[deliverable/linux.git] / net / ipv4 / tcp_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.
 *
 *	TCPv4 GSO/GRO support
 */

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

struct sk_buff *tcp_gso_segment(struct sk_buff *skb,
				netdev_features_t features)
{
	struct sk_buff *segs = ERR_PTR(-EINVAL);
	unsigned int sum_truesize = 0;
	struct tcphdr *th;
	unsigned int thlen;
	unsigned int seq;
	__be32 delta;
	unsigned int oldlen;
	unsigned int mss;
	struct sk_buff *gso_skb = skb;
	__sum16 newcheck;
	bool ooo_okay, copy_destructor;

	if (!pskb_may_pull(skb, sizeof(*th)))
		goto out;

	th = tcp_hdr(skb);
	thlen = th->doff * 4;
	if (thlen < sizeof(*th))
		goto out;

	if (!pskb_may_pull(skb, thlen))
		goto out;

	oldlen = (u16)~skb->len;
	__skb_pull(skb, thlen);

	mss = tcp_skb_mss(skb);
	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_TCPV4 |
			       SKB_GSO_DODGY |
			       SKB_GSO_TCP_ECN |
			       SKB_GSO_TCPV6 |
			       SKB_GSO_GRE |
			       SKB_GSO_GRE_CSUM |
			       SKB_GSO_IPIP |
			       SKB_GSO_SIT |
			       SKB_GSO_MPLS |
			       SKB_GSO_UDP_TUNNEL |
			       SKB_GSO_UDP_TUNNEL_CSUM |
			       0) ||
			     !(type & (SKB_GSO_TCPV4 | SKB_GSO_TCPV6))))
			goto out;

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

		segs = NULL;
		goto out;
	}

	copy_destructor = gso_skb->destructor == tcp_wfree;
	ooo_okay = gso_skb->ooo_okay;
	/* All segments but the first should have ooo_okay cleared */
	skb->ooo_okay = 0;

	segs = skb_segment(skb, features);
	if (IS_ERR(segs))
		goto out;

	/* Only first segment might have ooo_okay set */
	segs->ooo_okay = ooo_okay;

	delta = htonl(oldlen + (thlen + mss));

	skb = segs;
	th = tcp_hdr(skb);
	seq = ntohl(th->seq);

	newcheck = ~csum_fold((__force __wsum)((__force u32)th->check +
					       (__force u32)delta));

	do {
		th->fin = th->psh = 0;
		th->check = newcheck;

		if (skb->ip_summed != CHECKSUM_PARTIAL)
			th->check = gso_make_checksum(skb, ~th->check);

		seq += mss;
		if (copy_destructor) {
			skb->destructor = gso_skb->destructor;
			skb->sk = gso_skb->sk;
			sum_truesize += skb->truesize;
		}
		skb = skb->next;
		th = tcp_hdr(skb);

		th->seq = htonl(seq);
		th->cwr = 0;
	} while (skb->next);

	/* Following permits TCP Small Queues to work well with GSO :
	 * The callback to TCP stack will be called at the time last frag
	 * is freed at TX completion, and not right now when gso_skb
	 * is freed by GSO engine
	 */
	if (copy_destructor) {
		swap(gso_skb->sk, skb->sk);
		swap(gso_skb->destructor, skb->destructor);
		sum_truesize += skb->truesize;
		atomic_add(sum_truesize - gso_skb->truesize,
			   &skb->sk->sk_wmem_alloc);
	}

	delta = htonl(oldlen + (skb_tail_pointer(skb) -
				skb_transport_header(skb)) +
		      skb->data_len);
	th->check = ~csum_fold((__force __wsum)((__force u32)th->check +
				(__force u32)delta));
	if (skb->ip_summed != CHECKSUM_PARTIAL)
		th->check = gso_make_checksum(skb, ~th->check);
out:
	return segs;
}

struct sk_buff **tcp_gro_receive(struct sk_buff **head, struct sk_buff *skb)
{
	struct sk_buff **pp = NULL;
	struct sk_buff *p;
	struct tcphdr *th;
	struct tcphdr *th2;
	unsigned int len;
	unsigned int thlen;
	__be32 flags;
	unsigned int mss = 1;
	unsigned int hlen;
	unsigned int off;
	int flush = 1;
	int i;

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

	thlen = th->doff * 4;
	if (thlen < sizeof(*th))
		goto out;

	hlen = off + thlen;
	if (skb_gro_header_hard(skb, hlen)) {
		th = skb_gro_header_slow(skb, hlen, off);
		if (unlikely(!th))
			goto out;
	}

	skb_gro_pull(skb, thlen);

	len = skb_gro_len(skb);
	flags = tcp_flag_word(th);

