[deliverable/linux.git] / net / ipv4 / udp_offload.c

/*
 *	IPV4 GSO/GRO offload support
 *	Linux INET 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.
 *
 *	UDPv4 GSO support
 */

#include <linux/skbuff.h>
#include <net/udp.h>
#include <net/protocol.h>

static DEFINE_SPINLOCK(udp_offload_lock);
static struct udp_offload_priv __rcu *udp_offload_base __read_mostly;

#define udp_deref_protected(X) rcu_dereference_protected(X, lockdep_is_held(&udp_offload_lock))

struct udp_offload_priv {
	struct udp_offload	*offload;
	possible_net_t	net;
	struct rcu_head		rcu;
	struct udp_offload_priv __rcu *next;
};

static struct sk_buff *__skb_udp_tunnel_segment(struct sk_buff *skb,
	netdev_features_t features,
	struct sk_buff *(*gso_inner_segment)(struct sk_buff *skb,
					     netdev_features_t features),
	__be16 new_protocol, bool is_ipv6)
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	u16 mac_offset = skb->mac_header;
	int mac_len = skb->mac_len;
	int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb);
	__be16 protocol = skb->protocol;
	netdev_features_t enc_features;
	int udp_offset, outer_hlen;
	unsigned int oldlen;
	bool need_csum = !!(skb_shinfo(skb)->gso_type &
			    SKB_GSO_UDP_TUNNEL_CSUM);
	bool remcsum = !!(skb_shinfo(skb)->gso_type & SKB_GSO_TUNNEL_REMCSUM);
	bool offload_csum = false, dont_encap = (need_csum || remcsum);

	oldlen = (u16)~skb->len;

	if (unlikely(!pskb_may_pull(skb, tnl_hlen)))
		goto out;

	skb->encapsulation = 0;
	__skb_pull(skb, tnl_hlen);
	skb_reset_mac_header(skb);
	skb_set_network_header(skb, skb_inner_network_offset(skb));
	skb->mac_len = skb_inner_network_offset(skb);
	skb->protocol = new_protocol;
	skb->encap_hdr_csum = need_csum;
	skb->remcsum_offload = remcsum;

	/* Try to offload checksum if possible */
	offload_csum = !!(need_csum &&
			  ((skb->dev->features & NETIF_F_HW_CSUM) ||
			   (skb->dev->features & (is_ipv6 ?
			    NETIF_F_IPV6_CSUM : NETIF_F_IP_CSUM))));

	/* segment inner packet. */
	enc_features = skb->dev->hw_enc_features & features;
	segs = gso_inner_segment(skb, enc_features);
	if (IS_ERR_OR_NULL(segs)) {
		skb_gso_error_unwind(skb, protocol, tnl_hlen, mac_offset,
				     mac_len);
		goto out;
	}

	outer_hlen = skb_tnl_header_len(skb);
	udp_offset = outer_hlen - tnl_hlen;
	skb = segs;
	do {
		struct udphdr *uh;
		int len;
		__be32 delta;

		if (dont_encap) {
			skb->encapsulation = 0;
			skb->ip_summed = CHECKSUM_NONE;
		} else {
			/* Only set up inner headers if we might be offloading
			 * inner checksum.
			 */
			skb_reset_inner_headers(skb);
			skb->encapsulation = 1;
		}

		skb->mac_len = mac_len;
		skb->protocol = protocol;

		skb_push(skb, outer_hlen);
		skb_reset_mac_header(skb);
		skb_set_network_header(skb, mac_len);
		skb_set_transport_header(skb, udp_offset);
		len = skb->len - udp_offset;
		uh = udp_hdr(skb);
		uh->len = htons(len);

		if (!need_csum)
			continue;

		delta = htonl(oldlen + len);

		uh->check = ~csum_fold((__force __wsum)
				       ((__force u32)uh->check +
					(__force u32)delta));
		if (offload_csum) {
			skb->ip_summed = CHECKSUM_PARTIAL;
			skb->csum_start = skb_transport_header(skb) - skb->head;
			skb->csum_offset = offsetof(struct udphdr, check);
		} else if (remcsum) {
			/* Need to calculate checksum from scratch,
			 * inner checksums are never when doing
			 * remote_checksum_offload.
			 */

