[linux-2.6-block.git] / net / ipv6 / ip6_offload.c

/*
 *	IPV6 GSO/GRO offload support
 *	Linux INET6 implementation
 *
 *	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/socket.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/printk.h>

#include <net/protocol.h>
#include <net/ipv6.h>
#include <net/inet_common.h>

#include "ip6_offload.h"

/* All GRO functions are always builtin, except UDP over ipv6, which lays in
 * ipv6 module, as it depends on UDPv6 lookup function, so we need special care
 * when ipv6 is built as a module
 */
#if IS_BUILTIN(CONFIG_IPV6)
#define INDIRECT_CALL_L4(f, f2, f1, ...) INDIRECT_CALL_2(f, f2, f1, __VA_ARGS__)
#else
#define INDIRECT_CALL_L4(f, f2, f1, ...) INDIRECT_CALL_1(f, f2, __VA_ARGS__)
#endif

#define indirect_call_gro_receive_l4(f2, f1, cb, head, skb)	\
({								\
	unlikely(gro_recursion_inc_test(skb)) ?			\
		NAPI_GRO_CB(skb)->flush |= 1, NULL :		\
		INDIRECT_CALL_L4(cb, f2, f1, head, skb);	\
})

static int ipv6_gso_pull_exthdrs(struct sk_buff *skb, int proto)
{
	const struct net_offload *ops = NULL;

	for (;;) {
		struct ipv6_opt_hdr *opth;
		int len;

		if (proto != NEXTHDR_HOP) {
			ops = rcu_dereference(inet6_offloads[proto]);

			if (unlikely(!ops))
				break;

			if (!(ops->flags & INET6_PROTO_GSO_EXTHDR))
				break;
		}

		if (unlikely(!pskb_may_pull(skb, 8)))
			break;

		opth = (void *)skb->data;
		len = ipv6_optlen(opth);

		if (unlikely(!pskb_may_pull(skb, len)))
			break;

		opth = (void *)skb->data;
		proto = opth->nexthdr;
		__skb_pull(skb, len);
	}

	return proto;
}

static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,
	netdev_features_t features)
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	struct ipv6hdr *ipv6h;
	const struct net_offload *ops;
	int proto;
	struct frag_hdr *fptr;
	unsigned int payload_len;
	u8 *prevhdr;
	int offset = 0;
	bool encap, udpfrag;
	int nhoff;
	bool gso_partial;

	skb_reset_network_header(skb);
	nhoff = skb_network_header(skb) - skb_mac_header(skb);
	if (unlikely(!pskb_may_pull(skb, sizeof(*ipv6h))))
		goto out;

	encap = SKB_GSO_CB(skb)->encap_level > 0;
	if (encap)
		features &= skb->dev->hw_enc_features;
	SKB_GSO_CB(skb)->encap_level += sizeof(*ipv6h);

	ipv6h = ipv6_hdr(skb);
	__skb_pull(skb, sizeof(*ipv6h));
	segs = ERR_PTR(-EPROTONOSUPPORT);

	proto = ipv6_gso_pull_exthdrs(skb, ipv6h->nexthdr);

	if (skb->encapsulation &&
	    skb_shinfo(skb)->gso_type & (SKB_GSO_IPXIP4 | SKB_GSO_IPXIP6))
		udpfrag = proto == IPPROTO_UDP && encap &&
			  (skb_shinfo(skb)->gso_type & SKB_GSO_UDP);
	else
		udpfrag = proto == IPPROTO_UDP && !skb->encapsulation &&
			  (skb_shinfo(skb)->gso_type & SKB_GSO_UDP);

	ops = rcu_dereference(inet6_offloads[proto]);
	if (likely(ops && ops->callbacks.gso_segment)) {
		skb_reset_transport_header(skb);
		segs = ops->callbacks.gso_segment(skb, features);
	}

	if (IS_ERR_OR_NULL(segs))
		goto out;

	gso_partial = !!(skb_shinfo(segs)->gso_type & SKB_GSO_PARTIAL);

	for (skb = segs; skb; skb = skb->next) {
		ipv6h = (struct ipv6hdr *)(skb_mac_header(skb) + nhoff);
		if (gso_partial && skb_is_gso(skb))
			payload_len = skb_shinfo(skb)->gso_size +
				      SKB_GSO_CB(skb)->data_offset +
				      skb->head - (unsigned char *)(ipv6h + 1);
		else
			payload_len = skb->len - nhoff - sizeof(*ipv6h);
		ipv6h->payload_len = htons(payload_len);
		skb->network_header = (u8 *)ipv6h - skb->head;
		skb_reset_mac_len(skb);

