freebsd-nq/sys/netinet/sctp_pcb.h
Randall Stewart a99b67833a - Cleanup checksum code.
- Prepare for CRC offloading, add MIB counters (RS/MT).
- Bugfix: Disable CRC computation for IPv6 addresses with local scope (MT).
- Bugfix: Handle close() with SO_LINGER correctly when notifications
          are generated during the close() call(MT).
- Bugfix: Generate DRY event when sender is dry during subscription.
          Only for 1-to-1 style sockets (RS/MT)
- Bugfix: Put vtags for the correct amount of time into time-wait (MT).
- Bugfix: Clear vtag entries correctly on expiration (MT).
- Bugfix: shutdown() indicates ENOTCONN when called for unconnected
          1-to-1 style sockets (MT).
- Bugfix: In sctp Auth code (PL).
- Add support for devices that support SCTP csum offload (igb).
- Add missing sctp_associd to mib sysctl xsctp_tcb structure (RS)
Obtained from:	With help from Peter Lei and Michael Tuexen
2009-02-03 11:04:03 +00:00

629 lines
19 KiB
C

/*-
* Copyright (c) 2001-2007, by Cisco Systems, Inc. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* a) Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* b) Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the distribution.
*
* c) Neither the name of Cisco Systems, Inc. nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
* THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
* THE POSSIBILITY OF SUCH DAMAGE.
*/
/* $KAME: sctp_pcb.h,v 1.21 2005/07/16 01:18:47 suz Exp $ */
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#ifndef __sctp_pcb_h__
#define __sctp_pcb_h__
#include <netinet/sctp_os.h>
#include <netinet/sctp.h>
#include <netinet/sctp_constants.h>
#include <netinet/sctp_sysctl.h>
LIST_HEAD(sctppcbhead, sctp_inpcb);
LIST_HEAD(sctpasochead, sctp_tcb);
LIST_HEAD(sctpladdr, sctp_laddr);
LIST_HEAD(sctpvtaghead, sctp_tagblock);
LIST_HEAD(sctp_vrflist, sctp_vrf);
LIST_HEAD(sctp_ifnlist, sctp_ifn);
LIST_HEAD(sctp_ifalist, sctp_ifa);
TAILQ_HEAD(sctp_readhead, sctp_queued_to_read);
TAILQ_HEAD(sctp_streamhead, sctp_stream_queue_pending);
#include <netinet/sctp_structs.h>
#include <netinet/sctp_auth.h>
#define SCTP_PCBHASH_ALLADDR(port, mask) (port & mask)
#define SCTP_PCBHASH_ASOC(tag, mask) (tag & mask)
struct sctp_vrf {
LIST_ENTRY(sctp_vrf) next_vrf;
struct sctp_ifalist *vrf_addr_hash;
struct sctp_ifnlist ifnlist;
uint32_t vrf_id;
uint32_t tbl_id_v4; /* default v4 table id */
uint32_t tbl_id_v6; /* default v6 table id */
uint32_t total_ifa_count;
u_long vrf_addr_hashmark;
uint32_t refcount;
};
struct sctp_ifn {
struct sctp_ifalist ifalist;
struct sctp_vrf *vrf;
LIST_ENTRY(sctp_ifn) next_ifn;
LIST_ENTRY(sctp_ifn) next_bucket;
void *ifn_p; /* never access without appropriate lock */
uint32_t ifn_mtu;
uint32_t ifn_type;
uint32_t ifn_index; /* shorthand way to look at ifn for reference */
uint32_t refcount; /* number of reference held should be >=
* ifa_count */
uint32_t ifa_count; /* IFA's we hold (in our list - ifalist) */
uint32_t num_v6; /* number of v6 addresses */
uint32_t num_v4; /* number of v4 addresses */
uint32_t registered_af; /* registered address family for i/f events */
char ifn_name[SCTP_IFNAMSIZ];
};
/* SCTP local IFA flags */
#define SCTP_ADDR_VALID 0x00000001 /* its up and active */
#define SCTP_BEING_DELETED 0x00000002 /* being deleted, when
* refcount = 0. Note that it
* is pulled from the ifn list
* and ifa_p is nulled right
* away but it cannot be freed
* until the last *net
* pointing to it is deleted. */
#define SCTP_ADDR_DEFER_USE 0x00000004 /* Hold off using this one */
#define SCTP_ADDR_IFA_UNUSEABLE 0x00000008
struct sctp_ifa {
LIST_ENTRY(sctp_ifa) next_ifa;
LIST_ENTRY(sctp_ifa) next_bucket;
