/*

 * CDDL HEADER START

 *

 * The contents of this file are subject to the terms of the

 * Common Development and Distribution License, Version 1.0 only

 * (the "License").  You may not use this file except in compliance

 * with the License.

 *

 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE

 * or http://www.opensolaris.org/os/licensing.

 * See the License for the specific language governing permissions

 * and limitations under the License.

 *

 * When distributing Covered Code, include this CDDL HEADER in each

 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.

 * If applicable, add the following below this CDDL HEADER, with the

 * fields enclosed by brackets "[]" replaced with your own identifying

 * information: Portions Copyright [yyyy] [name of copyright owner]

 *

 * CDDL HEADER END

 */

/*

 * Copyright 2005 Sun Microsystems, Inc.  All rights reserved.

 * Use is subject to license terms.

 */

/* Copyright (c) 1990 Mentat Inc. */

#ifndef	_INET_IP_H

#define	_INET_IP_H

#pragma ident	"%Z%%M%	%I%	%E% SMI"

#ifdef	__cplusplus

extern "C" {

#endif

#include <sys/isa_defs.h>

#include <sys/types.h>

#include <inet/mib2.h>

#include <inet/nd.h>

#include <sys/atomic.h>

#include <sys/socket.h>

#include <net/if_dl.h>

#include <net/if.h>

#include <netinet/ip.h>

#include <sys/dlpi.h>

#include <netinet/igmp.h>

#ifdef _KERNEL

#include <netinet/ip6.h>

#include <sys/avl.h>

#include <sys/vmem.h>

#include <sys/squeue.h>

#include <sys/systm.h>

#include <sys/multidata.h>

#ifdef DEBUG

#define	ILL_DEBUG

#define	IRE_DEBUG

#define	NCE_DEBUG

#define	CONN_DEBUG

#endif

#define	IP_DEBUG

/*

 * The mt-streams(9F) flags for the IP module; put here so that other

 * "drivers" that are actually IP (e.g., ICMP, UDP) can use the same set

 * of flags.

 */

#define	IP_DEVMTFLAGS D_MP

#endif	/* _KERNEL */

#define	IP_MOD_NAME	"ip"

#define	IP_DEV_NAME	"/dev/ip"

#define	IP6_DEV_NAME	"/dev/ip6"

#define	UDP_MOD_NAME	"udp"

#define	UDP_DEV_NAME	"/dev/udp"

#define	UDP6_DEV_NAME	"/dev/udp6"

#define	TCP_MOD_NAME	"tcp"

#define	TCP_DEV_NAME	"/dev/tcp"

#define	TCP6_DEV_NAME	"/dev/tcp6"

/* Minor numbers */

#define	IPV4_MINOR	0

#define	IPV6_MINOR	1

#define	TCP_MINOR	2

#define	TCP_MINOR6	3

#ifndef	_IPADDR_T

#define	_IPADDR_T

typedef uint32_t ipaddr_t;

#endif

/* Number of bits in an address */

#define	IP_ABITS		32

#define	IPV6_ABITS		128

#define	IP_HOST_MASK		(ipaddr_t)0xffffffffU

#define	IP_CSUM(mp, off, sum)		(~ip_cksum(mp, off, sum) & 0xFFFF)

#define	IP_CSUM_PARTIAL(mp, off, sum)	ip_cksum(mp, off, sum)

#define	IP_BCSUM_PARTIAL(bp, len, sum)	bcksum(bp, len, sum)

#define	IP_MD_CSUM(pd, off, sum)	(~ip_md_cksum(pd, off, sum) & 0xffff)

#define	IP_MD_CSUM_PARTIAL(pd, off, sum) ip_md_cksum(pd, off, sum)

/*

 * Flag to IP write side to indicate that the appln has sent in a pre-built

 * IP header. Stored in ipha_ident (which is otherwise zero).

 */

#define	IP_HDR_INCLUDED			0xFFFF

#define	ILL_FRAG_HASH_TBL_COUNT	((unsigned int)64)

#define	ILL_FRAG_HASH_TBL_SIZE	(ILL_FRAG_HASH_TBL_COUNT * sizeof (ipfb_t))

#define	IPV4_ADDR_LEN			4

#define	IP_ADDR_LEN			IPV4_ADDR_LEN

#define	IP_ARP_PROTO_TYPE		0x0800

#define	IPV4_VERSION			4

#define	IP_VERSION			IPV4_VERSION

#define	IP_SIMPLE_HDR_LENGTH_IN_WORDS	5

#define	IP_SIMPLE_HDR_LENGTH		20

#define	IP_MAX_HDR_LENGTH		60

#define	IP_MIN_MTU			(IP_MAX_HDR_LENGTH + 8)	/* 68 bytes */

/*

 * XXX IP_MAXPACKET is defined in <netinet/ip.h> as well. At some point the

 * 2 files should be cleaned up to remove all redundant definitions.

