECN_flow_xmit(sk, label) do {				\
	if (INET_ECN_is_capable(inet6_sk(sk)->tclass))			\
		(label) |= htonl(INET_ECN_ECT_0 << 20);			\
    } while (0)

static inline int IP_ECN_set_ce(struct iphdr *iph)
{
	u32 check = (__force u32)iph->check;
	u32 ecn = (iph->tos + 1) & INET_ECN_MASK;

	/*
	 * After the last operation we have (in binary):
	 * INET_ECN_NOT_ECT => 01
	 * INET_ECN_ECT_1   => 10
	 * INET_ECN_ECT_0   => 11
	 * INET_ECN_CE      => 00
	 */
	if (!(ecn & 2))
		return !ecn;

	/*
	 * The following gives us:
	 * INET_ECN_ECT_1 => check += htons(0xFFFD)
	 * INET_ECN_ECT_0 => check += htons(0xFFFE)
	 */
	check += (__force u16)htons(0xFFFB) + (__force u16)htons(ecn);

	iph->check = (__force __sum16)(check + (check>=0xFFFF));
	iph->tos |= INET_ECN_CE;
	return 1;
}

static inline void IP_ECN_clear(struct iphdr *iph)
{
	iph->tos &= ~INET_ECN_MASK;
}

static inline void ipv4_copy_dscp(unsigned int dscp, struct iphdr *inner)
{
	dscp &= ~INET_ECN_MASK;
	ipv4_change_dsfield(inner, INET_ECN_MASK, dscp);
}

struct ipv6hdr;

/* Note:
 * IP_ECN_set_ce() has to tweak IPV4 checksum when setting CE,
 * meaning both changes have no effect on skb->csum if/when CHECKSUM_COMPLETE
 * In IPv6 case, no checksum compensates the change in IPv6 header,
 * so we have to update skb->csum.
 */
static inline int IP6_ECN_set_ce(struct sk_buff *skb, struct ipv6hdr *iph)
{
	__be32 from, to;

	if (INET_ECN_is_not_ect(ipv6_get_dsfield(iph)))
		return 0;

	from = *(__be32 *)iph;
	to = from | htonl(INET_ECN_CE << 20);
	*(__be32 *)iph = to;
	if (skb->ip_summed == CHECKSUM_COMPLETE)
		skb->csum = csum_add(csum_sub(skb->csum, (__force __wsum)from),
				     (__force __wsum)to);
	return 1;
}

static inline void ipv6_copy_dscp(unsigned int dscp, struct ipv6hdr *inner)
{
	dscp &= ~INET_ECN_MASK;
	ipv6_change_dsfield(inner, INET_ECN_MASK, dscp);
}

static inline int INET_ECN_set_ce(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case cpu_to_be16(ETH_P_IP):
		if (skb_network_header(skb) + sizeof(struct iphdr) <=
		    skb_tail_pointer(skb))
			return IP_ECN_set_ce(ip_hdr(skb));
		break;

	case cpu_to_be16(ETH_P_IPV6):
		if (skb_network_header(skb) + sizeof(struct ipv6hdr) <=
		    skb_tail_pointer(skb))
			return IP6_ECN_set_ce(skb, ipv6_hdr(skb));
		break;
	}

	return 0;
}

/*
 * RFC 6040 4.2
 *  To decapsulate the inner header at the tunnel egress, a compliant
 *  tunnel egress MUST set the outgoing ECN field to the codepoint at the
 *  intersection of the appropriate arriving inner header (row) and outer
 *  header (column) in Figure 4
 *
 *      +---------+------------------------------------------------+
 *      |Arriving |            Arriving Outer Header               |
 *      |   Inner +---------+------------+------------+------------+
 *      |  Header | Not-ECT | ECT(0)     | ECT(1)     |     CE     |
 *      +---------+---------+------------+------------+------------+
 *      | Not-ECT | Not-ECT |Not-ECT(!!!)|Not-ECT(!!!)| <drop>(!!!)|
 *      |  ECT(0) |  ECT(0) | ECT(0)     | ECT(1)     |     CE     |
 *      |  ECT(1) |  ECT(1) | ECT(1) (!) | ECT(1)     |     CE     |
 *      |    CE   |      CE |     CE     |     CE(!!!)|     CE     |
 *      +---------+---------+------------+------------+------------+
 *
 *             Figure 4: New IP in IP Decapsulation Behaviour
 *
 *  returns 0 on success
 *          1 if something is broken and should be logged (!!! above)
 *          2 if packet should be dropped
 */
static inline int INET_ECN_decapsulate(struct sk_buff *skb,
				       __u8 outer, __u8 inner)
{
	if (INET_ECN_is_not_ect(inner)) {
		switch (outer & INET_ECN_MASK) {
		case INET_ECN_NOT_ECT:
			return 0;
		case INET_ECN_ECT_0:
		case INET_ECN_ECT_1:
			return 1;
		case INET_ECN_CE:
			return 2;
		}
	}

