[mirror_ubuntu-jammy-kernel.git] / net / mptcp / options.c

// SPDX-License-Identifier: GPL-2.0
/* Multipath TCP
 *
 * Copyright (c) 2017 - 2019, Intel Corporation.
 */

#include <linux/kernel.h>
#include <net/tcp.h>
#include <net/mptcp.h>
#include "protocol.h"

static bool mptcp_cap_flag_sha256(u8 flags)
{
	return (flags & MPTCP_CAP_FLAG_MASK) == MPTCP_CAP_HMAC_SHA256;
}

void mptcp_parse_option(const struct sk_buff *skb, const unsigned char *ptr,
			int opsize, struct tcp_options_received *opt_rx)
{
	struct mptcp_options_received *mp_opt = &opt_rx->mptcp;
	u8 subtype = *ptr >> 4;
	int expected_opsize;
	u8 version;
	u8 flags;

	switch (subtype) {
	case MPTCPOPT_MP_CAPABLE:
		/* strict size checking */
		if (!(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) {
			if (skb->len > tcp_hdr(skb)->doff << 2)
				expected_opsize = TCPOLEN_MPTCP_MPC_ACK_DATA;
			else
				expected_opsize = TCPOLEN_MPTCP_MPC_ACK;
		} else {
			if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_ACK)
				expected_opsize = TCPOLEN_MPTCP_MPC_SYNACK;
			else
				expected_opsize = TCPOLEN_MPTCP_MPC_SYN;
		}
		if (opsize != expected_opsize)
			break;

		/* try to be gentle vs future versions on the initial syn */
		version = *ptr++ & MPTCP_VERSION_MASK;
		if (opsize != TCPOLEN_MPTCP_MPC_SYN) {
			if (version != MPTCP_SUPPORTED_VERSION)
				break;
		} else if (version < MPTCP_SUPPORTED_VERSION) {
			break;
		}

		flags = *ptr++;
		if (!mptcp_cap_flag_sha256(flags) ||
		    (flags & MPTCP_CAP_EXTENSIBILITY))
			break;

		/* RFC 6824, Section 3.1:
		 * "For the Checksum Required bit (labeled "A"), if either
		 * host requires the use of checksums, checksums MUST be used.
		 * In other words, the only way for checksums not to be used
		 * is if both hosts in their SYNs set A=0."
		 *
		 * Section 3.3.0:
		 * "If a checksum is not present when its use has been
		 * negotiated, the receiver MUST close the subflow with a RST as
		 * it is considered broken."
		 *
		 * We don't implement DSS checksum - fall back to TCP.
		 */
		if (flags & MPTCP_CAP_CHECKSUM_REQD)
			break;

		mp_opt->mp_capable = 1;
		if (opsize >= TCPOLEN_MPTCP_MPC_SYNACK) {
			mp_opt->sndr_key = get_unaligned_be64(ptr);
			ptr += 8;
		}
		if (opsize >= TCPOLEN_MPTCP_MPC_ACK) {
			mp_opt->rcvr_key = get_unaligned_be64(ptr);
			ptr += 8;
		}
		if (opsize == TCPOLEN_MPTCP_MPC_ACK_DATA) {
			/* Section 3.1.:
			 * "the data parameters in a MP_CAPABLE are semantically
			 * equivalent to those in a DSS option and can be used
			 * interchangeably."
			 */
			mp_opt->dss = 1;
			mp_opt->use_map = 1;
			mp_opt->mpc_map = 1;
			mp_opt->data_len = get_unaligned_be16(ptr);
			ptr += 2;
		}
		pr_debug("MP_CAPABLE version=%x, flags=%x, optlen=%d sndr=%llu, rcvr=%llu len=%d",
			 version, flags, opsize, mp_opt->sndr_key,
			 mp_opt->rcvr_key, mp_opt->data_len);
		break;

	case MPTCPOPT_DSS:
		pr_debug("DSS");
		ptr++;

		/* we must clear 'mpc_map' be able to detect MP_CAPABLE
		 * map vs DSS map in mptcp_incoming_options(), and reconstruct
		 * map info accordingly
		 */
		mp_opt->mpc_map = 0;
		flags = (*ptr++) & MPTCP_DSS_FLAG_MASK;
		mp_opt->data_fin = (flags & MPTCP_DSS_DATA_FIN) != 0;
		mp_opt->dsn64 = (flags & MPTCP_DSS_DSN64) != 0;
		mp_opt->use_map = (flags & MPTCP_DSS_HAS_MAP) != 0;
		mp_opt->ack64 = (flags & MPTCP_DSS_ACK64) != 0;
		mp_opt->use_ack = (flags & MPTCP_DSS_HAS_ACK);

		pr_debug("data_fin=%d dsn64=%d use_map=%d ack64=%d use_ack=%d",
			 mp_opt->data_fin, mp_opt->dsn64,
			 mp_opt->use_map, mp_opt->ack64,
			 mp_opt->use_ack);

		expected_opsize = TCPOLEN_MPTCP_DSS_BASE;

		if (mp_opt->use_ack) {
			if (mp_opt->ack64)
				expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
			else
				expected_opsize += TCPOLEN_MPTCP_DSS_ACK32;
		}

		if (mp_opt->use_map) {
			if (mp_opt->dsn64)
				expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
			else
				expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
		}