 */

#define	IP_MAXPACKET			65535

#define	IP_SIMPLE_HDR_VERSION \

	((IP_VERSION << 4) | IP_SIMPLE_HDR_LENGTH_IN_WORDS)

/*

 * Constants and type definitions to support IP IOCTL commands

 */

#define	IP_IOCTL			(('i'<<8)|'p')

#define	IP_IOC_IRE_DELETE		4

#define	IP_IOC_IRE_DELETE_NO_REPLY	5

#define	IP_IOC_IRE_ADVISE_NO_REPLY	6

#define	IP_IOC_RTS_REQUEST		7

/* Common definitions used by IP IOCTL data structures */

typedef struct ipllcmd_s {

	uint_t	ipllc_cmd;

	uint_t	ipllc_name_offset;

	uint_t	ipllc_name_length;

} ipllc_t;

/* IP IRE Change Command Structure. */

typedef struct ipic_s {

	ipllc_t	ipic_ipllc;

	uint_t	ipic_ire_type;

	uint_t	ipic_max_frag;

	uint_t	ipic_addr_offset;

	uint_t	ipic_addr_length;

	uint_t	ipic_mask_offset;

	uint_t	ipic_mask_length;

	uint_t	ipic_src_addr_offset;

	uint_t	ipic_src_addr_length;

	uint_t	ipic_ll_hdr_offset;

	uint_t	ipic_ll_hdr_length;

	uint_t	ipic_gateway_addr_offset;

	uint_t	ipic_gateway_addr_length;

	clock_t	ipic_rtt;

	uint32_t ipic_ssthresh;

	clock_t	ipic_rtt_sd;

	uchar_t ipic_ire_marks;

} ipic_t;

#define	ipic_cmd		ipic_ipllc.ipllc_cmd

#define	ipic_ll_name_length	ipic_ipllc.ipllc_name_length

#define	ipic_ll_name_offset	ipic_ipllc.ipllc_name_offset

/* IP IRE Delete Command Structure. */

typedef struct ipid_s {

	ipllc_t	ipid_ipllc;

	uint_t	ipid_ire_type;

	uint_t	ipid_addr_offset;

	uint_t	ipid_addr_length;

	uint_t	ipid_mask_offset;

	uint_t	ipid_mask_length;

} ipid_t;

#define	ipid_cmd		ipid_ipllc.ipllc_cmd

#ifdef _KERNEL

/*

 * Temporary state for ip options parser.

 */

typedef struct ipoptp_s

{

	uint8_t		*ipoptp_next;	/* next option to look at */

	uint8_t		*ipoptp_end;	/* end of options */

	uint8_t		*ipoptp_cur;	/* start of current option */

	uint8_t		ipoptp_len;	/* length of current option */

	uint32_t	ipoptp_flags;

} ipoptp_t;

/*

 * Flag(s) for ipoptp_flags

 */

#define	IPOPTP_ERROR	0x00000001

#endif	/* _KERNEL */

/* Controls forwarding of IP packets, set via ndd */

#define	IP_FORWARD_NEVER	0

#define	IP_FORWARD_ALWAYS	1

#define	WE_ARE_FORWARDING	(ip_g_forward == IP_FORWARD_ALWAYS)

#define	IPH_HDR_LENGTH(ipha)						\

	((int)(((ipha_t *)ipha)->ipha_version_and_hdr_length & 0xF) << 2)

#define	IPH_HDR_VERSION(ipha)						\

	((int)(((ipha_t *)ipha)->ipha_version_and_hdr_length) >> 4)

#ifdef _KERNEL

/*

 * IP reassembly macros.  We hide starting and ending offsets in b_next and

 * b_prev of messages on the reassembly queue.	The messages are chained using

 * b_cont.  These macros are used in ip_reassemble() so we don't have to see

 * the ugly casts and assignments.

 * Note that the offsets are <= 64k i.e. a uint_t is sufficient to represent

 * them.

 */

#define	IP_REASS_START(mp)		((uint_t)(uintptr_t)((mp)->b_next))

#define	IP_REASS_SET_START(mp, u)	\

	((mp)->b_next = (mblk_t *)(uintptr_t)(u))

#define	IP_REASS_END(mp)		((uint_t)(uintptr_t)((mp)->b_prev))

#define	IP_REASS_SET_END(mp, u)		\

	((mp)->b_prev = (mblk_t *)(uintptr_t)(u))

#define	IP_REASS_COMPLETE	0x1

#define	IP_REASS_PARTIAL	0x2

#define	IP_REASS_FAILED		0x4

/*

 * Test to determine whether this is a module instance of IP or a

 * driver instance of IP.