	if (INET_ECN_is_ce(outer))
		INET_ECN_set_ce(skb);

	return 0;
}

static inline int IP_ECN_decapsulate(const struct iphdr *oiph,
				     struct sk_buff *skb)
{
	__u8 inner;

	if (skb->protocol == htons(ETH_P_IP))
		inner = ip_hdr(skb)->tos;
	else if (skb->protocol == htons(ETH_P_IPV6))
		inner = ipv6_get_dsfield(ipv6_hdr(skb));
	else
		return 0;

	return INET_ECN_decapsulate(skb, oiph->tos, inner);
}

static inline int IP6_ECN_decapsulate(const struct ipv6hdr *oipv6h,
				      struct sk_buff *skb)
{
	__u8 inner;

	if (skb->protocol == htons(ETH_P_IP))
		inner = ip_hdr(skb)->tos;
	else if (skb->protocol == htons(ETH_P_IPV6))
		inner = ipv6_get_dsfield(ipv6_hdr(skb));
	else
		return 0;

	return INET_ECN_decapsulate(skb, ipv6_get_dsfield(oipv6h), inner);
}
#endif
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
1da177e4 LT	2	#ifndef _INET_ECN_H_
	3	#define _INET_ECN_H_
	4
	5	#include <linux/ip.h>
2566a509	6	#include <linux/skbuff.h>
14c85021 ACM	7
14c85021 ACM	8	#include <net/inet_sock.h>
1da177e4 LT	9	#include <net/dsfield.h>
	10
	11	enum {
	12	INET_ECN_NOT_ECT = 0,
	13	INET_ECN_ECT_1 = 1,
	14	INET_ECN_ECT_0 = 2,
	15	INET_ECN_CE = 3,
	16	INET_ECN_MASK = 3,
	17	};
	18
eccc1bb8	19	extern int sysctl_tunnel_ecn_log;
eccc1bb8	20
1da177e4 LT	21	static inline int INET_ECN_is_ce(__u8 dsfield)
	22	{
	23	return (dsfield & INET_ECN_MASK) == INET_ECN_CE;
	24	}
	25
	26	static inline int INET_ECN_is_not_ect(__u8 dsfield)
	27	{
	28	return (dsfield & INET_ECN_MASK) == INET_ECN_NOT_ECT;
	29	}
	30
	31	static inline int INET_ECN_is_capable(__u8 dsfield)
	32	{
a02cec21	33	return dsfield & INET_ECN_ECT_0;
1da177e4 LT	34	}
1da177e4 LT	35
b5d9c9c2 ED	36	/*
	37	* RFC 3168 9.1.1
	38	* The full-functionality option for ECN encapsulation is to copy the
	39	* ECN codepoint of the inside header to the outside header on
	40	* encapsulation if the inside header is not-ECT or ECT, and to set the
	41	* ECN codepoint of the outside header to ECT(0) if the ECN codepoint of
	42	* the inside header is CE.
	43	*/
1da177e4 LT	44	static inline __u8 INET_ECN_encapsulate(__u8 outer, __u8 inner)
	45	{
	46	outer &= ~INET_ECN_MASK;
	47	outer \|= !INET_ECN_is_ce(inner) ? (inner & INET_ECN_MASK) :
	48	INET_ECN_ECT_0;
	49	return outer;
	50	}
	51
ca067070 SG	52	static inline void INET_ECN_xmit(struct sock *sk)
	53	{
	54	inet_sk(sk)->tos \|= INET_ECN_ECT_0;
	55	if (inet6_sk(sk) != NULL)
	56	inet6_sk(sk)->tclass \|= INET_ECN_ECT_0;
	57	}
	58
	59	static inline void INET_ECN_dontxmit(struct sock *sk)
	60	{
	61	inet_sk(sk)->tos &= ~INET_ECN_MASK;
	62	if (inet6_sk(sk) != NULL)
	63	inet6_sk(sk)->tclass &= ~INET_ECN_MASK;
	64	}
1da177e4 LT	65
	66	#define IP6_ECN_flow_init(label) do { \
	67	(label) &= ~htonl(INET_ECN_MASK << 20); \
