Lines Matching defs:to
19 * contributors may be used to endorse or promote products derived
81 * We really don't need to lock this, but I will just because it
97 * Addresses are added to VRF's (Virtual Router's). For BSD we
101 * to a VRF we lookup the ifn/ifn_index, if the ifn does
102 * not exist we create it and add it to the list of IFN's
104 * address to the list. So we look something like:
113 * point to these from its source address selection nets structure.
119 * selection pointing to the sctp_ifa, eventually the refcount
134 * - An endpoint may add multiple VRF's to it.
182 /* Add it to the hash table */
388 * Add an ifa to an ifn.
561 /* easy to solve, just switch back to active */
580 SCTPDBG(SCTP_DEBUG_PCB4, "Moving ifa %p from %s (0x%x) to %s (0x%x)\n",
586 /* move the address over to the new ifn */
632 /* ok to use deprecated addresses? */
767 * hang around in EP's that have bound specific to it until
823 sctp_does_stcb_own_this_addr(struct sctp_tcb *stcb, struct sockaddr *to)
866 * we have sent an asconf-add to be
871 if (sctp_ifa->address.sa.sa_family != to->sa_family) {
882 rsin = (struct sockaddr_in *)to;
905 rsin6 = (struct sockaddr_in6 *)to;
947 * sent an asconf-add to be considered
952 if (laddr->ifa->address.sa.sa_family != to->sa_family) {
955 switch (to->sa_family) {
962 rsin = (struct sockaddr_in *)to;
976 rsin6 = (struct sockaddr_in6 *)to;
997 struct sockaddr *to, struct sctp_nets **netp, uint32_t vrf_id)
1001 * If we support the TCP model, then we must now dig through to see
1011 if ((to == NULL) || (from == NULL)) {
1015 switch (to->sa_family) {
1019 lport = ((struct sockaddr_in *)to)->sin_port;
1029 lport = ((struct sockaddr_in6 *)to)->sin6_port;
1043 * addresses match this address (to). If so we update p_ep to point
1044 * to this ep and return the tcb from it.
1056 switch (to->sa_family) {
1062 sin = (struct sockaddr_in *)to;
1076 sin6 = (struct sockaddr_in6 *)to;
1093 /* check to see if the ep has one of the addresses */
1108 to->sa_family) {
1116 sin = (struct sockaddr_in *)to;
1130 to;
1159 if (!sctp_does_stcb_own_this_addr(stcb, to)) {
1175 if (!sctp_does_stcb_own_this_addr(stcb, to)) {
1294 * it's only chance is to be the first TCB in the list. If
1295 * it is the acceptor, then do the special_lookup to hash
1299 /* to is peer addr, from is my addr */
1540 * Use my the assoc_id to find a endpoint
1714 * ok, not bound to all so see if we can find a EP bound to this
1827 /* else fall through to conflict */
1854 /* Kick out all non-listeners to the TCP hash */
1889 * First we check the hash table to see if someone has this port
1920 * I could cheat here and just cast to one of the types but we will
1940 * its open to bind, so this is good enough.
1961 * Find an association for an endpoint with the pointer to whom you want to
1962 * send to and the endpoint pointer. The address can be IPv4 or IPv6. We may
1963 * need to change the *to to some other struct like a mbuf...
1966 sctp_findassociation_addr_sa(struct sockaddr *from, struct sockaddr *to,
1976 stcb = sctp_tcb_special_locate(inp_p, from, to, netp,
1979 stcb = sctp_tcb_special_locate(&inp, from, to, netp,
1987 inp = sctp_pcb_findep(to, 0, 1, vrf_id);
1997 * we now place the source address or from in the to of the find
2002 stcb = sctp_findassociation_ep_addr(inp_p, from, netp, to,
2005 stcb = sctp_findassociation_ep_addr(&inp, from, netp, to,
2014 * address will be used to lookup the TCB and see if one exits.
2109 sctp_findassoc_by_vtag(struct sockaddr *from, struct sockaddr *to, uint32_t vtag,
2114 * Use my vtag to hash. If we find it we then verify the source addr
2152 if (sctp_does_stcb_own_this_addr(stcb, to) == 0) {
2206 * Find an association with the pointer to the inbound IP packet. This can be
2242 * association that is linked to an existing
2267 * if the lookup address is 0.0.0.0 or ::0, use the vtag to do the lookup
2290 SCTPDBG(SCTP_DEBUG_INPUT3, "%s: failed to get asconf lookup addr\n",
2310 SCTPDBG(SCTP_DEBUG_INPUT3, "%s: failed to get asconf v6 lookup addr\n",
2337 SCTPDBG(SCTP_DEBUG_INPUT3, "%s: failed to get asconf v4 lookup addr\n",
2371 * allocate a sctp_inpcb and setup a temporary binding to a port/all
2379 * we get called when a new endpoint starts up. We need to allocate
2436 /* init the small hash table we use to track asocid <-> tcb */
2506 /* add it to the info area */
2552 /* number of streams to pre-open on a association */
2668 /* Its safe to access */
2674 * Question, do we even need to worry about the ep-hash since we
2722 /* Now any running timers need to be adjusted. */
2819 * sctp_ifap is used to bypass normal local address validation checks.
2825 /* bind a ep to a socket address */
2882 * will transmute the ip address to the
2913 * will transmute the ipv6 address to the
2941 /* Setup a vrf_id to be the default for the non-bind-all case. */
2949 /* got to be root to get at low ports */
3084 /* ok we look clear to give out this port, so lets setup the binding */
3086 /* binding to all addresses, so just set in the proper flags */
3126 * address structure to the sctp_addr_list inside the ep
3129 * We will need to allocate one and insert it at the head.
