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

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 *	IPV6 GSO/GRO offload support
 *	Linux INET6 implementation
 */

#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 <net/tcp.h>
#include <net/udp.h>
#include <net/gro.h>
#include <net/gso.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, err;
	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);
	err = ipv6_hopopt_jumbo_remove(skb);
	if (err)
		return ERR_PTR(err);
	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 (!segs)
			skb->network_header = skb_mac_header(skb) + nhoff - skb->head;
	}

	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_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(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);

	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;

		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) ||
		     iph->nexthdr != 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)) |
			(__force __be32)(iph->hop_limit ^ iph2->hop_limit));
		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:
	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_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff)
{
	const struct net_offload *ops;
	struct ipv6hdr *iph;
	int err = -ENOSYS;
	u32 payload_len;

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

	payload_len = skb->len - nhoff - sizeof(*iph);
	if (unlikely(payload_len > IPV6_MAXPLEN)) {
		struct hop_jumbo_hdr *hop_jumbo;
		int hoplen = sizeof(*hop_jumbo);

		/* Move network header left */
		memmove(skb_mac_header(skb) - hoplen, skb_mac_header(skb),
			skb->transport_header - skb->mac_header);
		skb->data -= hoplen;
		skb->len += hoplen;
		skb->mac_header -= hoplen;
		skb->network_header -= hoplen;
		iph = (struct ipv6hdr *)(skb->data + nhoff);
		hop_jumbo = (struct hop_jumbo_hdr *)(iph + 1);

		/* Build hop-by-hop options */
		hop_jumbo->nexthdr = iph->nexthdr;
		hop_jumbo->hdrlen = 0;
		hop_jumbo->tlv_type = IPV6_TLV_JUMBO;
		hop_jumbo->tlv_len = 4;
		hop_jumbo->jumbo_payload_len = htonl(payload_len + hoplen);

		iph->nexthdr = NEXTHDR_HOP;
		iph->payload_len = 0;
	} else {
		iph = (struct ipv6hdr *)(skb->data + nhoff);
		iph->payload_len = htons(payload_len);
	}