	68	} while (0)
	69
	70	#define IP6_ECN_flow_xmit(sk, label) do { \
e9df2e8f	71	if (INET_ECN_is_capable(inet6_sk(sk)->tclass)) \
95026cd2	72	(label) \|= htonl(INET_ECN_ECT_0 << 20); \
1da177e4 LT	73	} while (0)
1da177e4 LT	74
2566a509	75	static inline int IP_ECN_set_ce(struct iphdr *iph)
1da177e4	76	{
5c78f275	77	u32 check = (__force u32)iph->check;
1da177e4 LT	78	u32 ecn = (iph->tos + 1) & INET_ECN_MASK;
	79
	80	/*
	81	* After the last operation we have (in binary):
	82	* INET_ECN_NOT_ECT => 01
	83	* INET_ECN_ECT_1 => 10
	84	* INET_ECN_ECT_0 => 11
	85	* INET_ECN_CE => 00
	86	*/
	87	if (!(ecn & 2))
2566a509	88	return !ecn;
1da177e4 LT	89
	90	/*
	91	* The following gives us:
	92	* INET_ECN_ECT_1 => check += htons(0xFFFD)
	93	* INET_ECN_ECT_0 => check += htons(0xFFFE)
	94	*/
5c78f275	95	check += (__force u16)htons(0xFFFB) + (__force u16)htons(ecn);
1da177e4	96
5c78f275	97	iph->check = (__force __sum16)(check + (check>=0xFFFF));
1da177e4	98	iph->tos \|= INET_ECN_CE;
2566a509	99	return 1;
1da177e4 LT	100	}
	101
	102	static inline void IP_ECN_clear(struct iphdr *iph)
	103	{
	104	iph->tos &= ~INET_ECN_MASK;
	105	}
	106
29bb43b4	107	static inline void ipv4_copy_dscp(unsigned int dscp, struct iphdr *inner)
1da177e4	108	{
29bb43b4	109	dscp &= ~INET_ECN_MASK;
1da177e4 LT	110	ipv4_change_dsfield(inner, INET_ECN_MASK, dscp);
	111	}
	112
	113	struct ipv6hdr;
	114
34ae6a1a ED	115	/* Note:
	116	* IP_ECN_set_ce() has to tweak IPV4 checksum when setting CE,
	117	* meaning both changes have no effect on skb->csum if/when CHECKSUM_COMPLETE
	118	* In IPv6 case, no checksum compensates the change in IPv6 header,
	119	* so we have to update skb->csum.
	120	*/
	121	static inline int IP6_ECN_set_ce(struct sk_buff skb, struct ipv6hdr iph)
1da177e4	122	{
34ae6a1a ED	123	__be32 from, to;
34ae6a1a ED	124
1da177e4	125	if (INET_ECN_is_not_ect(ipv6_get_dsfield(iph)))
2566a509	126	return 0;
34ae6a1a ED	127
	128	from = (__be32 )iph;
	129	to = from \| htonl(INET_ECN_CE << 20);
	130	(__be32 )iph = to;
	131	if (skb->ip_summed == CHECKSUM_COMPLETE)
c15c0ab1 JB	132	skb->csum = csum_add(csum_sub(skb->csum, (__force __wsum)from),
c15c0ab1 JB	133	(__force __wsum)to);
2566a509	134	return 1;
1da177e4 LT	135	}
1da177e4 LT	136
29bb43b4	137	static inline void ipv6_copy_dscp(unsigned int dscp, struct ipv6hdr *inner)
1da177e4	138	{
29bb43b4	139	dscp &= ~INET_ECN_MASK;
1da177e4 LT	140	ipv6_change_dsfield(inner, INET_ECN_MASK, dscp);
	141	}
	142
2566a509 TG	143	static inline int INET_ECN_set_ce(struct sk_buff *skb)
	144	{
	145	switch (skb->protocol) {
f3a7c66b	146	case cpu_to_be16(ETH_P_IP):
ced14f68 SH	147	if (skb_network_header(skb) + sizeof(struct iphdr) <=
ced14f68 SH	148	skb_tail_pointer(skb))