3131 * as well. It will also have to do the embed scope kame
3155 * first find the interface with the bound address need to
3156 * zero out the port to find the address! yuck! can't do
3188 /* allow bindx() to send ASCONF's for binding changes */
3193 /* add this address to the endpoint list */
3209 SCTPDBG(SCTP_DEBUG_PCB1, "Main hash to bind at head:%p, bound port:%d - in tcp_pool=%d\n",
3256 * iterator will try to operate on inp. We need to stop that
3262 * iterator flag. Otherwise we set the iterator to go to the
3273 * Now go through and remove any single reference to our inp that
3282 /* This one points to me is it inp specific? */
3315 * complete then any timer has to be stopped. Then start the actual
3381 * kill timer on the asoc due to it
3396 * to just free since the app may have done,
3398 * it wants the data to get across first.
3437 * there is nothing queued to send,
3513 * If there is a timer running to kill us, forget it, since it may
3514 * have a contest on the INP lock.. which would cause us to die ...
3617 * no need to free the net count, since at this point all
3719 * The following is redundant to the same lines in the
3726 SCTPDBG(SCTP_DEBUG_PCB1, "Adding an address (from:%d) to the peer: ",
3733 * Lie and return ok, we don't want to make the association
3739 * been added... and we don't want to end up removing it
3885 * We set this to 0, the timer code knows that this means its an
4044 * CMT: CUC algo - set find_pseudo_cumack to TRUE (1) at beginning
4105 /* got to have a primary set */
4111 /* No route to current primary adopt new primary */
4144 * We don't allow assoc id to be one of SCTP_FUTURE_ASSOC,
4163 * allocate an association and add it to the endpoint. The caller must be
4164 * careful to add all additional addresses once they are know right away or
4187 * sctp_findassociation_ep_addr(ep, addr's); to make sure the
4210 * If its in the TCP pool, its NOT allowed to create an
4211 * association. The parent listener needs to call
4301 * If you have not performed a bind, then we need to do the
4338 /* now that my_vtag is set, add it to the hash */
4476 /* Try to find a confirmed primary */
4511 * Here we need to remove a remote address. This is quite simple, we
4514 * on that item. Note we do not allow it to be removed if there are
4612 * We only do up to the block where we can place our
4618 /* Need to add a new block to chain */
4674 * expect that the input to this function IS a locked TCB.
4751 /* Now the read queue needs to be cleaned up (only once) */
4774 /* Add an end to wake them */
4818 * When I reach here, no others want to kill the assoc yet.. and I
4820 * lock exchange below to grab all the locks to do the final take
4821 * out. to prevent this we increment the count, which will start a
4845 * connected. We also include the peel'ed off ones to.
4866 * Make it invalid too, that way if its about to run it will abort
4904 * Now restop the timers to be sure this is paranoia at is finest!
4912 /* anything on the wheel needs to be removed */
5097 panic("no net's left alloc'ed, or list points to itself");
5172 * this is the only safe way to do this since
5195 * to the current endpoint (e.g. only v4 or v6 currently bound)
5198 * FIX: if we allow assoc-level bindx(), then this needs to be fixed to use
5214 * only want to read the sctp_flags, which is either bound-all or
5223 * are bound-all you still might need to obey the V4--V6
5224 * flags??? IMO this bound-all stuff needs to be removed!
5287 * Add the address to the endpoint local address list There is nothing to be
5288 * done if we are bound to all addresses
5300 /* You are already bound to all. You have it already */
5349 * when we deleted an ep addr that is the only usable source address to reach
5372 * to be done if we are bound to all addresses
5382 /* You are already bound to all. You have it already */
5419 * to laddr
5442 * Add the address to the TCB local address restricted list.
5453 * Assumes TCB is locked.. and possibly the INP. May need to
5473 /* add to the list */
5531 /* Queue a packet to a processor for the specified core */
5585 /* Bind to our cpu */
5742 * doomed to panic anyways... add later maybe.
5854 * still add the ifdef to make it compile on old versions.
5861 * continue and free/destroy it. While holding the lock, spin, to
5862 * avoid the race condition as sctp_iterator_worker() will have to
5863 * wait to re-acquire the lock.
5987 * grub through the INIT pulling addresses and loading them to the
5990 * with either INIT or INIT-ACK's as long as the m points to the IP
5991 * packet and the offset points to the beginning of the parameters.
6095 /* It belongs to another association? */
6558 * with the call to
6574 /* Now check to see if we need to purge any addresses */
6667 /* failed to get memory for the key */
6693 /* Must be confirmed, so queue to set */
6728 * We choose not to lock anything here. TCB's can't be
6900 /* We must back down to see what the new highest is */
6916 * a p-d-api up. An attacker COULD send enough in to
6917 * initiate the PD-API and then send a bunch of stuff to
6919 * queue.. and then we would not get to them. But in order
6920 * to do this I then have to back-track and un-deliver
6922 * we will NOT look at the delivery queue and leave it to be
6923 * something to consider later. An alternative would be to
6925 * the data.... Or another method might be to keep track of
6933 * Now do we need to find a new
6946 * two things to occur. It may cause us to do extra work in checking
6948 * to SACK every datagram. This may not be a bad thing though since
7090 * Atomically add flags to the sctp_flags of an inp.