			skb->csum = skb_checksum(skb, udp_offset,
						 skb->len - udp_offset,
						 0);
			uh->check = csum_fold(skb->csum);
			if (uh->check == 0)
				uh->check = CSUM_MANGLED_0;
		} else {
			uh->check = gso_make_checksum(skb, ~uh->check);

			if (uh->check == 0)
				uh->check = CSUM_MANGLED_0;
		}
	} while ((skb = skb->next));
out:
	return segs;
}

struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
				       netdev_features_t features,
				       bool is_ipv6)
{
	__be16 protocol = skb->protocol;
	const struct net_offload **offloads;
	const struct net_offload *ops;
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	struct sk_buff *(*gso_inner_segment)(struct sk_buff *skb,
					     netdev_features_t features);

	rcu_read_lock();

	switch (skb->inner_protocol_type) {
	case ENCAP_TYPE_ETHER:
		protocol = skb->inner_protocol;
		gso_inner_segment = skb_mac_gso_segment;
		break;
	case ENCAP_TYPE_IPPROTO:
		offloads = is_ipv6 ? inet6_offloads : inet_offloads;
		ops = rcu_dereference(offloads[skb->inner_ipproto]);
		if (!ops || !ops->callbacks.gso_segment)
			goto out_unlock;
		gso_inner_segment = ops->callbacks.gso_segment;
		break;
	default:
		goto out_unlock;
	}

	segs = __skb_udp_tunnel_segment(skb, features, gso_inner_segment,
					protocol, is_ipv6);

out_unlock:
	rcu_read_unlock();

	return segs;
}

static struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
					 netdev_features_t features)
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	unsigned int mss;
	__wsum csum;
	struct udphdr *uh;
	struct iphdr *iph;

	if (skb->encapsulation &&
	    (skb_shinfo(skb)->gso_type &
	     (SKB_GSO_UDP_TUNNEL|SKB_GSO_UDP_TUNNEL_CSUM))) {
		segs = skb_udp_tunnel_segment(skb, features, false);
		goto out;
	}

	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
		goto out;

	mss = skb_shinfo(skb)->gso_size;
	if (unlikely(skb->len <= mss))
		goto out;

	if (skb_gso_ok(skb, features | NETIF_F_GSO_ROBUST)) {
		/* Packet is from an untrusted source, reset gso_segs. */
		int type = skb_shinfo(skb)->gso_type;

		if (unlikely(type & ~(SKB_GSO_UDP | SKB_GSO_DODGY |
				      SKB_GSO_UDP_TUNNEL |
				      SKB_GSO_UDP_TUNNEL_CSUM |
				      SKB_GSO_TUNNEL_REMCSUM |
				      SKB_GSO_IPIP |
				      SKB_GSO_GRE | SKB_GSO_GRE_CSUM) ||
			     !(type & (SKB_GSO_UDP))))
			goto out;

		skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);

		segs = NULL;
		goto out;
	}

	/* Do software UFO. Complete and fill in the UDP checksum as
	 * HW cannot do checksum of UDP packets sent as multiple
	 * IP fragments.
	 */

	uh = udp_hdr(skb);
	iph = ip_hdr(skb);

	uh->check = 0;
	csum = skb_checksum(skb, 0, skb->len, 0);
	uh->check = udp_v4_check(skb->len, iph->saddr, iph->daddr, csum);
	if (uh->check == 0)
		uh->check = CSUM_MANGLED_0;

	skb->ip_summed = CHECKSUM_NONE;

	/* Fragment the skb. IP headers of the fragments are updated in
	 * inet_gso_segment()
	 */
	segs = skb_segment(skb, features);
out:
	return segs;
}

int udp_add_offload(struct net *net, struct udp_offload *uo)
{
	struct udp_offload_priv *new_offload = kzalloc(sizeof(*new_offload), GFP_ATOMIC);

	if (!new_offload)
		return -ENOMEM;

	write_pnet(&new_offload->net, net);
	new_offload->offload = uo;

	spin_lock(&udp_offload_lock);
	new_offload->next = udp_offload_base;
	rcu_assign_pointer(udp_offload_base, new_offload);
	spin_unlock(&udp_offload_lock);

	return 0;
}
EXPORT_SYMBOL(udp_add_offload);

static void udp_offload_free_routine(struct rcu_head *head)
{
	struct udp_offload_priv *ou_priv = container_of(head, struct udp_offload_priv, rcu);
	kfree(ou_priv);
}