		/* RFC 6824, Section 3.3:
		 * If a checksum is present, but its use had
		 * not been negotiated in the MP_CAPABLE handshake,
		 * the checksum field MUST be ignored.
		 */
		if (opsize != expected_opsize &&
		    opsize != expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM)
			break;

		mp_opt->dss = 1;

		if (mp_opt->use_ack) {
			if (mp_opt->ack64) {
				mp_opt->data_ack = get_unaligned_be64(ptr);
				ptr += 8;
			} else {
				mp_opt->data_ack = get_unaligned_be32(ptr);
				ptr += 4;
			}

			pr_debug("data_ack=%llu", mp_opt->data_ack);
		}

		if (mp_opt->use_map) {
			if (mp_opt->dsn64) {
				mp_opt->data_seq = get_unaligned_be64(ptr);
				ptr += 8;
			} else {
				mp_opt->data_seq = get_unaligned_be32(ptr);
				ptr += 4;
			}

			mp_opt->subflow_seq = get_unaligned_be32(ptr);
			ptr += 4;

			mp_opt->data_len = get_unaligned_be16(ptr);
			ptr += 2;

			pr_debug("data_seq=%llu subflow_seq=%u data_len=%u",
				 mp_opt->data_seq, mp_opt->subflow_seq,
				 mp_opt->data_len);
		}

		break;

	default:
		break;
	}
}

void mptcp_get_options(const struct sk_buff *skb,
		       struct tcp_options_received *opt_rx)
{
	const unsigned char *ptr;
	const struct tcphdr *th = tcp_hdr(skb);
	int length = (th->doff * 4) - sizeof(struct tcphdr);

	ptr = (const unsigned char *)(th + 1);

	while (length > 0) {
		int opcode = *ptr++;
		int opsize;

		switch (opcode) {
		case TCPOPT_EOL:
			return;
		case TCPOPT_NOP:	/* Ref: RFC 793 section 3.1 */
			length--;
			continue;
		default:
			opsize = *ptr++;
			if (opsize < 2) /* "silly options" */
				return;
			if (opsize > length)
				return;	/* don't parse partial options */
			if (opcode == TCPOPT_MPTCP)
				mptcp_parse_option(skb, ptr, opsize, opt_rx);
			ptr += opsize - 2;
			length -= opsize;
		}
	}
}

bool mptcp_syn_options(struct sock *sk, const struct sk_buff *skb,
		       unsigned int *size, struct mptcp_out_options *opts)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);

	/* we will use snd_isn to detect first pkt [re]transmission
	 * in mptcp_established_options_mp()
	 */
	subflow->snd_isn = TCP_SKB_CB(skb)->end_seq;
	if (subflow->request_mptcp) {
		pr_debug("local_key=%llu", subflow->local_key);
		opts->suboptions = OPTION_MPTCP_MPC_SYN;
		opts->sndr_key = subflow->local_key;
		*size = TCPOLEN_MPTCP_MPC_SYN;
		return true;
	}
	return false;
}

void mptcp_rcv_synsent(struct sock *sk)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	struct tcp_sock *tp = tcp_sk(sk);

	pr_debug("subflow=%p", subflow);
	if (subflow->request_mptcp && tp->rx_opt.mptcp.mp_capable) {
		subflow->mp_capable = 1;
		subflow->can_ack = 1;
		subflow->remote_key = tp->rx_opt.mptcp.sndr_key;
	} else {
		tcp_sk(sk)->is_mptcp = 0;
	}
}

static bool mptcp_established_options_mp(struct sock *sk, struct sk_buff *skb,
					 unsigned int *size,
					 unsigned int remaining,
					 struct mptcp_out_options *opts)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	struct mptcp_ext *mpext;
	unsigned int data_len;

	pr_debug("subflow=%p fourth_ack=%d seq=%x:%x remaining=%d", subflow,
		 subflow->fourth_ack, subflow->snd_isn,
		 skb ? TCP_SKB_CB(skb)->seq : 0, remaining);

	if (subflow->mp_capable && !subflow->fourth_ack && skb &&
	    subflow->snd_isn == TCP_SKB_CB(skb)->seq) {
		/* When skb is not available, we better over-estimate the
		 * emitted options len. A full DSS option is longer than
		 * TCPOLEN_MPTCP_MPC_ACK_DATA, so let's the caller try to fit
		 * that.
		 */
		mpext = mptcp_get_ext(skb);
		data_len = mpext ? mpext->data_len : 0;

		/* we will check ext_copy.data_len in mptcp_write_options() to
		 * discriminate between TCPOLEN_MPTCP_MPC_ACK_DATA and
		 * TCPOLEN_MPTCP_MPC_ACK
		 */
		opts->ext_copy.data_len = data_len;
		opts->suboptions = OPTION_MPTCP_MPC_ACK;
		opts->sndr_key = subflow->local_key;
		opts->rcvr_key = subflow->remote_key;

		/* Section 3.1.
		 * The MP_CAPABLE option is carried on the SYN, SYN/ACK, and ACK
		 * packets that start the first subflow of an MPTCP connection,
		 * as well as the first packet that carries data
		 */
		if (data_len > 0)
			*size = ALIGN(TCPOLEN_MPTCP_MPC_ACK_DATA, 4);
		else
			*size = TCPOLEN_MPTCP_MPC_ACK;

		pr_debug("subflow=%p, local_key=%llu, remote_key=%llu map_len=%d",
			 subflow, subflow->local_key, subflow->remote_key,
			 data_len);

		return true;
	}
	return false;
}

static void mptcp_write_data_fin(struct mptcp_subflow_context *subflow,
				 struct mptcp_ext *ext)
{
	if (!ext->use_map) {
		/* RFC6824 requires a DSS mapping with specific values
		 * if DATA_FIN is set but no data payload is mapped
		 */
		ext->data_fin = 1;
		ext->use_map = 1;
		ext->dsn64 = 1;
		ext->data_seq = subflow->data_fin_tx_seq;
		ext->subflow_seq = 0;
		ext->data_len = 1;
	} else if (ext->data_seq + ext->data_len == subflow->data_fin_tx_seq) {
		/* If there's an existing DSS mapping and it is the
		 * final mapping, DATA_FIN consumes 1 additional byte of
		 * mapping space.
		 */
		ext->data_fin = 1;
		ext->data_len++;
	}
}

static bool mptcp_established_options_dss(struct sock *sk, struct sk_buff *skb,
					  unsigned int *size,
					  unsigned int remaining,
					  struct mptcp_out_options *opts)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	unsigned int dss_size = 0;
	struct mptcp_ext *mpext;
	struct mptcp_sock *msk;
	unsigned int ack_size;
	bool ret = false;
	bool can_ack;
	u64 ack_seq;
	u8 tcp_fin;

	if (skb) {
		mpext = mptcp_get_ext(skb);
		tcp_fin = TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN;
	} else {
		mpext = NULL;
		tcp_fin = 0;
	}

	if (!skb || (mpext && mpext->use_map) || tcp_fin) {
		unsigned int map_size;

		map_size = TCPOLEN_MPTCP_DSS_BASE + TCPOLEN_MPTCP_DSS_MAP64;

		remaining -= map_size;
		dss_size = map_size;
		if (mpext)
			opts->ext_copy = *mpext;

		if (skb && tcp_fin && subflow->data_fin_tx_enable)
			mptcp_write_data_fin(subflow, &opts->ext_copy);
		ret = true;
	}