		if (udpfrag) {
			int err = ip6_find_1stfragopt(skb, &prevhdr);
			if (err < 0) {
				kfree_skb_list(segs);
				return ERR_PTR(err);
			}
			fptr = (struct frag_hdr *)((u8 *)ipv6h + err);
			fptr->frag_off = htons(offset);
			if (skb->next)
				fptr->frag_off |= htons(IP6_MF);
			offset += (ntohs(ipv6h->payload_len) -
				   sizeof(struct frag_hdr));
		}
		if (encap)
			skb_reset_inner_headers(skb);
	}

out:
	return segs;
}

/* Return the total length of all the extension hdrs, following the same
 * logic in ipv6_gso_pull_exthdrs() when parsing ext-hdrs.
 */
static int ipv6_exthdrs_len(struct ipv6hdr *iph,
			    const struct net_offload **opps)
{
	struct ipv6_opt_hdr *opth = (void *)iph;
	int len = 0, proto, optlen = sizeof(*iph);

	proto = iph->nexthdr;
	for (;;) {
		if (proto != NEXTHDR_HOP) {
			*opps = rcu_dereference(inet6_offloads[proto]);
			if (unlikely(!(*opps)))
				break;
			if (!((*opps)->flags & INET6_PROTO_GSO_EXTHDR))
				break;
		}
		opth = (void *)opth + optlen;
		optlen = ipv6_optlen(opth);
		len += optlen;
		proto = opth->nexthdr;
	}
	return len;
}

INDIRECT_CALLABLE_DECLARE(struct sk_buff *tcp6_gro_receive(struct list_head *,
							   struct sk_buff *));
INDIRECT_CALLABLE_DECLARE(struct sk_buff *udp6_gro_receive(struct list_head *,
							   struct sk_buff *));
INDIRECT_CALLABLE_SCOPE struct sk_buff *ipv6_gro_receive(struct list_head *head,
							 struct sk_buff *skb)
{
	const struct net_offload *ops;
	struct sk_buff *pp = NULL;
	struct sk_buff *p;
	struct ipv6hdr *iph;
	unsigned int nlen;
	unsigned int hlen;
	unsigned int off;
	u16 flush = 1;
	int proto;

	off = skb_gro_offset(skb);
	hlen = off + sizeof(*iph);
	iph = skb_gro_header_fast(skb, off);
	if (skb_gro_header_hard(skb, hlen)) {
		iph = skb_gro_header_slow(skb, hlen, off);
		if (unlikely(!iph))
			goto out;
	}

	skb_set_network_header(skb, off);
	skb_gro_pull(skb, sizeof(*iph));
	skb_set_transport_header(skb, skb_gro_offset(skb));

	flush += ntohs(iph->payload_len) != skb_gro_len(skb);

	rcu_read_lock();
	proto = iph->nexthdr;
	ops = rcu_dereference(inet6_offloads[proto]);
	if (!ops || !ops->callbacks.gro_receive) {
		__pskb_pull(skb, skb_gro_offset(skb));
		skb_gro_frag0_invalidate(skb);
		proto = ipv6_gso_pull_exthdrs(skb, proto);
		skb_gro_pull(skb, -skb_transport_offset(skb));
		skb_reset_transport_header(skb);
		__skb_push(skb, skb_gro_offset(skb));

		ops = rcu_dereference(inet6_offloads[proto]);
		if (!ops || !ops->callbacks.gro_receive)
			goto out_unlock;

		iph = ipv6_hdr(skb);
	}

	NAPI_GRO_CB(skb)->proto = proto;

	flush--;
	nlen = skb_network_header_len(skb);

	list_for_each_entry(p, head, list) {
		const struct ipv6hdr *iph2;
		__be32 first_word; /* <Version:4><Traffic_Class:8><Flow_Label:20> */

		if (!NAPI_GRO_CB(p)->same_flow)
			continue;

		iph2 = (struct ipv6hdr *)(p->data + off);
		first_word = *(__be32 *)iph ^ *(__be32 *)iph2;

		/* All fields must match except length and Traffic Class.
		 * XXX skbs on the gro_list have all been parsed and pulled
		 * already so we don't need to compare nlen
		 * (nlen != (sizeof(*iph2) + ipv6_exthdrs_len(iph2, &ops)))
		 * memcmp() alone below is sufficient, right?
		 */
		 if ((first_word & htonl(0xF00FFFFF)) ||
		    !ipv6_addr_equal(&iph->saddr, &iph2->saddr) ||
		    !ipv6_addr_equal(&iph->daddr, &iph2->daddr) ||
		    *(u16 *)&iph->nexthdr != *(u16 *)&iph2->nexthdr) {
not_same_flow:
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}
		if (unlikely(nlen > sizeof(struct ipv6hdr))) {
			if (memcmp(iph + 1, iph2 + 1,
				   nlen - sizeof(struct ipv6hdr)))
				goto not_same_flow;
		}
		/* flush if Traffic Class fields are different */
		NAPI_GRO_CB(p)->flush |= !!(first_word & htonl(0x0FF00000));
		NAPI_GRO_CB(p)->flush |= flush;

