diff --git a/lib/libc/sys/send.2 b/lib/libc/sys/send.2 index 6a61426bd7a0..8fa2c64354d0 100644 --- a/lib/libc/sys/send.2 +++ b/lib/libc/sys/send.2 @@ -1,232 +1,232 @@ .\" Copyright (c) 1983, 1991, 1993 .\" The Regents of the University of California. All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. 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. .\" 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. .\" .\" From: @(#)send.2 8.2 (Berkeley) 2/21/94 .\" $FreeBSD$ .\" -.Dd September 13, 2006 +.Dd February 5, 2009 .Dt SEND 2 .Os .Sh NAME .Nm send , .Nm sendto , .Nm sendmsg .Nd send a message from a socket .Sh LIBRARY .Lb libc .Sh SYNOPSIS .In sys/types.h .In sys/socket.h .Ft ssize_t .Fn send "int s" "const void *msg" "size_t len" "int flags" .Ft ssize_t .Fn sendto "int s" "const void *msg" "size_t len" "int flags" "const struct sockaddr *to" "socklen_t tolen" .Ft ssize_t .Fn sendmsg "int s" "const struct msghdr *msg" "int flags" .Sh DESCRIPTION The .Fn send function, and .Fn sendto and .Fn sendmsg system calls are used to transmit a message to another socket. The .Fn send function may be used only when the socket is in a .Em connected state, while .Fn sendto and .Fn sendmsg may be used at any time. .Pp The address of the target is given by .Fa to with .Fa tolen specifying its size. The length of the message is given by .Fa len . If the message is too long to pass atomically through the underlying protocol, the error .Er EMSGSIZE is returned, and the message is not transmitted. .Pp No indication of failure to deliver is implicit in a .Fn send . Locally detected errors are indicated by a return value of -1. .Pp If no messages space is available at the socket to hold the message to be transmitted, then .Fn send normally blocks, unless the socket has been placed in non-blocking I/O mode. The .Xr select 2 system call may be used to determine when it is possible to send more data. .Pp The .Fa flags argument may include one or more of the following: .Bd -literal #define MSG_OOB 0x00001 /* process out-of-band data */ #define MSG_DONTROUTE 0x00004 /* bypass routing, use direct interface */ #define MSG_EOR 0x00008 /* data completes record */ #define MSG_EOF 0x00100 /* data completes transaction */ #define MSG_NOSIGNAL 0x20000 /* do not generate SIGPIPE on EOF */ .Ed .Pp The flag .Dv MSG_OOB is used to send .Dq out-of-band data on sockets that support this notion (e.g.\& .Dv SOCK_STREAM ) ; the underlying protocol must also support .Dq out-of-band data. .Dv MSG_EOR is used to indicate a record mark for protocols which support the concept. .Dv MSG_EOF requests that the sender side of a socket be shut down, and that an appropriate indication be sent at the end of the specified data; this flag is only implemented for .Dv SOCK_STREAM sockets in the .Dv PF_INET protocol family. .Dv MSG_DONTROUTE is usually used only by diagnostic or routing programs. .Dv MSG_NOSIGNAL is used to prevent .Dv SIGPIPE generation when writing a socket that may be closed. .Pp See .Xr recv 2 for a description of the .Fa msghdr structure. .Sh RETURN VALUES The call returns the number of characters sent, or -1 if an error occurred. .Sh ERRORS The .Fn send function and .Fn sendto and .Fn sendmsg system calls fail if: .Bl -tag -width Er .It Bq Er EBADF An invalid descriptor was specified. .It Bq Er EACCES The destination address is a broadcast address, and .Dv SO_BROADCAST has not been set on the socket. .It Bq Er ENOTSOCK The argument .Fa s is not a socket. .It Bq Er EFAULT An invalid user space address was specified for an argument. .It Bq Er EMSGSIZE The socket requires that message be sent atomically, and the size of the message to be sent made this impossible. .It Bq Er EAGAIN The socket is marked non-blocking and the requested operation would block. .It Bq Er ENOBUFS The system was unable to allocate an internal buffer. The operation may succeed when buffers become available. .It Bq Er ENOBUFS The output queue for a network interface was full. This generally indicates that the interface has stopped sending, but may be caused by transient congestion. .It Bq Er EHOSTUNREACH The remote host was unreachable. .It Bq Er EISCONN A destination address was specified and the socket is already connected. .It Bq Er ECONNREFUSED The socket received an ICMP destination unreachable message from the last message sent. This typically means that the receiver is not listening on the remote port. .It Bq Er EHOSTDOWN The remote host was down. .It Bq Er ENETDOWN The remote network was down. -.It Bq Er EPERM +.It Bq Er EADDRNOTAVAIL The process using a .Dv SOCK_RAW socket was jailed and the source address specified in the IP header did not match the IP address bound to the prison. .It Bq Er EPIPE The socket is unable to send anymore data .Dv ( SBS_CANTSENDMORE has been set on the socket). This typically means that the socket is not connected. .El .Sh SEE ALSO .Xr fcntl 2 , .Xr getsockopt 2 , .Xr recv 2 , .Xr select 2 , .Xr socket 2 , .Xr write 2 .Sh HISTORY The .Fn send function appeared in .Bx 4.2 . .Sh BUGS Because .Fn sendmsg does not necessarily block until the data has been transferred, it is possible to transfer an open file descriptor across an .Dv AF_UNIX domain socket (see .Xr recv 2 ) , then .Fn close it before it has actually been sent, the result being that the receiver gets a closed file descriptor. It is left to the application to implement an acknowledgment mechanism to prevent this from happening. diff --git a/sys/kern/kern_jail.c b/sys/kern/kern_jail.c index b87de0b9d715..9b2f6d8a8f85 100644 --- a/sys/kern/kern_jail.c +++ b/sys/kern/kern_jail.c @@ -1,1605 +1,1609 @@ /*- * Copyright (c) 1999 Poul-Henning Kamp. * Copyright (c) 2008 Bjoern A. Zeeb. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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. */ #include __FBSDID("$FreeBSD$"); #include "opt_ddb.h" #include "opt_inet.h" #include "opt_inet6.h" #include "opt_mac.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef DDB #include #ifdef INET6 #include #endif /* INET6 */ #endif /* DDB */ #include MALLOC_DEFINE(M_PRISON, "prison", "Prison structures"); SYSCTL_NODE(_security, OID_AUTO, jail, CTLFLAG_RW, 0, "Jail rules"); int jail_set_hostname_allowed = 1; SYSCTL_INT(_security_jail, OID_AUTO, set_hostname_allowed, CTLFLAG_RW, &jail_set_hostname_allowed, 0, "Processes in jail can set their hostnames"); int jail_socket_unixiproute_only = 1; SYSCTL_INT(_security_jail, OID_AUTO, socket_unixiproute_only, CTLFLAG_RW, &jail_socket_unixiproute_only, 0, "Processes in jail are limited to creating UNIX/IP/route sockets only"); int jail_sysvipc_allowed = 0; SYSCTL_INT(_security_jail, OID_AUTO, sysvipc_allowed, CTLFLAG_RW, &jail_sysvipc_allowed, 0, "Processes in jail can use System V IPC primitives"); static int jail_enforce_statfs = 2; SYSCTL_INT(_security_jail, OID_AUTO, enforce_statfs, CTLFLAG_RW, &jail_enforce_statfs, 0, "Processes in jail cannot see all mounted file systems"); int jail_allow_raw_sockets = 0; SYSCTL_INT(_security_jail, OID_AUTO, allow_raw_sockets, CTLFLAG_RW, &jail_allow_raw_sockets, 0, "Prison root can create raw sockets"); int jail_chflags_allowed = 0; SYSCTL_INT(_security_jail, OID_AUTO, chflags_allowed, CTLFLAG_RW, &jail_chflags_allowed, 0, "Processes in jail can alter system file flags"); int jail_mount_allowed = 0; SYSCTL_INT(_security_jail, OID_AUTO, mount_allowed, CTLFLAG_RW, &jail_mount_allowed, 0, "Processes in jail can mount/unmount jail-friendly file systems"); int jail_max_af_ips = 255; SYSCTL_INT(_security_jail, OID_AUTO, jail_max_af_ips, CTLFLAG_RW, &jail_max_af_ips, 0, "Number of IP addresses a jail may have at most per address family"); /* allprison, lastprid, and prisoncount are protected by allprison_lock. */ struct prisonlist allprison; struct sx allprison_lock; int lastprid = 0; int prisoncount = 0; static void init_prison(void *); static void prison_complete(void *context, int pending); static int sysctl_jail_list(SYSCTL_HANDLER_ARGS); #ifdef INET static int _prison_check_ip4(struct prison *, struct in_addr *); #endif #ifdef INET6 static int _prison_check_ip6(struct prison *, struct in6_addr *); #endif static void init_prison(void *data __unused) { sx_init(&allprison_lock, "allprison"); LIST_INIT(&allprison); } SYSINIT(prison, SI_SUB_INTRINSIC, SI_ORDER_ANY, init_prison, NULL); #ifdef INET static int qcmp_v4(const void *ip1, const void *ip2) { in_addr_t iaa, iab; /* * We need to compare in HBO here to get the list sorted as expected * by the result of the code. Sorting NBO addresses gives you * interesting results. If you do not understand, do not try. */ iaa = ntohl(((const struct in_addr *)ip1)->s_addr); iab = ntohl(((const struct in_addr *)ip2)->s_addr); /* * Do not simply return the difference of the two numbers, the int is * not wide enough. */ if (iaa > iab) return (1); else if (iaa < iab) return (-1); else return (0); } #endif #ifdef INET6 static int qcmp_v6(const void *ip1, const void *ip2) { const struct in6_addr *ia6a, *ia6b; int i, rc; ia6a = (const struct in6_addr *)ip1; ia6b = (const struct in6_addr *)ip2; rc = 0; for (i=0; rc == 0 && i < sizeof(struct in6_addr); i++) { if (ia6a->s6_addr[i] > ia6b->s6_addr[i]) rc = 1; else if (ia6a->s6_addr[i] < ia6b->s6_addr[i]) rc = -1; } return (rc); } #endif #if defined(INET) || defined(INET6) static int prison_check_conflicting_ips(struct prison *p) { struct prison *pr; int i; sx_assert(&allprison_lock, SX_LOCKED); if (p->pr_ip4s == 0 && p->pr_ip6s == 0) return (0); LIST_FOREACH(pr, &allprison, pr_list) { /* * Skip 'dying' prisons to avoid problems when * restarting multi-IP jails. */ if (pr->pr_state == PRISON_STATE_DYING) continue; /* * We permit conflicting IPs if there is no * more than 1 IP on eeach jail. * In case there is one duplicate on a jail with * more than one IP stop checking and return error. */ #ifdef INET if ((p->pr_ip4s >= 1 && pr->pr_ip4s > 1) || (p->pr_ip4s > 1 && pr->pr_ip4s >= 1)) { for (i = 0; i < p->pr_ip4s; i++) { - if (_prison_check_ip4(pr, &p->pr_ip4[i])) + if (_prison_check_ip4(pr, &p->pr_ip4[i]) == 0) return (EINVAL); } } #endif #ifdef INET6 if ((p->pr_ip6s >= 1 && pr->pr_ip6s > 1) || (p->pr_ip6s > 1 && pr->pr_ip6s >= 1)) { for (i = 0; i < p->pr_ip6s; i++) { - if (_prison_check_ip6(pr, &p->pr_ip6[i])) + if (_prison_check_ip6(pr, &p->pr_ip6[i]) == 0) return (EINVAL); } } #endif } return (0); } static int jail_copyin_ips(struct jail *j) { #ifdef INET struct in_addr *ip4; #endif #ifdef INET6 struct in6_addr *ip6; #endif int error, i; /* * Copy in addresses, check for duplicate addresses and do some * simple 0 and broadcast checks. If users give other bogus addresses * it is their problem. * * IP addresses are all sorted but ip[0] to preserve the primary IP * address as given from userland. This special IP is used for * unbound outgoing connections as well for "loopback" traffic. */ #ifdef INET ip4 = NULL; #endif #ifdef INET6 ip6 = NULL; #endif #ifdef INET if (j->ip4s > 0) { ip4 = (struct in_addr *)malloc(j->ip4s * sizeof(struct in_addr), M_PRISON, M_WAITOK | M_ZERO); error = copyin(j->ip4, ip4, j->ip4s * sizeof(struct in_addr)); if (error) goto e_free_ip; /* Sort all but the first IPv4 address. */ if (j->ip4s > 1) qsort((ip4 + 1), j->ip4s - 1, sizeof(struct in_addr), qcmp_v4); /* * We do not have to care about byte order for these checks * so we will do them in NBO. */ for (i=0; iip4s; i++) { if (ip4[i].s_addr == htonl(INADDR_ANY) || ip4[i].s_addr == htonl(INADDR_BROADCAST)) { error = EINVAL; goto e_free_ip; } if ((i+1) < j->ip4s && (ip4[0].s_addr == ip4[i+1].s_addr || ip4[i].s_addr == ip4[i+1].s_addr)) { error = EINVAL; goto e_free_ip; } } j->ip4 = ip4; } else j->ip4 = NULL; #endif #ifdef INET6 if (j->ip6s > 0) { ip6 = (struct in6_addr *)malloc(j->ip6s * sizeof(struct in6_addr), M_PRISON, M_WAITOK | M_ZERO); error = copyin(j->ip6, ip6, j->ip6s * sizeof(struct in6_addr)); if (error) goto e_free_ip; /* Sort all but the first IPv6 address. */ if (j->ip6s > 1) qsort((ip6 + 1), j->ip6s - 1, sizeof(struct in6_addr), qcmp_v6); for (i=0; iip6s; i++) { if (IN6_IS_ADDR_UNSPECIFIED(&ip6[i])) { error = EINVAL; goto e_free_ip; } if ((i+1) < j->ip6s && (IN6_ARE_ADDR_EQUAL(&ip6[0], &ip6[i+1]) || IN6_ARE_ADDR_EQUAL(&ip6[i], &ip6[i+1]))) { error = EINVAL; goto e_free_ip; } } j->ip6 = ip6; } else j->ip6 = NULL; #endif return (0); e_free_ip: #ifdef INET6 free(ip6, M_PRISON); #endif #ifdef INET free(ip4, M_PRISON); #endif return (error); } #endif /* INET || INET6 */ static int jail_handle_ips(struct jail *j) { #if defined(INET) || defined(INET6) int error; #endif /* * Finish conversion for older versions, copyin and setup IPs. */ switch (j->version) { case 0: { #ifdef INET /* FreeBSD single IPv4 jails. */ struct in_addr *ip4; if (j->ip4s == INADDR_ANY || j->ip4s == INADDR_BROADCAST) return (EINVAL); ip4 = (struct in_addr *)malloc(sizeof(struct in_addr), M_PRISON, M_WAITOK | M_ZERO); /* * Jail version 0 still used HBO for the IPv4 address. */ ip4->s_addr = htonl(j->ip4s); j->ip4s = 1; j->ip4 = ip4; break; #else return (EINVAL); #endif } case 1: /* * Version 1 was used by multi-IPv4 jail implementations * that never made it into the official kernel. * We should never hit this here; jail() should catch it. */ return (EINVAL); case 2: /* JAIL_API_VERSION */ /* FreeBSD multi-IPv4/IPv6,noIP jails. */ #if defined(INET) || defined(INET6) #ifdef INET if (j->ip4s > jail_max_af_ips) return (EINVAL); #else if (j->ip4s != 0) return (EINVAL); #endif #ifdef INET6 if (j->ip6s > jail_max_af_ips) return (EINVAL); #else if (j->ip6s != 0) return (EINVAL); #endif error = jail_copyin_ips(j); if (error) return (error); #endif break; default: /* Sci-Fi jails are not supported, sorry. */ return (EINVAL); } return (0); } /* * struct jail_args { * struct jail *jail; * }; */ int jail(struct thread *td, struct jail_args *uap) { uint32_t version; int error; struct jail j; error = copyin(uap->jail, &version, sizeof(uint32_t)); if (error) return (error); switch (version) { case 0: /* FreeBSD single IPv4 jails. */ { struct jail_v0 j0; bzero(&j, sizeof(struct jail)); error = copyin(uap->jail, &j0, sizeof(struct jail_v0)); if (error) return (error); j.version = j0.version; j.path = j0.path; j.hostname = j0.hostname; j.ip4s = j0.ip_number; break; } case 1: /* * Version 1 was used by multi-IPv4 jail implementations * that never made it into the official kernel. */ return (EINVAL); case 2: /* JAIL_API_VERSION */ /* FreeBSD multi-IPv4/IPv6,noIP jails. */ error = copyin(uap->jail, &j, sizeof(struct jail)); if (error) return (error); break; default: /* Sci-Fi jails are not supported, sorry. */ return (EINVAL); } return (kern_jail(td, &j)); } int kern_jail(struct thread *td, struct jail *j) { struct nameidata nd; struct prison *pr, *tpr; struct jail_attach_args jaa; int vfslocked, error, tryprid; KASSERT(j != NULL, ("%s: j is NULL", __func__)); /* Handle addresses - convert old structs, copyin, check IPs. */ error = jail_handle_ips(j); if (error) return (error); /* Allocate struct prison and fill it with life. */ pr = malloc(sizeof(*pr), M_PRISON, M_WAITOK | M_ZERO); mtx_init(&pr->pr_mtx, "jail mutex", NULL, MTX_DEF); pr->pr_ref = 1; error = copyinstr(j->path, &pr->pr_path, sizeof(pr->pr_path), NULL); if (error) goto e_killmtx; NDINIT(&nd, LOOKUP, MPSAFE | FOLLOW | LOCKLEAF, UIO_SYSSPACE, pr->pr_path, td); error = namei(&nd); if (error) goto e_killmtx; vfslocked = NDHASGIANT(&nd); pr->pr_root = nd.ni_vp; VOP_UNLOCK(nd.ni_vp, 0); NDFREE(&nd, NDF_ONLY_PNBUF); VFS_UNLOCK_GIANT(vfslocked); error = copyinstr(j->hostname, &pr->pr_host, sizeof(pr->pr_host), NULL); if (error) goto e_dropvnref; if (j->jailname != NULL) { error = copyinstr(j->jailname, &pr->pr_name, sizeof(pr->pr_name), NULL); if (error) goto e_dropvnref; } if (j->ip4s > 0) { pr->pr_ip4 = j->ip4; pr->pr_ip4s = j->ip4s; } #ifdef INET6 if (j->ip6s > 0) { pr->pr_ip6 = j->ip6; pr->pr_ip6s = j->ip6s; } #endif pr->pr_linux = NULL; pr->pr_securelevel = securelevel; bzero(&pr->pr_osd, sizeof(pr->pr_osd)); /* * Pre-set prison state to ALIVE upon cration. This is needed so we * can later attach the process to it, etc (avoiding another extra * state for ther process of creation, complicating things). */ pr->pr_state = PRISON_STATE_ALIVE; /* Allocate a dedicated cpuset for each jail. */ error = cpuset_create_root(td, &pr->pr_cpuset); if (error) goto e_dropvnref; sx_xlock(&allprison_lock); /* Make sure we cannot run into problems with ambiguous bind()ings. */ #if defined(INET) || defined(INET6) error = prison_check_conflicting_ips(pr); if (error) { sx_xunlock(&allprison_lock); goto e_dropcpuset; } #endif /* Determine next pr_id and add prison to allprison list. */ tryprid = lastprid + 1; if (tryprid == JAIL_MAX) tryprid = 1; next: LIST_FOREACH(tpr, &allprison, pr_list) { if (tpr->pr_id == tryprid) { tryprid++; if (tryprid == JAIL_MAX) { sx_xunlock(&allprison_lock); error = EAGAIN; goto e_dropcpuset; } goto next; } } pr->pr_id = jaa.jid = lastprid = tryprid; LIST_INSERT_HEAD(&allprison, pr, pr_list); prisoncount++; sx_xunlock(&allprison_lock); error = jail_attach(td, &jaa); if (error) goto e_dropprref; mtx_lock(&pr->pr_mtx); pr->pr_ref--; mtx_unlock(&pr->pr_mtx); td->td_retval[0] = jaa.jid; return (0); e_dropprref: sx_xlock(&allprison_lock); LIST_REMOVE(pr, pr_list); prisoncount--; sx_xunlock(&allprison_lock); e_dropcpuset: cpuset_rel(pr->pr_cpuset); e_dropvnref: vfslocked = VFS_LOCK_GIANT(pr->pr_root->v_mount); vrele(pr->pr_root); VFS_UNLOCK_GIANT(vfslocked); e_killmtx: mtx_destroy(&pr->pr_mtx); free(pr, M_PRISON); #ifdef INET6 free(j->ip6, M_PRISON); #endif #ifdef INET free(j->ip4, M_PRISON); #endif return (error); } /* * struct jail_attach_args { * int jid; * }; */ int jail_attach(struct thread *td, struct jail_attach_args *uap) { struct proc *p; struct ucred *newcred, *oldcred; struct prison *pr; int vfslocked, error; /* * XXX: Note that there is a slight race here if two threads * in the same privileged process attempt to attach to two * different jails at the same time. It is important for * user processes not to do this, or they might end up with * a process root from one prison, but attached to the jail * of another. */ error = priv_check(td, PRIV_JAIL_ATTACH); if (error) return (error); p = td->td_proc; sx_slock(&allprison_lock); pr = prison_find(uap->jid); if (pr == NULL) { sx_sunlock(&allprison_lock); return (EINVAL); } /* * Do not allow a process to attach to a prison that is not * considered to be "ALIVE". */ if (pr->pr_state != PRISON_STATE_ALIVE) { mtx_unlock(&pr->pr_mtx); sx_sunlock(&allprison_lock); return (EINVAL); } pr->pr_ref++; mtx_unlock(&pr->pr_mtx); sx_sunlock(&allprison_lock); /* * Reparent the newly attached process to this jail. */ error = cpuset_setproc_update_set(p, pr->pr_cpuset); if (error) goto e_unref; vfslocked = VFS_LOCK_GIANT(pr->pr_root->v_mount); vn_lock(pr->pr_root, LK_EXCLUSIVE | LK_RETRY); if ((error = change_dir(pr->pr_root, td)) != 0) goto e_unlock; #ifdef MAC if ((error = mac_vnode_check_chroot(td->td_ucred, pr->pr_root))) goto e_unlock; #endif VOP_UNLOCK(pr->pr_root, 0); change_root(pr->pr_root, td); VFS_UNLOCK_GIANT(vfslocked); newcred = crget(); PROC_LOCK(p); oldcred = p->p_ucred; setsugid(p); crcopy(newcred, oldcred); newcred->cr_prison = pr; p->p_ucred = newcred; prison_proc_hold(pr); PROC_UNLOCK(p); crfree(oldcred); return (0); e_unlock: VOP_UNLOCK(pr->pr_root, 0); VFS_UNLOCK_GIANT(vfslocked); e_unref: mtx_lock(&pr->pr_mtx); pr->pr_ref--; mtx_unlock(&pr->pr_mtx); return (error); } /* * Returns a locked prison instance, or NULL on failure. */ struct prison * prison_find(int prid) { struct prison *pr; sx_assert(&allprison_lock, SX_LOCKED); LIST_FOREACH(pr, &allprison, pr_list) { if (pr->pr_id == prid) { mtx_lock(&pr->pr_mtx); if (pr->pr_ref == 0) { mtx_unlock(&pr->pr_mtx); break; } return (pr); } } return (NULL); } void prison_free_locked(struct prison *pr) { mtx_assert(&pr->pr_mtx, MA_OWNED); pr->pr_ref--; if (pr->pr_ref == 0) { mtx_unlock(&pr->pr_mtx); TASK_INIT(&pr->pr_task, 0, prison_complete, pr); taskqueue_enqueue(taskqueue_thread, &pr->pr_task); return; } mtx_unlock(&pr->pr_mtx); } void prison_free(struct prison *pr) { mtx_lock(&pr->pr_mtx); prison_free_locked(pr); } static void prison_complete(void *context, int pending) { struct prison *pr; int vfslocked; pr = (struct prison *)context; sx_xlock(&allprison_lock); LIST_REMOVE(pr, pr_list); prisoncount--; sx_xunlock(&allprison_lock); cpuset_rel(pr->pr_cpuset); /* Free all OSD associated to this jail. */ osd_jail_exit(pr); vfslocked = VFS_LOCK_GIANT(pr->pr_root->v_mount); vrele(pr->pr_root); VFS_UNLOCK_GIANT(vfslocked); mtx_destroy(&pr->pr_mtx); free(pr->pr_linux, M_PRISON); #ifdef INET6 free(pr->pr_ip6, M_PRISON); #endif #ifdef INET free(pr->pr_ip4, M_PRISON); #endif free(pr, M_PRISON); } void prison_hold_locked(struct prison *pr) { mtx_assert(&pr->pr_mtx, MA_OWNED); KASSERT(pr->pr_ref > 0, ("Trying to hold dead prison (id=%d).", pr->pr_id)); pr->pr_ref++; } void prison_hold(struct prison *pr) { mtx_lock(&pr->pr_mtx); prison_hold_locked(pr); mtx_unlock(&pr->pr_mtx); } void prison_proc_hold(struct prison *pr) { mtx_lock(&pr->pr_mtx); KASSERT(pr->pr_state == PRISON_STATE_ALIVE, ("Cannot add a process to a non-alive prison (id=%d).", pr->pr_id)); pr->pr_nprocs++; mtx_unlock(&pr->pr_mtx); } void prison_proc_free(struct prison *pr) { mtx_lock(&pr->pr_mtx); KASSERT(pr->pr_state == PRISON_STATE_ALIVE && pr->pr_nprocs > 0, ("Trying to kill a process in a dead prison (id=%d).", pr->pr_id)); pr->pr_nprocs--; if (pr->pr_nprocs == 0) pr->pr_state = PRISON_STATE_DYING; mtx_unlock(&pr->pr_mtx); } #ifdef INET /* * Pass back primary IPv4 address of this jail. * * If not jailed return success but do not alter the address. Caller has to - * make sure to intialize it correctly (INADDR_ANY). + * make sure to intialize it correctly (e.g. INADDR_ANY). * - * Returns 0 on success, 1 on error. Address returned in NBO. + * Returns 0 on success, EAFNOSUPPORT if the jail doesn't allow IPv4. + * Address returned in NBO. */ int prison_get_ip4(struct ucred *cred, struct in_addr *ia) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia != NULL, ("%s: ia is NULL", __func__)); if (!jailed(cred)) /* Do not change address passed in. */ return (0); if (cred->cr_prison->pr_ip4 == NULL) - return (1); + return (EAFNOSUPPORT); ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr; return (0); } /* * Make sure our (source) address is set to something meaningful to this * jail. * - * Returns 0 on success, 1 on error. Address passed in in NBO and returned - * in NBO. + * Returns 0 if not jailed or if address belongs to jail, EADDRNOTAVAIL if + * the address doesn't belong, or EAFNOSUPPORT if the jail doesn't allow IPv4. + * Address passed in in NBO and returned in NBO. */ int prison_local_ip4(struct ucred *cred, struct in_addr *ia) { struct in_addr ia0; KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia != NULL, ("%s: ia is NULL", __func__)); if (!jailed(cred)) return (0); if (cred->cr_prison->pr_ip4 == NULL) - return (1); + return (EAFNOSUPPORT); ia0.s_addr = ntohl(ia->s_addr); if (ia0.s_addr == INADDR_LOOPBACK) { ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr; return (0); } - /* - * In case there is only 1 IPv4 address, bind directly. - */ - if (ia0.s_addr == INADDR_ANY && cred->cr_prison->pr_ip4s == 1) { - ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr; + if (ia0.s_addr == INADDR_ANY) { + /* + * In case there is only 1 IPv4 address, bind directly. + */ + if (cred->cr_prison->pr_ip4s == 1) + ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr; return (0); } - if (ia0.s_addr == INADDR_ANY || prison_check_ip4(cred, ia)) - return (0); - - return (1); + return (_prison_check_ip4(cred->cr_prison, ia)); } /* * Rewrite destination address in case we will connect to loopback address. * - * Returns 0 on success, 1 on error. Address passed in in NBO and returned - * in NBO. + * Returns 0 on success, EAFNOSUPPORT if the jail doesn't allow IPv4. + * Address passed in in NBO and returned in NBO. */ int prison_remote_ip4(struct ucred *cred, struct in_addr *ia) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia != NULL, ("%s: ia is NULL", __func__)); if (!jailed(cred)) return (0); if (cred->cr_prison->pr_ip4 == NULL) - return (1); + return (EAFNOSUPPORT); + if (ntohl(ia->s_addr) == INADDR_LOOPBACK) { ia->s_addr = cred->cr_prison->pr_ip4[0].s_addr; return (0); } /* * Return success because nothing had to be changed. */ return (0); } /* - * Check if given address belongs to the jail referenced by cred. + * Check if given address belongs to the jail referenced by cred/prison. * - * Returns 1 if address belongs to jail, 0 if not. Address passed in in NBO. + * Returns 0 if not jailed or if address belongs to jail, EADDRNOTAVAIL if + * the address doesn't belong, or EAFNOSUPPORT if the jail doesn't allow IPv4. + * Address passed in in NBO. */ static int _prison_check_ip4(struct prison *pr, struct in_addr *ia) { int i, a, z, d; - if (pr->pr_ip4 == NULL) - return (0); - /* * Check the primary IP. */ if (pr->pr_ip4[0].s_addr == ia->s_addr) - return (1); + return (0); /* * All the other IPs are sorted so we can do a binary search. */ a = 0; z = pr->pr_ip4s - 2; while (a <= z) { i = (a + z) / 2; d = qcmp_v4(&pr->pr_ip4[i+1], ia); if (d > 0) z = i - 1; else if (d < 0) a = i + 1; else - return (1); + return (0); } - return (0); + + return (EADDRNOTAVAIL); } int prison_check_ip4(struct ucred *cred, struct in_addr *ia) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia != NULL, ("%s: ia is NULL", __func__)); if (!jailed(cred)) - return (1); + return (0); + if (cred->cr_prison->pr_ip4 == NULL) + return (EAFNOSUPPORT); return (_prison_check_ip4(cred->cr_prison, ia)); } #endif #ifdef INET6 /* * Pass back primary IPv6 address for this jail. * * If not jailed return success but do not alter the address. Caller has to - * make sure to intialize it correctly (IN6ADDR_ANY_INIT). + * make sure to intialize it correctly (e.g. IN6ADDR_ANY_INIT). * - * Returns 0 on success, 1 on error. + * Returns 0 on success, EAFNOSUPPORT if the jail doesn't allow IPv6. */ int prison_get_ip6(struct ucred *cred, struct in6_addr *ia6) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__)); if (!jailed(cred)) return (0); if (cred->cr_prison->pr_ip6 == NULL) - return (1); + return (EAFNOSUPPORT); + bcopy(&cred->cr_prison->pr_ip6[0], ia6, sizeof(struct in6_addr)); return (0); } /* * Make sure our (source) address is set to something meaningful to this jail. * * v6only should be set based on (inp->inp_flags & IN6P_IPV6_V6ONLY != 0) * when needed while binding. * - * Returns 0 on success, 1 on error. + * Returns 0 if not jailed or if address belongs to jail, EADDRNOTAVAIL if + * the address doesn't belong, or EAFNOSUPPORT if the jail doesn't allow IPv6. */ int prison_local_ip6(struct ucred *cred, struct in6_addr *ia6, int v6only) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__)); if (!jailed(cred)) return (0); if (cred->cr_prison->pr_ip6 == NULL) - return (1); + return (EAFNOSUPPORT); + if (IN6_IS_ADDR_LOOPBACK(ia6)) { bcopy(&cred->cr_prison->pr_ip6[0], ia6, sizeof(struct in6_addr)); return (0); } - /* - * In case there is only 1 IPv6 address, and v6only is true, then - * bind directly. - */ - if (v6only != 0 && IN6_IS_ADDR_UNSPECIFIED(ia6) && - cred->cr_prison->pr_ip6s == 1) { - bcopy(&cred->cr_prison->pr_ip6[0], ia6, - sizeof(struct in6_addr)); + if (IN6_IS_ADDR_UNSPECIFIED(ia6)) { + /* + * In case there is only 1 IPv6 address, and v6only is true, + * then bind directly. + */ + if (v6only != 0 && cred->cr_prison->pr_ip6s == 1) + bcopy(&cred->cr_prison->pr_ip6[0], ia6, + sizeof(struct in6_addr)); return (0); } - if (IN6_IS_ADDR_UNSPECIFIED(ia6) || prison_check_ip6(cred, ia6)) - return (0); - return (1); + + return (_prison_check_ip6(cred->cr_prison, ia6)); } /* * Rewrite destination address in case we will connect to loopback address. * - * Returns 0 on success, 1 on error. + * Returns 0 on success, EAFNOSUPPORT if the jail doesn't allow IPv6. */ int prison_remote_ip6(struct ucred *cred, struct in6_addr *ia6) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__)); if (!jailed(cred)) return (0); if (cred->cr_prison->pr_ip6 == NULL) - return (1); + return (EAFNOSUPPORT); + if (IN6_IS_ADDR_LOOPBACK(ia6)) { bcopy(&cred->cr_prison->pr_ip6[0], ia6, sizeof(struct in6_addr)); return (0); } /* * Return success because nothing had to be changed. */ return (0); } /* - * Check if given address belongs to the jail referenced by cred. + * Check if given address belongs to the jail referenced by cred/prison. * - * Returns 1 if address belongs to jail, 0 if not. + * Returns 0 if not jailed or if address belongs to jail, EADDRNOTAVAIL if + * the address doesn't belong, or EAFNOSUPPORT if the jail doesn't allow IPv6. */ static int _prison_check_ip6(struct prison *pr, struct in6_addr *ia6) { int i, a, z, d; - if (pr->pr_ip6 == NULL) - return (0); - /* * Check the primary IP. */ if (IN6_ARE_ADDR_EQUAL(&pr->pr_ip6[0], ia6)) - return (1); + return (0); /* * All the other IPs are sorted so we can do a binary search. */ a = 0; z = pr->pr_ip6s - 2; while (a <= z) { i = (a + z) / 2; d = qcmp_v6(&pr->pr_ip6[i+1], ia6); if (d > 0) z = i - 1; else if (d < 0) a = i + 1; else - return (1); + return (0); } - return (0); + + return (EADDRNOTAVAIL); } int prison_check_ip6(struct ucred *cred, struct in6_addr *ia6) { KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(ia6 != NULL, ("%s: ia6 is NULL", __func__)); if (!jailed(cred)) - return (1); + return (0); + if (cred->cr_prison->pr_ip6 == NULL) + return (EAFNOSUPPORT); return (_prison_check_ip6(cred->cr_prison, ia6)); } #endif /* * Check if given address belongs to the jail referenced by cred (wrapper to * prison_check_ip[46]). * - * Returns 1 if address belongs to jail, 0 if not. IPv4 Address passed in in - * NBO. + * Returns 0 if not jailed or if address belongs to jail, EADDRNOTAVAIL if + * the address doesn't belong, or EAFNOSUPPORT if the jail doesn't allow + * the address family. IPv4 Address passed in in NBO. */ int prison_if(struct ucred *cred, struct sockaddr *sa) { #ifdef INET struct sockaddr_in *sai; #endif #ifdef INET6 struct sockaddr_in6 *sai6; #endif - int ok; + int error; KASSERT(cred != NULL, ("%s: cred is NULL", __func__)); KASSERT(sa != NULL, ("%s: sa is NULL", __func__)); - ok = 0; - switch(sa->sa_family) + error = 0; + switch (sa->sa_family) { #ifdef INET case AF_INET: sai = (struct sockaddr_in *)sa; - if (prison_check_ip4(cred, &sai->sin_addr)) - ok = 1; + error = prison_check_ip4(cred, &sai->sin_addr); break; - #endif #ifdef INET6 case AF_INET6: sai6 = (struct sockaddr_in6 *)sa; - if (prison_check_ip6(cred, (struct in6_addr *)&sai6->sin6_addr)) - ok = 1; + error = prison_check_ip6(cred, &sai6->sin6_addr); break; - #endif