[linux-2.6-block.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;
	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 &
			   (is_ipv6 ? NETIF_F_V6_CSUM : NETIF_F_V4_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 udp_offload *uo)
{
	struct udp_offload_priv *new_offload = kzalloc(sizeof(*new_offload), GFP_ATOMIC);

	if (!new_offload)
		return -ENOMEM;

	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 != NULL)
		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 (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);

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 (uo_priv->offload->port == uh->dest &&
		    uo_priv->offload->callbacks.gro_complete)
			break;
	}

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

	rcu_read_unlock();
	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)
		uh->check = ~udp_v4_check(skb->len - nhoff, iph->saddr,
					  iph->daddr, 0);

	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 {
	23	struct udp_offload *offload;
	24	struct rcu_head rcu;
	25	struct udp_offload_priv __rcu *next;
	26	};
	27
8bce6d7d TH	28	static struct sk_buff __skb_udp_tunnel_segment(struct sk_buff skb,
	29	netdev_features_t features,
	30	struct sk_buff (gso_inner_segment)(struct sk_buff *skb,
	31	netdev_features_t features),
4bcb877d	32	__be16 new_protocol, bool is_ipv6)
155e010e TH	33	{
	34	struct sk_buff *segs = ERR_PTR(-EINVAL);
	35	u16 mac_offset = skb->mac_header;
	36	int mac_len = skb->mac_len;
	37	int tnl_hlen = skb_inner_mac_header(skb) - skb_transport_header(skb);
	38	__be16 protocol = skb->protocol;
	39	netdev_features_t enc_features;
	40	int udp_offset, outer_hlen;
	41	unsigned int oldlen;
4bcb877d TH	42	bool need_csum = !!(skb_shinfo(skb)->gso_type &
4bcb877d TH	43	SKB_GSO_UDP_TUNNEL_CSUM);
e585f236 TH	44	bool remcsum = !!(skb_shinfo(skb)->gso_type & SKB_GSO_TUNNEL_REMCSUM);
e585f236 TH	45	bool offload_csum = false, dont_encap = (need_csum \|\| remcsum);
155e010e TH	46
	47	oldlen = (u16)~skb->len;
	48
	49	if (unlikely(!pskb_may_pull(skb, tnl_hlen)))
	50	goto out;
	51
	52	skb->encapsulation = 0;
	53	__skb_pull(skb, tnl_hlen);
	54	skb_reset_mac_header(skb);
	55	skb_set_network_header(skb, skb_inner_network_offset(skb));
	56	skb->mac_len = skb_inner_network_offset(skb);
8bce6d7d	57	skb->protocol = new_protocol;
4bcb877d	58	skb->encap_hdr_csum = need_csum;
e585f236	59	skb->remcsum_offload = remcsum;
155e010e	60
4bcb877d TH	61	/* Try to offload checksum if possible */
	62	offload_csum = !!(need_csum &&
	63	(skb->dev->features &
	64	(is_ipv6 ? NETIF_F_V6_CSUM : NETIF_F_V4_CSUM)));
155e010e TH	65
155e010e TH	66	/* segment inner packet. */
1e16aa3d	67	enc_features = skb->dev->hw_enc_features & features;
8bce6d7d	68	segs = gso_inner_segment(skb, enc_features);
27446442	69	if (IS_ERR_OR_NULL(segs)) {
155e010e TH	70	skb_gso_error_unwind(skb, protocol, tnl_hlen, mac_offset,
	71	mac_len);
	72	goto out;
	73	}
	74
	75	outer_hlen = skb_tnl_header_len(skb);
	76	udp_offset = outer_hlen - tnl_hlen;
	77	skb = segs;
	78	do {
	79	struct udphdr *uh;
	80	int len;
4bcb877d TH	81	__be32 delta;
	82
	83	if (dont_encap) {
	84	skb->encapsulation = 0;
	85	skb->ip_summed = CHECKSUM_NONE;
	86	} else {
	87	/* Only set up inner headers if we might be offloading
	88	* inner checksum.
	89	*/
	90	skb_reset_inner_headers(skb);
	91	skb->encapsulation = 1;
	92	}
155e010e TH	93
155e010e TH	94	skb->mac_len = mac_len;
4bcb877d	95	skb->protocol = protocol;
155e010e TH	96
	97	skb_push(skb, outer_hlen);
	98	skb_reset_mac_header(skb);
	99	skb_set_network_header(skb, mac_len);
	100	skb_set_transport_header(skb, udp_offset);
	101	len = skb->len - udp_offset;
	102	uh = udp_hdr(skb);
	103	uh->len = htons(len);
	104
4bcb877d TH	105	if (!need_csum)
	106	continue;
	107
	108	delta = htonl(oldlen + len);
155e010e	109
4bcb877d TH	110	uh->check = ~csum_fold((__force __wsum)
	111	((__force u32)uh->check +
	112	(__force u32)delta));
4bcb877d TH	113	if (offload_csum) {
	114	skb->ip_summed = CHECKSUM_PARTIAL;
	115	skb->csum_start = skb_transport_header(skb) - skb->head;
	116	skb->csum_offset = offsetof(struct udphdr, check);
e585f236 TH	117	} else if (remcsum) {
	118	/* Need to calculate checksum from scratch,
	119	* inner checksums are never when doing
	120	* remote_checksum_offload.