		/* If the previous IP ID value was based on an atomic
		 * datagram we can overwrite the value and ignore it.
		 */
		if (NAPI_GRO_CB(skb)->is_atomic)
			NAPI_GRO_CB(p)->flush_id = 0;
	}

	NAPI_GRO_CB(skb)->is_atomic = true;
	NAPI_GRO_CB(skb)->flush |= flush;

	skb_gro_postpull_rcsum(skb, iph, nlen);

	pp = indirect_call_gro_receive_l4(tcp6_gro_receive, udp6_gro_receive,
					 ops->callbacks.gro_receive, head, skb);

out_unlock:
	rcu_read_unlock();

out:
	skb_gro_flush_final(skb, pp, flush);

	return pp;
}

static struct sk_buff *sit_ip6ip6_gro_receive(struct list_head *head,
					      struct sk_buff *skb)
{
	/* Common GRO receive for SIT and IP6IP6 */

	if (NAPI_GRO_CB(skb)->encap_mark) {
		NAPI_GRO_CB(skb)->flush = 1;
		return NULL;
	}

	NAPI_GRO_CB(skb)->encap_mark = 1;

	return ipv6_gro_receive(head, skb);
}

static struct sk_buff *ip4ip6_gro_receive(struct list_head *head,
					  struct sk_buff *skb)
{
	/* Common GRO receive for SIT and IP6IP6 */

	if (NAPI_GRO_CB(skb)->encap_mark) {
		NAPI_GRO_CB(skb)->flush = 1;
		return NULL;
	}

	NAPI_GRO_CB(skb)->encap_mark = 1;

	return inet_gro_receive(head, skb);
}

INDIRECT_CALLABLE_DECLARE(int tcp6_gro_complete(struct sk_buff *, int));
INDIRECT_CALLABLE_DECLARE(int udp6_gro_complete(struct sk_buff *, int));
INDIRECT_CALLABLE_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff)
{
	const struct net_offload *ops;
	struct ipv6hdr *iph = (struct ipv6hdr *)(skb->data + nhoff);
	int err = -ENOSYS;

	if (skb->encapsulation) {
		skb_set_inner_protocol(skb, cpu_to_be16(ETH_P_IPV6));
		skb_set_inner_network_header(skb, nhoff);
	}

	iph->payload_len = htons(skb->len - nhoff - sizeof(*iph));

	rcu_read_lock();

	nhoff += sizeof(*iph) + ipv6_exthdrs_len(iph, &ops);
	if (WARN_ON(!ops || !ops->callbacks.gro_complete))
		goto out_unlock;

	err = INDIRECT_CALL_L4(ops->callbacks.gro_complete, tcp6_gro_complete,
			       udp6_gro_complete, skb, nhoff);

out_unlock:
	rcu_read_unlock();

	return err;
}

static int sit_gro_complete(struct sk_buff *skb, int nhoff)
{
	skb->encapsulation = 1;
	skb_shinfo(skb)->gso_type |= SKB_GSO_IPXIP4;
	return ipv6_gro_complete(skb, nhoff);
}

static int ip6ip6_gro_complete(struct sk_buff *skb, int nhoff)
{
	skb->encapsulation = 1;
	skb_shinfo(skb)->gso_type |= SKB_GSO_IPXIP6;
	return ipv6_gro_complete(skb, nhoff);
}

static int ip4ip6_gro_complete(struct sk_buff *skb, int nhoff)
{
	skb->encapsulation = 1;
	skb_shinfo(skb)->gso_type |= SKB_GSO_IPXIP6;
	return inet_gro_complete(skb, nhoff);
}

static struct packet_offload ipv6_packet_offload __read_mostly = {
	.type = cpu_to_be16(ETH_P_IPV6),
	.callbacks = {
		.gso_segment = ipv6_gso_segment,
		.gro_receive = ipv6_gro_receive,
		.gro_complete = ipv6_gro_complete,
	},
};

static struct sk_buff *sit_gso_segment(struct sk_buff *skb,
				       netdev_features_t features)
{
	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP4))
		return ERR_PTR(-EINVAL);

	return ipv6_gso_segment(skb, features);
}

static struct sk_buff *ip4ip6_gso_segment(struct sk_buff *skb,
					  netdev_features_t features)
{
	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP6))
		return ERR_PTR(-EINVAL);

	return inet_gso_segment(skb, features);
}

static struct sk_buff *ip6ip6_gso_segment(struct sk_buff *skb,
					  netdev_features_t features)
{
	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP6))
		return ERR_PTR(-EINVAL);