struct sctp_ifn *ifn_p; /* back pointer to parent ifn */
void *ifa; /* pointer to ifa, needed for flag update for
* that we MUST lock appropriate locks. This
* is for V6. */
union sctp_sockstore address;
uint32_t refcount; /* number of folks refering to this */
uint32_t flags;
uint32_t localifa_flags;
uint32_t vrf_id; /* vrf_id of this addr (for deleting) */
uint8_t src_is_loop;
uint8_t src_is_priv;
uint8_t src_is_glob;
uint8_t resv;
};
struct sctp_laddr {
LIST_ENTRY(sctp_laddr) sctp_nxt_addr; /* next in list */
struct sctp_ifa *ifa;
uint32_t action; /* Used during asconf and adding if no-zero
* src-addr selection will not consider this
* address. */
struct timeval start_time; /* time when this address was created */
};
struct sctp_block_entry {
int error;
};
struct sctp_timewait {
uint32_t tv_sec_at_expire; /* the seconds from boot to expire */
uint32_t v_tag; /* the vtag that can not be reused */
uint16_t lport; /* the local port used in vtag */
uint16_t rport; /* the remote port used in vtag */
};
struct sctp_tagblock {
LIST_ENTRY(sctp_tagblock) sctp_nxt_tagblock;
struct sctp_timewait vtag_block[SCTP_NUMBER_IN_VTAG_BLOCK];
};
struct sctp_epinfo {
struct socket *udp_tun_socket;
struct sctpasochead *sctp_asochash;
u_long hashasocmark;
struct sctppcbhead *sctp_ephash;
u_long hashmark;
/*-
* The TCP model represents a substantial overhead in that we get an
* additional hash table to keep explicit connections in. The
* listening TCP endpoint will exist in the usual ephash above and
* accept only INIT's. It will be incapable of sending off an INIT.
* When a dg arrives we must look in the normal ephash. If we find a
* TCP endpoint that will tell us to go to the specific endpoint
* hash and re-hash to find the right assoc/socket. If we find a UDP
* model socket we then must complete the lookup. If this fails,
* i.e. no association can be found then we must continue to see if
* a sctp_peeloff()'d socket is in the tcpephash (a spun off socket
* acts like a TCP model connected socket).
*/
struct sctppcbhead *sctp_tcpephash;
u_long hashtcpmark;
uint32_t hashtblsize;
struct sctp_vrflist *sctp_vrfhash;
u_long hashvrfmark;
struct sctp_ifnlist *vrf_ifn_hash;
u_long vrf_ifn_hashmark;
struct sctppcbhead listhead;
struct sctpladdr addr_wq;
struct sctpiterators iteratorhead;
int threads_must_exit;
/* ep zone info */
sctp_zone_t ipi_zone_ep;
sctp_zone_t ipi_zone_asoc;
sctp_zone_t ipi_zone_laddr;
sctp_zone_t ipi_zone_net;
sctp_zone_t ipi_zone_chunk;
sctp_zone_t ipi_zone_readq;
sctp_zone_t ipi_zone_strmoq;
sctp_zone_t ipi_zone_asconf;
sctp_zone_t ipi_zone_asconf_ack;
struct rwlock ipi_ep_mtx;
struct mtx it_mtx;
struct mtx ipi_iterator_wq_mtx;
struct rwlock ipi_addr_mtx;
struct mtx ipi_pktlog_mtx;
uint32_t ipi_count_ep;
/* assoc/tcb zone info */
uint32_t ipi_count_asoc;
/* local addrlist zone info */
uint32_t ipi_count_laddr;
/* remote addrlist zone info */
uint32_t ipi_count_raddr;
/* chunk structure list for output */
uint32_t ipi_count_chunk;
/* socket queue zone info */
uint32_t ipi_count_readq;
/* socket queue zone info */
uint32_t ipi_count_strmoq;
/* Number of vrfs */
uint32_t ipi_count_vrfs;
/* Number of ifns */
uint32_t ipi_count_ifns;
/* Number of ifas */
uint32_t ipi_count_ifas;
/* system wide number of free chunks hanging around */
uint32_t ipi_free_chunks;
uint32_t ipi_free_strmoq;
struct sctpvtaghead vtag_timewait[SCTP_STACK_VTAG_HASH_SIZE];
/* address work queue handling */
#if defined(SCTP_USE_THREAD_BASED_ITERATOR)
uint32_t iterator_running;
SCTP_PROCESS_STRUCT thread_proc;
#endif
struct sctp_timer addr_wq_timer;
};
struct sctp_base_info {
/*
* All static structures that anchor the system must be here.