	/* passive sockets msk will set the 'can_ack' after accept(), even
	 * if the first subflow may have the already the remote key handy
	 */
	can_ack = true;
	opts->ext_copy.use_ack = 0;
	msk = mptcp_sk(subflow->conn);
	if (likely(msk && READ_ONCE(msk->can_ack))) {
		ack_seq = msk->ack_seq;
	} else if (subflow->can_ack) {
		mptcp_crypto_key_sha(subflow->remote_key, NULL, &ack_seq);
		ack_seq++;
	} else {
		can_ack = false;
	}

	if (unlikely(!can_ack)) {
		*size = ALIGN(dss_size, 4);
		return ret;
	}

	ack_size = TCPOLEN_MPTCP_DSS_ACK64;

	/* Add kind/length/subtype/flag overhead if mapping is not populated */
	if (dss_size == 0)
		ack_size += TCPOLEN_MPTCP_DSS_BASE;

	dss_size += ack_size;

	opts->ext_copy.data_ack = ack_seq;
	opts->ext_copy.ack64 = 1;
	opts->ext_copy.use_ack = 1;

	*size = ALIGN(dss_size, 4);
	return true;
}

bool mptcp_established_options(struct sock *sk, struct sk_buff *skb,
			       unsigned int *size, unsigned int remaining,
			       struct mptcp_out_options *opts)
{
	unsigned int opt_size = 0;
	bool ret = false;

	if (mptcp_established_options_mp(sk, skb, &opt_size, remaining, opts))
		ret = true;
	else if (mptcp_established_options_dss(sk, skb, &opt_size, remaining,
					       opts))
		ret = true;

	/* we reserved enough space for the above options, and exceeding the
	 * TCP option space would be fatal
	 */
	if (WARN_ON_ONCE(opt_size > remaining))
		return false;

	*size += opt_size;
	remaining -= opt_size;

	return ret;
}

bool mptcp_synack_options(const struct request_sock *req, unsigned int *size,
			  struct mptcp_out_options *opts)
{
	struct mptcp_subflow_request_sock *subflow_req = mptcp_subflow_rsk(req);

	if (subflow_req->mp_capable) {
		opts->suboptions = OPTION_MPTCP_MPC_SYNACK;
		opts->sndr_key = subflow_req->local_key;
		*size = TCPOLEN_MPTCP_MPC_SYNACK;
		pr_debug("subflow_req=%p, local_key=%llu",
			 subflow_req, subflow_req->local_key);
		return true;
	}
	return false;
}

static bool check_fourth_ack(struct mptcp_subflow_context *subflow,
			     struct sk_buff *skb,
			     struct mptcp_options_received *mp_opt)
{
	/* here we can process OoO, in-window pkts, only in-sequence 4th ack
	 * are relevant
	 */
	if (likely(subflow->fourth_ack ||
		   TCP_SKB_CB(skb)->seq != subflow->ssn_offset + 1))
		return true;

	if (mp_opt->use_ack)
		subflow->fourth_ack = 1;

	if (subflow->can_ack)
		return true;

	/* If the first established packet does not contain MP_CAPABLE + data
	 * then fallback to TCP
	 */
	if (!mp_opt->mp_capable) {
		subflow->mp_capable = 0;
		tcp_sk(mptcp_subflow_tcp_sock(subflow))->is_mptcp = 0;
		return false;
	}
	subflow->remote_key = mp_opt->sndr_key;
	subflow->can_ack = 1;
	return true;
}

void mptcp_incoming_options(struct sock *sk, struct sk_buff *skb,
			    struct tcp_options_received *opt_rx)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	struct mptcp_options_received *mp_opt;
	struct mptcp_ext *mpext;

	mp_opt = &opt_rx->mptcp;
	if (!check_fourth_ack(subflow, skb, mp_opt))
		return;

	if (!mp_opt->dss)
		return;

	mpext = skb_ext_add(skb, SKB_EXT_MPTCP);
	if (!mpext)
		return;

	memset(mpext, 0, sizeof(*mpext));

	if (mp_opt->use_map) {
		if (mp_opt->mpc_map) {
			/* this is an MP_CAPABLE carrying MPTCP data
			 * we know this map the first chunk of data
			 */
			mptcp_crypto_key_sha(subflow->remote_key, NULL,
					     &mpext->data_seq);
			mpext->data_seq++;
			mpext->subflow_seq = 1;
			mpext->dsn64 = 1;
			mpext->mpc_map = 1;
		} else {
			mpext->data_seq = mp_opt->data_seq;
			mpext->subflow_seq = mp_opt->subflow_seq;
			mpext->dsn64 = mp_opt->dsn64;
		}
		mpext->data_len = mp_opt->data_len;
		mpext->use_map = 1;
	}

	if (mp_opt->use_ack) {
		mpext->data_ack = mp_opt->data_ack;
		mpext->use_ack = 1;
		mpext->ack64 = mp_opt->ack64;
	}

	mpext->data_fin = mp_opt->data_fin;
}

void mptcp_write_options(__be32 *ptr, struct mptcp_out_options *opts)
{
	if ((OPTION_MPTCP_MPC_SYN | OPTION_MPTCP_MPC_SYNACK |
	     OPTION_MPTCP_MPC_ACK) & opts->suboptions) {
		u8 len;

		if (OPTION_MPTCP_MPC_SYN & opts->suboptions)
			len = TCPOLEN_MPTCP_MPC_SYN;
		else if (OPTION_MPTCP_MPC_SYNACK & opts->suboptions)
			len = TCPOLEN_MPTCP_MPC_SYNACK;
		else if (opts->ext_copy.data_len)
			len = TCPOLEN_MPTCP_MPC_ACK_DATA;
		else
			len = TCPOLEN_MPTCP_MPC_ACK;