	return ipv6_gso_segment(skb, features);
}

static const struct net_offload sit_offload = {
	.callbacks = {
		.gso_segment	= sit_gso_segment,
		.gro_receive    = sit_ip6ip6_gro_receive,
		.gro_complete   = sit_gro_complete,
	},
};

static const struct net_offload ip4ip6_offload = {
	.callbacks = {
		.gso_segment	= ip4ip6_gso_segment,
		.gro_receive    = ip4ip6_gro_receive,
		.gro_complete   = ip4ip6_gro_complete,
	},
};

static const struct net_offload ip6ip6_offload = {
	.callbacks = {
		.gso_segment	= ip6ip6_gso_segment,
		.gro_receive    = sit_ip6ip6_gro_receive,
		.gro_complete   = ip6ip6_gro_complete,
	},
};
static int __init ipv6_offload_init(void)
{

	if (tcpv6_offload_init() < 0)
		pr_crit("%s: Cannot add TCP protocol offload\n", __func__);
	if (ipv6_exthdrs_offload_init() < 0)
		pr_crit("%s: Cannot add EXTHDRS protocol offload\n", __func__);

	dev_add_offload(&ipv6_packet_offload);

	inet_add_offload(&sit_offload, IPPROTO_IPV6);
	inet6_add_offload(&ip6ip6_offload, IPPROTO_IPV6);
	inet6_add_offload(&ip4ip6_offload, IPPROTO_IPIP);