void udp_del_offload(struct udp_offload *uo)
{
	struct udp_offload_priv __rcu **head = &udp_offload_base;
	struct udp_offload_priv *uo_priv;

	spin_lock(&udp_offload_lock);

	uo_priv = udp_deref_protected(*head);
	for (; uo_priv != NULL;
	     uo_priv = udp_deref_protected(*head)) {
		if (uo_priv->offload == uo) {
			rcu_assign_pointer(*head,
					   udp_deref_protected(uo_priv->next));
			goto unlock;
		}
		head = &uo_priv->next;
	}
	pr_warn("udp_del_offload: didn't find offload for port %d\n", ntohs(uo->port));
unlock:
	spin_unlock(&udp_offload_lock);
	if (uo_priv)
		call_rcu(&uo_priv->rcu, udp_offload_free_routine);
}
EXPORT_SYMBOL(udp_del_offload);

struct sk_buff **udp_gro_receive(struct sk_buff **head, struct sk_buff *skb,
				 struct udphdr *uh)
{
	struct udp_offload_priv *uo_priv;
	struct sk_buff *p, **pp = NULL;
	struct udphdr *uh2;
	unsigned int off = skb_gro_offset(skb);
	int flush = 1;

	if (NAPI_GRO_CB(skb)->udp_mark ||
	    (skb->ip_summed != CHECKSUM_PARTIAL &&
	     NAPI_GRO_CB(skb)->csum_cnt == 0 &&
	     !NAPI_GRO_CB(skb)->csum_valid))
		goto out;

	/* mark that this skb passed once through the udp gro layer */
	NAPI_GRO_CB(skb)->udp_mark = 1;

	rcu_read_lock();
	uo_priv = rcu_dereference(udp_offload_base);
	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
		if (net_eq(read_pnet(&uo_priv->net), dev_net(skb->dev)) &&
		    uo_priv->offload->port == uh->dest &&
		    uo_priv->offload->callbacks.gro_receive)
			goto unflush;
	}
	goto out_unlock;

unflush:
	flush = 0;

	for (p = *head; p; p = p->next) {
		if (!NAPI_GRO_CB(p)->same_flow)
			continue;

		uh2 = (struct udphdr   *)(p->data + off);

		/* Match ports and either checksums are either both zero
		 * or nonzero.
		 */
		if ((*(u32 *)&uh->source != *(u32 *)&uh2->source) ||
		    (!uh->check ^ !uh2->check)) {
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}
	}

	skb_gro_pull(skb, sizeof(struct udphdr)); /* pull encapsulating udp header */
	skb_gro_postpull_rcsum(skb, uh, sizeof(struct udphdr));
	NAPI_GRO_CB(skb)->proto = uo_priv->offload->ipproto;
	pp = uo_priv->offload->callbacks.gro_receive(head, skb,
						     uo_priv->offload);

out_unlock:
	rcu_read_unlock();
out:
	NAPI_GRO_CB(skb)->flush |= flush;
	return pp;
}

static struct sk_buff **udp4_gro_receive(struct sk_buff **head,
					 struct sk_buff *skb)
{
	struct udphdr *uh = udp_gro_udphdr(skb);

	if (unlikely(!uh))
		goto flush;

	/* Don't bother verifying checksum if we're going to flush anyway. */
	if (NAPI_GRO_CB(skb)->flush)
		goto skip;

	if (skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
						 inet_gro_compute_pseudo))
		goto flush;
	else if (uh->check)
		skb_gro_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
					     inet_gro_compute_pseudo);
skip:
	NAPI_GRO_CB(skb)->is_ipv6 = 0;
	return udp_gro_receive(head, skb, uh);

flush:
	NAPI_GRO_CB(skb)->flush = 1;
	return NULL;
}

int udp_gro_complete(struct sk_buff *skb, int nhoff)
{
	struct udp_offload_priv *uo_priv;
	__be16 newlen = htons(skb->len - nhoff);
	struct udphdr *uh = (struct udphdr *)(skb->data + nhoff);
	int err = -ENOSYS;

	uh->len = newlen;

	rcu_read_lock();

	uo_priv = rcu_dereference(udp_offload_base);
	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
		if (net_eq(read_pnet(&uo_priv->net), dev_net(skb->dev)) &&
		    uo_priv->offload->port == uh->dest &&
		    uo_priv->offload->callbacks.gro_complete)
			break;
	}