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

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

out:
	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
2874c5fd	1	// SPDX-License-Identifier: GPL-2.0-or-later
d1da932e VY	2	/*
	3	* IPV6 GSO/GRO offload support
	4	* Linux INET6 implementation
d1da932e VY	5	*/
	6
	7	#include <linux/kernel.h>
	8	#include <linux/socket.h>
	9	#include <linux/netdevice.h>
	10	#include <linux/skbuff.h>
c6b641a4	11	#include <linux/printk.h>
d1da932e VY	12
	13	#include <net/protocol.h>
	14	#include <net/ipv6.h>
b8921ca8	15	#include <net/inet_common.h>
5521d95e	16	#include <net/tcp.h>
6db69328	17	#include <net/udp.h>
04f00ab2	18	#include <net/gro.h>
d457a0e3	19	#include <net/gso.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;
89300468	81	int proto, err;
d1da932e	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	90	skb_reset_network_header(skb);
89300468 CL	91	err = ipv6_hopopt_jumbo_remove(skb);
	92	if (err)
	93	return ERR_PTR(err);
d3e5e006	94	nhoff = skb_network_header(skb) - skb_mac_header(skb);
d1da932e VY	95	if (unlikely(!pskb_may_pull(skb, sizeof(*ipv6h))))
	96	goto out;
	97
91a48a2e HFS	98	encap = SKB_GSO_CB(skb)->encap_level > 0;
91a48a2e HFS	99	if (encap)
1e16aa3d	100	features &= skb->dev->hw_enc_features;
d3e5e006 ED	101	SKB_GSO_CB(skb)->encap_level += sizeof(*ipv6h);
d3e5e006 ED	102
d1da932e VY	103	ipv6h = ipv6_hdr(skb);
	104	__skb_pull(skb, sizeof(*ipv6h));
	105	segs = ERR_PTR(-EPROTONOSUPPORT);
	106
	107	proto = ipv6_gso_pull_exthdrs(skb, ipv6h->nexthdr);
b917eb15	108
91a48a2e	109	if (skb->encapsulation &&
7e13318d	110	skb_shinfo(skb)->gso_type & (SKB_GSO_IPXIP4 \| SKB_GSO_IPXIP6))
ee80d1eb WB	111	udpfrag = proto == IPPROTO_UDP && encap &&
ee80d1eb WB	112	(skb_shinfo(skb)->gso_type & SKB_GSO_UDP);
91a48a2e	113	else
ee80d1eb WB	114	udpfrag = proto == IPPROTO_UDP && !skb->encapsulation &&
ee80d1eb WB	115	(skb_shinfo(skb)->gso_type & SKB_GSO_UDP);
91a48a2e	116
d1da932e	117	ops = rcu_dereference(inet6_offloads[proto]);
f191a1d1	118	if (likely(ops && ops->callbacks.gso_segment)) {
d1da932e	119	skb_reset_transport_header(skb);
f191a1d1	120	segs = ops->callbacks.gso_segment(skb, features);
cc20cced TL	121	if (!segs)
cc20cced TL	122	skb->network_header = skb_mac_header(skb) + nhoff - skb->head;
d1da932e	123	}
d1da932e	124
6b6ebb6b	125	if (IS_ERR_OR_NULL(segs))
d1da932e VY	126	goto out;
d1da932e VY	127
07b26c94 SK	128	gso_partial = !!(skb_shinfo(segs)->gso_type & SKB_GSO_PARTIAL);
07b26c94 SK	129
d1da932e	130	for (skb = segs; skb; skb = skb->next) {
d3e5e006	131	ipv6h = (struct ipv6hdr *)(skb_mac_header(skb) + nhoff);
3d0241d5	132	if (gso_partial && skb_is_gso(skb))
802ab55a AD	133	payload_len = skb_shinfo(skb)->gso_size +
	134	SKB_GSO_CB(skb)->data_offset +
	135	skb->head - (unsigned char *)(ipv6h + 1);
	136	else
	137	payload_len = skb->len - nhoff - sizeof(*ipv6h);
	138	ipv6h->payload_len = htons(payload_len);
d3e5e006	139	skb->network_header = (u8 *)ipv6h - skb->head;
c56cae23	140	skb_reset_mac_len(skb);
d3e5e006	141
91a48a2e	142	if (udpfrag) {
7dd7eb95	143	int err = ip6_find_1stfragopt(skb, &prevhdr);
e3e86b51 DM	144	if (err < 0) {
e3e86b51 DM	145	kfree_skb_list(segs);
7dd7eb95	146	return ERR_PTR(err);
e3e86b51	147	}
7dd7eb95	148	fptr = (struct frag_hdr )((u8 )ipv6h + err);
d1da932e	149	fptr->frag_off = htons(offset);
53b24b8f	150	if (skb->next)
d1da932e VY	151	fptr->frag_off \|= htons(IP6_MF);
	152	offset += (ntohs(ipv6h->payload_len) -
	153	sizeof(struct frag_hdr));
	154	}
91a48a2e HFS	155	if (encap)