	return 0;
}

fs_initcall(ipv6_offload_init);
Commit	Line	Data
d1da932e VY	1	/*
	2	* IPV6 GSO/GRO offload support
	3	* Linux INET6 implementation
	4	*
	5	* This program is free software; you can redistribute it and/or
	6	* modify it under the terms of the GNU General Public License
	7	* as published by the Free Software Foundation; either version
	8	* 2 of the License, or (at your option) any later version.
	9	*/
	10
	11	#include <linux/kernel.h>
	12	#include <linux/socket.h>
	13	#include <linux/netdevice.h>
	14	#include <linux/skbuff.h>
c6b641a4	15	#include <linux/printk.h>
d1da932e VY	16
	17	#include <net/protocol.h>
	18	#include <net/ipv6.h>
b8921ca8	19	#include <net/inet_common.h>
d1da932e VY	20
	21	#include "ip6_offload.h"
	22
028e0a47 PA	23	/* All GRO functions are always builtin, except UDP over ipv6, which lays in
	24	* ipv6 module, as it depends on UDPv6 lookup function, so we need special care
	25	* when ipv6 is built as a module
	26	*/
	27	#if IS_BUILTIN(CONFIG_IPV6)
	28	#define INDIRECT_CALL_L4(f, f2, f1, ...) INDIRECT_CALL_2(f, f2, f1, __VA_ARGS__)
	29	#else
	30	#define INDIRECT_CALL_L4(f, f2, f1, ...) INDIRECT_CALL_1(f, f2, __VA_ARGS__)
	31	#endif
	32
	33	#define indirect_call_gro_receive_l4(f2, f1, cb, head, skb) \
	34	({ \
	35	unlikely(gro_recursion_inc_test(skb)) ? \
	36	NAPI_GRO_CB(skb)->flush \|= 1, NULL : \
	37	INDIRECT_CALL_L4(cb, f2, f1, head, skb); \
	38	})
	39
d1da932e VY	40	static int ipv6_gso_pull_exthdrs(struct sk_buff *skb, int proto)
	41	{
	42	const struct net_offload *ops = NULL;
	43
	44	for (;;) {
	45	struct ipv6_opt_hdr *opth;
	46	int len;
	47
	48	if (proto != NEXTHDR_HOP) {
	49	ops = rcu_dereference(inet6_offloads[proto]);
	50
	51	if (unlikely(!ops))
	52	break;
	53
	54	if (!(ops->flags & INET6_PROTO_GSO_EXTHDR))
	55	break;
	56	}
	57
	58	if (unlikely(!pskb_may_pull(skb, 8)))
	59	break;
	60
	61	opth = (void *)skb->data;
	62	len = ipv6_optlen(opth);
	63
	64	if (unlikely(!pskb_may_pull(skb, len)))
	65	break;
	66
fc6fb41c	67	opth = (void *)skb->data;
d1da932e VY	68	proto = opth->nexthdr;
	69	__skb_pull(skb, len);
	70	}
	71
	72	return proto;
	73	}
	74
d1da932e VY	75	static struct sk_buff ipv6_gso_segment(struct sk_buff skb,
	76	netdev_features_t features)
	77	{
	78	struct sk_buff *segs = ERR_PTR(-EINVAL);
	79	struct ipv6hdr *ipv6h;
	80	const struct net_offload *ops;
	81	int proto;
	82	struct frag_hdr *fptr;
802ab55a	83	unsigned int payload_len;
d1da932e VY	84	u8 *prevhdr;
d1da932e VY	85	int offset = 0;
91a48a2e	86	bool encap, udpfrag;
d3e5e006	87	int nhoff;
07b26c94	88	bool gso_partial;
d1da932e	89
d3e5e006 ED	90	skb_reset_network_header(skb);
d3e5e006 ED	91	nhoff = skb_network_header(skb) - skb_mac_header(skb);
d1da932e VY	92	if (unlikely(!pskb_may_pull(skb, sizeof(*ipv6h))))
	93	goto out;
	94
91a48a2e HFS	95	encap = SKB_GSO_CB(skb)->encap_level > 0;
91a48a2e HFS	96	if (encap)
1e16aa3d	97	features &= skb->dev->hw_enc_features;
d3e5e006 ED	98	SKB_GSO_CB(skb)->encap_level += sizeof(*ipv6h);
d3e5e006 ED	99
d1da932e VY	100	ipv6h = ipv6_hdr(skb);
	101	__skb_pull(skb, sizeof(*ipv6h));
	102	segs = ERR_PTR(-EPROTONOSUPPORT);
	103
	104	proto = ipv6_gso_pull_exthdrs(skb, ipv6h->nexthdr);
b917eb15	105
91a48a2e	106	if (skb->encapsulation &&
7e13318d	107	skb_shinfo(skb)->gso_type & (SKB_GSO_IPXIP4 \| SKB_GSO_IPXIP6))
ee80d1eb WB	108	udpfrag = proto == IPPROTO_UDP && encap &&
ee80d1eb WB	109	(skb_shinfo(skb)->gso_type & SKB_GSO_UDP);
91a48a2e	110	else
ee80d1eb WB	111	udpfrag = proto == IPPROTO_UDP && !skb->encapsulation &&
ee80d1eb WB	112	(skb_shinfo(skb)->gso_type & SKB_GSO_UDP);
91a48a2e	113
d1da932e	114	ops = rcu_dereference(inet6_offloads[proto]);
f191a1d1	115	if (likely(ops && ops->callbacks.gso_segment)) {
d1da932e	116	skb_reset_transport_header(skb);
f191a1d1	117	segs = ops->callbacks.gso_segment(skb, features);