	if (uo_priv) {
		NAPI_GRO_CB(skb)->proto = uo_priv->offload->ipproto;
		err = uo_priv->offload->callbacks.gro_complete(skb,
				nhoff + sizeof(struct udphdr),
				uo_priv->offload);
	}

	rcu_read_unlock();

	if (skb->remcsum_offload)
		skb_shinfo(skb)->gso_type |= SKB_GSO_TUNNEL_REMCSUM;

	skb->encapsulation = 1;
	skb_set_inner_mac_header(skb, nhoff + sizeof(struct udphdr));

	return err;
}

static int udp4_gro_complete(struct sk_buff *skb, int nhoff)
{
	const struct iphdr *iph = ip_hdr(skb);
	struct udphdr *uh = (struct udphdr *)(skb->data + nhoff);

	if (uh->check) {
		skb_shinfo(skb)->gso_type |= SKB_GSO_UDP_TUNNEL_CSUM;
		uh->check = ~udp_v4_check(skb->len - nhoff, iph->saddr,
					  iph->daddr, 0);
	} else {
		skb_shinfo(skb)->gso_type |= SKB_GSO_UDP_TUNNEL;
	}

	return udp_gro_complete(skb, nhoff);
}

static const struct net_offload udpv4_offload = {
	.callbacks = {
		.gso_segment = udp4_ufo_fragment,
		.gro_receive  =	udp4_gro_receive,
		.gro_complete =	udp4_gro_complete,
	},
};