	121	*/
	122
	123	skb->csum = skb_checksum(skb, udp_offset,
	124	skb->len - udp_offset,
	125	0);
	126	uh->check = csum_fold(skb->csum);
	127	if (uh->check == 0)
	128	uh->check = CSUM_MANGLED_0;
4bcb877d	129	} else {
155e010e TH	130	uh->check = gso_make_checksum(skb, ~uh->check);
	131
	132	if (uh->check == 0)
	133	uh->check = CSUM_MANGLED_0;
	134	}
155e010e TH	135	} while ((skb = skb->next));
	136	out:
	137	return segs;
	138	}
	139
8bce6d7d TH	140	struct sk_buff skb_udp_tunnel_segment(struct sk_buff skb,
	141	netdev_features_t features,
	142	bool is_ipv6)
	143	{
	144	__be16 protocol = skb->protocol;
	145	const struct net_offload **offloads;
	146	const struct net_offload *ops;
	147	struct sk_buff *segs = ERR_PTR(-EINVAL);
	148	struct sk_buff (gso_inner_segment)(struct sk_buff *skb,
	149	netdev_features_t features);
	150
	151	rcu_read_lock();
	152
	153	switch (skb->inner_protocol_type) {
	154	case ENCAP_TYPE_ETHER:
	155	protocol = skb->inner_protocol;
	156	gso_inner_segment = skb_mac_gso_segment;
	157	break;
	158	case ENCAP_TYPE_IPPROTO:
	159	offloads = is_ipv6 ? inet6_offloads : inet_offloads;
	160	ops = rcu_dereference(offloads[skb->inner_ipproto]);
	161	if (!ops \|\| !ops->callbacks.gso_segment)
	162	goto out_unlock;
	163	gso_inner_segment = ops->callbacks.gso_segment;
	164	break;
	165	default:
	166	goto out_unlock;
	167	}
	168
	169	segs = __skb_udp_tunnel_segment(skb, features, gso_inner_segment,
4bcb877d	170	protocol, is_ipv6);
8bce6d7d TH	171
	172	out_unlock:
	173	rcu_read_unlock();
	174
	175	return segs;
	176	}
	177
da5bab07 DB	178	static struct sk_buff udp4_ufo_fragment(struct sk_buff skb,
	179	netdev_features_t features)
	180	{
	181	struct sk_buff *segs = ERR_PTR(-EINVAL);
	182	unsigned int mss;
7a7ffbab	183	__wsum csum;
f71470b3 TH	184	struct udphdr *uh;
f71470b3 TH	185	struct iphdr *iph;
7a7ffbab WCC	186
7a7ffbab WCC	187	if (skb->encapsulation &&
0f4f4ffa TH	188	(skb_shinfo(skb)->gso_type &
0f4f4ffa TH	189	(SKB_GSO_UDP_TUNNEL\|SKB_GSO_UDP_TUNNEL_CSUM))) {
8bce6d7d	190	segs = skb_udp_tunnel_segment(skb, features, false);
7a7ffbab WCC	191	goto out;
7a7ffbab WCC	192	}
da5bab07	193
f71470b3 TH	194	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
	195	goto out;
	196
da5bab07 DB	197	mss = skb_shinfo(skb)->gso_size;
	198	if (unlikely(skb->len <= mss))
	199	goto out;
	200
	201	if (skb_gso_ok(skb, features \| NETIF_F_GSO_ROBUST)) {
	202	/* Packet is from an untrusted source, reset gso_segs. */
	203	int type = skb_shinfo(skb)->gso_type;
	204
	205	if (unlikely(type & ~(SKB_GSO_UDP \| SKB_GSO_DODGY \|
	206	SKB_GSO_UDP_TUNNEL \|
0f4f4ffa	207	SKB_GSO_UDP_TUNNEL_CSUM \|
e585f236	208	SKB_GSO_TUNNEL_REMCSUM \|
cb32f511	209	SKB_GSO_IPIP \|
59b93b41	210	SKB_GSO_GRE \| SKB_GSO_GRE_CSUM) \|\|
da5bab07 DB	211	!(type & (SKB_GSO_UDP))))
	212	goto out;
	213
	214	skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);
	215
	216	segs = NULL;
	217	goto out;
	218	}
	219
7a7ffbab WCC	220	/* Do software UFO. Complete and fill in the UDP checksum as
	221	* HW cannot do checksum of UDP packets sent as multiple
	222	* IP fragments.