eddc9ec5	149	return IP_ECN_set_ce(ip_hdr(skb));
2566a509 TG	150	break;
2566a509 TG	151
f3a7c66b	152	case cpu_to_be16(ETH_P_IPV6):
ced14f68 SH	153	if (skb_network_header(skb) + sizeof(struct ipv6hdr) <=
ced14f68 SH	154	skb_tail_pointer(skb))
34ae6a1a	155	return IP6_ECN_set_ce(skb, ipv6_hdr(skb));
2566a509 TG	156	break;
	157	}
	158
	159	return 0;
	160	}
	161
eccc1bb8	162	/*
d28071d1	163	* RFC 6040 4.2
eccc1bb8	164	* To decapsulate the inner header at the tunnel egress, a compliant
	165	* tunnel egress MUST set the outgoing ECN field to the codepoint at the
	166	* intersection of the appropriate arriving inner header (row) and outer
	167	* header (column) in Figure 4
	168	*
	169	* +---------+------------------------------------------------+
	170	* \|Arriving \| Arriving Outer Header \|
	171	* \| Inner +---------+------------+------------+------------+
	172	* \| Header \| Not-ECT \| ECT(0) \| ECT(1) \| CE \|
	173	* +---------+---------+------------+------------+------------+
	174	* \| Not-ECT \| Not-ECT \|Not-ECT(!!!)\|Not-ECT(!!!)\| <drop>(!!!)\|
	175	* \| ECT(0) \| ECT(0) \| ECT(0) \| ECT(1) \| CE \|
	176	* \| ECT(1) \| ECT(1) \| ECT(1) (!) \| ECT(1) \| CE \|
	177	* \| CE \| CE \| CE \| CE(!!!)\| CE \|
	178	* +---------+---------+------------+------------+------------+
	179	*
	180	* Figure 4: New IP in IP Decapsulation Behaviour
	181	*
	182	* returns 0 on success
	183	* 1 if something is broken and should be logged (!!! above)
	184	* 2 if packet should be dropped
	185	*/
	186	static inline int INET_ECN_decapsulate(struct sk_buff *skb,
	187	__u8 outer, __u8 inner)
	188	{
	189	if (INET_ECN_is_not_ect(inner)) {
	190	switch (outer & INET_ECN_MASK) {
	191	case INET_ECN_NOT_ECT:
	192	return 0;
	193	case INET_ECN_ECT_0:
	194	case INET_ECN_ECT_1:
	195	return 1;
	196	case INET_ECN_CE:
	197	return 2;
	198	}
	199	}
	200
	201	if (INET_ECN_is_ce(outer))
	202	INET_ECN_set_ce(skb);
	203
	204	return 0;
	205	}
	206
	207	static inline int IP_ECN_decapsulate(const struct iphdr *oiph,
	208	struct sk_buff *skb)
	209	{
	210	__u8 inner;
	211
	212	if (skb->protocol == htons(ETH_P_IP))
	213	inner = ip_hdr(skb)->tos;
	214	else if (skb->protocol == htons(ETH_P_IPV6))
	215	inner = ipv6_get_dsfield(ipv6_hdr(skb));
	216	else
	217	return 0;
	218
	219	return INET_ECN_decapsulate(skb, oiph->tos, inner);
	220	}
	221
	222	static inline int IP6_ECN_decapsulate(const struct ipv6hdr *oipv6h,
	223	struct sk_buff *skb)
	224	{
	225	__u8 inner;
	226
	227	if (skb->protocol == htons(ETH_P_IP))
228	inner = ip_hdr(skb)->tos;
229	else if (skb->protocol == htons(ETH_P_IPV6))
230	inner = ipv6_get_dsfield(ipv6_hdr(skb));
231	else
232	return 0;
233
234	return INET_ECN_decapsulate(skb, ipv6_get_dsfield(oipv6h), inner);
235	}
1da177e4	236	#endif