		*ptr++ = htonl((TCPOPT_MPTCP << 24) | (len << 16) |
			       (MPTCPOPT_MP_CAPABLE << 12) |
			       (MPTCP_SUPPORTED_VERSION << 8) |
			       MPTCP_CAP_HMAC_SHA256);

		if (!((OPTION_MPTCP_MPC_SYNACK | OPTION_MPTCP_MPC_ACK) &
		    opts->suboptions))
			goto mp_capable_done;

		put_unaligned_be64(opts->sndr_key, ptr);
		ptr += 2;
		if (!((OPTION_MPTCP_MPC_ACK) & opts->suboptions))
			goto mp_capable_done;

		put_unaligned_be64(opts->rcvr_key, ptr);
		ptr += 2;
		if (!opts->ext_copy.data_len)
			goto mp_capable_done;

		put_unaligned_be32(opts->ext_copy.data_len << 16 |
				   TCPOPT_NOP << 8 | TCPOPT_NOP, ptr);
		ptr += 1;
	}

mp_capable_done:
	if (opts->ext_copy.use_ack || opts->ext_copy.use_map) {
		struct mptcp_ext *mpext = &opts->ext_copy;
		u8 len = TCPOLEN_MPTCP_DSS_BASE;
		u8 flags = 0;

		if (mpext->use_ack) {
			len += TCPOLEN_MPTCP_DSS_ACK64;
			flags = MPTCP_DSS_HAS_ACK | MPTCP_DSS_ACK64;
		}

		if (mpext->use_map) {
			len += TCPOLEN_MPTCP_DSS_MAP64;

			/* Use only 64-bit mapping flags for now, add
			 * support for optional 32-bit mappings later.
			 */
			flags |= MPTCP_DSS_HAS_MAP | MPTCP_DSS_DSN64;
			if (mpext->data_fin)
				flags |= MPTCP_DSS_DATA_FIN;
		}

		*ptr++ = htonl((TCPOPT_MPTCP << 24) |
			       (len  << 16) |
			       (MPTCPOPT_DSS << 12) |
			       (flags));

		if (mpext->use_ack) {
			put_unaligned_be64(mpext->data_ack, ptr);
			ptr += 2;
		}