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

		th2 = tcp_hdr(p);

		if (*(u32 *)&th->source ^ *(u32 *)&th2->source) {
			NAPI_GRO_CB(p)->same_flow = 0;
			continue;
		}

		goto found;
	}

	goto out_check_final;

found:
	/* Include the IP ID check below from the inner most IP hdr */
	flush = NAPI_GRO_CB(p)->flush | NAPI_GRO_CB(p)->flush_id;
	flush |= (__force int)(flags & TCP_FLAG_CWR);
	flush |= (__force int)((flags ^ tcp_flag_word(th2)) &
		  ~(TCP_FLAG_CWR | TCP_FLAG_FIN | TCP_FLAG_PSH));
	flush |= (__force int)(th->ack_seq ^ th2->ack_seq);
	for (i = sizeof(*th); i < thlen; i += 4)
		flush |= *(u32 *)((u8 *)th + i) ^
			 *(u32 *)((u8 *)th2 + i);

	mss = tcp_skb_mss(p);

	flush |= (len - 1) >= mss;
	flush |= (ntohl(th2->seq) + skb_gro_len(p)) ^ ntohl(th->seq);

	if (flush || skb_gro_receive(head, skb)) {
		mss = 1;
		goto out_check_final;
	}

	p = *head;
	th2 = tcp_hdr(p);
	tcp_flag_word(th2) |= flags & (TCP_FLAG_FIN | TCP_FLAG_PSH);

out_check_final:
	flush = len < mss;
	flush |= (__force int)(flags & (TCP_FLAG_URG | TCP_FLAG_PSH |
					TCP_FLAG_RST | TCP_FLAG_SYN |
					TCP_FLAG_FIN));

	if (p && (!NAPI_GRO_CB(skb)->same_flow || flush))
		pp = head;

out:
	NAPI_GRO_CB(skb)->flush |= (flush != 0);

	return pp;
}

int tcp_gro_complete(struct sk_buff *skb)
{
	struct tcphdr *th = tcp_hdr(skb);

	skb->csum_start = (unsigned char *)th - skb->head;
	skb->csum_offset = offsetof(struct tcphdr, check);
	skb->ip_summed = CHECKSUM_PARTIAL;

	skb_shinfo(skb)->gso_segs = NAPI_GRO_CB(skb)->count;

	if (th->cwr)
		skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;

	return 0;
}
EXPORT_SYMBOL(tcp_gro_complete);

static int tcp_v4_gso_send_check(struct sk_buff *skb)
{
	const struct iphdr *iph;
	struct tcphdr *th;

	if (!pskb_may_pull(skb, sizeof(*th)))
		return -EINVAL;

	iph = ip_hdr(skb);
	th = tcp_hdr(skb);

	th->check = 0;
	skb->ip_summed = CHECKSUM_PARTIAL;
	__tcp_v4_send_check(skb, iph->saddr, iph->daddr);
	return 0;
}

static struct sk_buff **tcp4_gro_receive(struct sk_buff **head, struct sk_buff *skb)
{
	/* Use the IP hdr immediately proceeding for this transport */
	const struct iphdr *iph = skb_gro_network_header(skb);
	__wsum wsum;

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

	wsum = NAPI_GRO_CB(skb)->csum;

	switch (skb->ip_summed) {
	case CHECKSUM_NONE:
		wsum = skb_checksum(skb, skb_gro_offset(skb), skb_gro_len(skb),
				    0);

		/* fall through */

	case CHECKSUM_COMPLETE:
		if (!tcp_v4_check(skb_gro_len(skb), iph->saddr, iph->daddr,
				  wsum)) {
			skb->ip_summed = CHECKSUM_UNNECESSARY;
			break;
		}