 */

#define	CONN_Q(q)	(WR(q)->q_next == NULL)

#define	Q_TO_CONN(q)	((conn_t *)(q)->q_ptr)

#define	Q_TO_TCP(q)	(Q_TO_CONN((q))->conn_tcp)

#define	Q_TO_UDP(q)	(Q_TO_CONN((q))->conn_udp)

/*

 * The following two macros are used by IP to get the appropriate

 * wq and rq for a conn. If it is a TCP conn, then we need

 * tcp_wq/tcp_rq else, conn_wq/conn_rq. IP can use conn_wq and conn_rq

 * from a conn directly if it knows that the conn is not TCP.

 */

#define	CONNP_TO_WQ(connp)	\

	(IPCL_IS_TCP(connp) ? (connp)->conn_tcp->tcp_wq : (connp)->conn_wq)

#define	CONNP_TO_RQ(connp)	RD(CONNP_TO_WQ(connp))

#define	GRAB_CONN_LOCK(q)	{				\

	if (q != NULL && CONN_Q(q))				\

		mutex_enter(&(Q_TO_CONN(q))->conn_lock);	\

}

#define	RELEASE_CONN_LOCK(q)	{				\

	if (q != NULL && CONN_Q(q))				\

		mutex_exit(&(Q_TO_CONN(q))->conn_lock);		\

}

/* "Congestion controlled" protocol */

#define	IP_FLOW_CONTROLLED_ULP(p)   ((p) == IPPROTO_TCP || (p) == IPPROTO_SCTP)

/*

 * Complete the pending operation. Usually an ioctl. Can also

 * be a bind or option management request that got enqueued

 * in an ipsq_t. Called on completion of the operation.

 */

#define	CONN_OPER_PENDING_DONE(connp)	{			\

	mutex_enter(&(connp)->conn_lock);			\

	(connp)->conn_oper_pending_ill = NULL;			\

	cv_broadcast(&(connp)->conn_refcv);			\

	mutex_exit(&(connp)->conn_lock);			\

	CONN_DEC_REF(connp);					\

}

/* Get the credential of an IP queue of unknown type */

#define	GET_QUEUE_CRED(wq)						\

	((wq)->q_next ? (((ill_t *)(wq)->q_ptr)->ill_credp) \

	    : ((Q_TO_CONN((wq)))->conn_cred))

/*

 * Flags for the various ip_fanout_* routines.

 */

#define	IP_FF_SEND_ICMP		0x01	/* Send an ICMP error */

#define	IP_FF_HDR_COMPLETE	0x02	/* Call ip_hdr_complete if error */

#define	IP_FF_CKSUM		0x04	/* Recompute ipha_cksum if error */

#define	IP_FF_RAWIP		0x08	/* Use rawip mib variable */

#define	IP_FF_SRC_QUENCH	0x10	/* OK to send ICMP_SOURCE_QUENCH */

#define	IP_FF_SYN_ADDIRE	0x20	/* Add IRE if TCP syn packet */

#define	IP_FF_IP6INFO		0x80	/* Add ip6i_t if needed */

#define	IP_FF_SEND_SLLA		0x100	/* Send source link layer info ? */

#define	IPV6_REACHABILITY_CONFIRMATION	0x200	/* Flags for ip_xmit_v6 */

#define	IP_FF_NO_MCAST_LOOP	0x400	/* No multicasts for sending zone */

/*

 * Following flags are used by IPQoS to determine if policy processing is

 * required.

 */

#define	IP6_NO_IPPOLICY		0x800	/* Don't do IPQoS processing */

#define	IP6_IN_LLMCAST		0x1000	/* Multicast */

#define	IP_FF_LOOPBACK		0x2000	/* Loopback fanout */

#ifndef	IRE_DB_TYPE

#define	IRE_DB_TYPE	M_SIG

#endif

#ifndef	IRE_DB_REQ_TYPE

#define	IRE_DB_REQ_TYPE	M_PCSIG

#endif

/*

 * Values for squeue switch:

 */

#define	IP_SQUEUE_ENTER_NODRAIN	1

#define	IP_SQUEUE_ENTER	2

/*

 * This is part of the interface between Transport provider and

 * IP which can be used to set policy information. This is usually

 * accompanied with O_T_BIND_REQ/T_BIND_REQ.ip_bind assumes that

 * only IPSEC_POLICY_SET is there when it is found in the chain.