91a48a2e HFS	156	skb_reset_inner_headers(skb);
d1da932e VY	157	}
	158
	159	out:
	160	return segs;
	161	}
	162
299603e8 JC	163	/* Return the total length of all the extension hdrs, following the same
	164	* logic in ipv6_gso_pull_exthdrs() when parsing ext-hdrs.
	165	*/
	166	static int ipv6_exthdrs_len(struct ipv6hdr *iph,
	167	const struct net_offload **opps)
	168	{
810c23a3 JC	169	struct ipv6_opt_hdr opth = (void )iph;
810c23a3 JC	170	int len = 0, proto, optlen = sizeof(*iph);
299603e8 JC	171
	172	proto = iph->nexthdr;
	173	for (;;) {
	174	if (proto != NEXTHDR_HOP) {
	175	*opps = rcu_dereference(inet6_offloads[proto]);
	176	if (unlikely(!(*opps)))
	177	break;
	178	if (!((*opps)->flags & INET6_PROTO_GSO_EXTHDR))
	179	break;
	180	}
810c23a3 JC	181	opth = (void *)opth + optlen;
810c23a3 JC	182	optlen = ipv6_optlen(opth);
299603e8 JC	183	len += optlen;
	184	proto = opth->nexthdr;
	185	}
	186	return len;
	187	}
	188
aaa5d90b PA	189	INDIRECT_CALLABLE_SCOPE struct sk_buff ipv6_gro_receive(struct list_head head,
aaa5d90b PA	190	struct sk_buff *skb)
d1da932e VY	191	{
d1da932e VY	192	const struct net_offload *ops;
d4546c25	193	struct sk_buff *pp = NULL;
d1da932e VY	194	struct sk_buff *p;
	195	struct ipv6hdr *iph;
	196	unsigned int nlen;
	197	unsigned int hlen;
	198	unsigned int off;
bf5a755f	199	u16 flush = 1;
d1da932e	200	int proto;
d1da932e VY	201
	202	off = skb_gro_offset(skb);
	203	hlen = off + sizeof(*iph);
cb628a9a	204	iph = skb_gro_header(skb, hlen, off);
35ffb665 RG	205	if (unlikely(!iph))
35ffb665 RG	206	goto out;
d1da932e	207
299603e8	208	skb_set_network_header(skb, off);
d1da932e VY	209	skb_gro_pull(skb, sizeof(*iph));
	210	skb_set_transport_header(skb, skb_gro_offset(skb));
	211
	212	flush += ntohs(iph->payload_len) != skb_gro_len(skb);
	213
d1da932e VY	214	proto = iph->nexthdr;
d1da932e VY	215	ops = rcu_dereference(inet6_offloads[proto]);
f191a1d1	216	if (!ops \|\| !ops->callbacks.gro_receive) {
d427c899	217	pskb_pull(skb, skb_gro_offset(skb));
57ea52a8	218	skb_gro_frag0_invalidate(skb);
d1da932e VY	219	proto = ipv6_gso_pull_exthdrs(skb, proto);
	220	skb_gro_pull(skb, -skb_transport_offset(skb));
	221	skb_reset_transport_header(skb);
	222	__skb_push(skb, skb_gro_offset(skb));
	223
	224	ops = rcu_dereference(inet6_offloads[proto]);
f191a1d1	225	if (!ops \|\| !ops->callbacks.gro_receive)
fc1ca334	226	goto out;
d1da932e VY	227
	228	iph = ipv6_hdr(skb);
	229	}
	230
	231	NAPI_GRO_CB(skb)->proto = proto;
	232
	233	flush--;
	234	nlen = skb_network_header_len(skb);
	235
d4546c25	236	list_for_each_entry(p, head, list) {
d1da932e VY	237	const struct ipv6hdr *iph2;
	238	__be32 first_word; /* <Version:4><Traffic_Class:8><Flow_Label:20> */
	239
	240	if (!NAPI_GRO_CB(p)->same_flow)
	241	continue;
	242
299603e8	243	iph2 = (struct ipv6hdr *)(p->data + off);
67ba4152	244	first_word = (__be32 )iph ^ (__be32 )iph2;
d1da932e	245
299603e8 JC	246	/* All fields must match except length and Traffic Class.
	247	* XXX skbs on the gro_list have all been parsed and pulled
	248	* already so we don't need to compare nlen
	249	* (nlen != (sizeof(*iph2) + ipv6_exthdrs_len(iph2, &ops)))
0b215b97	250	* memcmp() alone below is sufficient, right?
299603e8 JC	251	*/
299603e8 JC	252	if ((first_word & htonl(0xF00FFFFF)) \|\|
1ebb87cc JC	253	!ipv6_addr_equal(&iph->saddr, &iph2->saddr) \|\|
1ebb87cc JC	254	!ipv6_addr_equal(&iph->daddr, &iph2->daddr) \|\|
6fc2f383	255	iph->nexthdr != iph2->nexthdr) {
0b215b97	256	not_same_flow:
d1da932e VY	257	NAPI_GRO_CB(p)->same_flow = 0;
	258	continue;
	259	}
