[mirror_ubuntu-zesty-kernel.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 int udp4_ufo_send_check(struct sk_buff *skb)
{
	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
		return -EINVAL;

	if (likely(!skb->encapsulation)) {
		const struct iphdr *iph;
		struct udphdr *uh;

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

		uh->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, skb->len,
				IPPROTO_UDP, 0);
		skb->csum_start = skb_transport_header(skb) - skb->head;
		skb->csum_offset = offsetof(struct udphdr, check);
		skb->ip_summed = CHECKSUM_PARTIAL;
	}

	return 0;
}

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;
	int offset;
	__wsum csum;

	if (skb->encapsulation &&
	    skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL) {
		segs = skb_udp_tunnel_segment(skb, features);
		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_IPIP |
				      SKB_GSO_GRE | SKB_GSO_MPLS) ||
			     !(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.
	 */
	offset = skb_checksum_start_offset(skb);
	csum = skb_checksum(skb, offset, skb->len - offset, 0);
	offset += skb->csum_offset;
	*(__sum16 *)(skb->data + offset) = csum_fold(csum);
	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);

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

	if (NAPI_GRO_CB(skb)->udp_mark ||
	    (!skb->encapsulation && skb->ip_summed != CHECKSUM_COMPLETE))
		goto out;

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

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

	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);
		if ((*(u32 *)&uh->source != *(u32 *)&uh2->source)) {
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}
	}

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

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

static 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)
		err = uo_priv->offload->callbacks.gro_complete(skb, nhoff + sizeof(struct udphdr));