	223	*/
f71470b3 TH	224
	225	uh = udp_hdr(skb);
	226	iph = ip_hdr(skb);
	227
	228	uh->check = 0;
	229	csum = skb_checksum(skb, 0, skb->len, 0);
	230	uh->check = udp_v4_check(skb->len, iph->saddr, iph->daddr, csum);
	231	if (uh->check == 0)
	232	uh->check = CSUM_MANGLED_0;
	233
7a7ffbab WCC	234	skb->ip_summed = CHECKSUM_NONE;
7a7ffbab WCC	235
da5bab07 DB	236	/* Fragment the skb. IP headers of the fragments are updated in
	237	* inet_gso_segment()
	238	*/
7a7ffbab	239	segs = skb_segment(skb, features);
da5bab07 DB	240	out:
	241	return segs;
	242	}
	243
b582ef09 OG	244	int udp_add_offload(struct udp_offload *uo)
b582ef09 OG	245	{
b5aaab12	246	struct udp_offload_priv new_offload = kzalloc(sizeof(new_offload), GFP_ATOMIC);
b582ef09 OG	247
	248	if (!new_offload)
	249	return -ENOMEM;
	250
	251	new_offload->offload = uo;
	252
	253	spin_lock(&udp_offload_lock);
a664a4f7 SP	254	new_offload->next = udp_offload_base;
a664a4f7 SP	255	rcu_assign_pointer(udp_offload_base, new_offload);
b582ef09 OG	256	spin_unlock(&udp_offload_lock);
	257
	258	return 0;
	259	}
	260	EXPORT_SYMBOL(udp_add_offload);
	261
	262	static void udp_offload_free_routine(struct rcu_head *head)
	263	{
	264	struct udp_offload_priv *ou_priv = container_of(head, struct udp_offload_priv, rcu);
	265	kfree(ou_priv);
	266	}
	267
	268	void udp_del_offload(struct udp_offload *uo)
	269	{
	270	struct udp_offload_priv __rcu **head = &udp_offload_base;
	271	struct udp_offload_priv *uo_priv;
	272
	273	spin_lock(&udp_offload_lock);
	274
a664a4f7	275	uo_priv = udp_deref_protected(*head);
b582ef09	276	for (; uo_priv != NULL;
a664a4f7	277	uo_priv = udp_deref_protected(*head)) {
b582ef09	278	if (uo_priv->offload == uo) {
a664a4f7 SP	279	rcu_assign_pointer(*head,
a664a4f7 SP	280	udp_deref_protected(uo_priv->next));
b582ef09 OG	281	goto unlock;
	282	}
	283	head = &uo_priv->next;
	284	}
a1d0cd8e	285	pr_warn("udp_del_offload: didn't find offload for port %d\n", ntohs(uo->port));
b582ef09 OG	286	unlock:
	287	spin_unlock(&udp_offload_lock);
	288	if (uo_priv != NULL)
	289	call_rcu(&uo_priv->rcu, udp_offload_free_routine);
	290	}
	291	EXPORT_SYMBOL(udp_del_offload);
	292
57c67ff4 TH	293	struct sk_buff udp_gro_receive(struct sk_buff head, struct sk_buff *skb,
57c67ff4 TH	294	struct udphdr *uh)
b582ef09 OG	295	{
	296	struct udp_offload_priv *uo_priv;
	297	struct sk_buff p, *pp = NULL;
57c67ff4 TH	298	struct udphdr *uh2;
57c67ff4 TH	299	unsigned int off = skb_gro_offset(skb);
b582ef09 OG	300	int flush = 1;
	301
	302	if (NAPI_GRO_CB(skb)->udp_mark \|\|
662880f4 TH	303	(skb->ip_summed != CHECKSUM_PARTIAL &&
	304	NAPI_GRO_CB(skb)->csum_cnt == 0 &&
	305	!NAPI_GRO_CB(skb)->csum_valid))
b582ef09 OG	306	goto out;
	307
	308	/* mark that this skb passed once through the udp gro layer */
	309	NAPI_GRO_CB(skb)->udp_mark = 1;
b582ef09 OG	310
	311	rcu_read_lock();
	312	uo_priv = rcu_dereference(udp_offload_base);
	313	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
	314	if (uo_priv->offload->port == uh->dest &&
	315	uo_priv->offload->callbacks.gro_receive)
	316	goto unflush;
	317	}
	318	goto out_unlock;
	319
	320	unflush:
	321	flush = 0;
	322
	323	for (p = *head; p; p = p->next) {
	324	if (!NAPI_GRO_CB(p)->same_flow)
	325	continue;
	326
	327	uh2 = (struct udphdr *)(p->data + off);
57c67ff4 TH	328
	329	/* Match ports and either checksums are either both zero
	330	* or nonzero.