int __init udpv4_offload_init(void)
{
	return inet_add_offload(&udpv4_offload, IPPROTO_UDP);
}
Commit	Line	Data
da5bab07 DB	1	/*
	2	* IPV4 GSO/GRO offload support
	3	* Linux INET 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	* UDPv4 GSO support
	11	*/
	12
	13	#include <linux/skbuff.h>
	14	#include <net/udp.h>
	15	#include <net/protocol.h>
	16
b582ef09	17	static DEFINE_SPINLOCK(udp_offload_lock);
a1d0cd8e	18	static struct udp_offload_priv __rcu *udp_offload_base __read_mostly;
b582ef09	19
a664a4f7 SP	20	#define udp_deref_protected(X) rcu_dereference_protected(X, lockdep_is_held(&udp_offload_lock))
a664a4f7 SP	21
b582ef09 OG	22	struct udp_offload_priv {
b582ef09 OG	23	struct udp_offload *offload;
787d7ac3	24	possible_net_t net;
b582ef09 OG	25	struct rcu_head rcu;
	26	struct udp_offload_priv __rcu *next;
	27	};
	28
8bce6d7d TH	29	static struct sk_buff __skb_udp_tunnel_segment(struct sk_buff skb,
	30	netdev_features_t features,
	31	struct sk_buff (gso_inner_segment)(struct sk_buff *skb,
	32	netdev_features_t features),
4bcb877d	33	__be16 new_protocol, bool is_ipv6)
155e010e TH	34	{
	35	struct sk_buff *segs = ERR_PTR(-EINVAL);
	36	u16 mac_offset = skb->mac_header;
	37	int mac_len = skb->mac_len;
	38	int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb);
	39	__be16 protocol = skb->protocol;
	40	netdev_features_t enc_features;
	41	int udp_offset, outer_hlen;
	42	unsigned int oldlen;
4bcb877d TH	43	bool need_csum = !!(skb_shinfo(skb)->gso_type &
4bcb877d TH	44	SKB_GSO_UDP_TUNNEL_CSUM);
e585f236 TH	45	bool remcsum = !!(skb_shinfo(skb)->gso_type & SKB_GSO_TUNNEL_REMCSUM);
e585f236 TH	46	bool offload_csum = false, dont_encap = (need_csum \|\| remcsum);
155e010e TH	47
	48	oldlen = (u16)~skb->len;
	49
	50	if (unlikely(!pskb_may_pull(skb, tnl_hlen)))
	51	goto out;
	52
	53	skb->encapsulation = 0;
	54	__skb_pull(skb, tnl_hlen);
	55	skb_reset_mac_header(skb);
	56	skb_set_network_header(skb, skb_inner_network_offset(skb));
	57	skb->mac_len = skb_inner_network_offset(skb);
8bce6d7d	58	skb->protocol = new_protocol;
4bcb877d	59	skb->encap_hdr_csum = need_csum;
e585f236	60	skb->remcsum_offload = remcsum;
155e010e	61
4bcb877d TH	62	/* Try to offload checksum if possible */
4bcb877d TH	63	offload_csum = !!(need_csum &&
c8cd0989 TH	64	((skb->dev->features & NETIF_F_HW_CSUM) \|\|
	65	(skb->dev->features & (is_ipv6 ?
	66	NETIF_F_IPV6_CSUM : NETIF_F_IP_CSUM))));
155e010e TH	67
155e010e TH	68	/* segment inner packet. */
1e16aa3d	69	enc_features = skb->dev->hw_enc_features & features;
8bce6d7d	70	segs = gso_inner_segment(skb, enc_features);
27446442	71	if (IS_ERR_OR_NULL(segs)) {
155e010e TH	72	skb_gso_error_unwind(skb, protocol, tnl_hlen, mac_offset,
	73	mac_len);
	74	goto out;
	75	}
	76
	77	outer_hlen = skb_tnl_header_len(skb);
	78	udp_offset = outer_hlen - tnl_hlen;
	79	skb = segs;
	80	do {
	81	struct udphdr *uh;
	82	int len;
4bcb877d TH	83	__be32 delta;
	84
	85	if (dont_encap) {
	86	skb->encapsulation = 0;
	87	skb->ip_summed = CHECKSUM_NONE;
	88	} else {
	89	/* Only set up inner headers if we might be offloading
	90	* inner checksum.
	91	*/
	92	skb_reset_inner_headers(skb);
	93	skb->encapsulation = 1;
	94	}
155e010e TH	95
155e010e TH	96	skb->mac_len = mac_len;
4bcb877d	97	skb->protocol = protocol;
155e010e TH	98
	99	skb_push(skb, outer_hlen);
	100	skb_reset_mac_header(skb);
	101	skb_set_network_header(skb, mac_len);
	102	skb_set_transport_header(skb, udp_offset);
	103	len = skb->len - udp_offset;
	104	uh = udp_hdr(skb);
	105	uh->len = htons(len);
	106
4bcb877d TH	107	if (!need_csum)
	108	continue;
	109