*/
struct sctp_epinfo sctppcbinfo;
struct sctpstat sctpstat;
struct sctp_sysctl sctpsysctl;
uint8_t first_time;
char sctp_pcb_initialized;
#if defined(SCTP_PACKET_LOGGING)
int packet_log_writers;
int packet_log_end;
uint8_t packet_log_buffer[SCTP_PACKET_LOG_SIZE];
#endif
};
/*-
* Here we have all the relevant information for each SCTP entity created. We
* will need to modify this as approprate. We also need to figure out how to
* access /dev/random.
*/
struct sctp_pcb {
unsigned int time_of_secret_change; /* number of seconds from
* timeval.tv_sec */
uint32_t secret_key[SCTP_HOW_MANY_SECRETS][SCTP_NUMBER_OF_SECRETS];
unsigned int size_of_a_cookie;
unsigned int sctp_timeoutticks[SCTP_NUM_TMRS];
unsigned int sctp_minrto;
unsigned int sctp_maxrto;
unsigned int initial_rto;
int initial_init_rto_max;
unsigned int sctp_sack_freq;
uint32_t sctp_sws_sender;
uint32_t sctp_sws_receiver;
uint32_t sctp_default_cc_module;
/* authentication related fields */
struct sctp_keyhead shared_keys;
sctp_auth_chklist_t *local_auth_chunks;
sctp_hmaclist_t *local_hmacs;
uint16_t default_keyid;
/* various thresholds */
/* Max times I will init at a guy */
uint16_t max_init_times;
/* Max times I will send before we consider someone dead */
uint16_t max_send_times;
uint16_t def_net_failure;
/* number of streams to pre-open on a association */
uint16_t pre_open_stream_count;
uint16_t max_open_streams_intome;
/* random number generator */
uint32_t random_counter;
uint8_t random_numbers[SCTP_SIGNATURE_ALOC_SIZE];
uint8_t random_store[SCTP_SIGNATURE_ALOC_SIZE];
/*
* This timer is kept running per endpoint. When it fires it will
* change the secret key. The default is once a hour
*/
struct sctp_timer signature_change;
/* Zero copy full buffer timer */
struct sctp_timer zero_copy_timer;
/* Zero copy app to transport (sendq) read repulse timer */
struct sctp_timer zero_copy_sendq_timer;
uint32_t def_cookie_life;
/* defaults to 0 */
int auto_close_time;
uint32_t initial_sequence_debug;
uint32_t adaptation_layer_indicator;
uint32_t store_at;
uint8_t max_burst;
char current_secret_number;
char last_secret_number;
};
#ifndef SCTP_ALIGNMENT
#define SCTP_ALIGNMENT 32
#endif
#ifndef SCTP_ALIGNM1
#define SCTP_ALIGNM1 (SCTP_ALIGNMENT-1)
#endif
#define sctp_lport ip_inp.inp.inp_lport
struct sctp_pcbtsn_rlog {
uint32_t vtag;
uint16_t strm;
uint16_t seq;
uint16_t sz;
uint16_t flgs;
};
#define SCTP_READ_LOG_SIZE 135 /* we choose the number to make a pcb a page */
struct sctp_inpcb {
/*-
* put an inpcb in front of it all, kind of a waste but we need to
* for compatability with all the other stuff.
*/
union {
struct inpcb inp;
char align[(sizeof(struct in6pcb) + SCTP_ALIGNM1) &
~SCTP_ALIGNM1];
} ip_inp;
/* Socket buffer lock protects read_queue and of course sb_cc */
struct sctp_readhead read_queue;
LIST_ENTRY(sctp_inpcb) sctp_list; /* lists all endpoints */
/* hash of all endpoints for model */
LIST_ENTRY(sctp_inpcb) sctp_hash;
/* count of local addresses bound, 0 if bound all */
int laddr_count;
/* list of addrs in use by the EP, NULL if bound-all */
struct sctpladdr sctp_addr_list;
/*
* used for source address selection rotation when we are subset
* bound
*/
struct sctp_laddr *next_addr_touse;
/* back pointer to our socket */
struct socket *sctp_socket;
uint32_t sctp_flags; /* INP state flag set */
uint32_t sctp_features; /* Feature flags */
uint32_t sctp_mobility_features; /* Mobility Feature flags */
struct sctp_pcb sctp_ep;/* SCTP ep data */
/* head of the hash of all associations */
struct sctpasochead *sctp_tcbhash;
u_long sctp_hashmark;
/* head of the list of all associations */
struct sctpasochead sctp_asoc_list;
#ifdef SCTP_TRACK_FREED_ASOCS
struct sctpasochead sctp_asoc_free_list;
#endif
struct sctp_iterator *inp_starting_point_for_iterator;
uint32_t sctp_frag_point;
uint32_t partial_delivery_point;
uint32_t sctp_context;
struct sctp_nonpad_sndrcvinfo def_send;
/*-
* These three are here for the sosend_dgram
* (pkt, pkt_last and control).