		if (mpext->use_map) {
			put_unaligned_be64(mpext->data_seq, ptr);
			ptr += 2;
			put_unaligned_be32(mpext->subflow_seq, ptr);
			ptr += 1;
			put_unaligned_be32(mpext->data_len << 16 |
					   TCPOPT_NOP << 8 | TCPOPT_NOP, ptr);
		}
	}
}
Commit	Line	Data
eda7acdd PK	1	// SPDX-License-Identifier: GPL-2.0
	2	/* Multipath TCP
	3	*
	4	* Copyright (c) 2017 - 2019, Intel Corporation.
	5	*/
	6
	7	#include <linux/kernel.h>
	8	#include <net/tcp.h>
	9	#include <net/mptcp.h>
	10	#include "protocol.h"
	11
65492c5a PA	12	static bool mptcp_cap_flag_sha256(u8 flags)
	13	{
	14	return (flags & MPTCP_CAP_FLAG_MASK) == MPTCP_CAP_HMAC_SHA256;
	15	}
	16
cc7972ea CP	17	void mptcp_parse_option(const struct sk_buff skb, const unsigned char ptr,
cc7972ea CP	18	int opsize, struct tcp_options_received *opt_rx)
eda7acdd PK	19	{
	20	struct mptcp_options_received *mp_opt = &opt_rx->mptcp;
	21	u8 subtype = *ptr >> 4;
648ef4b8	22	int expected_opsize;
eda7acdd PK	23	u8 version;
	24	u8 flags;
	25
	26	switch (subtype) {
	27	case MPTCPOPT_MP_CAPABLE:
cc7972ea CP	28	/* strict size checking */
	29	if (!(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) {
	30	if (skb->len > tcp_hdr(skb)->doff << 2)
	31	expected_opsize = TCPOLEN_MPTCP_MPC_ACK_DATA;
	32	else
	33	expected_opsize = TCPOLEN_MPTCP_MPC_ACK;
	34	} else {
	35	if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_ACK)
	36	expected_opsize = TCPOLEN_MPTCP_MPC_SYNACK;
	37	else
	38	expected_opsize = TCPOLEN_MPTCP_MPC_SYN;
	39	}
	40	if (opsize != expected_opsize)
eda7acdd PK	41	break;
eda7acdd PK	42
cc7972ea	43	/* try to be gentle vs future versions on the initial syn */
eda7acdd	44	version = *ptr++ & MPTCP_VERSION_MASK;
cc7972ea CP	45	if (opsize != TCPOLEN_MPTCP_MPC_SYN) {
	46	if (version != MPTCP_SUPPORTED_VERSION)
	47	break;
	48	} else if (version < MPTCP_SUPPORTED_VERSION) {
eda7acdd	49	break;
cc7972ea	50	}
eda7acdd PK	51
eda7acdd PK	52	flags = *ptr++;
65492c5a	53	if (!mptcp_cap_flag_sha256(flags) \|\|
eda7acdd PK	54	(flags & MPTCP_CAP_EXTENSIBILITY))
	55	break;
	56
	57	/* RFC 6824, Section 3.1:
	58	* "For the Checksum Required bit (labeled "A"), if either
	59	* host requires the use of checksums, checksums MUST be used.
	60	* In other words, the only way for checksums not to be used
	61	* is if both hosts in their SYNs set A=0."
	62	*
	63	* Section 3.3.0:
	64	* "If a checksum is not present when its use has been
	65	* negotiated, the receiver MUST close the subflow with a RST as
	66	* it is considered broken."
	67	*
	68	* We don't implement DSS checksum - fall back to TCP.
	69	*/
	70	if (flags & MPTCP_CAP_CHECKSUM_REQD)
	71	break;
	72
	73	mp_opt->mp_capable = 1;
cc7972ea CP	74	if (opsize >= TCPOLEN_MPTCP_MPC_SYNACK) {
	75	mp_opt->sndr_key = get_unaligned_be64(ptr);
	76	ptr += 8;
	77	}
	78	if (opsize >= TCPOLEN_MPTCP_MPC_ACK) {
eda7acdd PK	79	mp_opt->rcvr_key = get_unaligned_be64(ptr);
eda7acdd PK	80	ptr += 8;
eda7acdd	81	}
cc7972ea CP	82	if (opsize == TCPOLEN_MPTCP_MPC_ACK_DATA) {
	83	/* Section 3.1.:
	84	* "the data parameters in a MP_CAPABLE are semantically
	85	* equivalent to those in a DSS option and can be used
	86	* interchangeably."
	87	*/
	88	mp_opt->dss = 1;
	89	mp_opt->use_map = 1;
	90	mp_opt->mpc_map = 1;
	91	mp_opt->data_len = get_unaligned_be16(ptr);
	92	ptr += 2;
	93	}
	94	pr_debug("MP_CAPABLE version=%x, flags=%x, optlen=%d sndr=%llu, rcvr=%llu len=%d",
	95	version, flags, opsize, mp_opt->sndr_key,
	96	mp_opt->rcvr_key, mp_opt->data_len);
eda7acdd PK	97	break;
	98
	99	case MPTCPOPT_DSS:
	100	pr_debug("DSS");
648ef4b8 MM	101	ptr++;
648ef4b8 MM	102
cc7972ea CP	103	/* we must clear 'mpc_map' be able to detect MP_CAPABLE
	104	* map vs DSS map in mptcp_incoming_options(), and reconstruct
	105	* map info accordingly
	106	*/
	107	mp_opt->mpc_map = 0;
648ef4b8 MM	108	flags = (*ptr++) & MPTCP_DSS_FLAG_MASK;
	109	mp_opt->data_fin = (flags & MPTCP_DSS_DATA_FIN) != 0;
	110	mp_opt->dsn64 = (flags & MPTCP_DSS_DSN64) != 0;
	111	mp_opt->use_map = (flags & MPTCP_DSS_HAS_MAP) != 0;
	112	mp_opt->ack64 = (flags & MPTCP_DSS_ACK64) != 0;
	113	mp_opt->use_ack = (flags & MPTCP_DSS_HAS_ACK);
	114
	115	pr_debug("data_fin=%d dsn64=%d use_map=%d ack64=%d use_ack=%d",
	116	mp_opt->data_fin, mp_opt->dsn64,
	117	mp_opt->use_map, mp_opt->ack64,
	118	mp_opt->use_ack);
	119
	120	expected_opsize = TCPOLEN_MPTCP_DSS_BASE;
	121
	122	if (mp_opt->use_ack) {
	123	if (mp_opt->ack64)
	124	expected_opsize += TCPOLEN_MPTCP_DSS_ACK64;
	125	else
	126	expected_opsize += TCPOLEN_MPTCP_DSS_ACK32;
	127	}
	128
	129	if (mp_opt->use_map) {
	130	if (mp_opt->dsn64)
	131	expected_opsize += TCPOLEN_MPTCP_DSS_MAP64;
	132	else
	133	expected_opsize += TCPOLEN_MPTCP_DSS_MAP32;
	134	}
	135
	136	/* RFC 6824, Section 3.3:
	137	* If a checksum is present, but its use had
	138	* not been negotiated in the MP_CAPABLE handshake,
	139	* the checksum field MUST be ignored.
	140	*/
	141	if (opsize != expected_opsize &&
	142	opsize != expected_opsize + TCPOLEN_MPTCP_DSS_CHECKSUM)
	143	break;
	144
eda7acdd	145	mp_opt->dss = 1;
648ef4b8 MM	146
	147	if (mp_opt->use_ack) {
	148	if (mp_opt->ack64) {
	149	mp_opt->data_ack = get_unaligned_be64(ptr);
	150	ptr += 8;
	151	} else {
	152	mp_opt->data_ack = get_unaligned_be32(ptr);
	153	ptr += 4;
	154	}
	155
	156	pr_debug("data_ack=%llu", mp_opt->data_ack);
	157	}
	158
	159	if (mp_opt->use_map) {
	160	if (mp_opt->dsn64) {
	161	mp_opt->data_seq = get_unaligned_be64(ptr);
	162	ptr += 8;
	163	} else {
	164	mp_opt->data_seq = get_unaligned_be32(ptr);
	165	ptr += 4;
	166	}
	167
	168	mp_opt->subflow_seq = get_unaligned_be32(ptr);
	169	ptr += 4;
	170
	171	mp_opt->data_len = get_unaligned_be16(ptr);
	172	ptr += 2;
	173
	174	pr_debug("data_seq=%llu subflow_seq=%u data_len=%u",
	175	mp_opt->data_seq, mp_opt->subflow_seq,
	176	mp_opt->data_len);
	177	}
	178
eda7acdd PK	179	break;
	180
	181	default:
	182	break;
	183	}
	184	}
	185
cec37a6e PK	186	void mptcp_get_options(const struct sk_buff *skb,
	187	struct tcp_options_received *opt_rx)
	188	{
	189	const unsigned char *ptr;
	190	const struct tcphdr *th = tcp_hdr(skb);
	191	int length = (th->doff * 4) - sizeof(struct tcphdr);
	192
	193	ptr = (const unsigned char *)(th + 1);
	194
	195	while (length > 0) {
	196	int opcode = *ptr++;
	197	int opsize;
	198
	199	switch (opcode) {
	200	case TCPOPT_EOL:
	201	return;
	202	case TCPOPT_NOP: /* Ref: RFC 793 section 3.1 */
	203	length--;
	204	continue;
	205	default:
	206	opsize = *ptr++;
	207	if (opsize < 2) /* "silly options" */
	208	return;
	209	if (opsize > length)
	210	return; /* don't parse partial options */
	211	if (opcode == TCPOPT_MPTCP)
cc7972ea	212	mptcp_parse_option(skb, ptr, opsize, opt_rx);
cec37a6e PK	213	ptr += opsize - 2;
	214	length -= opsize;
	215	}
	216	}
	217	}
	218
cc7972ea CP	219	bool mptcp_syn_options(struct sock sk, const struct sk_buff skb,
cc7972ea CP	220	unsigned int size, struct mptcp_out_options opts)
cec37a6e PK	221	{
	222	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	223
cc7972ea CP	224	/* we will use snd_isn to detect first pkt [re]transmission
	225	* in mptcp_established_options_mp()
	226	*/
	227	subflow->snd_isn = TCP_SKB_CB(skb)->end_seq;
cec37a6e PK	228	if (subflow->request_mptcp) {
	229	pr_debug("local_key=%llu", subflow->local_key);
	230	opts->suboptions = OPTION_MPTCP_MPC_SYN;
	231	opts->sndr_key = subflow->local_key;
	232	*size = TCPOLEN_MPTCP_MPC_SYN;
	233	return true;
	234	}
	235	return false;
	236	}
	237
	238	void mptcp_rcv_synsent(struct sock *sk)
	239	{
	240	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	241	struct tcp_sock *tp = tcp_sk(sk);
	242
	243	pr_debug("subflow=%p", subflow);
	244	if (subflow->request_mptcp && tp->rx_opt.mptcp.mp_capable) {
	245	subflow->mp_capable = 1;
d22f4988	246	subflow->can_ack = 1;
cec37a6e PK	247	subflow->remote_key = tp->rx_opt.mptcp.sndr_key;
	248	} else {
	249	tcp_sk(sk)->is_mptcp = 0;
	250	}
	251	}
	252
cc7972ea CP	253	static bool mptcp_established_options_mp(struct sock sk, struct sk_buff skb,
cc7972ea CP	254	unsigned int *size,
6d0060f6 MM	255	unsigned int remaining,
6d0060f6 MM	256	struct mptcp_out_options *opts)
cec37a6e PK	257	{
cec37a6e PK	258	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
cc7972ea CP	259	struct mptcp_ext *mpext;
	260	unsigned int data_len;
	261
	262	pr_debug("subflow=%p fourth_ack=%d seq=%x:%x remaining=%d", subflow,
	263	subflow->fourth_ack, subflow->snd_isn,
	264	skb ? TCP_SKB_CB(skb)->seq : 0, remaining);
	265
	266	if (subflow->mp_capable && !subflow->fourth_ack && skb &&
	267	subflow->snd_isn == TCP_SKB_CB(skb)->seq) {
	268	/* When skb is not available, we better over-estimate the
	269	* emitted options len. A full DSS option is longer than
	270	* TCPOLEN_MPTCP_MPC_ACK_DATA, so let's the caller try to fit
	271	* that.
	272	*/
	273	mpext = mptcp_get_ext(skb);
	274	data_len = mpext ? mpext->data_len : 0;
cec37a6e	275
cc7972ea CP	276	/* we will check ext_copy.data_len in mptcp_write_options() to
	277	* discriminate between TCPOLEN_MPTCP_MPC_ACK_DATA and
	278	* TCPOLEN_MPTCP_MPC_ACK
	279	*/
	280	opts->ext_copy.data_len = data_len;
cec37a6e PK	281	opts->suboptions = OPTION_MPTCP_MPC_ACK;
	282	opts->sndr_key = subflow->local_key;
	283	opts->rcvr_key = subflow->remote_key;
cc7972ea CP	284
	285	/* Section 3.1.
	286	* The MP_CAPABLE option is carried on the SYN, SYN/ACK, and ACK
	287	* packets that start the first subflow of an MPTCP connection,
	288	* as well as the first packet that carries data