 * The information contained is an struct ipsec_req_t. On success

 * or failure, either the T_BIND_ACK or the T_ERROR_ACK is returned.

 * IPSEC_POLICY_SET is never returned.

 */

#define	IPSEC_POLICY_SET	M_SETOPTS

#define	IRE_IS_LOCAL(ire)	((ire != NULL) && \

				((ire)->ire_type & (IRE_LOCAL | IRE_LOOPBACK)))

#define	IRE_IS_TARGET(ire)	((ire != NULL) && \

				((ire)->ire_type != IRE_BROADCAST))

/* IP Fragmentation Reassembly Header */

typedef struct ipf_s {

	struct ipf_s	*ipf_hash_next;

	struct ipf_s	**ipf_ptphn;	/* Pointer to previous hash next. */

	uint32_t	ipf_ident;	/* Ident to match. */

	uint8_t		ipf_protocol;	/* Protocol to match. */

	uchar_t		ipf_last_frag_seen : 1;	/* Last fragment seen ? */

	time_t		ipf_timestamp;	/* Reassembly start time. */

	mblk_t		*ipf_mp;	/* mblk we live in. */

	mblk_t		*ipf_tail_mp;	/* Frag queue tail pointer. */

	int		ipf_hole_cnt;	/* Number of holes (hard-case). */

	int		ipf_end;	/* Tail end offset (0 -> hard-case). */

	uint_t		ipf_gen;	/* Frag queue generation */

	size_t		ipf_count;	/* Count of bytes used by frag */

	uint_t		ipf_nf_hdr_len; /* Length of nonfragmented header */

	in6_addr_t	ipf_v6src;	/* IPv6 source address */

	in6_addr_t	ipf_v6dst;	/* IPv6 dest address */

	uint_t		ipf_prev_nexthdr_offset; /* Offset for nexthdr value */

	uint8_t		ipf_ecn;	/* ECN info for the fragments */

	uint8_t		ipf_num_dups;	/* Number of times dup frags recvd */

	uint16_t	ipf_checksum_flags; /* Hardware checksum flags */

	uint32_t	ipf_checksum;	/* Partial checksum of fragment data */

} ipf_t;

#define	ipf_src	V4_PART_OF_V6(ipf_v6src)

#define	ipf_dst	V4_PART_OF_V6(ipf_v6dst)

typedef enum {

	IB_PKT =  0x01,

	OB_PKT = 0x02

} ip_pkt_t;