d1da932e	118	}
d1da932e	119
6b6ebb6b	120	if (IS_ERR_OR_NULL(segs))
d1da932e VY	121	goto out;
d1da932e VY	122
07b26c94 SK	123	gso_partial = !!(skb_shinfo(segs)->gso_type & SKB_GSO_PARTIAL);
07b26c94 SK	124
d1da932e	125	for (skb = segs; skb; skb = skb->next) {
d3e5e006	126	ipv6h = (struct ipv6hdr *)(skb_mac_header(skb) + nhoff);
3d0241d5	127	if (gso_partial && skb_is_gso(skb))
802ab55a AD	128	payload_len = skb_shinfo(skb)->gso_size +
	129	SKB_GSO_CB(skb)->data_offset +
	130	skb->head - (unsigned char *)(ipv6h + 1);
	131	else
	132	payload_len = skb->len - nhoff - sizeof(*ipv6h);
	133	ipv6h->payload_len = htons(payload_len);
d3e5e006	134	skb->network_header = (u8 *)ipv6h - skb->head;
c56cae23	135	skb_reset_mac_len(skb);
d3e5e006	136
91a48a2e	137	if (udpfrag) {
7dd7eb95	138	int err = ip6_find_1stfragopt(skb, &prevhdr);
e3e86b51 DM	139	if (err < 0) {
e3e86b51 DM	140	kfree_skb_list(segs);
7dd7eb95	141	return ERR_PTR(err);
e3e86b51	142	}
7dd7eb95	143	fptr = (struct frag_hdr )((u8 )ipv6h + err);
d1da932e	144	fptr->frag_off = htons(offset);
53b24b8f	145	if (skb->next)
d1da932e VY	146	fptr->frag_off \|= htons(IP6_MF);
	147	offset += (ntohs(ipv6h->payload_len) -
	148	sizeof(struct frag_hdr));
	149	}
91a48a2e HFS	150	if (encap)
91a48a2e HFS	151	skb_reset_inner_headers(skb);
d1da932e VY	152	}
	153
	154	out:
	155	return segs;
	156	}
	157
299603e8 JC	158	/* Return the total length of all the extension hdrs, following the same
	159	* logic in ipv6_gso_pull_exthdrs() when parsing ext-hdrs.
	160	*/
	161	static int ipv6_exthdrs_len(struct ipv6hdr *iph,
	162	const struct net_offload **opps)
	163	{
810c23a3 JC	164	struct ipv6_opt_hdr opth = (void )iph;
810c23a3 JC	165	int len = 0, proto, optlen = sizeof(*iph);
299603e8 JC	166
	167	proto = iph->nexthdr;
	168	for (;;) {
	169	if (proto != NEXTHDR_HOP) {
	170	*opps = rcu_dereference(inet6_offloads[proto]);
	171	if (unlikely(!(*opps)))
	172	break;
	173	if (!((*opps)->flags & INET6_PROTO_GSO_EXTHDR))
	174	break;
	175	}
810c23a3 JC	176	opth = (void *)opth + optlen;
810c23a3 JC	177	optlen = ipv6_optlen(opth);
299603e8 JC	178	len += optlen;
	179	proto = opth->nexthdr;
	180	}
	181	return len;
	182	}
	183
028e0a47 PA	184	INDIRECT_CALLABLE_DECLARE(struct sk_buff tcp6_gro_receive(struct list_head ,
	185	struct sk_buff *));
	186	INDIRECT_CALLABLE_DECLARE(struct sk_buff udp6_gro_receive(struct list_head ,
	187	struct sk_buff *));
aaa5d90b PA	188	INDIRECT_CALLABLE_SCOPE struct sk_buff ipv6_gro_receive(struct list_head head,
aaa5d90b PA	189	struct sk_buff *skb)
d1da932e VY	190	{
d1da932e VY	191	const struct net_offload *ops;
d4546c25	192	struct sk_buff *pp = NULL;
d1da932e VY	193	struct sk_buff *p;
	194	struct ipv6hdr *iph;
	195	unsigned int nlen;
	196	unsigned int hlen;
	197	unsigned int off;
bf5a755f	198	u16 flush = 1;
d1da932e	199	int proto;
d1da932e VY	200
	201	off = skb_gro_offset(skb);
	202	hlen = off + sizeof(*iph);
	203	iph = skb_gro_header_fast(skb, off);
	204	if (skb_gro_header_hard(skb, hlen)) {
	205	iph = skb_gro_header_slow(skb, hlen, off);
	206	if (unlikely(!iph))
	207	goto out;
	208	}
	209
299603e8	210	skb_set_network_header(skb, off);
d1da932e VY	211	skb_gro_pull(skb, sizeof(*iph));
	212	skb_set_transport_header(skb, skb_gro_offset(skb));
	213
	214	flush += ntohs(iph->payload_len) != skb_gro_len(skb);
	215
	216	rcu_read_lock();
	217	proto = iph->nexthdr;
	218	ops = rcu_dereference(inet6_offloads[proto]);
f191a1d1	219	if (!ops \|\| !ops->callbacks.gro_receive) {
d1da932e	220	__pskb_pull(skb, skb_gro_offset(skb));
57ea52a8	221	skb_gro_frag0_invalidate(skb);
d1da932e VY	222	proto = ipv6_gso_pull_exthdrs(skb, proto);
	223	skb_gro_pull(skb, -skb_transport_offset(skb));
	224	skb_reset_transport_header(skb);
	225	__skb_push(skb, skb_gro_offset(skb));
	226
	227	ops = rcu_dereference(inet6_offloads[proto]);