	289	*/
	290	if (data_len > 0)
	291	*size = ALIGN(TCPOLEN_MPTCP_MPC_ACK_DATA, 4);
	292	else
	293	*size = TCPOLEN_MPTCP_MPC_ACK;
	294
	295	pr_debug("subflow=%p, local_key=%llu, remote_key=%llu map_len=%d",
	296	subflow, subflow->local_key, subflow->remote_key,
	297	data_len);
	298
cec37a6e PK	299	return true;
	300	}
	301	return false;
	302	}
	303
6d0060f6 MM	304	static void mptcp_write_data_fin(struct mptcp_subflow_context *subflow,
	305	struct mptcp_ext *ext)
	306	{
6d0060f6 MM	307	if (!ext->use_map) {
	308	/* RFC6824 requires a DSS mapping with specific values
	309	* if DATA_FIN is set but no data payload is mapped
	310	*/
6d37a0b8	311	ext->data_fin = 1;
6d0060f6 MM	312	ext->use_map = 1;
6d0060f6 MM	313	ext->dsn64 = 1;
76c42a29	314	ext->data_seq = subflow->data_fin_tx_seq;
6d0060f6 MM	315	ext->subflow_seq = 0;
6d0060f6 MM	316	ext->data_len = 1;
6d37a0b8 MM	317	} else if (ext->data_seq + ext->data_len == subflow->data_fin_tx_seq) {
	318	/* If there's an existing DSS mapping and it is the
	319	* final mapping, DATA_FIN consumes 1 additional byte of
	320	* mapping space.
6d0060f6	321	*/
6d37a0b8	322	ext->data_fin = 1;
6d0060f6 MM	323	ext->data_len++;
	324	}
	325	}
	326
	327	static bool mptcp_established_options_dss(struct sock sk, struct sk_buff skb,
	328	unsigned int *size,
	329	unsigned int remaining,
	330	struct mptcp_out_options *opts)
	331	{
	332	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	333	unsigned int dss_size = 0;
	334	struct mptcp_ext *mpext;
	335	struct mptcp_sock *msk;
	336	unsigned int ack_size;
d22f4988	337	bool ret = false;
2398e399 PA	338	bool can_ack;
2398e399 PA	339	u64 ack_seq;
6d0060f6 MM	340	u8 tcp_fin;
	341
	342	if (skb) {
	343	mpext = mptcp_get_ext(skb);
	344	tcp_fin = TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN;
	345	} else {
	346	mpext = NULL;
	347	tcp_fin = 0;
	348	}
	349
	350	if (!skb \|\| (mpext && mpext->use_map) \|\| tcp_fin) {
	351	unsigned int map_size;
	352
	353	map_size = TCPOLEN_MPTCP_DSS_BASE + TCPOLEN_MPTCP_DSS_MAP64;
	354
	355	remaining -= map_size;
	356	dss_size = map_size;
	357	if (mpext)
	358	opts->ext_copy = *mpext;
	359
76c42a29	360	if (skb && tcp_fin && subflow->data_fin_tx_enable)
6d0060f6	361	mptcp_write_data_fin(subflow, &opts->ext_copy);
d22f4988 CP	362	ret = true;
	363	}
	364
2398e399 PA	365	/* passive sockets msk will set the 'can_ack' after accept(), even
	366	* if the first subflow may have the already the remote key handy
	367	*/
	368	can_ack = true;
d22f4988 CP	369	opts->ext_copy.use_ack = 0;
d22f4988 CP	370	msk = mptcp_sk(subflow->conn);
2398e399 PA	371	if (likely(msk && READ_ONCE(msk->can_ack))) {
	372	ack_seq = msk->ack_seq;
	373	} else if (subflow->can_ack) {
	374	mptcp_crypto_key_sha(subflow->remote_key, NULL, &ack_seq);
	375	ack_seq++;
	376	} else {
	377	can_ack = false;
	378	}
	379
	380	if (unlikely(!can_ack)) {
d22f4988 CP	381	*size = ALIGN(dss_size, 4);
d22f4988 CP	382	return ret;
6d0060f6 MM	383	}
	384
	385	ack_size = TCPOLEN_MPTCP_DSS_ACK64;
	386
	387	/* Add kind/length/subtype/flag overhead if mapping is not populated */
	388	if (dss_size == 0)
	389	ack_size += TCPOLEN_MPTCP_DSS_BASE;
	390
	391	dss_size += ack_size;
	392
2398e399	393	opts->ext_copy.data_ack = ack_seq;
6d0060f6 MM	394	opts->ext_copy.ack64 = 1;
	395	opts->ext_copy.use_ack = 1;
	396
	397	*size = ALIGN(dss_size, 4);
	398	return true;
	399	}
	400
	401	bool mptcp_established_options(struct sock sk, struct sk_buff skb,
	402	unsigned int *size, unsigned int remaining,
	403	struct mptcp_out_options *opts)
	404	{
	405	unsigned int opt_size = 0;
	406	bool ret = false;
	407
cc7972ea	408	if (mptcp_established_options_mp(sk, skb, &opt_size, remaining, opts))
6d0060f6 MM	409	ret = true;
	410	else if (mptcp_established_options_dss(sk, skb, &opt_size, remaining,
	411	opts))
	412	ret = true;
	413
	414	/* we reserved enough space for the above options, and exceeding the
	415	* TCP option space would be fatal
	416	*/
	417	if (WARN_ON_ONCE(opt_size > remaining))
	418	return false;
	419
	420	*size += opt_size;
	421	remaining -= opt_size;
	422
	423	return ret;
	424	}
	425
cec37a6e PK	426	bool mptcp_synack_options(const struct request_sock req, unsigned int size,
	427	struct mptcp_out_options *opts)
	428	{
	429	struct mptcp_subflow_request_sock *subflow_req = mptcp_subflow_rsk(req);
	430
	431	if (subflow_req->mp_capable) {
	432	opts->suboptions = OPTION_MPTCP_MPC_SYNACK;
	433	opts->sndr_key = subflow_req->local_key;
	434	*size = TCPOLEN_MPTCP_MPC_SYNACK;
	435	pr_debug("subflow_req=%p, local_key=%llu",
	436	subflow_req, subflow_req->local_key);
	437	return true;
	438	}
	439	return false;
	440	}
	441
d22f4988 CP	442	static bool check_fourth_ack(struct mptcp_subflow_context *subflow,