0b215b97 ED	260	if (unlikely(nlen > sizeof(struct ipv6hdr))) {
	261	if (memcmp(iph + 1, iph2 + 1,
	262	nlen - sizeof(struct ipv6hdr)))
	263	goto not_same_flow;
	264	}
d1da932e	265	/* flush if Traffic Class fields are different */
6fc2f383 JK	266	NAPI_GRO_CB(p)->flush \|= !!((first_word & htonl(0x0FF00000)) \|
6fc2f383 JK	267	(__force __be32)(iph->hop_limit ^ iph2->hop_limit));
d1da932e	268	NAPI_GRO_CB(p)->flush \|= flush;
03d56daa	269
1530545e AD	270	/* If the previous IP ID value was based on an atomic
	271	* datagram we can overwrite the value and ignore it.
	272	*/
	273	if (NAPI_GRO_CB(skb)->is_atomic)
	274	NAPI_GRO_CB(p)->flush_id = 0;
d1da932e VY	275	}
d1da932e VY	276
1530545e	277	NAPI_GRO_CB(skb)->is_atomic = true;
d1da932e VY	278	NAPI_GRO_CB(skb)->flush \|= flush;
d1da932e VY	279
4de462ab	280	skb_gro_postpull_rcsum(skb, iph, nlen);
d1da932e	281
028e0a47 PA	282	pp = indirect_call_gro_receive_l4(tcp6_gro_receive, udp6_gro_receive,
028e0a47 PA	283	ops->callbacks.gro_receive, head, skb);
d1da932e	284
d1da932e	285	out:
5f114163	286	skb_gro_flush_final(skb, pp, flush);
d1da932e VY	287
	288	return pp;
	289	}
	290
d4546c25 DM	291	static struct sk_buff sit_ip6ip6_gro_receive(struct list_head head,
d4546c25 DM	292	struct sk_buff *skb)
fac8e0f5	293	{
815d22e5 TH	294	/* Common GRO receive for SIT and IP6IP6 */
815d22e5 TH	295
fac8e0f5 JG	296	if (NAPI_GRO_CB(skb)->encap_mark) {
	297	NAPI_GRO_CB(skb)->flush = 1;
	298	return NULL;
	299	}
	300
	301	NAPI_GRO_CB(skb)->encap_mark = 1;
	302
	303	return ipv6_gro_receive(head, skb);
	304	}
	305
d4546c25 DM	306	static struct sk_buff ip4ip6_gro_receive(struct list_head head,
d4546c25 DM	307	struct sk_buff *skb)
b8921ca8 TH	308	{
	309	/* Common GRO receive for SIT and IP6IP6 */
	310
	311	if (NAPI_GRO_CB(skb)->encap_mark) {
	312	NAPI_GRO_CB(skb)->flush = 1;
	313	return NULL;
	314	}
	315
	316	NAPI_GRO_CB(skb)->encap_mark = 1;
	317
	318	return inet_gro_receive(head, skb);
	319	}
	320
aaa5d90b	321	INDIRECT_CALLABLE_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff)
d1da932e VY	322	{
d1da932e VY	323	const struct net_offload *ops;
81fbc812	324	struct ipv6hdr *iph;
d1da932e	325	int err = -ENOSYS;
81fbc812	326	u32 payload_len;
d1da932e	327
294acf1c PA	328	if (skb->encapsulation) {
294acf1c PA	329	skb_set_inner_protocol(skb, cpu_to_be16(ETH_P_IPV6));
feec0cb3	330	skb_set_inner_network_header(skb, nhoff);
294acf1c	331	}
feec0cb3	332
81fbc812 ED	333	payload_len = skb->len - nhoff - sizeof(*iph);
	334	if (unlikely(payload_len > IPV6_MAXPLEN)) {
	335	struct hop_jumbo_hdr *hop_jumbo;
	336	int hoplen = sizeof(*hop_jumbo);
	337
	338	/* Move network header left */
	339	memmove(skb_mac_header(skb) - hoplen, skb_mac_header(skb),
	340	skb->transport_header - skb->mac_header);
	341	skb->data -= hoplen;
	342	skb->len += hoplen;
	343	skb->mac_header -= hoplen;
	344	skb->network_header -= hoplen;
	345	iph = (struct ipv6hdr *)(skb->data + nhoff);
	346	hop_jumbo = (struct hop_jumbo_hdr *)(iph + 1);
	347
	348	/* Build hop-by-hop options */
	349	hop_jumbo->nexthdr = iph->nexthdr;
	350	hop_jumbo->hdrlen = 0;
	351	hop_jumbo->tlv_type = IPV6_TLV_JUMBO;
	352	hop_jumbo->tlv_len = 4;
	353	hop_jumbo->jumbo_payload_len = htonl(payload_len + hoplen);
	354
	355	iph->nexthdr = NEXTHDR_HOP;
	356	iph->payload_len = 0;
	357	} else {
	358	iph = (struct ipv6hdr *)(skb->data + nhoff);
	359	iph->payload_len = htons(payload_len);
	360	}
d1da932e	361
299603e8	362	nhoff += sizeof(*iph) + ipv6_exthdrs_len(iph, &ops);
f191a1d1	363	if (WARN_ON(!ops \|\| !ops->callbacks.gro_complete))