	331	*/
	332	if (((u32 )&uh->source != (u32 )&uh2->source) \|\|
	333	(!uh->check ^ !uh2->check)) {
b582ef09 OG	334	NAPI_GRO_CB(p)->same_flow = 0;
	335	continue;
	336	}
	337	}
	338
	339	skb_gro_pull(skb, sizeof(struct udphdr)); /* pull encapsulating udp header */
6bae1d4c	340	skb_gro_postpull_rcsum(skb, uh, sizeof(struct udphdr));
afe93325	341	NAPI_GRO_CB(skb)->proto = uo_priv->offload->ipproto;
b582ef09 OG	342	pp = uo_priv->offload->callbacks.gro_receive(head, skb);
	343
	344	out_unlock:
	345	rcu_read_unlock();
	346	out:
	347	NAPI_GRO_CB(skb)->flush \|= flush;
	348	return pp;
	349	}
	350
57c67ff4 TH	351	static struct sk_buff udp4_gro_receive(struct sk_buff head,
	352	struct sk_buff *skb)
	353	{
	354	struct udphdr *uh = udp_gro_udphdr(skb);
	355
2abb7cdc TH	356	if (unlikely(!uh))
2abb7cdc TH	357	goto flush;
57c67ff4	358
2abb7cdc	359	/* Don't bother verifying checksum if we're going to flush anyway. */
2d8f7e2c	360	if (NAPI_GRO_CB(skb)->flush)
2abb7cdc TH	361	goto skip;
	362
	363	if (skb_gro_checksum_validate_zero_check(skb, IPPROTO_UDP, uh->check,
	364	inet_gro_compute_pseudo))
	365	goto flush;
	366	else if (uh->check)
	367	skb_gro_checksum_try_convert(skb, IPPROTO_UDP, uh->check,
	368	inet_gro_compute_pseudo);
	369	skip:
efc98d08	370	NAPI_GRO_CB(skb)->is_ipv6 = 0;
57c67ff4	371	return udp_gro_receive(head, skb, uh);
2abb7cdc TH	372
	373	flush:
	374	NAPI_GRO_CB(skb)->flush = 1;
	375	return NULL;
57c67ff4 TH	376	}
	377
	378	int udp_gro_complete(struct sk_buff *skb, int nhoff)
b582ef09 OG	379	{
	380	struct udp_offload_priv *uo_priv;
	381	__be16 newlen = htons(skb->len - nhoff);
	382	struct udphdr uh = (struct udphdr )(skb->data + nhoff);
	383	int err = -ENOSYS;
	384
	385	uh->len = newlen;
	386
	387	rcu_read_lock();
	388
	389	uo_priv = rcu_dereference(udp_offload_base);
	390	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
	391	if (uo_priv->offload->port == uh->dest &&
	392	uo_priv->offload->callbacks.gro_complete)
	393	break;
	394	}
	395
afe93325 TH	396	if (uo_priv != NULL) {
afe93325 TH	397	NAPI_GRO_CB(skb)->proto = uo_priv->offload->ipproto;
b582ef09	398	err = uo_priv->offload->callbacks.gro_complete(skb, nhoff + sizeof(struct udphdr));
afe93325	399	}
b582ef09 OG	400
	401	rcu_read_unlock();
	402	return err;
	403	}
	404
72bb17b3	405	static int udp4_gro_complete(struct sk_buff *skb, int nhoff)
57c67ff4 TH	406	{
	407	const struct iphdr *iph = ip_hdr(skb);
	408	struct udphdr uh = (struct udphdr )(skb->data + nhoff);
	409
	410	if (uh->check)
	411	uh->check = ~udp_v4_check(skb->len - nhoff, iph->saddr,
	412	iph->daddr, 0);
	413
	414	return udp_gro_complete(skb, nhoff);
	415	}
	416
da5bab07 DB	417	static const struct net_offload udpv4_offload = {
da5bab07 DB	418	.callbacks = {
da5bab07	419	.gso_segment = udp4_ufo_fragment,
57c67ff4 TH	420	.gro_receive = udp4_gro_receive,
57c67ff4 TH	421	.gro_complete = udp4_gro_complete,
da5bab07 DB	422	},
	423	};
	424
	425	int __init udpv4_offload_init(void)
	426	{
	427	return inet_add_offload(&udpv4_offload, IPPROTO_UDP);
	428	}