	443	struct sk_buff *skb,
	444	struct mptcp_options_received *mp_opt)
	445	{
	446	/* here we can process OoO, in-window pkts, only in-sequence 4th ack
	447	* are relevant
	448	*/
	449	if (likely(subflow->fourth_ack \|\|
	450	TCP_SKB_CB(skb)->seq != subflow->ssn_offset + 1))
	451	return true;
	452
	453	if (mp_opt->use_ack)
	454	subflow->fourth_ack = 1;
	455
	456	if (subflow->can_ack)
	457	return true;
	458
	459	/* If the first established packet does not contain MP_CAPABLE + data
	460	* then fallback to TCP
	461	*/
	462	if (!mp_opt->mp_capable) {
	463	subflow->mp_capable = 0;
	464	tcp_sk(mptcp_subflow_tcp_sock(subflow))->is_mptcp = 0;
	465	return false;
	466	}
	467	subflow->remote_key = mp_opt->sndr_key;
	468	subflow->can_ack = 1;
	469	return true;
	470	}
	471
648ef4b8 MM	472	void mptcp_incoming_options(struct sock sk, struct sk_buff skb,
	473	struct tcp_options_received *opt_rx)
	474	{
d22f4988	475	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
648ef4b8 MM	476	struct mptcp_options_received *mp_opt;
	477	struct mptcp_ext *mpext;
	478
	479	mp_opt = &opt_rx->mptcp;
d22f4988 CP	480	if (!check_fourth_ack(subflow, skb, mp_opt))
d22f4988 CP	481	return;
648ef4b8 MM	482
	483	if (!mp_opt->dss)
	484	return;
	485
	486	mpext = skb_ext_add(skb, SKB_EXT_MPTCP);
	487	if (!mpext)
	488	return;
	489
	490	memset(mpext, 0, sizeof(*mpext));
	491
	492	if (mp_opt->use_map) {
cc7972ea	493	if (mp_opt->mpc_map) {
cc7972ea CP	494	/* this is an MP_CAPABLE carrying MPTCP data
	495	* we know this map the first chunk of data
	496	*/
	497	mptcp_crypto_key_sha(subflow->remote_key, NULL,
	498	&mpext->data_seq);
	499	mpext->data_seq++;
	500	mpext->subflow_seq = 1;
	501	mpext->dsn64 = 1;
	502	mpext->mpc_map = 1;
	503	} else {
	504	mpext->data_seq = mp_opt->data_seq;
	505	mpext->subflow_seq = mp_opt->subflow_seq;
	506	mpext->dsn64 = mp_opt->dsn64;
	507	}
648ef4b8 MM	508	mpext->data_len = mp_opt->data_len;
648ef4b8 MM	509	mpext->use_map = 1;
648ef4b8 MM	510	}
	511
	512	if (mp_opt->use_ack) {
	513	mpext->data_ack = mp_opt->data_ack;
	514	mpext->use_ack = 1;
	515	mpext->ack64 = mp_opt->ack64;
	516	}
	517
	518	mpext->data_fin = mp_opt->data_fin;
	519	}
	520
eda7acdd PK	521	void mptcp_write_options(__be32 ptr, struct mptcp_out_options opts)
eda7acdd PK	522	{
cc7972ea	523	if ((OPTION_MPTCP_MPC_SYN \| OPTION_MPTCP_MPC_SYNACK \|
eda7acdd PK	524	OPTION_MPTCP_MPC_ACK) & opts->suboptions) {
	525	u8 len;
	526
	527	if (OPTION_MPTCP_MPC_SYN & opts->suboptions)
	528	len = TCPOLEN_MPTCP_MPC_SYN;
cec37a6e PK	529	else if (OPTION_MPTCP_MPC_SYNACK & opts->suboptions)
cec37a6e PK	530	len = TCPOLEN_MPTCP_MPC_SYNACK;
cc7972ea CP	531	else if (opts->ext_copy.data_len)
cc7972ea CP	532	len = TCPOLEN_MPTCP_MPC_ACK_DATA;
eda7acdd PK	533	else
	534	len = TCPOLEN_MPTCP_MPC_ACK;
	535
	536	*ptr++ = htonl((TCPOPT_MPTCP << 24) \| (len << 16) \|
	537	(MPTCPOPT_MP_CAPABLE << 12) \|
	538	(MPTCP_SUPPORTED_VERSION << 8) \|
65492c5a	539	MPTCP_CAP_HMAC_SHA256);
cc7972ea CP	540
	541	if (!((OPTION_MPTCP_MPC_SYNACK \| OPTION_MPTCP_MPC_ACK) &
	542	opts->suboptions))
	543	goto mp_capable_done;
	544
eda7acdd PK	545	put_unaligned_be64(opts->sndr_key, ptr);
eda7acdd PK	546	ptr += 2;
cc7972ea CP	547	if (!((OPTION_MPTCP_MPC_ACK) & opts->suboptions))
	548	goto mp_capable_done;
	549
	550	put_unaligned_be64(opts->rcvr_key, ptr);
	551	ptr += 2;
	552	if (!opts->ext_copy.data_len)
	553	goto mp_capable_done;
	554
	555	put_unaligned_be32(opts->ext_copy.data_len << 16 \|
	556	TCPOPT_NOP << 8 \| TCPOPT_NOP, ptr);
	557	ptr += 1;
eda7acdd	558	}
6d0060f6	559
cc7972ea	560	mp_capable_done:
6d0060f6 MM	561	if (opts->ext_copy.use_ack \|\| opts->ext_copy.use_map) {
	562	struct mptcp_ext *mpext = &opts->ext_copy;
	563	u8 len = TCPOLEN_MPTCP_DSS_BASE;
	564	u8 flags = 0;
	565
	566	if (mpext->use_ack) {
	567	len += TCPOLEN_MPTCP_DSS_ACK64;
	568	flags = MPTCP_DSS_HAS_ACK \| MPTCP_DSS_ACK64;
	569	}
	570
	571	if (mpext->use_map) {
	572	len += TCPOLEN_MPTCP_DSS_MAP64;
	573
	574	/* Use only 64-bit mapping flags for now, add
	575	* support for optional 32-bit mappings later.
	576	*/
	577	flags \|= MPTCP_DSS_HAS_MAP \| MPTCP_DSS_DSN64;
	578	if (mpext->data_fin)
	579	flags \|= MPTCP_DSS_DATA_FIN;
	580	}
	581
	582	*ptr++ = htonl((TCPOPT_MPTCP << 24) \|
	583	(len << 16) \|
	584	(MPTCPOPT_DSS << 12) \|
	585	(flags));
	586
	587	if (mpext->use_ack) {
	588	put_unaligned_be64(mpext->data_ack, ptr);
	589	ptr += 2;
	590	}
	591
	592	if (mpext->use_map) {
	593	put_unaligned_be64(mpext->data_seq, ptr);
	594	ptr += 2;
	595	put_unaligned_be32(mpext->subflow_seq, ptr);
	596	ptr += 1;
	597	put_unaligned_be32(mpext->data_len << 16 \|
	598	TCPOPT_NOP << 8 \| TCPOPT_NOP, ptr);
	599	}
	600	}
eda7acdd	601	}