	110	delta = htonl(oldlen + len);
155e010e	111
4bcb877d TH	112	uh->check = ~csum_fold((__force __wsum)
	113	((__force u32)uh->check +
	114	(__force u32)delta));
4bcb877d TH	115	if (offload_csum) {
	116	skb->ip_summed = CHECKSUM_PARTIAL;
	117	skb->csum_start = skb_transport_header(skb) - skb->head;
	118	skb->csum_offset = offsetof(struct udphdr, check);
e585f236 TH	119	} else if (remcsum) {
	120	/* Need to calculate checksum from scratch,
	121	* inner checksums are never when doing
	122	* remote_checksum_offload.
	123	*/
	124
	125	skb->csum = skb_checksum(skb, udp_offset,
	126	skb->len - udp_offset,
	127	0);
	128	uh->check = csum_fold(skb->csum);
	129	if (uh->check == 0)
	130	uh->check = CSUM_MANGLED_0;
4bcb877d	131	} else {
155e010e TH	132	uh->check = gso_make_checksum(skb, ~uh->check);
	133
	134	if (uh->check == 0)
	135	uh->check = CSUM_MANGLED_0;
	136	}
155e010e TH	137	} while ((skb = skb->next));
	138	out:
	139	return segs;
	140	}
	141
8bce6d7d TH	142	struct sk_buff skb_udp_tunnel_segment(struct sk_buff skb,
	143	netdev_features_t features,
	144	bool is_ipv6)
	145	{
	146	__be16 protocol = skb->protocol;
	147	const struct net_offload **offloads;
	148	const struct net_offload *ops;
	149	struct sk_buff *segs = ERR_PTR(-EINVAL);
	150	struct sk_buff (gso_inner_segment)(struct sk_buff *skb,
	151	netdev_features_t features);
	152
	153	rcu_read_lock();
	154
	155	switch (skb->inner_protocol_type) {
	156	case ENCAP_TYPE_ETHER:
	157	protocol = skb->inner_protocol;
	158	gso_inner_segment = skb_mac_gso_segment;
	159	break;
	160	case ENCAP_TYPE_IPPROTO:
	161	offloads = is_ipv6 ? inet6_offloads : inet_offloads;
	162	ops = rcu_dereference(offloads[skb->inner_ipproto]);
	163	if (!ops \|\| !ops->callbacks.gso_segment)
	164	goto out_unlock;
	165	gso_inner_segment = ops->callbacks.gso_segment;
	166	break;
	167	default:
	168	goto out_unlock;
	169	}
	170
	171	segs = __skb_udp_tunnel_segment(skb, features, gso_inner_segment,
4bcb877d	172	protocol, is_ipv6);
8bce6d7d TH	173
	174	out_unlock:
	175	rcu_read_unlock();
	176
	177	return segs;
	178	}
	179
da5bab07 DB	180	static struct sk_buff udp4_ufo_fragment(struct sk_buff skb,
	181	netdev_features_t features)
	182	{
	183	struct sk_buff *segs = ERR_PTR(-EINVAL);
	184	unsigned int mss;
7a7ffbab	185	__wsum csum;
f71470b3 TH	186	struct udphdr *uh;
f71470b3 TH	187	struct iphdr *iph;
7a7ffbab WCC	188
7a7ffbab WCC	189	if (skb->encapsulation &&
0f4f4ffa TH	190	(skb_shinfo(skb)->gso_type &
0f4f4ffa TH	191	(SKB_GSO_UDP_TUNNEL\|SKB_GSO_UDP_TUNNEL_CSUM))) {
8bce6d7d	192	segs = skb_udp_tunnel_segment(skb, features, false);
7a7ffbab WCC	193	goto out;
7a7ffbab WCC	194	}
da5bab07	195
f71470b3 TH	196	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
	197	goto out;
	198
da5bab07 DB	199	mss = skb_shinfo(skb)->gso_size;
	200	if (unlikely(skb->len <= mss))
	201	goto out;
	202
	203	if (skb_gso_ok(skb, features \| NETIF_F_GSO_ROBUST)) {
	204	/* Packet is from an untrusted source, reset gso_segs. */
	205	int type = skb_shinfo(skb)->gso_type;
	206
	207	if (unlikely(type & ~(SKB_GSO_UDP \| SKB_GSO_DODGY \|
	208	SKB_GSO_UDP_TUNNEL \|
0f4f4ffa	209	SKB_GSO_UDP_TUNNEL_CSUM \|
e585f236	210	SKB_GSO_TUNNEL_REMCSUM \|
cb32f511	211	SKB_GSO_IPIP \|
59b93b41	212	SKB_GSO_GRE \| SKB_GSO_GRE_CSUM) \|\|
da5bab07 DB	213	!(type & (SKB_GSO_UDP))))
	214	goto out;
	215
	216	skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);
	217
	218	segs = NULL;
	219	goto out;
	220	}
	221
7a7ffbab WCC	222	/* Do software UFO. Complete and fill in the UDP checksum as
	223	* HW cannot do checksum of UDP packets sent as multiple
	224	* IP fragments.