	rcu_read_unlock();
	return err;
}

static const struct net_offload udpv4_offload = {
	.callbacks = {
		.gso_send_check = udp4_ufo_send_check,
		.gso_segment = udp4_ufo_fragment,
		.gro_receive  =	udp_gro_receive,
		.gro_complete =	udp_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
da5bab07 DB	28	static int udp4_ufo_send_check(struct sk_buff *skb)
	29	{
	30	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
	31	return -EINVAL;
	32
	33	if (likely(!skb->encapsulation)) {
	34	const struct iphdr *iph;
	35	struct udphdr *uh;
	36
	37	iph = ip_hdr(skb);
	38	uh = udp_hdr(skb);
	39
	40	uh->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, skb->len,
	41	IPPROTO_UDP, 0);
	42	skb->csum_start = skb_transport_header(skb) - skb->head;
	43	skb->csum_offset = offsetof(struct udphdr, check);
	44	skb->ip_summed = CHECKSUM_PARTIAL;
	45	}
	46
	47	return 0;
	48	}
	49
	50	static struct sk_buff udp4_ufo_fragment(struct sk_buff skb,
	51	netdev_features_t features)
	52	{
	53	struct sk_buff *segs = ERR_PTR(-EINVAL);
	54	unsigned int mss;
7a7ffbab WCC	55	int offset;
	56	__wsum csum;
	57
	58	if (skb->encapsulation &&
	59	skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL) {
	60	segs = skb_udp_tunnel_segment(skb, features);
	61	goto out;
	62	}
da5bab07 DB	63
	64	mss = skb_shinfo(skb)->gso_size;
	65	if (unlikely(skb->len <= mss))
	66	goto out;
	67
	68	if (skb_gso_ok(skb, features \| NETIF_F_GSO_ROBUST)) {
	69	/* Packet is from an untrusted source, reset gso_segs. */
	70	int type = skb_shinfo(skb)->gso_type;
	71
	72	if (unlikely(type & ~(SKB_GSO_UDP \| SKB_GSO_DODGY \|
	73	SKB_GSO_UDP_TUNNEL \|
cb32f511	74	SKB_GSO_IPIP \|
da5bab07 DB	75	SKB_GSO_GRE \| SKB_GSO_MPLS) \|\|
	76	!(type & (SKB_GSO_UDP))))
	77	goto out;
	78
	79	skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);
	80
	81	segs = NULL;
	82	goto out;
	83	}
	84
7a7ffbab WCC	85	/* Do software UFO. Complete and fill in the UDP checksum as
	86	* HW cannot do checksum of UDP packets sent as multiple
	87	* IP fragments.
	88	*/
	89	offset = skb_checksum_start_offset(skb);
	90	csum = skb_checksum(skb, offset, skb->len - offset, 0);
	91	offset += skb->csum_offset;
	92	(__sum16 )(skb->data + offset) = csum_fold(csum);
	93	skb->ip_summed = CHECKSUM_NONE;
	94
da5bab07 DB	95	/* Fragment the skb. IP headers of the fragments are updated in
	96	* inet_gso_segment()
	97	*/
7a7ffbab	98	segs = skb_segment(skb, features);
da5bab07 DB	99	out:
	100	return segs;
	101	}
	102
b582ef09 OG	103	int udp_add_offload(struct udp_offload *uo)
b582ef09 OG	104	{
b5aaab12	105	struct udp_offload_priv new_offload = kzalloc(sizeof(new_offload), GFP_ATOMIC);
b582ef09 OG	106
	107	if (!new_offload)
	108	return -ENOMEM;
	109
	110	new_offload->offload = uo;
	111
	112	spin_lock(&udp_offload_lock);
a664a4f7 SP	113	new_offload->next = udp_offload_base;
a664a4f7 SP	114	rcu_assign_pointer(udp_offload_base, new_offload);
b582ef09 OG	115	spin_unlock(&udp_offload_lock);
	116
	117	return 0;
	118	}
	119	EXPORT_SYMBOL(udp_add_offload);
	120
	121	static void udp_offload_free_routine(struct rcu_head *head)
	122	{
	123	struct udp_offload_priv *ou_priv = container_of(head, struct udp_offload_priv, rcu);
	124	kfree(ou_priv);
	125	}
	126
	127	void udp_del_offload(struct udp_offload *uo)
	128	{
	129	struct udp_offload_priv __rcu **head = &udp_offload_base;
	130	struct udp_offload_priv *uo_priv;
	131
	132	spin_lock(&udp_offload_lock);
	133
a664a4f7	134	uo_priv = udp_deref_protected(*head);
b582ef09	135	for (; uo_priv != NULL;
a664a4f7	136	uo_priv = udp_deref_protected(*head)) {
b582ef09	137	if (uo_priv->offload == uo) {
a664a4f7 SP	138	rcu_assign_pointer(*head,
a664a4f7 SP	139	udp_deref_protected(uo_priv->next));
b582ef09 OG	140	goto unlock;
	141	}
	142	head = &uo_priv->next;
	143	}
a1d0cd8e	144	pr_warn("udp_del_offload: didn't find offload for port %d\n", ntohs(uo->port));
b582ef09 OG	145	unlock:
	146	spin_unlock(&udp_offload_lock);
	147	if (uo_priv != NULL)
	148	call_rcu(&uo_priv->rcu, udp_offload_free_routine);
	149	}
	150	EXPORT_SYMBOL(udp_del_offload);
	151
	152	static struct sk_buff udp_gro_receive(struct sk_buff head, struct sk_buff *skb)
	153	{
	154	struct udp_offload_priv *uo_priv;
	155	struct sk_buff p, *pp = NULL;
	156	struct udphdr uh, uh2;
	157	unsigned int hlen, off;
	158	int flush = 1;
	159
	160	if (NAPI_GRO_CB(skb)->udp_mark \|\|
	161	(!skb->encapsulation && skb->ip_summed != CHECKSUM_COMPLETE))
	162	goto out;
	163
	164	/* mark that this skb passed once through the udp gro layer */
	165	NAPI_GRO_CB(skb)->udp_mark = 1;
	166
	167	off = skb_gro_offset(skb);
	168	hlen = off + sizeof(*uh);
	169	uh = skb_gro_header_fast(skb, off);
	170	if (skb_gro_header_hard(skb, hlen)) {
	171	uh = skb_gro_header_slow(skb, hlen, off);
	172	if (unlikely(!uh))
	173	goto out;
	174	}
	175
	176	rcu_read_lock();
	177	uo_priv = rcu_dereference(udp_offload_base);
	178	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
	179	if (uo_priv->offload->port == uh->dest &&
	180	uo_priv->offload->callbacks.gro_receive)
	181	goto unflush;
	182	}
	183	goto out_unlock;
	184
	185	unflush:
	186	flush = 0;
	187
	188	for (p = *head; p; p = p->next) {
	189	if (!NAPI_GRO_CB(p)->same_flow)
	190	continue;
	191
	192	uh2 = (struct udphdr *)(p->data + off);
	193	if (((u32 )&uh->source != (u32 )&uh2->source)) {
	194	NAPI_GRO_CB(p)->same_flow = 0;
	195	continue;
	196	}
	197	}
	198
	199	skb_gro_pull(skb, sizeof(struct udphdr)); /* pull encapsulating udp header */
	200	pp = uo_priv->offload->callbacks.gro_receive(head, skb);
	201
	202	out_unlock:
	203	rcu_read_unlock();
	204	out:
	205	NAPI_GRO_CB(skb)->flush \|= flush;
	206	return pp;
	207	}
	208
209	static int udp_gro_complete(struct sk_buff *skb, int nhoff)
210	{
211	struct udp_offload_priv *uo_priv;
212	__be16 newlen = htons(skb->len - nhoff);
213	struct udphdr uh = (struct udphdr )(skb->data + nhoff);
214	int err = -ENOSYS;
215
216	uh->len = newlen;
217
218	rcu_read_lock();
219
220	uo_priv = rcu_dereference(udp_offload_base);
221	for (; uo_priv != NULL; uo_priv = rcu_dereference(uo_priv->next)) {
222	if (uo_priv->offload->port == uh->dest &&
223	uo_priv->offload->callbacks.gro_complete)
224	break;
225	}
226
227	if (uo_priv != NULL)
228	err = uo_priv->offload->callbacks.gro_complete(skb, nhoff + sizeof(struct udphdr));
229
230	rcu_read_unlock();
231	return err;
232	}
233
da5bab07 DB	234	static const struct net_offload udpv4_offload = {
	235	.callbacks = {
	236	.gso_send_check = udp4_ufo_send_check,
	237	.gso_segment = udp4_ufo_fragment,
b582ef09 OG	238	.gro_receive = udp_gro_receive,
b582ef09 OG	239	.gro_complete = udp_gro_complete,
da5bab07 DB	240	},
	241	};
	242
	243	int __init udpv4_offload_init(void)
	244	{
	245	return inet_add_offload(&udpv4_offload, IPPROTO_UDP);
	246	}