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

skip_csum:
	return tcp_gro_receive(head, skb);
}

static int tcp4_gro_complete(struct sk_buff *skb, int thoff)
{
	const struct iphdr *iph = ip_hdr(skb);
	struct tcphdr *th = tcp_hdr(skb);

	th->check = ~tcp_v4_check(skb->len - thoff, iph->saddr,
				  iph->daddr, 0);
	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;

	return tcp_gro_complete(skb);
}

static const struct net_offload tcpv4_offload = {
	.callbacks = {
		.gso_send_check	=	tcp_v4_gso_send_check,
		.gso_segment	=	tcp_gso_segment,
		.gro_receive	=	tcp4_gro_receive,
		.gro_complete	=	tcp4_gro_complete,
	},
};

int __init tcpv4_offload_init(void)
{
	return inet_add_offload(&tcpv4_offload, IPPROTO_TCP);
}
Commit	Line	Data
28850dc7 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	* TCPv4 GSO/GRO support
	11	*/
	12
	13	#include <linux/skbuff.h>
	14	#include <net/tcp.h>
	15	#include <net/protocol.h>
	16
28be6e07	17	struct sk_buff tcp_gso_segment(struct sk_buff skb,
28850dc7 DB	18	netdev_features_t features)
	19	{
	20	struct sk_buff *segs = ERR_PTR(-EINVAL);
0d08c42c	21	unsigned int sum_truesize = 0;
28850dc7 DB	22	struct tcphdr *th;
	23	unsigned int thlen;
	24	unsigned int seq;
	25	__be32 delta;
	26	unsigned int oldlen;
	27	unsigned int mss;
	28	struct sk_buff *gso_skb = skb;
	29	__sum16 newcheck;
	30	bool ooo_okay, copy_destructor;
	31
	32	if (!pskb_may_pull(skb, sizeof(*th)))
	33	goto out;
	34
	35	th = tcp_hdr(skb);
	36	thlen = th->doff * 4;
	37	if (thlen < sizeof(*th))
	38	goto out;
	39
	40	if (!pskb_may_pull(skb, thlen))
	41	goto out;
	42
	43	oldlen = (u16)~skb->len;
	44	__skb_pull(skb, thlen);
	45
	46	mss = tcp_skb_mss(skb);
	47	if (unlikely(skb->len <= mss))
	48	goto out;
	49
	50	if (skb_gso_ok(skb, features \| NETIF_F_GSO_ROBUST)) {
	51	/* Packet is from an untrusted source, reset gso_segs. */
	52	int type = skb_shinfo(skb)->gso_type;
	53
	54	if (unlikely(type &
	55	~(SKB_GSO_TCPV4 \|
	56	SKB_GSO_DODGY \|
	57	SKB_GSO_TCP_ECN \|
	58	SKB_GSO_TCPV6 \|
	59	SKB_GSO_GRE \|
4749c09c	60	SKB_GSO_GRE_CSUM \|
cb32f511	61	SKB_GSO_IPIP \|
61c1db7f	62	SKB_GSO_SIT \|
28850dc7 DB	63	SKB_GSO_MPLS \|
28850dc7 DB	64	SKB_GSO_UDP_TUNNEL \|
0f4f4ffa	65	SKB_GSO_UDP_TUNNEL_CSUM \|
28850dc7 DB	66	0) \|\|
	67	!(type & (SKB_GSO_TCPV4 \| SKB_GSO_TCPV6))))
	68	goto out;
	69
	70	skb_shinfo(skb)->gso_segs = DIV_ROUND_UP(skb->len, mss);
	71
	72	segs = NULL;
	73	goto out;
	74	}
	75
	76	copy_destructor = gso_skb->destructor == tcp_wfree;
	77	ooo_okay = gso_skb->ooo_okay;
	78	/* All segments but the first should have ooo_okay cleared */
	79	skb->ooo_okay = 0;
	80
	81	segs = skb_segment(skb, features);
	82	if (IS_ERR(segs))
	83	goto out;
	84
	85	/* Only first segment might have ooo_okay set */
	86	segs->ooo_okay = ooo_okay;
	87
	88	delta = htonl(oldlen + (thlen + mss));
	89
	90	skb = segs;
	91	th = tcp_hdr(skb);
	92	seq = ntohl(th->seq);
	93
	94	newcheck = ~csum_fold((__force __wsum)((__force u32)th->check +
	95	(__force u32)delta));
	96
	97	do {
	98	th->fin = th->psh = 0;
	99	th->check = newcheck;
	100
	101	if (skb->ip_summed != CHECKSUM_PARTIAL)
e9c3a24b	102	th->check = gso_make_checksum(skb, ~th->check);
28850dc7 DB	103
	104	seq += mss;
	105	if (copy_destructor) {
