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