* routine. However, I don't think anyone in
* the current FreeBSD kernel calls this. So
* they are candidates with sctp_sendm for
* de-supporting.
*/
struct mbuf *pkt, *pkt_last;
struct mbuf *control;
struct mtx inp_mtx;
struct mtx inp_create_mtx;
struct mtx inp_rdata_mtx;
int32_t refcount;
uint32_t def_vrf_id;
uint32_t total_sends;
uint32_t total_recvs;
uint32_t last_abort_code;
uint32_t total_nospaces;
struct sctpasochead *sctp_asocidhash;
u_long hashasocidmark;
uint32_t sctp_associd_counter;
#ifdef SCTP_ASOCLOG_OF_TSNS
struct sctp_pcbtsn_rlog readlog[SCTP_READ_LOG_SIZE];
uint32_t readlog_index;
#endif
};
struct sctp_tcb {
struct socket *sctp_socket; /* back pointer to socket */
struct sctp_inpcb *sctp_ep; /* back pointer to ep */
LIST_ENTRY(sctp_tcb) sctp_tcbhash; /* next link in hash
* table */
LIST_ENTRY(sctp_tcb) sctp_tcblist; /* list of all of the
* TCB's */
LIST_ENTRY(sctp_tcb) sctp_tcbasocidhash; /* next link in asocid
* hash table */
LIST_ENTRY(sctp_tcb) sctp_asocs; /* vtag hash list */
struct sctp_block_entry *block_entry; /* pointer locked by socket
* send buffer */
struct sctp_association asoc;
/*-
* freed_by_sorcv_sincelast is protected by the sockbuf_lock NOT the
* tcb_lock. Its special in this way to help avoid extra mutex calls
* in the reading of data.
*/
uint32_t freed_by_sorcv_sincelast;
uint32_t total_sends;
uint32_t total_recvs;
int freed_from_where;
uint16_t rport; /* remote port in network format */
uint16_t resv;
struct mtx tcb_mtx;
struct mtx tcb_send_mtx;
};
#include <netinet/sctp_lock_bsd.h>
/* TODO where to put non-_KERNEL things for __Userspace__? */
#if defined(_KERNEL) || defined(__Userspace__)
/* Attention Julian, this is the extern that
* goes with the base info. sctp_pcb.c has
* the real definition.
*/
extern struct sctp_base_info system_base_info;
#ifdef INET6
int SCTP6_ARE_ADDR_EQUAL(struct sockaddr_in6 *a, struct sockaddr_in6 *b);
#endif
void sctp_fill_pcbinfo(struct sctp_pcbinfo *);
struct sctp_ifn *
sctp_find_ifn(void *ifn, uint32_t ifn_index);
struct sctp_vrf *sctp_allocate_vrf(int vrfid);
struct sctp_vrf *sctp_find_vrf(uint32_t vrfid);
void sctp_free_vrf(struct sctp_vrf *vrf);
/*-
* Change address state, can be used if
* O/S supports telling transports about
* changes to IFA/IFN's (link layer triggers).
* If a ifn goes down, we will do src-addr-selection
* and NOT use that, as a source address. This does
* not stop the routing system from routing out
* that interface, but we won't put it as a source.
*/
void sctp_mark_ifa_addr_down(uint32_t vrf_id, struct sockaddr *addr, const char *if_name, uint32_t ifn_index);
void sctp_mark_ifa_addr_up(uint32_t vrf_id, struct sockaddr *addr, const char *if_name, uint32_t ifn_index);
struct sctp_ifa *
sctp_add_addr_to_vrf(uint32_t vrfid,
void *ifn, uint32_t ifn_index, uint32_t ifn_type,
const char *if_name,
void *ifa, struct sockaddr *addr, uint32_t ifa_flags,
int dynamic_add);
void sctp_update_ifn_mtu(uint32_t ifn_index, uint32_t mtu);
void sctp_free_ifn(struct sctp_ifn *sctp_ifnp);
void sctp_free_ifa(struct sctp_ifa *sctp_ifap);
void
sctp_del_addr_from_vrf(uint32_t vrfid, struct sockaddr *addr,
uint32_t ifn_index, const char *if_name);
struct sctp_nets *sctp_findnet(struct sctp_tcb *, struct sockaddr *);
struct sctp_inpcb *sctp_pcb_findep(struct sockaddr *, int, int, uint32_t);
int
sctp_inpcb_bind(struct socket *, struct sockaddr *,
struct sctp_ifa *, struct thread *);
struct sctp_tcb *
sctp_findassociation_addr(struct mbuf *, int, int,
struct sctphdr *, struct sctp_chunkhdr *, struct sctp_inpcb **,
struct sctp_nets **, uint32_t vrf_id);
struct sctp_tcb *
sctp_findassociation_addr_sa(struct sockaddr *,
struct sockaddr *, struct sctp_inpcb **, struct sctp_nets **, int, uint32_t);
void
sctp_move_pcb_and_assoc(struct sctp_inpcb *, struct sctp_inpcb *,
struct sctp_tcb *);
/*-
* For this call ep_addr, the to is the destination endpoint address of the
* peer (relative to outbound). The from field is only used if the TCP model
* is enabled and helps distingush amongst the subset bound (non-boundall).