627b94f7	364	goto out;
d1da932e	365
028e0a47 PA	366	err = INDIRECT_CALL_L4(ops->callbacks.gro_complete, tcp6_gro_complete,
028e0a47 PA	367	udp6_gro_complete, skb, nhoff);
d1da932e	368
627b94f7	369	out:
d1da932e VY	370	return err;
	371	}
	372
feec0cb3 ED	373	static int sit_gro_complete(struct sk_buff *skb, int nhoff)
	374	{
	375	skb->encapsulation = 1;
7e13318d	376	skb_shinfo(skb)->gso_type \|= SKB_GSO_IPXIP4;
feec0cb3 ED	377	return ipv6_gro_complete(skb, nhoff);
	378	}
	379
815d22e5 TH	380	static int ip6ip6_gro_complete(struct sk_buff *skb, int nhoff)
	381	{
	382	skb->encapsulation = 1;
	383	skb_shinfo(skb)->gso_type \|= SKB_GSO_IPXIP6;
	384	return ipv6_gro_complete(skb, nhoff);
	385	}
	386
b8921ca8 TH	387	static int ip4ip6_gro_complete(struct sk_buff *skb, int nhoff)
	388	{
	389	skb->encapsulation = 1;
	390	skb_shinfo(skb)->gso_type \|= SKB_GSO_IPXIP6;
	391	return inet_gro_complete(skb, nhoff);
	392	}
	393
d1da932e VY	394	static struct packet_offload ipv6_packet_offload __read_mostly = {
d1da932e VY	395	.type = cpu_to_be16(ETH_P_IPV6),
f191a1d1	396	.callbacks = {
f191a1d1 VY	397	.gso_segment = ipv6_gso_segment,
	398	.gro_receive = ipv6_gro_receive,
	399	.gro_complete = ipv6_gro_complete,
	400	},
d1da932e VY	401	};
d1da932e VY	402
418e897e WB	403	static struct sk_buff sit_gso_segment(struct sk_buff skb,
	404	netdev_features_t features)
	405	{
	406	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP4))
	407	return ERR_PTR(-EINVAL);
	408
	409	return ipv6_gso_segment(skb, features);
	410	}
	411
	412	static struct sk_buff ip4ip6_gso_segment(struct sk_buff skb,
	413	netdev_features_t features)
	414	{
	415	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP6))
	416	return ERR_PTR(-EINVAL);
	417
	418	return inet_gso_segment(skb, features);
	419	}
	420
	421	static struct sk_buff ip6ip6_gso_segment(struct sk_buff skb,
	422	netdev_features_t features)
	423	{
	424	if (!(skb_shinfo(skb)->gso_type & SKB_GSO_IPXIP6))
	425	return ERR_PTR(-EINVAL);
	426
	427	return ipv6_gso_segment(skb, features);
	428	}
	429
61c1db7f ED	430	static const struct net_offload sit_offload = {
61c1db7f ED	431	.callbacks = {
418e897e	432	.gso_segment = sit_gso_segment,
815d22e5	433	.gro_receive = sit_ip6ip6_gro_receive,
feec0cb3	434	.gro_complete = sit_gro_complete,
61c1db7f ED	435	},
	436	};
	437
b8921ca8 TH	438	static const struct net_offload ip4ip6_offload = {
b8921ca8 TH	439	.callbacks = {
418e897e	440	.gso_segment = ip4ip6_gso_segment,
b8921ca8 TH	441	.gro_receive = ip4ip6_gro_receive,
	442	.gro_complete = ip4ip6_gro_complete,
	443	},
	444	};
	445
815d22e5 TH	446	static const struct net_offload ip6ip6_offload = {
815d22e5 TH	447	.callbacks = {
418e897e	448	.gso_segment = ip6ip6_gso_segment,
815d22e5 TH	449	.gro_receive = sit_ip6ip6_gro_receive,
	450	.gro_complete = ip6ip6_gro_complete,
	451	},
	452	};
c6b641a4	453	static int __init ipv6_offload_init(void)
d1da932e	454	{
c6b641a4 VY	455
	456	if (tcpv6_offload_init() < 0)
	457	pr_crit("%s: Cannot add TCP protocol offload\n", __func__);
c6b641a4 VY	458	if (ipv6_exthdrs_offload_init() < 0)
	459	pr_crit("%s: Cannot add EXTHDRS protocol offload\n", __func__);
	460
d1da932e	461	dev_add_offload(&ipv6_packet_offload);
61c1db7f ED	462
61c1db7f ED	463	inet_add_offload(&sit_offload, IPPROTO_IPV6);
815d22e5	464	inet6_add_offload(&ip6ip6_offload, IPPROTO_IPV6);
b8921ca8	465	inet6_add_offload(&ip4ip6_offload, IPPROTO_IPIP);
61c1db7f	466
c6b641a4	467	return 0;
d1da932e VY	468	}
d1da932e VY	469
c6b641a4	470	fs_initcall(ipv6_offload_init);