default: - if (!jail_socket_unixiproute_only) - ok = 1; + if (jailed(cred) && jail_socket_unixiproute_only) + error = EAFNOSUPPORT; } - return (ok); + return (error); } /* * Return 0 if jails permit p1 to frob p2, otherwise ESRCH. */ int prison_check(struct ucred *cred1, struct ucred *cred2) { if (jailed(cred1)) { if (!jailed(cred2)) return (ESRCH); if (cred2->cr_prison != cred1->cr_prison) return (ESRCH); } return (0); } /* * Return 1 if the passed credential is in a jail, otherwise 0. */ int jailed(struct ucred *cred) { return (cred->cr_prison != NULL); } /* * Return the correct hostname for the passed credential. */ void getcredhostname(struct ucred *cred, char *buf, size_t size) { INIT_VPROCG(cred->cr_vimage->v_procg); if (jailed(cred)) { mtx_lock(&cred->cr_prison->pr_mtx); strlcpy(buf, cred->cr_prison->pr_host, size); mtx_unlock(&cred->cr_prison->pr_mtx); } else { mtx_lock(&hostname_mtx); strlcpy(buf, V_hostname, size); mtx_unlock(&hostname_mtx); } } /* * Determine whether the subject represented by cred can "see" * status of a mount point. * Returns: 0 for permitted, ENOENT otherwise. * XXX: This function should be called cr_canseemount() and should be * placed in kern_prot.c. */ int prison_canseemount(struct ucred *cred, struct mount *mp) { struct prison *pr; struct statfs *sp; size_t len; if (!jailed(cred) || jail_enforce_statfs == 0) return (0); pr = cred->cr_prison; if (pr->pr_root->v_mount == mp) return (0); if (jail_enforce_statfs == 2) return (ENOENT); /* * If jail's chroot directory is set to "/" we should be able to see * all mount-points from inside a jail. * This is ugly check, but this is the only situation when jail's * directory ends with '/'. */ if (strcmp(pr->pr_path, "/") == 0) return (0); len = strlen(pr->pr_path); sp = &mp->mnt_stat; if (strncmp(pr->pr_path, sp->f_mntonname, len) != 0) return (ENOENT); /* * Be sure that we don't have situation where jail's root directory * is "/some/path" and mount point is "/some/pathpath". */ if (sp->f_mntonname[len] != '\0' && sp->f_mntonname[len] != '/') return (ENOENT); return (0); } void prison_enforce_statfs(struct ucred *cred, struct mount *mp, struct statfs *sp) { char jpath[MAXPATHLEN]; struct prison *pr; size_t len; if (!jailed(cred) || jail_enforce_statfs == 0) return; pr = cred->cr_prison; if (prison_canseemount(cred, mp) != 0) { bzero(sp->f_mntonname, sizeof(sp->f_mntonname)); strlcpy(sp->f_mntonname, "[restricted]", sizeof(sp->f_mntonname)); return; } if (pr->pr_root->v_mount == mp) { /* * Clear current buffer data, so we are sure nothing from * the valid path left there. */ bzero(sp->f_mntonname, sizeof(sp->f_mntonname)); *sp->f_mntonname = '/'; return; } /* * If jail's chroot directory is set to "/" we should be able to see * all mount-points from inside a jail. */ if (strcmp(pr->pr_path, "/") == 0) return; len = strlen(pr->pr_path); strlcpy(jpath, sp->f_mntonname + len, sizeof(jpath)); /* * Clear current buffer data, so we are sure nothing from * the valid path left there. */ bzero(sp->f_mntonname, sizeof(sp->f_mntonname)); if (*jpath == '\0') { /* Should never happen. */ *sp->f_mntonname = '/'; } else { strlcpy(sp->f_mntonname, jpath, sizeof(sp->f_mntonname)); } } /* * Check with permission for a specific privilege is granted within jail. We * have a specific list of accepted privileges; the rest are denied. */ int prison_priv_check(struct ucred *cred, int priv) { if (!jailed(cred)) return (0); switch (priv) { /* * Allow ktrace privileges for root in jail. */ case PRIV_KTRACE: #if 0 /* * Allow jailed processes to configure audit identity and * submit audit records (login, etc). In the future we may * want to further refine the relationship between audit and * jail. */ case PRIV_AUDIT_GETAUDIT: case PRIV_AUDIT_SETAUDIT: case PRIV_AUDIT_SUBMIT: #endif /* * Allow jailed processes to manipulate process UNIX * credentials in any way they see fit. */ case PRIV_CRED_SETUID: case PRIV_CRED_SETEUID: case PRIV_CRED_SETGID: case PRIV_CRED_SETEGID: case PRIV_CRED_SETGROUPS: case PRIV_CRED_SETREUID: case PRIV_CRED_SETREGID: case PRIV_CRED_SETRESUID: case PRIV_CRED_SETRESGID: /* * Jail implements visibility constraints already, so allow * jailed root to override uid/gid-based constraints. */ case PRIV_SEEOTHERGIDS: case PRIV_SEEOTHERUIDS: /* * Jail implements inter-process debugging limits already, so * allow jailed root various debugging privileges. */ case PRIV_DEBUG_DIFFCRED: case PRIV_DEBUG_SUGID: case PRIV_DEBUG_UNPRIV: /* * Allow jail to set various resource limits and login * properties, and for now, exceed process resource limits. */ case PRIV_PROC_LIMIT: case PRIV_PROC_SETLOGIN: case PRIV_PROC_SETRLIMIT: /* * System V and POSIX IPC privileges are granted in jail. */ case PRIV_IPC_READ: case PRIV_IPC_WRITE: case PRIV_IPC_ADMIN: case PRIV_IPC_MSGSIZE: case PRIV_MQ_ADMIN: /* * Jail implements its own inter-process limits, so allow * root processes in jail to change scheduling on other * processes in the same jail. Likewise for signalling. */ case PRIV_SCHED_DIFFCRED: case PRIV_SCHED_CPUSET: case PRIV_SIGNAL_DIFFCRED: case PRIV_SIGNAL_SUGID: /* * Allow jailed processes to write to sysctls marked as jail * writable. */ case PRIV_SYSCTL_WRITEJAIL: /* * Allow root in jail to manage a variety of quota * properties. These should likely be conditional on a * configuration option. */ case PRIV_VFS_GETQUOTA: case PRIV_VFS_SETQUOTA: /* * Since Jail relies on chroot() to implement file system * protections, grant many VFS privileges to root in jail. * Be careful to exclude mount-related and NFS-related * privileges. */ case PRIV_VFS_READ: case PRIV_VFS_WRITE: case PRIV_VFS_ADMIN: case PRIV_VFS_EXEC: case PRIV_VFS_LOOKUP: case PRIV_VFS_BLOCKRESERVE: /* XXXRW: Slightly surprising. */ case PRIV_VFS_CHFLAGS_DEV: case PRIV_VFS_CHOWN: case PRIV_VFS_CHROOT: case PRIV_VFS_RETAINSUGID: case PRIV_VFS_FCHROOT: case PRIV_VFS_LINK: case PRIV_VFS_SETGID: case PRIV_VFS_STAT: case PRIV_VFS_STICKYFILE: return (0); /* * Depending on the global setting, allow privilege of * setting system flags. */ case PRIV_VFS_SYSFLAGS: if (jail_chflags_allowed) return (0); else return (EPERM); /* * Depending on the global setting, allow privilege of * mounting/unmounting file systems. */ case PRIV_VFS_MOUNT: case PRIV_VFS_UNMOUNT: case PRIV_VFS_MOUNT_NONUSER: case PRIV_VFS_MOUNT_OWNER: if (jail_mount_allowed) return (0); else return (EPERM); /* * Allow jailed root to bind reserved ports and reuse in-use * ports. */ case PRIV_NETINET_RESERVEDPORT: case PRIV_NETINET_REUSEPORT: return (0); /* * Allow jailed root to set certian IPv4/6 (option) headers. */ case PRIV_NETINET_SETHDROPTS: return (0); /* * Conditionally allow creating raw sockets in jail. */ case PRIV_NETINET_RAW: if (jail_allow_raw_sockets) return (0); else return (EPERM); /* * Since jail implements its own visibility limits on netstat * sysctls, allow getcred. This allows identd to work in * jail. */ case PRIV_NETINET_GETCRED: return (0); default: /* * In all remaining cases, deny the privilege request. This * includes almost all network privileges, many system * configuration privileges. */ return (EPERM); } } static int sysctl_jail_list(SYSCTL_HANDLER_ARGS) { struct xprison *xp, *sxp; struct prison *pr; char *p; size_t len; int count, error; if (jailed(req->td->td_ucred)) return (0); sx_slock(&allprison_lock); if ((count = prisoncount) == 0) { sx_sunlock(&allprison_lock); return (0); } len = sizeof(*xp) * count; LIST_FOREACH(pr, &allprison, pr_list) { #ifdef INET len += pr->pr_ip4s * sizeof(struct in_addr); #endif #ifdef INET6 len += pr->pr_ip6s * sizeof(struct in6_addr); #endif } sxp = xp = malloc(len, M_TEMP, M_WAITOK | M_ZERO); LIST_FOREACH(pr, &allprison, pr_list) { xp->pr_version = XPRISON_VERSION; xp->pr_id = pr->pr_id; xp->pr_state = pr->pr_state; xp->pr_cpusetid = pr->pr_cpuset->cs_id; strlcpy(xp->pr_path, pr->pr_path, sizeof(xp->pr_path)); mtx_lock(&pr->pr_mtx); strlcpy(xp->pr_host, pr->pr_host, sizeof(xp->pr_host)); strlcpy(xp->pr_name, pr->pr_name, sizeof(xp->pr_name)); mtx_unlock(&pr->pr_mtx); #ifdef INET xp->pr_ip4s = pr->pr_ip4s; #endif #ifdef INET6 xp->pr_ip6s = pr->pr_ip6s; #endif p = (char *)(xp + 1); #ifdef INET if (pr->pr_ip4s > 0) { bcopy(pr->pr_ip4, (struct in_addr *)p, pr->pr_ip4s * sizeof(struct in_addr)); p += (pr->pr_ip4s * sizeof(struct in_addr)); } #endif #ifdef INET6 if (pr->pr_ip6s > 0) { bcopy(pr->pr_ip6, (struct in6_addr *)p, pr->pr_ip6s * sizeof(struct in6_addr)); p += (pr->pr_ip6s * sizeof(struct in6_addr)); } #endif xp = (struct xprison *)p; } sx_sunlock(&allprison_lock); error = SYSCTL_OUT(req, sxp, len); free(sxp, M_TEMP); return (error); } SYSCTL_OID(_security_jail, OID_AUTO, list, CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, 0, sysctl_jail_list, "S", "List of active jails"); static int sysctl_jail_jailed(SYSCTL_HANDLER_ARGS) { int error, injail; injail = jailed(req->td->td_ucred); error = SYSCTL_OUT(req, &injail, sizeof(injail)); return (error); } SYSCTL_PROC(_security_jail, OID_AUTO, jailed, CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, 0, sysctl_jail_jailed, "I", "Process in jail?"); #ifdef DDB DB_SHOW_COMMAND(jails, db_show_jails) { struct prison *pr; #ifdef INET struct in_addr ia; #endif #ifdef INET6 char ip6buf[INET6_ADDRSTRLEN]; #endif const char *state; #if defined(INET) || defined(INET6) int i; #endif db_printf( " JID pr_ref pr_nprocs pr_ip4s pr_ip6s\n"); db_printf( " Hostname Path\n"); db_printf( " Name State\n"); db_printf( " Cpusetid\n"); db_printf( " IP Address(es)\n"); LIST_FOREACH(pr, &allprison, pr_list) { db_printf("%6d %6d %9d %7d %7d\n", pr->pr_id, pr->pr_ref, pr->pr_nprocs, pr->pr_ip4s, pr->pr_ip6s); db_printf("%6s %-29.29s %.74s\n", "", pr->pr_host, pr->pr_path); if (pr->pr_state < 0 || pr->pr_state >= (int)((sizeof( prison_states) / sizeof(struct prison_state)))) state = "(bogus)"; else state = prison_states[pr->pr_state].state_name; db_printf("%6s %-29.29s %.74s\n", "", (pr->pr_name[0] != '\0') ? pr->pr_name : "", state); db_printf("%6s %-6d\n", "", pr->pr_cpuset->cs_id); #ifdef INET for (i=0; i < pr->pr_ip4s; i++) { ia.s_addr = pr->pr_ip4[i].s_addr; db_printf("%6s %s\n", "", inet_ntoa(ia)); } #endif #ifdef INET6 for (i=0; i < pr->pr_ip6s; i++) db_printf("%6s %s\n", "", ip6_sprintf(ip6buf, &pr->pr_ip6[i])); #endif /* INET6 */ if (db_pager_quit) break; } } #endif /* DDB */ diff --git a/sys/net/if.c b/sys/net/if.c index 2fb8e236fa72..6966a51462b2 100644 --- a/sys/net/if.c +++ b/sys/net/if.c @@ -1,2939 +1,2938 @@ /*- * Copyright (c) 1980, 1986, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)if.c 8.5 (Berkeley) 1/9/95 * $FreeBSD$ */ #include "opt_compat.h" #include "opt_inet6.h" #include "opt_inet.h" #include "opt_mac.h" #include "opt_carp.