* The TCP model MAY change the actual ep field, this is why it is passed.
*/
struct sctp_tcb *
sctp_findassociation_ep_addr(struct sctp_inpcb **,
struct sockaddr *, struct sctp_nets **, struct sockaddr *,
struct sctp_tcb *);
struct sctp_tcb *
sctp_findasoc_ep_asocid_locked(struct sctp_inpcb *inp, sctp_assoc_t asoc_id, int want_lock);
struct sctp_tcb *
sctp_findassociation_ep_asocid(struct sctp_inpcb *,
sctp_assoc_t, int);
struct sctp_tcb *
sctp_findassociation_ep_asconf(struct mbuf *, int, int,
struct sctphdr *, struct sctp_inpcb **, struct sctp_nets **, uint32_t vrf_id);
int sctp_inpcb_alloc(struct socket *so, uint32_t vrf_id);
int sctp_is_address_on_local_host(struct sockaddr *addr, uint32_t vrf_id);
void sctp_inpcb_free(struct sctp_inpcb *, int, int);
struct sctp_tcb *
sctp_aloc_assoc(struct sctp_inpcb *, struct sockaddr *,
int, int *, uint32_t, uint32_t, struct thread *);
int sctp_free_assoc(struct sctp_inpcb *, struct sctp_tcb *, int, int);
void sctp_delete_from_timewait(uint32_t, uint16_t, uint16_t);
int sctp_is_in_timewait(uint32_t tag, uint16_t lport, uint16_t rport);
void
sctp_add_vtag_to_timewait(uint32_t tag, uint32_t time, uint16_t lport, uint16_t rport);
void sctp_add_local_addr_ep(struct sctp_inpcb *, struct sctp_ifa *, uint32_t);
int sctp_insert_laddr(struct sctpladdr *, struct sctp_ifa *, uint32_t);
void sctp_remove_laddr(struct sctp_laddr *);
void sctp_del_local_addr_ep(struct sctp_inpcb *, struct sctp_ifa *);
int sctp_add_remote_addr(struct sctp_tcb *, struct sockaddr *, int, int);
void sctp_remove_net(struct sctp_tcb *, struct sctp_nets *);
int sctp_del_remote_addr(struct sctp_tcb *, struct sockaddr *);
void sctp_pcb_init(void);
void sctp_pcb_finish(void);
void sctp_add_local_addr_restricted(struct sctp_tcb *, struct sctp_ifa *);
void sctp_del_local_addr_restricted(struct sctp_tcb *, struct sctp_ifa *);
int
sctp_load_addresses_from_init(struct sctp_tcb *, struct mbuf *, int, int,
int, struct sctphdr *, struct sockaddr *);
int
sctp_set_primary_addr(struct sctp_tcb *, struct sockaddr *,
struct sctp_nets *);
int sctp_is_vtag_good(struct sctp_inpcb *, uint32_t, uint16_t lport, uint16_t rport, struct timeval *, int);
/* void sctp_drain(void); */
int sctp_destination_is_reachable(struct sctp_tcb *, struct sockaddr *);
int sctp_swap_inpcb_for_listen(struct sctp_inpcb *inp);
/*-
* Null in last arg inpcb indicate run on ALL ep's. Specific inp in last arg
* indicates run on ONLY assoc's of the specified endpoint.
*/
int
sctp_initiate_iterator(inp_func inpf,
asoc_func af,
inp_func inpe,
uint32_t, uint32_t,
uint32_t, void *,
uint32_t,
end_func ef,
struct sctp_inpcb *,
uint8_t co_off);
#endif /* _KERNEL */
#endif /* !__sctp_pcb_h__ */