	225	*/
f71470b3 TH	226
	227	uh = udp_hdr(skb);
	228	iph = ip_hdr(skb);
	229
	230	uh->check = 0;
	231	csum = skb_checksum(skb, 0, skb->len, 0);
	232	uh->check = udp_v4_check(skb->len, iph->saddr, iph->daddr, csum);
	233	if (uh->check == 0)
	234	uh->check = CSUM_MANGLED_0;
	235
7a7ffbab WCC	236	skb->ip_summed = CHECKSUM_NONE;
7a7ffbab WCC	237
da5bab07 DB	238	/* Fragment the skb. IP headers of the fragments are updated in
	239	* inet_gso_segment()
	240	*/
7a7ffbab	241	segs = skb_segment(skb, features);
da5bab07 DB	242	out:
	243	return segs;
	244	}
	245
787d7ac3	246	int udp_add_offload(struct net net, struct udp_offload uo)
b582ef09	247	{
b5aaab12	248	struct udp_offload_priv new_offload = kzalloc(sizeof(new_offload), GFP_ATOMIC);
b582ef09 OG	249
	250	if (!new_offload)
	251	return -ENOMEM;
	252
787d7ac3	253	write_pnet(&new_offload->net, net);
b582ef09 OG	254	new_offload->offload = uo;
	255
	256	spin_lock(&udp_offload_lock);
a664a4f7 SP	257	new_offload->next = udp_offload_base;
a664a4f7 SP	258	rcu_assign_pointer(udp_offload_base, new_offload);
b582ef09 OG	259	spin_unlock(&udp_offload_lock);
	260
	261	return 0;
	262	}
	263	EXPORT_SYMBOL(udp_add_offload);
	264
	265	static void udp_offload_free_routine(struct rcu_head *head)
	266	{
	267	struct udp_offload_priv *ou_priv = container_of(head, struct udp_offload_priv, rcu);
	268	kfree(ou_priv);
	269	}
	270
	271	void udp_del_offload(struct udp_offload *uo)
	272	{
	273	struct udp_offload_priv __rcu **head = &udp_offload_base;
	274	struct udp_offload_priv *uo_priv;
	275
	276	spin_lock(&udp_offload_lock);
	277
a664a4f7	278	uo_priv = udp_deref_protected(*head);
b582ef09	279	for (; uo_priv != NULL;
a664a4f7	280	uo_priv = udp_deref_protected(*head)) {
b582ef09	281	if (uo_priv->offload == uo) {
a664a4f7 SP	282	rcu_assign_pointer(*head,
a664a4f7 SP	283	udp_deref_protected(uo_priv->next));
b582ef09 OG	284	goto unlock;
	285	}
	286	head = &uo_priv->next;
	287	}
a1d0cd8e	288	pr_warn("udp_del_offload: didn't find offload for port %d\n", ntohs(uo->port));
b582ef09 OG	289	unlock:
b582ef09 OG	290	spin_unlock(&udp_offload_lock);
00db4124	291	if (uo_priv)
b582ef09 OG	292	call_rcu(&uo_priv->rcu, udp_offload_free_routine);
	293	}
	294	EXPORT_SYMBOL(udp_del_offload);
	295
57c67ff4 TH	296	struct sk_buff udp_gro_receive(struct sk_buff head, struct sk_buff *skb,
57c67ff4 TH	297	struct udphdr *uh)
b582ef09 OG	298	{
	299	struct udp_offload_priv *uo_priv;
	300	struct sk_buff p, *pp = NULL;
57c67ff4 TH	301	struct udphdr *uh2;
57c67ff4 TH	302	unsigned int off = skb_gro_offset(skb);
b582ef09 OG	303	int flush = 1;
	304
	305	if (NAPI_GRO_CB(skb)->udp_mark \|\|
662880f4 TH	306	(skb->ip_summed != CHECKSUM_PARTIAL &&
	307	NAPI_GRO_CB(skb)->csum_cnt == 0 &&
	308	!NAPI_GRO_CB(skb)->csum_valid))
b582ef09 OG	309	goto out;
	310
	311	/* mark that this skb passed once through the udp gro layer */
	312	NAPI_GRO_CB(skb)->udp_mark = 1;
b582ef09 OG	313
	314	rcu_read_lock();
	315	uo_priv = rcu_dereference(udp_offload_base);
	316	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
787d7ac3 HFS	317	if (net_eq(read_pnet(&uo_priv->net), dev_net(skb->dev)) &&
787d7ac3 HFS	318	uo_priv->offload->port == uh->dest &&
b582ef09 OG	319	uo_priv->offload->callbacks.gro_receive)
	320	goto unflush;
	321	}
	322	goto out_unlock;
	323
	324	unflush:
	325	flush = 0;
	326
	327	for (p = *head; p; p = p->next) {
	328	if (!NAPI_GRO_CB(p)->same_flow)
	329	continue;
	330
	331	uh2 = (struct udphdr *)(p->data + off);
57c67ff4 TH	332
	333	/* Match ports and either checksums are either both zero
	334	* or nonzero.
	335	*/