#define	UPDATE_IB_PKT_COUNT(ire)\

	{ \

	(ire)->ire_ib_pkt_count++; \

	if ((ire)->ire_ipif != NULL) { \

		/* \

		 * forwarding packet \

		 */ \

		if ((ire)->ire_type & (IRE_LOCAL|IRE_BROADCAST)) \

			atomic_add_32(&(ire)->ire_ipif->ipif_ib_pkt_count, 1);\

		else \

			atomic_add_32(&(ire)->ire_ipif->ipif_fo_pkt_count, 1);\

	} \

	}

#define	UPDATE_OB_PKT_COUNT(ire)\

	{ \

	(ire)->ire_ob_pkt_count++;\

	if ((ire)->ire_ipif != NULL) { \

		atomic_add_32(&(ire)->ire_ipif->ipif_ob_pkt_count, 1); \

	} \

	}

#define	IP_RPUT_LOCAL(q, mp, ipha, ire, recv_ill) \

{ \

	switch (ipha->ipha_protocol) { \

		case IPPROTO_UDP: \

			ip_udp_input(q, mp, ipha, ire, recv_ill); \

			break; \

		default: \

			ip_proto_input(q, mp, ipha, ire, recv_ill); \

			break; \

	} \

}

#endif /* _KERNEL */

/* ICMP types */

#define	ICMP_ECHO_REPLY			0

#define	ICMP_DEST_UNREACHABLE		3

#define	ICMP_SOURCE_QUENCH		4

#define	ICMP_REDIRECT			5

#define	ICMP_ECHO_REQUEST		8

#define	ICMP_ROUTER_ADVERTISEMENT	9

#define	ICMP_ROUTER_SOLICITATION	10

#define	ICMP_TIME_EXCEEDED		11

#define	ICMP_PARAM_PROBLEM		12

#define	ICMP_TIME_STAMP_REQUEST		13

#define	ICMP_TIME_STAMP_REPLY		14

#define	ICMP_INFO_REQUEST		15

#define	ICMP_INFO_REPLY			16

#define	ICMP_ADDRESS_MASK_REQUEST	17

#define	ICMP_ADDRESS_MASK_REPLY		18

/* ICMP_TIME_EXCEEDED codes */

#define	ICMP_TTL_EXCEEDED		0

#define	ICMP_REASSEMBLY_TIME_EXCEEDED	1

/* ICMP_DEST_UNREACHABLE codes */

#define	ICMP_NET_UNREACHABLE		0

#define	ICMP_HOST_UNREACHABLE		1

#define	ICMP_PROTOCOL_UNREACHABLE	2

#define	ICMP_PORT_UNREACHABLE		3

#define	ICMP_FRAGMENTATION_NEEDED	4

#define	ICMP_SOURCE_ROUTE_FAILED	5

#define	ICMP_DEST_NET_UNKNOWN		6

#define	ICMP_DEST_HOST_UNKNOWN		7

#define	ICMP_SRC_HOST_ISOLATED		8

#define	ICMP_DEST_NET_UNREACH_ADMIN	9

#define	ICMP_DEST_HOST_UNREACH_ADMIN	10

#define	ICMP_DEST_NET_UNREACH_TOS	11

#define	ICMP_DEST_HOST_UNREACH_TOS	12

/* ICMP Header Structure */

typedef struct icmph_s {

	uint8_t		icmph_type;

	uint8_t		icmph_code;

	uint16_t	icmph_checksum;

	union {

		struct { /* ECHO request/response structure */

			uint16_t	u_echo_ident;

			uint16_t	u_echo_seqnum;

		} u_echo;

		struct { /* Destination unreachable structure */

			uint16_t	u_du_zero;

			uint16_t	u_du_mtu;

		} u_du;

		struct { /* Parameter problem structure */

			uint8_t		u_pp_ptr;

			uint8_t		u_pp_rsvd[3];

		} u_pp;

		struct { /* Redirect structure */

			ipaddr_t	u_rd_gateway;

		} u_rd;

	} icmph_u;

} icmph_t;

#define	icmph_echo_ident	icmph_u.u_echo.u_echo_ident

#define	icmph_echo_seqnum	icmph_u.u_echo.u_echo_seqnum

#define	icmph_du_zero		icmph_u.u_du.u_du_zero

#define	icmph_du_mtu		icmph_u.u_du.u_du_mtu

#define	icmph_pp_ptr		icmph_u.u_pp.u_pp_ptr

#define	icmph_rd_gateway	icmph_u.u_rd.u_rd_gateway

#define	ICMPH_SIZE	8

/*

 * Minimum length of transport layer header included in an ICMP error

 * message for it to be considered valid.

 */

#define	ICMP_MIN_TP_HDR_LEN	8

/* Aligned IP header */

typedef struct ipha_s {

	uint8_t		ipha_version_and_hdr_length;

	uint8_t		ipha_type_of_service;

	uint16_t	ipha_length;

	uint16_t	ipha_ident;

	uint16_t	ipha_fragment_offset_and_flags;

	uint8_t		ipha_ttl;

	uint8_t		ipha_protocol;

	uint16_t	ipha_hdr_checksum;

	ipaddr_t	ipha_src;

	ipaddr_t	ipha_dst;

} ipha_t;

#define	IPH_DF		0x4000	/* Don't fragment */

#define	IPH_MF		0x2000	/* More fragments to come */

#define	IPH_OFFSET	0x1FFF	/* Where the offset lives */

#define	IPH_FRAG_HDR	0x8000	/* IPv6 don't fragment bit */

/* ECN code points for IPv4 TOS byte and IPv6 traffic class octet. */

#define	IPH_ECN_NECT	0x0	/* Not ECN-Capabable Transport */

#define	IPH_ECN_ECT1	0x1	/* ECN-Capable Transport, ECT(1) */

#define	IPH_ECN_ECT0	0x2	/* ECN-Capable Transport, ECT(0) */

#define	IPH_ECN_CE	0x3	/* ECN-Congestion Experienced (CE) */

/* IP Mac info structure */

typedef struct ip_m_s {

	t_uscalar_t	ip_m_mac_type;	/* From <sys/dlpi.h> */

	int		ip_m_type;	/* From <net/if_types.h> */

	boolean_t	(*ip_m_v4mapinfo)(uint_t, uint8_t *, uint8_t *,

			    uint32_t *, ipaddr_t *);

	boolean_t	(*ip_m_v6mapinfo)(uint_t, uint8_t *, uint8_t *,

			    uint32_t *, in6_addr_t *);

	boolean_t	(*ip_m_v6intfid)(uint_t, uint8_t *, in6_addr_t *);

} ip_m_t;