f191a1d1	228	if (!ops \|\| !ops->callbacks.gro_receive)
d1da932e VY	229	goto out_unlock;
	230
	231	iph = ipv6_hdr(skb);
	232	}
	233
	234	NAPI_GRO_CB(skb)->proto = proto;
	235
	236	flush--;
	237	nlen = skb_network_header_len(skb);
	238
d4546c25	239	list_for_each_entry(p, head, list) {
d1da932e VY	240	const struct ipv6hdr *iph2;
	241	__be32 first_word; /* <Version:4><Traffic_Class:8><Flow_Label:20> */
	242
	243	if (!NAPI_GRO_CB(p)->same_flow)
	244	continue;
	245
299603e8	246	iph2 = (struct ipv6hdr *)(p->data + off);
67ba4152	247	first_word = (__be32 )iph ^ (__be32 )iph2;
d1da932e	248
299603e8 JC	249	/* All fields must match except length and Traffic Class.
	250	* XXX skbs on the gro_list have all been parsed and pulled
	251	* already so we don't need to compare nlen
	252	* (nlen != (sizeof(*iph2) + ipv6_exthdrs_len(iph2, &ops)))
0b215b97	253	* memcmp() alone below is sufficient, right?
299603e8 JC	254	*/
299603e8 JC	255	if ((first_word & htonl(0xF00FFFFF)) \|\|
0b215b97 ED	256	!ipv6_addr_equal(&iph->saddr, &iph2->saddr) \|\|
	257	!ipv6_addr_equal(&iph->daddr, &iph2->daddr) \|\|
	258	(u16 )&iph->nexthdr != (u16 )&iph2->nexthdr) {
	259	not_same_flow:
d1da932e VY	260	NAPI_GRO_CB(p)->same_flow = 0;
	261	continue;
	262	}
0b215b97 ED	263	if (unlikely(nlen > sizeof(struct ipv6hdr))) {
	264	if (memcmp(iph + 1, iph2 + 1,
	265	nlen - sizeof(struct ipv6hdr)))
	266	goto not_same_flow;
	267	}
d1da932e VY	268	/* flush if Traffic Class fields are different */
	269	NAPI_GRO_CB(p)->flush \|= !!(first_word & htonl(0x0FF00000));
	270	NAPI_GRO_CB(p)->flush \|= flush;
03d56daa	271
1530545e AD	272	/* If the previous IP ID value was based on an atomic
	273	* datagram we can overwrite the value and ignore it.
	274	*/
	275	if (NAPI_GRO_CB(skb)->is_atomic)
	276	NAPI_GRO_CB(p)->flush_id = 0;
d1da932e VY	277	}
d1da932e VY	278
1530545e	279	NAPI_GRO_CB(skb)->is_atomic = true;
d1da932e VY	280	NAPI_GRO_CB(skb)->flush \|= flush;
d1da932e VY	281
4de462ab	282	skb_gro_postpull_rcsum(skb, iph, nlen);
d1da932e	283
028e0a47 PA	284	pp = indirect_call_gro_receive_l4(tcp6_gro_receive, udp6_gro_receive,
028e0a47 PA	285	ops->callbacks.gro_receive, head, skb);
d1da932e	286
d1da932e VY	287	out_unlock:
	288	rcu_read_unlock();
	289
	290	out:
5f114163	291	skb_gro_flush_final(skb, pp, flush);
d1da932e VY	292
	293	return pp;
	294	}
	295
d4546c25 DM	296	static struct sk_buff sit_ip6ip6_gro_receive(struct list_head head,
d4546c25 DM	297	struct sk_buff *skb)
fac8e0f5	298	{
815d22e5 TH	299	/* Common GRO receive for SIT and IP6IP6 */
815d22e5 TH	300
fac8e0f5 JG	301	if (NAPI_GRO_CB(skb)->encap_mark) {
	302	NAPI_GRO_CB(skb)->flush = 1;
	303	return NULL;
	304	}
	305
	306	NAPI_GRO_CB(skb)->encap_mark = 1;
	307
	308	return ipv6_gro_receive(head, skb);
	309	}
	310
d4546c25 DM	311	static struct sk_buff ip4ip6_gro_receive(struct list_head head,
d4546c25 DM	312	struct sk_buff *skb)
b8921ca8 TH	313	{
	314	/* Common GRO receive for SIT and IP6IP6 */
	315
	316	if (NAPI_GRO_CB(skb)->encap_mark) {
	317	NAPI_GRO_CB(skb)->flush = 1;
	318	return NULL;
	319	}
	320
	321	NAPI_GRO_CB(skb)->encap_mark = 1;
	322
	323	return inet_gro_receive(head, skb);
	324	}
	325
028e0a47 PA	326	INDIRECT_CALLABLE_DECLARE(int tcp6_gro_complete(struct sk_buff *, int));
028e0a47 PA	327	INDIRECT_CALLABLE_DECLARE(int udp6_gro_complete(struct sk_buff *, int));
aaa5d90b	328	INDIRECT_CALLABLE_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff)
d1da932e VY	329	{
d1da932e VY	330	const struct net_offload *ops;
299603e8	331	struct ipv6hdr iph = (struct ipv6hdr )(skb->data + nhoff);
d1da932e VY	332	int err = -ENOSYS;
d1da932e VY	333
294acf1c PA	334	if (skb->encapsulation) {
294acf1c PA	335	skb_set_inner_protocol(skb, cpu_to_be16(ETH_P_IPV6));
feec0cb3	336	skb_set_inner_network_header(skb, nhoff);
294acf1c	337	}
feec0cb3	338