	336	if (((u32 )&uh->source != (u32 )&uh2->source) \|\|
	337	(!uh->check ^ !uh2->check)) {
b582ef09 OG	338	NAPI_GRO_CB(p)->same_flow = 0;
	339	continue;
	340	}
	341	}
	342
	343	skb_gro_pull(skb, sizeof(struct udphdr)); /* pull encapsulating udp header */
6bae1d4c	344	skb_gro_postpull_rcsum(skb, uh, sizeof(struct udphdr));
afe93325	345	NAPI_GRO_CB(skb)->proto = uo_priv->offload->ipproto;
a2b12f3c TH	346	pp = uo_priv->offload->callbacks.gro_receive(head, skb,
a2b12f3c TH	347	uo_priv->offload);
b582ef09 OG	348
	349	out_unlock:
	350	rcu_read_unlock();
	351	out:
	352	NAPI_GRO_CB(skb)->flush \|= flush;
	353	return pp;
	354	}
	355
57c67ff4 TH	356	static struct sk_buff udp4_gro_receive(struct sk_buff head,
	357	struct sk_buff *skb)
	358	{
	359	struct udphdr *uh = udp_gro_udphdr(skb);
	360
2abb7cdc TH	361	if (unlikely(!uh))
2abb7cdc TH	362	goto flush;
57c67ff4	363
2abb7cdc	364	/* Don't bother verifying checksum if we're going to flush anyway. */
2d8f7e2c	365	if (NAPI_GRO_CB(skb)->flush)
2abb7cdc TH	366	goto skip;
	367
	368	if (skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
	369	inet_gro_compute_pseudo))
	370	goto flush;
	371	else if (uh->check)
	372	skb_gro_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
	373	inet_gro_compute_pseudo);
	374	skip:
efc98d08	375	NAPI_GRO_CB(skb)->is_ipv6 = 0;
57c67ff4	376	return udp_gro_receive(head, skb, uh);
2abb7cdc TH	377
	378	flush:
	379	NAPI_GRO_CB(skb)->flush = 1;
	380	return NULL;
57c67ff4 TH	381	}
	382
	383	int udp_gro_complete(struct sk_buff *skb, int nhoff)
b582ef09 OG	384	{
	385	struct udp_offload_priv *uo_priv;
	386	__be16 newlen = htons(skb->len - nhoff);
	387	struct udphdr uh = (struct udphdr )(skb->data + nhoff);
	388	int err = -ENOSYS;
	389
	390	uh->len = newlen;
	391
	392	rcu_read_lock();
	393
	394	uo_priv = rcu_dereference(udp_offload_base);
	395	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
787d7ac3 HFS	396	if (net_eq(read_pnet(&uo_priv->net), dev_net(skb->dev)) &&
787d7ac3 HFS	397	uo_priv->offload->port == uh->dest &&
b582ef09 OG	398	uo_priv->offload->callbacks.gro_complete)
	399	break;
	400	}
	401
00db4124	402	if (uo_priv) {
afe93325	403	NAPI_GRO_CB(skb)->proto = uo_priv->offload->ipproto;
a2b12f3c TH	404	err = uo_priv->offload->callbacks.gro_complete(skb,
	405	nhoff + sizeof(struct udphdr),
	406	uo_priv->offload);
afe93325	407	}
b582ef09 OG	408
b582ef09 OG	409	rcu_read_unlock();
6db93ea1 TH	410
	411	if (skb->remcsum_offload)
	412	skb_shinfo(skb)->gso_type \|= SKB_GSO_TUNNEL_REMCSUM;
	413
	414	skb->encapsulation = 1;
	415	skb_set_inner_mac_header(skb, nhoff + sizeof(struct udphdr));
	416
b582ef09 OG	417	return err;
	418	}
	419
72bb17b3	420	static int udp4_gro_complete(struct sk_buff *skb, int nhoff)
57c67ff4 TH	421	{
	422	const struct iphdr *iph = ip_hdr(skb);
	423	struct udphdr uh = (struct udphdr )(skb->data + nhoff);
	424
6db93ea1 TH	425	if (uh->check) {
6db93ea1 TH	426	skb_shinfo(skb)->gso_type \|= SKB_GSO_UDP_TUNNEL_CSUM;
57c67ff4 TH	427	uh->check = ~udp_v4_check(skb->len - nhoff, iph->saddr,
57c67ff4 TH	428	iph->daddr, 0);
6db93ea1 TH	429	} else {
	430	skb_shinfo(skb)->gso_type \|= SKB_GSO_UDP_TUNNEL;
	431	}
57c67ff4 TH	432
	433	return udp_gro_complete(skb, nhoff);
	434	}
	435
da5bab07 DB	436	static const struct net_offload udpv4_offload = {
da5bab07 DB	437	.callbacks = {
da5bab07	438	.gso_segment = udp4_ufo_fragment,
57c67ff4 TH	439	.gro_receive = udp4_gro_receive,
57c67ff4 TH	440	.gro_complete = udp4_gro_complete,
da5bab07 DB	441	},
	442	};
	443
	444	int __init udpv4_offload_init(void)
	445	{
	446	return inet_add_offload(&udpv4_offload, IPPROTO_UDP);
	447	}