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #if defined(INET) || defined(INET6) /*XXX*/ #include #include #ifdef INET6 #include #include #endif #endif #ifdef INET #include #include #endif #ifdef DEV_CARP #include #endif #include #ifndef VIMAGE #ifndef VIMAGE_GLOBALS struct vnet_net vnet_net_0; #endif #endif static int slowtimo_started; SYSCTL_NODE(_net, PF_LINK, link, CTLFLAG_RW, 0, "Link layers"); SYSCTL_NODE(_net_link, 0, generic, CTLFLAG_RW, 0, "Generic link-management"); /* Log link state change events */ static int log_link_state_change = 1; SYSCTL_INT(_net_link, OID_AUTO, log_link_state_change, CTLFLAG_RW, &log_link_state_change, 0, "log interface link state change events"); void (*bstp_linkstate_p)(struct ifnet *ifp, int state); void (*ng_ether_link_state_p)(struct ifnet *ifp, int state); void (*lagg_linkstate_p)(struct ifnet *ifp, int state); struct mbuf *(*tbr_dequeue_ptr)(struct ifaltq *, int) = NULL; /* * XXX: Style; these should be sorted alphabetically, and unprototyped * static functions should be prototyped. Currently they are sorted by * declaration order. */ static void if_attachdomain(void *); static void if_attachdomain1(struct ifnet *); static int ifconf(u_long, caddr_t); static void if_freemulti(struct ifmultiaddr *); static void if_grow(void); static void if_init(void *); static void if_check(void *); static void if_qflush(struct ifnet *); static void if_route(struct ifnet *, int flag, int fam); static int if_setflag(struct ifnet *, int, int, int *, int); static void if_slowtimo(void *); static int if_transmit(struct ifnet *ifp, struct mbuf *m); static void if_unroute(struct ifnet *, int flag, int fam); static void link_rtrequest(int, struct rtentry *, struct rt_addrinfo *); static int if_rtdel(struct radix_node *, void *); static int ifhwioctl(u_long, struct ifnet *, caddr_t, struct thread *); static int if_delmulti_locked(struct ifnet *, struct ifmultiaddr *, int); static void if_start_deferred(void *context, int pending); static void do_link_state_change(void *, int); static int if_getgroup(struct ifgroupreq *, struct ifnet *); static int if_getgroupmembers(struct ifgroupreq *); #ifdef INET6 /* * XXX: declare here to avoid to include many inet6 related files.. * should be more generalized? */ extern void nd6_setmtu(struct ifnet *); #endif #ifdef VIMAGE_GLOBALS struct ifnethead ifnet; /* depend on static init XXX */ struct ifgrouphead ifg_head; int if_index; static int if_indexlim; /* Table of ifnet/cdev by index. Locked with ifnet_lock. */ static struct ifindex_entry *ifindex_table; static struct knlist ifklist; #endif int ifqmaxlen = IFQ_MAXLEN; struct rwlock ifnet_lock; static if_com_alloc_t *if_com_alloc[256]; static if_com_free_t *if_com_free[256]; static void filt_netdetach(struct knote *kn); static int filt_netdev(struct knote *kn, long hint); static struct filterops netdev_filtops = { 1, NULL, filt_netdetach, filt_netdev }; #ifndef VIMAGE_GLOBALS static struct vnet_symmap vnet_net_symmap[] = { VNET_SYMMAP(net, ifnet), VNET_SYMMAP(net, rt_tables), VNET_SYMMAP(net, rtstat), VNET_SYMMAP(net, rttrash), VNET_SYMMAP_END }; VNET_MOD_DECLARE(NET, net, vnet_net_iattach, vnet_net_idetach, NONE, vnet_net_symmap) #endif /* * System initialization */ SYSINIT(interfaces, SI_SUB_INIT_IF, SI_ORDER_FIRST, if_init, NULL); SYSINIT(interface_check, SI_SUB_PROTO_IF, SI_ORDER_FIRST, if_check, NULL); MALLOC_DEFINE(M_IFNET, "ifnet", "interface internals"); MALLOC_DEFINE(M_IFADDR, "ifaddr", "interface address"); MALLOC_DEFINE(M_IFMADDR, "ether_multi", "link-level multicast address"); static struct ifnet * ifnet_byindex_locked(u_short idx) { INIT_VNET_NET(curvnet); struct ifnet *ifp; ifp = V_ifindex_table[idx].ife_ifnet; return (ifp); } struct ifnet * ifnet_byindex(u_short idx) { struct ifnet *ifp; IFNET_RLOCK(); ifp = ifnet_byindex_locked(idx); IFNET_RUNLOCK(); return (ifp); } static void ifnet_setbyindex(u_short idx, struct ifnet *ifp) { INIT_VNET_NET(curvnet); IFNET_WLOCK_ASSERT(); V_ifindex_table[idx].ife_ifnet = ifp; } struct ifaddr * ifaddr_byindex(u_short idx) { struct ifaddr *ifa; IFNET_RLOCK(); ifa = ifnet_byindex_locked(idx)->if_addr; IFNET_RUNLOCK(); return (ifa); } struct cdev * ifdev_byindex(u_short idx) { INIT_VNET_NET(curvnet); struct cdev *cdev; IFNET_RLOCK(); cdev = V_ifindex_table[idx].ife_dev; IFNET_RUNLOCK(); return (cdev); } static void ifdev_setbyindex(u_short idx, struct cdev *cdev) { INIT_VNET_NET(curvnet); IFNET_WLOCK(); V_ifindex_table[idx].ife_dev = cdev; IFNET_WUNLOCK(); } static d_open_t netopen; static d_close_t netclose; static d_ioctl_t netioctl; static d_kqfilter_t netkqfilter; static struct cdevsw net_cdevsw = { .d_version = D_VERSION, .d_flags = D_NEEDGIANT, .d_open = netopen, .d_close = netclose, .d_ioctl = netioctl, .d_name = "net", .d_kqfilter = netkqfilter, }; static int netopen(struct cdev *dev, int flag, int mode, struct thread *td) { return (0); } static int netclose(struct cdev *dev, int flags, int fmt, struct thread *td) { return (0); } static int netioctl(struct cdev *dev, u_long cmd, caddr_t data, int flag, struct thread *td) { struct ifnet *ifp; int error, idx; /* only support interface specific ioctls */ if (IOCGROUP(cmd) != 'i') return (EOPNOTSUPP); idx = dev2unit(dev); if (idx == 0) { /* * special network device, not interface. */ if (cmd == SIOCGIFCONF) return (ifconf(cmd, data)); /* XXX remove cmd */ #ifdef __amd64__ if (cmd == SIOCGIFCONF32) return (ifconf(cmd, data)); /* XXX remove cmd */ #endif return (EOPNOTSUPP); } ifp = ifnet_byindex(idx); if (ifp == NULL) return (ENXIO); error = ifhwioctl(cmd, ifp, data, td); if (error == ENOIOCTL) error = EOPNOTSUPP; return (error); } static int netkqfilter(struct cdev *dev, struct knote *kn) { INIT_VNET_NET(curvnet); struct knlist *klist; struct ifnet *ifp; int idx; switch (kn->kn_filter) { case EVFILT_NETDEV: kn->kn_fop = &netdev_filtops; break; default: return (EINVAL); } idx = dev2unit(dev); if (idx == 0) { klist = &V_ifklist; } else { ifp = ifnet_byindex(idx); if (ifp == NULL) return (1); klist = &ifp->if_klist; } kn->kn_hook = (caddr_t)klist; knlist_add(klist, kn, 0); return (0); } static void filt_netdetach(struct knote *kn) { struct knlist *klist = (struct knlist *)kn->kn_hook; knlist_remove(klist, kn, 0); } static int filt_netdev(struct knote *kn, long hint) { struct knlist *klist = (struct knlist *)kn->kn_hook; /* * Currently NOTE_EXIT is abused to indicate device detach. */ if (hint == NOTE_EXIT) { kn->kn_data = NOTE_LINKINV; kn->kn_flags |= (EV_EOF | EV_ONESHOT); knlist_remove_inevent(klist, kn); return (1); } if (hint != 0) kn->kn_data = hint; /* current status */ if (kn->kn_sfflags & hint) kn->kn_fflags |= hint; return (kn->kn_fflags != 0); } /* * Network interface utility routines. * * Routines with ifa_ifwith* names take sockaddr *'s as * parameters. */ /* ARGSUSED*/ static void if_init(void *dummy __unused) { INIT_VNET_NET(curvnet); #ifndef VIMAGE_GLOBALS vnet_mod_register(&vnet_net_modinfo); #endif V_if_index = 0; V_ifindex_table = NULL; V_if_indexlim = 8; IFNET_LOCK_INIT(); TAILQ_INIT(&V_ifnet); TAILQ_INIT(&V_ifg_head); knlist_init(&V_ifklist, NULL, NULL, NULL, NULL); if_grow(); /* create initial table */ ifdev_setbyindex(0, make_dev(&net_cdevsw, 0, UID_ROOT, GID_WHEEL, 0600, "network")); if_clone_init(); } static void if_grow(void) { INIT_VNET_NET(curvnet); u_int n; struct ifindex_entry *e; V_if_indexlim <<= 1; n = V_if_indexlim * sizeof(*e); e = malloc(n, M_IFNET, M_WAITOK | M_ZERO); if (V_ifindex_table != NULL) { memcpy((caddr_t)e, (caddr_t)V_ifindex_table, n/2); free((caddr_t)V_ifindex_table, M_IFNET); } V_ifindex_table = e; } static void if_check(void *dummy __unused) { /* * If at least one interface added during boot uses * if_watchdog then start the timer. */ if (slowtimo_started) if_slowtimo(0); } /* * Allocate a struct ifnet and an index for an interface. A layer 2 * common structure will also be allocated if an allocation routine is * registered for the passed type. */ struct ifnet* if_alloc(u_char type) { INIT_VNET_NET(curvnet); struct ifnet *ifp; ifp = malloc(sizeof(struct ifnet), M_IFNET, M_WAITOK|M_ZERO); /* * Try to find an empty slot below if_index. If we fail, take * the next slot. * * XXX: should be locked! */ for (ifp->if_index = 1; ifp->if_index <= V_if_index; ifp->if_index++) { if (ifnet_byindex(ifp->if_index) == NULL) break; } /* Catch if_index overflow. */ if (ifp->if_index < 1) { free(ifp, M_IFNET); return (NULL); } if (ifp->if_index > V_if_index) V_if_index = ifp->if_index; if (V_if_index >= V_if_indexlim) if_grow(); ifp->if_type = type; if (if_com_alloc[type] != NULL) { ifp->if_l2com = if_com_alloc[type](type, ifp); if (ifp->if_l2com == NULL) { free(ifp, M_IFNET); return (NULL); } } IFNET_WLOCK(); ifnet_setbyindex(ifp->if_index, ifp); IFNET_WUNLOCK(); IF_ADDR_LOCK_INIT(ifp); return (ifp); } /* * Free the struct ifnet, the associated index, and the layer 2 common * structure if needed. All the work is done in if_free_type(). * * Do not add code to this function! Add it to if_free_type(). */ void if_free(struct ifnet *ifp) { if_free_type(ifp, ifp->if_type); } /* * Do the actual work of freeing a struct ifnet, associated index, and * layer 2 common structure. This version should only be called by * intefaces that switch their type after calling if_alloc(). */ void if_free_type(struct ifnet *ifp, u_char type) { INIT_VNET_NET(curvnet); /* ifp->if_vnet can be NULL here ! */ if (ifp != ifnet_byindex(ifp->if_index)) { if_printf(ifp, "%s: value was not if_alloced, skipping\n", __func__); return; } IFNET_WLOCK(); ifnet_setbyindex(ifp->if_index, NULL); /* XXX: should be locked with if_findindex() */ while (V_if_index > 0 && ifnet_byindex_locked(V_if_index) == NULL) V_if_index--; IFNET_WUNLOCK(); if (if_com_free[type] != NULL) if_com_free[type](ifp->if_l2com, type); IF_ADDR_LOCK_DESTROY(ifp); free(ifp, M_IFNET); }; void ifq_attach(struct ifaltq *ifq, struct ifnet *ifp) { mtx_init(&ifq->ifq_mtx, ifp->if_xname, "if send queue", MTX_DEF); if (ifq->ifq_maxlen == 0) ifq->ifq_maxlen = ifqmaxlen; ifq->altq_type = 0; ifq->altq_disc = NULL; ifq->altq_flags &= ALTQF_CANTCHANGE; ifq->altq_tbr = NULL; ifq->altq_ifp = ifp; } void ifq_detach(struct ifaltq *ifq) { mtx_destroy(&ifq->ifq_mtx); } /* * Perform generic interface initalization tasks and attach the interface * to the list of "active" interfaces. * * XXX: * - The decision to return void and thus require this function to * succeed is questionable. * - We do more initialization here then is probably a good idea. * Some of this should probably move to if_alloc(). * - We should probably do more sanity checking. For instance we don't * do anything to insure if_xname is unique or non-empty. */ void if_attach(struct ifnet *ifp) { INIT_VNET_NET(curvnet); unsigned socksize, ifasize; int namelen, masklen; struct sockaddr_dl *sdl; struct ifaddr *ifa; if (ifp->if_index == 0 || ifp != ifnet_byindex(ifp->if_index)) panic ("%s: BUG: if_attach called without if_alloc'd input()\n", ifp->if_xname); TASK_INIT(&ifp->if_starttask, 0, if_start_deferred, ifp); TASK_INIT(&ifp->if_linktask, 0, do_link_state_change, ifp); IF_AFDATA_LOCK_INIT(ifp); ifp->if_afdata_initialized = 0; TAILQ_INIT(&ifp->if_addrhead); TAILQ_INIT(&ifp->if_prefixhead); TAILQ_INIT(&ifp->if_multiaddrs); TAILQ_INIT(&ifp->if_groups); if_addgroup(ifp, IFG_ALL); knlist_init(&ifp->if_klist, NULL, NULL, NULL, NULL); getmicrotime(&ifp->if_lastchange); ifp->if_data.ifi_epoch = time_uptime; ifp->if_data.ifi_datalen = sizeof(struct if_data); ifp->if_transmit = if_transmit; ifp->if_qflush = if_qflush; #ifdef MAC mac_ifnet_init(ifp); mac_ifnet_create(ifp); #endif ifdev_setbyindex(ifp->if_index, make_dev(&net_cdevsw, ifp->if_index, UID_ROOT, GID_WHEEL, 0600, "%s/%s", net_cdevsw.d_name, ifp->if_xname)); make_dev_alias(ifdev_byindex(ifp->if_index), "%s%d", net_cdevsw.d_name, ifp->if_index); ifq_attach(&ifp->if_snd, ifp); /* * create a Link Level name for this device */ namelen = strlen(ifp->if_xname); /* * Always save enough space for any possiable name so we can do * a rename in place later. */ masklen = offsetof(struct sockaddr_dl, sdl_data[0]) + IFNAMSIZ; socksize = masklen + ifp->if_addrlen; if (socksize < sizeof(*sdl)) socksize = sizeof(*sdl); socksize = roundup2(socksize, sizeof(long)); ifasize = sizeof(*ifa) + 2 * socksize; ifa = malloc(ifasize, M_IFADDR, M_WAITOK | M_ZERO); IFA_LOCK_INIT(ifa); sdl = (struct sockaddr_dl *)(ifa + 1); sdl->sdl_len = socksize; sdl->sdl_family = AF_LINK; bcopy(ifp->if_xname, sdl->sdl_data, namelen); sdl->sdl_nlen = namelen; sdl->sdl_index = ifp->if_index; sdl->sdl_type = ifp->if_type; ifp->if_addr = ifa; ifa->ifa_ifp = ifp; ifa->ifa_rtrequest = link_rtrequest; ifa->ifa_addr = (struct sockaddr *)sdl; sdl = (struct sockaddr_dl *)(socksize + (caddr_t)sdl); ifa->ifa_netmask = (struct sockaddr *)sdl; sdl->sdl_len = masklen; while (namelen != 0) sdl->sdl_data[--namelen] = 0xff; ifa->ifa_refcnt = 1; TAILQ_INSERT_HEAD(&ifp->if_addrhead, ifa, ifa_link); ifp->if_broadcastaddr = NULL; /* reliably crash if used uninitialized */ IFNET_WLOCK(); TAILQ_INSERT_TAIL(&V_ifnet, ifp, if_link); IFNET_WUNLOCK(); if (domain_init_status >= 2) if_attachdomain1(ifp); EVENTHANDLER_INVOKE(ifnet_arrival_event, ifp); devctl_notify("IFNET", ifp->if_xname, "ATTACH", NULL); /* Announce the interface. */ rt_ifannouncemsg(ifp, IFAN_ARRIVAL); if (ifp->if_watchdog != NULL) { if_printf(ifp, "WARNING: using obsoleted if_watchdog interface\n"); /* * Note that we need if_slowtimo(). If this happens after * boot, then call if_slowtimo() directly. */ if (atomic_cmpset_int(&slowtimo_started, 0, 1) && !cold) if_slowtimo(0); } if (ifp->if_flags & IFF_NEEDSGIANT) if_printf(ifp, "WARNING: using obsoleted IFF_NEEDSGIANT flag\n"); } static void if_attachdomain(void *dummy) { INIT_VNET_NET(curvnet); struct ifnet *ifp; int s; s = splnet(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) if_attachdomain1(ifp); splx(s); } SYSINIT(domainifattach, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_SECOND, if_attachdomain, NULL); static void if_attachdomain1(struct ifnet *ifp) { struct domain *dp; int s; s = splnet(); /* * Since dp->dom_ifattach calls malloc() with M_WAITOK, we * cannot lock ifp->if_afdata initialization, entirely. */ if (IF_AFDATA_TRYLOCK(ifp) == 0) { splx(s); return; } if (ifp->if_afdata_initialized >= domain_init_status) { IF_AFDATA_UNLOCK(ifp); splx(s); printf("if_attachdomain called more than once on %s\n", ifp->if_xname); return; } ifp->if_afdata_initialized = domain_init_status; IF_AFDATA_UNLOCK(ifp); /* address family dependent data region */ bzero(ifp->if_afdata, sizeof(ifp->if_afdata)); for (dp = domains; dp; dp = dp->dom_next) { if (dp->dom_ifattach) ifp->if_afdata[dp->dom_family] = (*dp->dom_ifattach)(ifp); } splx(s); } /* * Remove any unicast or broadcast network addresses from an interface. */ void if_purgeaddrs(struct ifnet *ifp) { struct ifaddr *ifa, *next; TAILQ_FOREACH_SAFE(ifa, &ifp->if_addrhead, ifa_link, next) { if (ifa->ifa_addr->sa_family == AF_LINK) continue; #ifdef INET /* XXX: Ugly!! ad hoc just for INET */ if (ifa->ifa_addr->sa_family == AF_INET) { struct ifaliasreq ifr; bzero(&ifr, sizeof(ifr)); ifr.ifra_addr = *ifa->ifa_addr; if (ifa->ifa_dstaddr) ifr.ifra_broadaddr = *ifa->ifa_dstaddr; if (in_control(NULL, SIOCDIFADDR, (caddr_t)&ifr, ifp, NULL) == 0) continue; } #endif /* INET */ #ifdef INET6 if (ifa->ifa_addr->sa_family == AF_INET6) { in6_purgeaddr(ifa); /* ifp_addrhead is already updated */ continue; } #endif /* INET6 */ TAILQ_REMOVE(&ifp->if_addrhead, ifa, ifa_link); IFAFREE(ifa); } } /* * Remove any multicast network addresses from an interface. */ void if_purgemaddrs(struct ifnet *ifp) { struct ifmultiaddr *ifma; struct ifmultiaddr *next; IF_ADDR_LOCK(ifp); TAILQ_FOREACH_SAFE(ifma, &ifp->if_multiaddrs, ifma_link, next) if_delmulti_locked(ifp, ifma, 1); IF_ADDR_UNLOCK(ifp); } /* * Detach an interface, removing it from the * list of "active" interfaces. * * XXXRW: There are some significant questions about event ordering, and * how to prevent things from starting to use the interface during detach. */ void if_detach(struct ifnet *ifp) { INIT_VNET_NET(ifp->if_vnet); struct ifaddr *ifa; struct radix_node_head *rnh; int s, i, j; struct domain *dp; struct ifnet *iter; int found = 0; IFNET_WLOCK(); TAILQ_FOREACH(iter, &V_ifnet, if_link) if (iter == ifp) { TAILQ_REMOVE(&V_ifnet, ifp, if_link); found = 1; break; } IFNET_WUNLOCK(); if (!found) return; /* * Remove/wait for pending events. */ taskqueue_drain(taskqueue_swi, &ifp->if_linktask); /* * Remove routes and flush queues. */ s = splnet(); if_down(ifp); #ifdef ALTQ if (ALTQ_IS_ENABLED(&ifp->if_snd)) altq_disable(&ifp->if_snd); if (ALTQ_IS_ATTACHED(&ifp->if_snd)) altq_detach(&ifp->if_snd); #endif if_purgeaddrs(ifp); #ifdef INET in_ifdetach(ifp); #endif #ifdef INET6 /* * Remove all IPv6 kernel structs related to ifp. This should be done * before removing routing entries below, since IPv6 interface direct * routes are expected to be removed by the IPv6-specific kernel API. * Otherwise, the kernel will detect some inconsistency and bark it. */ in6_ifdetach(ifp); #endif if_purgemaddrs(ifp); /* * Remove link ifaddr pointer and maybe decrement if_index. * Clean up all addresses. */ ifp->if_addr = NULL; destroy_dev(ifdev_byindex(ifp->if_index)); ifdev_setbyindex(ifp->if_index, NULL); /* We can now free link ifaddr. */ if (!TAILQ_EMPTY(&ifp->if_addrhead)) { ifa = TAILQ_FIRST(&ifp->if_addrhead); TAILQ_REMOVE(&ifp->if_addrhead, ifa, ifa_link); IFAFREE(ifa); } /* * Delete all remaining routes using this interface * Unfortuneatly the only way to do this is to slog through * the entire routing table looking for routes which point * to this interface...oh well... */ for (i = 1; i <= AF_MAX; i++) { for (j = 0; j < rt_numfibs; j++) { if ((rnh = V_rt_tables[j][i]) == NULL) continue; RADIX_NODE_HEAD_LOCK(rnh); (void) rnh->rnh_walktree(rnh, if_rtdel, ifp); RADIX_NODE_HEAD_UNLOCK(rnh); } } /* Announce that the interface is gone. */ rt_ifannouncemsg(ifp, IFAN_DEPARTURE); EVENTHANDLER_INVOKE(ifnet_departure_event, ifp); devctl_notify("IFNET", ifp->if_xname, "DETACH", NULL); IF_AFDATA_LOCK(ifp); for (dp = domains; dp; dp = dp->dom_next) { if (dp->dom_ifdetach && ifp->if_afdata[dp->dom_family]) (*dp->dom_ifdetach)(ifp, ifp->if_afdata[dp->dom_family]); } IF_AFDATA_UNLOCK(ifp); #ifdef MAC mac_ifnet_destroy(ifp); #endif /* MAC */ KNOTE_UNLOCKED(&ifp->if_klist, NOTE_EXIT); knlist_clear(&ifp->if_klist, 0); knlist_destroy(&ifp->if_klist); ifq_detach(&ifp->if_snd); IF_AFDATA_DESTROY(ifp); splx(s); } /* * Add a group to an interface */ int if_addgroup(struct ifnet *ifp, const char *groupname) { INIT_VNET_NET(ifp->if_vnet); struct ifg_list *ifgl; struct ifg_group *ifg = NULL; struct ifg_member *ifgm; if (groupname[0] && groupname[strlen(groupname) - 1] >= '0' && groupname[strlen(groupname) - 1] <= '9') return (EINVAL); IFNET_WLOCK(); TAILQ_FOREACH(ifgl, &ifp->if_groups, ifgl_next) if (!strcmp(ifgl->ifgl_group->ifg_group, groupname)) { IFNET_WUNLOCK(); return (EEXIST); } if ((ifgl = (struct ifg_list *)malloc(sizeof(struct ifg_list), M_TEMP, M_NOWAIT)) == NULL) { IFNET_WUNLOCK(); return (ENOMEM); } if ((ifgm = (struct ifg_member *)malloc(sizeof(struct ifg_member), M_TEMP, M_NOWAIT)) == NULL) { free(ifgl, M_TEMP); IFNET_WUNLOCK(); return (ENOMEM); } TAILQ_FOREACH(ifg, &V_ifg_head, ifg_next) if (!strcmp(ifg->ifg_group, groupname)) break; if (ifg == NULL) { if ((ifg = (struct ifg_group *)malloc(sizeof(struct ifg_group), M_TEMP, M_NOWAIT)) == NULL) { free(ifgl, M_TEMP); free(ifgm, M_TEMP); IFNET_WUNLOCK(); return (ENOMEM); } strlcpy(ifg->ifg_group, groupname, sizeof(ifg->ifg_group)); ifg->ifg_refcnt = 0; TAILQ_INIT(&ifg->ifg_members); EVENTHANDLER_INVOKE(group_attach_event, ifg); TAILQ_INSERT_TAIL(&V_ifg_head, ifg, ifg_next); } ifg->ifg_refcnt++; ifgl->ifgl_group = ifg; ifgm->ifgm_ifp = ifp; IF_ADDR_LOCK(ifp); TAILQ_INSERT_TAIL(&ifg->ifg_members, ifgm, ifgm_next); TAILQ_INSERT_TAIL(&ifp->if_groups, ifgl, ifgl_next); IF_ADDR_UNLOCK(ifp); IFNET_WUNLOCK(); EVENTHANDLER_INVOKE(group_change_event, groupname); return (0); } /* * Remove a group from an interface */ int if_delgroup(struct ifnet *ifp, const char *groupname) { INIT_VNET_NET(ifp->if_vnet); struct ifg_list *ifgl; struct ifg_member *ifgm; IFNET_WLOCK(); TAILQ_FOREACH(ifgl, &ifp->if_groups, ifgl_next) if (!strcmp(ifgl->ifgl_group->ifg_group, groupname)) break; if (ifgl == NULL) { IFNET_WUNLOCK(); return (ENOENT); } IF_ADDR_LOCK(ifp); TAILQ_REMOVE(&ifp->if_groups, ifgl, ifgl_next); IF_ADDR_UNLOCK(ifp); TAILQ_FOREACH(ifgm, &ifgl->ifgl_group->ifg_members, ifgm_next) if (ifgm->ifgm_ifp == ifp) break; if (ifgm != NULL) { TAILQ_REMOVE(&ifgl->ifgl_group->ifg_members, ifgm, ifgm_next); free(ifgm, M_TEMP); } if (--ifgl->ifgl_group->ifg_refcnt == 0) { TAILQ_REMOVE(&V_ifg_head, ifgl->ifgl_group, ifg_next); EVENTHANDLER_INVOKE(group_detach_event, ifgl->ifgl_group); free(ifgl->ifgl_group, M_TEMP); } IFNET_WUNLOCK(); free(ifgl, M_TEMP); EVENTHANDLER_INVOKE(group_change_event, groupname); return (0); } /* * Stores all groups from an interface in memory pointed * to by data */ static int if_getgroup(struct ifgroupreq *data, struct ifnet *ifp) { int len, error; struct ifg_list *ifgl; struct ifg_req ifgrq, *ifgp; struct ifgroupreq *ifgr = data; if (ifgr->ifgr_len == 0) { IF_ADDR_LOCK(ifp); TAILQ_FOREACH(ifgl, &ifp->if_groups, ifgl_next) ifgr->ifgr_len += sizeof(struct ifg_req); IF_ADDR_UNLOCK(ifp); return (0); } len = ifgr->ifgr_len; ifgp = ifgr->ifgr_groups; /* XXX: wire */ IF_ADDR_LOCK(ifp); TAILQ_FOREACH(ifgl, &ifp->if_groups, ifgl_next) { if (len < sizeof(ifgrq)) { IF_ADDR_UNLOCK(ifp); return (EINVAL); } bzero(&ifgrq, sizeof ifgrq); strlcpy(ifgrq.ifgrq_group, ifgl->ifgl_group->ifg_group, sizeof(ifgrq.ifgrq_group)); if ((error = copyout(&ifgrq, ifgp, sizeof(struct ifg_req)))) { IF_ADDR_UNLOCK(ifp); return (error); } len -= sizeof(ifgrq); ifgp++; } IF_ADDR_UNLOCK(ifp); return (0); } /* * Stores all members of a group in memory pointed to by data */ static int if_getgroupmembers(struct ifgroupreq *data) { INIT_VNET_NET(curvnet); struct ifgroupreq *ifgr = data; struct ifg_group *ifg; struct ifg_member *ifgm; struct ifg_req ifgrq, *ifgp; int len, error; IFNET_RLOCK(); TAILQ_FOREACH(ifg, &V_ifg_head, ifg_next) if (!strcmp(ifg->ifg_group, ifgr->ifgr_name)) break; if (ifg == NULL) { IFNET_RUNLOCK(); return (ENOENT); } if (ifgr->ifgr_len == 0) { TAILQ_FOREACH(ifgm, &ifg->ifg_members, ifgm_next) ifgr->ifgr_len += sizeof(ifgrq); IFNET_RUNLOCK(); return (0); } len = ifgr->ifgr_len; ifgp = ifgr->ifgr_groups; TAILQ_FOREACH(ifgm, &ifg->ifg_members, ifgm_next) { if (len < sizeof(ifgrq)) { IFNET_RUNLOCK(); return (EINVAL); } bzero(&ifgrq, sizeof ifgrq); strlcpy(ifgrq.ifgrq_member, ifgm->ifgm_ifp->if_xname, sizeof(ifgrq.ifgrq_member)); if ((error = copyout(&ifgrq, ifgp, sizeof(struct ifg_req)))) { IFNET_RUNLOCK(); return (error); } len -= sizeof(ifgrq); ifgp++; } IFNET_RUNLOCK(); return (0); } /* * Delete Routes for a Network Interface * * Called for each routing entry via the rnh->rnh_walktree() call above * to delete all route entries referencing a detaching network interface. * * Arguments: * rn pointer to node in the routing table * arg argument passed to rnh->rnh_walktree() - detaching interface * * Returns: * 0 successful * errno failed - reason indicated * */ static int if_rtdel(struct radix_node *rn, void *arg) { struct rtentry *rt = (struct rtentry *)rn; struct ifnet *ifp = arg; int err; if (rt->rt_ifp == ifp) { /* * Protect (sorta) against walktree recursion problems * with cloned routes */ if ((rt->rt_flags & RTF_UP) == 0) return (0); err = rtrequest_fib(RTM_DELETE, rt_key(rt), rt->rt_gateway, rt_mask(rt), rt->rt_flags|RTF_RNH_LOCKED, (struct rtentry **) NULL, rt->rt_fibnum); if (err) { log(LOG_WARNING, "if_rtdel: error %d\n", err); } } return (0); } /* * XXX: Because sockaddr_dl has deeper structure than the sockaddr * structs used to represent other address families, it is necessary * to perform a different comparison. */ #define sa_equal(a1, a2) \ (bcmp((a1), (a2), ((a1))->sa_len) == 0) #define sa_dl_equal(a1, a2) \ ((((struct sockaddr_dl *)(a1))->sdl_len == \ ((struct sockaddr_dl *)(a2))->sdl_len) && \ (bcmp(LLADDR((struct sockaddr_dl *)(a1)), \ LLADDR((struct sockaddr_dl *)(a2)), \ ((struct sockaddr_dl *)(a1))->sdl_alen) == 0)) /* * Locate an interface based on a complete address. */ /*ARGSUSED*/ struct ifaddr * ifa_ifwithaddr(struct sockaddr *addr) { INIT_VNET_NET(curvnet); struct ifnet *ifp; struct ifaddr *ifa; IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != addr->sa_family) continue; if (sa_equal(addr, ifa->ifa_addr)) goto done; /* IP6 doesn't have broadcast */ if ((ifp->if_flags & IFF_BROADCAST) && ifa->ifa_broadaddr && ifa->ifa_broadaddr->sa_len != 0 && sa_equal(ifa->ifa_broadaddr, addr)) goto done; } ifa = NULL; done: IFNET_RUNLOCK(); return (ifa); } /* * Locate an interface based on the broadcast address. */ /* ARGSUSED */ struct ifaddr * ifa_ifwithbroadaddr(struct sockaddr *addr) { INIT_VNET_NET(curvnet); struct ifnet *ifp; struct ifaddr *ifa; IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != addr->sa_family) continue; if ((ifp->if_flags & IFF_BROADCAST) && ifa->ifa_broadaddr && ifa->ifa_broadaddr->sa_len != 0 && sa_equal(ifa->ifa_broadaddr, addr)) goto done; } ifa = NULL; done: IFNET_RUNLOCK(); return (ifa); } /* * Locate the point to point interface with a given destination address. */ /*ARGSUSED*/ struct ifaddr * ifa_ifwithdstaddr(struct sockaddr *addr) { INIT_VNET_NET(curvnet); struct ifnet *ifp; struct ifaddr *ifa; IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { if ((ifp->if_flags & IFF_POINTOPOINT) == 0) continue; TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != addr->sa_family) continue; if (ifa->ifa_dstaddr != NULL && sa_equal(addr, ifa->ifa_dstaddr)) goto done; } } ifa = NULL; done: IFNET_RUNLOCK(); return (ifa); } /* * Find an interface on a specific network. If many, choice * is most specific found. */ struct ifaddr * ifa_ifwithnet(struct sockaddr *addr) { INIT_VNET_NET(curvnet); struct ifnet *ifp; struct ifaddr *ifa; struct ifaddr *ifa_maybe = (struct ifaddr *) 0; u_int af = addr->sa_family; char *addr_data = addr->sa_data, *cplim; /* * AF_LINK addresses can be looked up directly by their index number, * so do that if we can. */ if (af == AF_LINK) { struct sockaddr_dl *sdl = (struct sockaddr_dl *)addr; if (sdl->sdl_index && sdl->sdl_index <= V_if_index) return (ifaddr_byindex(sdl->sdl_index)); } /* * Scan though each interface, looking for ones that have * addresses in this address family. */ IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { char *cp, *cp2, *cp3; if (ifa->ifa_addr->sa_family != af) next: continue; if (af == AF_INET && ifp->if_flags & IFF_POINTOPOINT) { /* * This is a bit broken as it doesn't * take into account that the remote end may * be a single node in the network we are * looking for. * The trouble is that we don't know the * netmask for the remote end. */ if (ifa->ifa_dstaddr != NULL && sa_equal(addr, ifa->ifa_dstaddr)) goto done; } else { /* * if we have a special address handler, * then use it instead of the generic one. */ if (ifa->ifa_claim_addr) { if ((*ifa->ifa_claim_addr)(ifa, addr)) goto done; continue; } /* * Scan all the bits in the ifa's address. * If a bit dissagrees with what we are * looking for, mask it with the netmask * to see if it really matters. * (A byte at a time) */ if (ifa->ifa_netmask == 0) continue; cp = addr_data; cp2 = ifa->ifa_addr->sa_data; cp3 = ifa->ifa_netmask->sa_data; cplim = ifa->ifa_netmask->sa_len + (char *)ifa->ifa_netmask; while (cp3 < cplim) if ((*cp++ ^ *cp2++) & *cp3++) goto next; /* next address! */ /* * If the netmask of what we just found * is more specific than what we had before * (if we had one) then remember the new one * before continuing to search * for an even better one. */ if (ifa_maybe == 0 || rn_refines((caddr_t)ifa->ifa_netmask, (caddr_t)ifa_maybe->ifa_netmask)) ifa_maybe = ifa; } } } ifa = ifa_maybe; done: IFNET_RUNLOCK(); return (ifa); } /* * Find an interface address specific to an interface best matching * a given address. */ struct ifaddr * ifaof_ifpforaddr(struct sockaddr *addr, struct ifnet *ifp) { struct ifaddr *ifa; char *cp, *cp2, *cp3; char *cplim; struct ifaddr *ifa_maybe = 0; u_int af = addr->sa_family; if (af >= AF_MAX) return (0); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != af) continue; if (ifa_maybe == 0) ifa_maybe = ifa; if (ifa->ifa_netmask == 0) { if (sa_equal(addr, ifa->ifa_addr) || (ifa->ifa_dstaddr && sa_equal(addr, ifa->ifa_dstaddr))) goto done; continue; } if (ifp->if_flags & IFF_POINTOPOINT) { if (sa_equal(addr, ifa->ifa_dstaddr)) goto done; } else { cp = addr->sa_data; cp2 = ifa->ifa_addr->sa_data; cp3 = ifa->ifa_netmask->sa_data; cplim = ifa->ifa_netmask->sa_len + (char *)ifa->ifa_netmask; for (; cp3 < cplim; cp3++) if ((*cp++ ^ *cp2++) & *cp3) break; if (cp3 == cplim) goto done; } } ifa = ifa_maybe; done: return (ifa); } #include #include /* * Default action when installing a route with a Link Level gateway. * Lookup an appropriate real ifa to point to. * This should be moved to /sys/net/link.c eventually. */ static void link_rtrequest(int cmd, struct rtentry *rt, struct rt_addrinfo *info) { struct ifaddr *ifa, *oifa; struct sockaddr *dst; struct ifnet *ifp; RT_LOCK_ASSERT(rt); if (cmd != RTM_ADD || ((ifa = rt->rt_ifa) == 0) || ((ifp = ifa->ifa_ifp) == 0) || ((dst = rt_key(rt)) == 0)) return; ifa = ifaof_ifpforaddr(dst, ifp); if (ifa) { IFAREF(ifa); /* XXX */ oifa = rt->rt_ifa; rt->rt_ifa = ifa; IFAFREE(oifa); if (ifa->ifa_rtrequest && ifa->ifa_rtrequest != link_rtrequest) ifa->ifa_rtrequest(cmd, rt, info); } } /* * Mark an interface down and notify protocols of * the transition. * NOTE: must be called at splnet or eqivalent. */ static void if_unroute(struct ifnet *ifp, int flag, int fam) { struct ifaddr *ifa; KASSERT(flag == IFF_UP, ("if_unroute: flag != IFF_UP")); ifp->if_flags &= ~flag; getmicrotime(&ifp->if_lastchange); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) if (fam == PF_UNSPEC || (fam == ifa->ifa_addr->sa_family)) pfctlinput(PRC_IFDOWN, ifa->ifa_addr); ifp->if_qflush(ifp); #ifdef DEV_CARP if (ifp->if_carp) carp_carpdev_state(ifp->if_carp); #endif rt_ifmsg(ifp); } /* * Mark an interface up and