	106	skb->destructor = gso_skb->destructor;
	107	skb->sk = gso_skb->sk;
0d08c42c	108	sum_truesize += skb->truesize;
28850dc7 DB	109	}
	110	skb = skb->next;
	111	th = tcp_hdr(skb);
	112
	113	th->seq = htonl(seq);
	114	th->cwr = 0;
	115	} while (skb->next);
	116
	117	/* Following permits TCP Small Queues to work well with GSO :
	118	* The callback to TCP stack will be called at the time last frag
	119	* is freed at TX completion, and not right now when gso_skb
	120	* is freed by GSO engine
	121	*/
	122	if (copy_destructor) {
	123	swap(gso_skb->sk, skb->sk);
	124	swap(gso_skb->destructor, skb->destructor);
0d08c42c ED	125	sum_truesize += skb->truesize;
	126	atomic_add(sum_truesize - gso_skb->truesize,
	127	&skb->sk->sk_wmem_alloc);
28850dc7 DB	128	}
	129
	130	delta = htonl(oldlen + (skb_tail_pointer(skb) -
	131	skb_transport_header(skb)) +
	132	skb->data_len);
	133	th->check = ~csum_fold((__force __wsum)((__force u32)th->check +
	134	(__force u32)delta));
	135	if (skb->ip_summed != CHECKSUM_PARTIAL)
e9c3a24b	136	th->check = gso_make_checksum(skb, ~th->check);
28850dc7 DB	137	out:
	138	return segs;
	139	}
28850dc7 DB	140
	141	struct sk_buff tcp_gro_receive(struct sk_buff head, struct sk_buff *skb)
	142	{
	143	struct sk_buff **pp = NULL;
	144	struct sk_buff *p;
	145	struct tcphdr *th;
	146	struct tcphdr *th2;
	147	unsigned int len;
	148	unsigned int thlen;
	149	__be32 flags;
	150	unsigned int mss = 1;
	151	unsigned int hlen;
	152	unsigned int off;
	153	int flush = 1;
	154	int i;
	155
	156	off = skb_gro_offset(skb);
	157	hlen = off + sizeof(*th);
	158	th = skb_gro_header_fast(skb, off);
	159	if (skb_gro_header_hard(skb, hlen)) {
	160	th = skb_gro_header_slow(skb, hlen, off);
	161	if (unlikely(!th))
	162	goto out;
	163	}
	164
	165	thlen = th->doff * 4;
	166	if (thlen < sizeof(*th))
	167	goto out;
	168
	169	hlen = off + thlen;
	170	if (skb_gro_header_hard(skb, hlen)) {
	171	th = skb_gro_header_slow(skb, hlen, off);
	172	if (unlikely(!th))
	173	goto out;
	174	}
	175
	176	skb_gro_pull(skb, thlen);
	177
	178	len = skb_gro_len(skb);
	179	flags = tcp_flag_word(th);
	180
	181	for (; (p = *head); head = &p->next) {
	182	if (!NAPI_GRO_CB(p)->same_flow)
	183	continue;
	184
	185	th2 = tcp_hdr(p);
	186
	187	if ((u32 )&th->source ^ (u32 )&th2->source) {
	188	NAPI_GRO_CB(p)->same_flow = 0;
	189	continue;
	190	}
	191
	192	goto found;
	193	}
	194
	195	goto out_check_final;
	196
	197	found:
bf5a755f JC	198	/* Include the IP ID check below from the inner most IP hdr */
bf5a755f JC	199	flush = NAPI_GRO_CB(p)->flush \| NAPI_GRO_CB(p)->flush_id;
28850dc7 DB	200	flush \|= (__force int)(flags & TCP_FLAG_CWR);
	201	flush \|= (__force int)((flags ^ tcp_flag_word(th2)) &
	202	~(TCP_FLAG_CWR \| TCP_FLAG_FIN \| TCP_FLAG_PSH));
	203	flush \|= (__force int)(th->ack_seq ^ th2->ack_seq);
	204	for (i = sizeof(*th); i < thlen; i += 4)
	205	flush \|= (u32 )((u8 *)th + i) ^
	206	(u32 )((u8 *)th2 + i);
	207
	208	mss = tcp_skb_mss(p);
	209
	210	flush \|= (len - 1) >= mss;
	211	flush \|= (ntohl(th2->seq) + skb_gro_len(p)) ^ ntohl(th->seq);
	212
	213	if (flush \|\| skb_gro_receive(head, skb)) {
	214	mss = 1;
	215	goto out_check_final;
	216	}
	217
	218	p = *head;
	219	th2 = tcp_hdr(p);