299603e8	339	iph->payload_len = htons(skb->len - nhoff - sizeof(*iph));
d1da932e VY	340
d1da932e VY	341	rcu_read_lock();
299603e8 JC	342
299603e8 JC	343	nhoff += sizeof(*iph) + ipv6_exthdrs_len(iph, &ops);
f191a1d1	344	if (WARN_ON(!ops \|\| !ops->callbacks.gro_complete))
d1da932e VY	345	goto out_unlock;
d1da932e VY	346
028e0a47 PA	347	err = INDIRECT_CALL_L4(ops->callbacks.gro_complete, tcp6_gro_complete,
028e0a47 PA	348	udp6_gro_complete, skb, nhoff);
d1da932e VY	349
	350	out_unlock:
	351	rcu_read_unlock();
	352
	353	return err;
	354	}
	355
feec0cb3 ED	356	static int sit_gro_complete(struct sk_buff *skb, int nhoff)
	357	{
	358	skb->encapsulation = 1;
7e13318d	359	skb_shinfo(skb)->gso_type \|= SKB_GSO_IPXIP4;
feec0cb3 ED	360	return ipv6_gro_complete(skb, nhoff);
	361	}
	362
815d22e5 TH	363	static int ip6ip6_gro_complete(struct sk_buff *skb, int nhoff)
	364	{
	365	skb->encapsulation = 1;
	366	skb_shinfo(skb)->gso_type \|= SKB_GSO_IPXIP6;
	367	return ipv6_gro_complete(skb, nhoff);
	368	}
	369
b8921ca8 TH	370	static int ip4ip6_gro_complete(struct sk_buff *skb, int nhoff)
	371	{
	372	skb->encapsulation = 1;
	373	skb_shinfo(skb)->gso_type \|= SKB_GSO_IPXIP6;
	374	return inet_gro_complete(skb, nhoff);
	375	}
	376
d1da932e VY	377	static struct packet_offload ipv6_packet_offload __read_mostly = {
d1da932e VY	378	.type = cpu_to_be16(ETH_P_IPV6),
f191a1d1	379	.callbacks = {
f191a1d1 VY	380	.gso_segment = ipv6_gso_segment,
	381	.gro_receive = ipv6_gro_receive,
	382	.gro_complete = ipv6_gro_complete,
	383	},
d1da932e VY	384	};
d1da932e VY	385
418e897e WB	386	static struct sk_buff sit_gso_segment(struct sk_buff skb,
	387	netdev_features_t features)
	388	{
	389	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP4))
	390	return ERR_PTR(-EINVAL);
	391
	392	return ipv6_gso_segment(skb, features);
	393	}
	394
	395	static struct sk_buff ip4ip6_gso_segment(struct sk_buff skb,
	396	netdev_features_t features)
	397	{
	398	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP6))
	399	return ERR_PTR(-EINVAL);
	400
	401	return inet_gso_segment(skb, features);
	402	}
	403
	404	static struct sk_buff ip6ip6_gso_segment(struct sk_buff skb,
	405	netdev_features_t features)
	406	{
	407	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP6))
	408	return ERR_PTR(-EINVAL);
	409
	410	return ipv6_gso_segment(skb, features);
	411	}
	412
61c1db7f ED	413	static const struct net_offload sit_offload = {
61c1db7f ED	414	.callbacks = {
418e897e	415	.gso_segment = sit_gso_segment,
815d22e5	416	.gro_receive = sit_ip6ip6_gro_receive,
feec0cb3	417	.gro_complete = sit_gro_complete,
61c1db7f ED	418	},
	419	};
	420
b8921ca8 TH	421	static const struct net_offload ip4ip6_offload = {
b8921ca8 TH	422	.callbacks = {
418e897e	423	.gso_segment = ip4ip6_gso_segment,
b8921ca8 TH	424	.gro_receive = ip4ip6_gro_receive,
	425	.gro_complete = ip4ip6_gro_complete,
	426	},
	427	};
	428
815d22e5 TH	429	static const struct net_offload ip6ip6_offload = {
815d22e5 TH	430	.callbacks = {
418e897e	431	.gso_segment = ip6ip6_gso_segment,
815d22e5 TH	432	.gro_receive = sit_ip6ip6_gro_receive,
	433	.gro_complete = ip6ip6_gro_complete,
	434	},
	435	};
c6b641a4	436	static int __init ipv6_offload_init(void)
d1da932e	437	{
c6b641a4 VY	438
	439	if (tcpv6_offload_init() < 0)
	440	pr_crit("%s: Cannot add TCP protocol offload\n", __func__);
c6b641a4 VY	441	if (ipv6_exthdrs_offload_init() < 0)
	442	pr_crit("%s: Cannot add EXTHDRS protocol offload\n", __func__);
	443
d1da932e	444	dev_add_offload(&ipv6_packet_offload);
61c1db7f ED	445
61c1db7f ED	446	inet_add_offload(&sit_offload, IPPROTO_IPV6);
815d22e5	447	inet6_add_offload(&ip6ip6_offload, IPPROTO_IPV6);
b8921ca8	448	inet6_add_offload(&ip4ip6_offload, IPPROTO_IPIP);
61c1db7f	449
c6b641a4	450	return 0;
d1da932e VY	451	}
d1da932e VY	452
c6b641a4	453	fs_initcall(ipv6_offload_init);