	220	tcp_flag_word(th2) \|= flags & (TCP_FLAG_FIN \| TCP_FLAG_PSH);
	221
	222	out_check_final:
	223	flush = len < mss;
	224	flush \|= (__force int)(flags & (TCP_FLAG_URG \| TCP_FLAG_PSH \|
	225	TCP_FLAG_RST \| TCP_FLAG_SYN \|
	226	TCP_FLAG_FIN));
	227
	228	if (p && (!NAPI_GRO_CB(skb)->same_flow \|\| flush))
	229	pp = head;
	230
	231	out:
bf5a755f	232	NAPI_GRO_CB(skb)->flush \|= (flush != 0);
28850dc7 DB	233
	234	return pp;
	235	}
28850dc7 DB	236
	237	int tcp_gro_complete(struct sk_buff *skb)
	238	{
	239	struct tcphdr *th = tcp_hdr(skb);
	240
299603e8	241	skb->csum_start = (unsigned char *)th - skb->head;
28850dc7 DB	242	skb->csum_offset = offsetof(struct tcphdr, check);
	243	skb->ip_summed = CHECKSUM_PARTIAL;
	244
	245	skb_shinfo(skb)->gso_segs = NAPI_GRO_CB(skb)->count;
	246
	247	if (th->cwr)
	248	skb_shinfo(skb)->gso_type \|= SKB_GSO_TCP_ECN;
	249
	250	return 0;
	251	}
	252	EXPORT_SYMBOL(tcp_gro_complete);
	253
	254	static int tcp_v4_gso_send_check(struct sk_buff *skb)
	255	{
	256	const struct iphdr *iph;
	257	struct tcphdr *th;
	258
	259	if (!pskb_may_pull(skb, sizeof(*th)))
	260	return -EINVAL;
	261
	262	iph = ip_hdr(skb);
	263	th = tcp_hdr(skb);
	264
	265	th->check = 0;
	266	skb->ip_summed = CHECKSUM_PARTIAL;
	267	__tcp_v4_send_check(skb, iph->saddr, iph->daddr);
	268	return 0;
	269	}
	270
	271	static struct sk_buff tcp4_gro_receive(struct sk_buff head, struct sk_buff *skb)
	272	{
299603e8	273	/* Use the IP hdr immediately proceeding for this transport */
28850dc7 DB	274	const struct iphdr *iph = skb_gro_network_header(skb);
28850dc7 DB	275	__wsum wsum;
28850dc7	276
cc5c00bb HX	277	/* Don't bother verifying checksum if we're going to flush anyway. */
	278	if (NAPI_GRO_CB(skb)->flush)
	279	goto skip_csum;
	280
bf5a755f	281	wsum = NAPI_GRO_CB(skb)->csum;
b8ee93ba	282
28850dc7	283	switch (skb->ip_summed) {
b8ee93ba HX	284	case CHECKSUM_NONE:
	285	wsum = skb_checksum(skb, skb_gro_offset(skb), skb_gro_len(skb),
	286	0);
	287
	288	/* fall through */
	289
28850dc7 DB	290	case CHECKSUM_COMPLETE:
28850dc7 DB	291	if (!tcp_v4_check(skb_gro_len(skb), iph->saddr, iph->daddr,
b8ee93ba	292	wsum)) {
28850dc7 DB	293	skb->ip_summed = CHECKSUM_UNNECESSARY;
	294	break;
	295	}
b8ee93ba	296
28850dc7 DB	297	NAPI_GRO_CB(skb)->flush = 1;
28850dc7 DB	298	return NULL;
28850dc7 DB	299	}
28850dc7 DB	300
cc5c00bb	301	skip_csum:
28850dc7 DB	302	return tcp_gro_receive(head, skb);
	303	}
	304
299603e8	305	static int tcp4_gro_complete(struct sk_buff *skb, int thoff)
28850dc7 DB	306	{
	307	const struct iphdr *iph = ip_hdr(skb);
	308	struct tcphdr *th = tcp_hdr(skb);
	309
299603e8 JC	310	th->check = ~tcp_v4_check(skb->len - thoff, iph->saddr,
299603e8 JC	311	iph->daddr, 0);
28850dc7 DB	312	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
	313
	314	return tcp_gro_complete(skb);
	315	}
	316
	317	static const struct net_offload tcpv4_offload = {
	318	.callbacks = {
	319	.gso_send_check = tcp_v4_gso_send_check,
28be6e07	320	.gso_segment = tcp_gso_segment,
28850dc7 DB	321	.gro_receive = tcp4_gro_receive,
	322	.gro_complete = tcp4_gro_complete,
	323	},
	324	};
	325
	326	int __init tcpv4_offload_init(void)
	327	{
	328	return inet_add_offload(&tcpv4_offload, IPPROTO_TCP);
	329	}