notify protocols of * the transition. * NOTE: must be called at splnet or eqivalent. */ static void if_route(struct ifnet *ifp, int flag, int fam) { struct ifaddr *ifa; KASSERT(flag == IFF_UP, ("if_route: flag != IFF_UP")); ifp->if_flags |= flag; getmicrotime(&ifp->if_lastchange); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) if (fam == PF_UNSPEC || (fam == ifa->ifa_addr->sa_family)) pfctlinput(PRC_IFUP, ifa->ifa_addr); #ifdef DEV_CARP if (ifp->if_carp) carp_carpdev_state(ifp->if_carp); #endif rt_ifmsg(ifp); #ifdef INET6 in6_if_up(ifp); #endif } void (*vlan_link_state_p)(struct ifnet *, int); /* XXX: private from if_vlan */ void (*vlan_trunk_cap_p)(struct ifnet *); /* XXX: private from if_vlan */ /* * Handle a change in the interface link state. To avoid LORs * between driver lock and upper layer locks, as well as possible * recursions, we post event to taskqueue, and all job * is done in static do_link_state_change(). */ void if_link_state_change(struct ifnet *ifp, int link_state) { /* Return if state hasn't changed. */ if (ifp->if_link_state == link_state) return; ifp->if_link_state = link_state; taskqueue_enqueue(taskqueue_swi, &ifp->if_linktask); } static void do_link_state_change(void *arg, int pending) { struct ifnet *ifp = (struct ifnet *)arg; int link_state = ifp->if_link_state; int link; CURVNET_SET(ifp->if_vnet); /* Notify that the link state has changed. */ rt_ifmsg(ifp); if (link_state == LINK_STATE_UP) link = NOTE_LINKUP; else if (link_state == LINK_STATE_DOWN) link = NOTE_LINKDOWN; else link = NOTE_LINKINV; KNOTE_UNLOCKED(&ifp->if_klist, link); if (ifp->if_vlantrunk != NULL) (*vlan_link_state_p)(ifp, link); if ((ifp->if_type == IFT_ETHER || ifp->if_type == IFT_L2VLAN) && IFP2AC(ifp)->ac_netgraph != NULL) (*ng_ether_link_state_p)(ifp, link_state); #ifdef DEV_CARP if (ifp->if_carp) carp_carpdev_state(ifp->if_carp); #endif if (ifp->if_bridge) { KASSERT(bstp_linkstate_p != NULL,("if_bridge bstp not loaded!")); (*bstp_linkstate_p)(ifp, link_state); } if (ifp->if_lagg) { KASSERT(lagg_linkstate_p != NULL,("if_lagg not loaded!")); (*lagg_linkstate_p)(ifp, link_state); } devctl_notify("IFNET", ifp->if_xname, (link_state == LINK_STATE_UP) ? "LINK_UP" : "LINK_DOWN", NULL); if (pending > 1) if_printf(ifp, "%d link states coalesced\n", pending); if (log_link_state_change) log(LOG_NOTICE, "%s: link state changed to %s\n", ifp->if_xname, (link_state == LINK_STATE_UP) ? "UP" : "DOWN" ); CURVNET_RESTORE(); } /* * Mark an interface down and notify protocols of * the transition. * NOTE: must be called at splnet or eqivalent. */ void if_down(struct ifnet *ifp) { if_unroute(ifp, IFF_UP, AF_UNSPEC); } /* * Mark an interface up and notify protocols of * the transition. * NOTE: must be called at splnet or eqivalent. */ void if_up(struct ifnet *ifp) { if_route(ifp, IFF_UP, AF_UNSPEC); } /* * Flush an interface queue. */ static void if_qflush(struct ifnet *ifp) { struct mbuf *m, *n; struct ifaltq *ifq; ifq = &ifp->if_snd; IFQ_LOCK(ifq); #ifdef ALTQ if (ALTQ_IS_ENABLED(ifq)) ALTQ_PURGE(ifq); #endif n = ifq->ifq_head; while ((m = n) != 0) { n = m->m_act; m_freem(m); } ifq->ifq_head = 0; ifq->ifq_tail = 0; ifq->ifq_len = 0; IFQ_UNLOCK(ifq); } /* * Handle interface watchdog timer routines. Called * from softclock, we decrement timers (if set) and * call the appropriate interface routine on expiration. * * XXXRW: Note that because timeouts run with Giant, if_watchdog() is called * holding Giant. If we switch to an MPSAFE callout, we likely need to grab * Giant before entering if_watchdog() on an IFF_NEEDSGIANT interface. */ static void if_slowtimo(void *arg) { VNET_ITERATOR_DECL(vnet_iter); struct ifnet *ifp; int s = splimp(); IFNET_RLOCK(); VNET_LIST_RLOCK(); VNET_FOREACH(vnet_iter) { CURVNET_SET(vnet_iter); INIT_VNET_NET(vnet_iter); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { if (ifp->if_timer == 0 || --ifp->if_timer) continue; if (ifp->if_watchdog) (*ifp->if_watchdog)(ifp); } CURVNET_RESTORE(); } VNET_LIST_RUNLOCK(); IFNET_RUNLOCK(); splx(s); timeout(if_slowtimo, (void *)0, hz / IFNET_SLOWHZ); } /* * Map interface name to * interface structure pointer. */ struct ifnet * ifunit(const char *name) { INIT_VNET_NET(curvnet); struct ifnet *ifp; IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { if (strncmp(name, ifp->if_xname, IFNAMSIZ) == 0) break; } IFNET_RUNLOCK(); return (ifp); } /* * Hardware specific interface ioctls. */ static int ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td) { struct ifreq *ifr; struct ifstat *ifs; int error = 0; int new_flags, temp_flags; size_t namelen, onamelen; char new_name[IFNAMSIZ]; struct ifaddr *ifa; struct sockaddr_dl *sdl; ifr = (struct ifreq *)data; switch (cmd) { case SIOCGIFINDEX: ifr->ifr_index = ifp->if_index; break; case SIOCGIFFLAGS: temp_flags = ifp->if_flags | ifp->if_drv_flags; ifr->ifr_flags = temp_flags & 0xffff; ifr->ifr_flagshigh = temp_flags >> 16; break; case SIOCGIFCAP: ifr->ifr_reqcap = ifp->if_capabilities; ifr->ifr_curcap = ifp->if_capenable; break; #ifdef MAC case SIOCGIFMAC: error = mac_ifnet_ioctl_get(td->td_ucred, ifr, ifp); break; #endif case SIOCGIFMETRIC: ifr->ifr_metric = ifp->if_metric; break; case SIOCGIFMTU: ifr->ifr_mtu = ifp->if_mtu; break; case SIOCGIFPHYS: ifr->ifr_phys = ifp->if_physical; break; case SIOCSIFFLAGS: error = priv_check(td, PRIV_NET_SETIFFLAGS); if (error) return (error); /* * Currently, no driver owned flags pass the IFF_CANTCHANGE * check, so we don't need special handling here yet. */ new_flags = (ifr->ifr_flags & 0xffff) | (ifr->ifr_flagshigh << 16); if (ifp->if_flags & IFF_SMART) { /* Smart drivers twiddle their own routes */ } else if (ifp->if_flags & IFF_UP && (new_flags & IFF_UP) == 0) { int s = splimp(); if_down(ifp); splx(s); } else if (new_flags & IFF_UP && (ifp->if_flags & IFF_UP) == 0) { int s = splimp(); if_up(ifp); splx(s); } /* See if permanently promiscuous mode bit is about to flip */ if ((ifp->if_flags ^ new_flags) & IFF_PPROMISC) { if (new_flags & IFF_PPROMISC) ifp->if_flags |= IFF_PROMISC; else if (ifp->if_pcount == 0) ifp->if_flags &= ~IFF_PROMISC; log(LOG_INFO, "%s: permanently promiscuous mode %s\n", ifp->if_xname, (new_flags & IFF_PPROMISC) ? "enabled" : "disabled"); } ifp->if_flags = (ifp->if_flags & IFF_CANTCHANGE) | (new_flags &~ IFF_CANTCHANGE); if (ifp->if_ioctl) { IFF_LOCKGIANT(ifp); (void) (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); } getmicrotime(&ifp->if_lastchange); break; case SIOCSIFCAP: error = priv_check(td, PRIV_NET_SETIFCAP); if (error) return (error); if (ifp->if_ioctl == NULL) return (EOPNOTSUPP); if (ifr->ifr_reqcap & ~ifp->if_capabilities) return (EINVAL); IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); if (error == 0) getmicrotime(&ifp->if_lastchange); break; #ifdef MAC case SIOCSIFMAC: error = mac_ifnet_ioctl_set(td->td_ucred, ifr, ifp); break; #endif case SIOCSIFNAME: error = priv_check(td, PRIV_NET_SETIFNAME); if (error) return (error); error = copyinstr(ifr->ifr_data, new_name, IFNAMSIZ, NULL); if (error != 0) return (error); if (new_name[0] == '\0') return (EINVAL); if (ifunit(new_name) != NULL) return (EEXIST); /* Announce the departure of the interface. */ rt_ifannouncemsg(ifp, IFAN_DEPARTURE); EVENTHANDLER_INVOKE(ifnet_departure_event, ifp); log(LOG_INFO, "%s: changing name to '%s'\n", ifp->if_xname, new_name); strlcpy(ifp->if_xname, new_name, sizeof(ifp->if_xname)); ifa = ifp->if_addr; IFA_LOCK(ifa); sdl = (struct sockaddr_dl *)ifa->ifa_addr; namelen = strlen(new_name); onamelen = sdl->sdl_nlen; /* * Move the address if needed. This is safe because we * allocate space for a name of length IFNAMSIZ when we * create this in if_attach(). */ if (namelen != onamelen) { bcopy(sdl->sdl_data + onamelen, sdl->sdl_data + namelen, sdl->sdl_alen); } bcopy(new_name, sdl->sdl_data, namelen); sdl->sdl_nlen = namelen; sdl = (struct sockaddr_dl *)ifa->ifa_netmask; bzero(sdl->sdl_data, onamelen); while (namelen != 0) sdl->sdl_data[--namelen] = 0xff; IFA_UNLOCK(ifa); EVENTHANDLER_INVOKE(ifnet_arrival_event, ifp); /* Announce the return of the interface. */ rt_ifannouncemsg(ifp, IFAN_ARRIVAL); break; case SIOCSIFMETRIC: error = priv_check(td, PRIV_NET_SETIFMETRIC); if (error) return (error); ifp->if_metric = ifr->ifr_metric; getmicrotime(&ifp->if_lastchange); break; case SIOCSIFPHYS: error = priv_check(td, PRIV_NET_SETIFPHYS); if (error) return (error); if (ifp->if_ioctl == NULL) return (EOPNOTSUPP); IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); if (error == 0) getmicrotime(&ifp->if_lastchange); break; case SIOCSIFMTU: { u_long oldmtu = ifp->if_mtu; error = priv_check(td, PRIV_NET_SETIFMTU); if (error) return (error); if (ifr->ifr_mtu < IF_MINMTU || ifr->ifr_mtu > IF_MAXMTU) return (EINVAL); if (ifp->if_ioctl == NULL) return (EOPNOTSUPP); IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); if (error == 0) { getmicrotime(&ifp->if_lastchange); rt_ifmsg(ifp); } /* * If the link MTU changed, do network layer specific procedure. */ if (ifp->if_mtu != oldmtu) { #ifdef INET6 nd6_setmtu(ifp); #endif } break; } case SIOCADDMULTI: case SIOCDELMULTI: if (cmd == SIOCADDMULTI) error = priv_check(td, PRIV_NET_ADDMULTI); else error = priv_check(td, PRIV_NET_DELMULTI); if (error) return (error); /* Don't allow group membership on non-multicast interfaces. */ if ((ifp->if_flags & IFF_MULTICAST) == 0) return (EOPNOTSUPP); /* Don't let users screw up protocols' entries. */ if (ifr->ifr_addr.sa_family != AF_LINK) return (EINVAL); if (cmd == SIOCADDMULTI) { struct ifmultiaddr *ifma; /* * Userland is only permitted to join groups once * via the if_addmulti() KPI, because it cannot hold * struct ifmultiaddr * between calls. It may also * lose a race while we check if the membership * already exists. */ IF_ADDR_LOCK(ifp); ifma = if_findmulti(ifp, &ifr->ifr_addr); IF_ADDR_UNLOCK(ifp); if (ifma != NULL) error = EADDRINUSE; else error = if_addmulti(ifp, &ifr->ifr_addr, &ifma); } else { error = if_delmulti(ifp, &ifr->ifr_addr); } if (error == 0) getmicrotime(&ifp->if_lastchange); break; case SIOCSIFPHYADDR: case SIOCDIFPHYADDR: #ifdef INET6 case SIOCSIFPHYADDR_IN6: #endif case SIOCSLIFPHYADDR: case SIOCSIFMEDIA: case SIOCSIFGENERIC: error = priv_check(td, PRIV_NET_HWIOCTL); if (error) return (error); if (ifp->if_ioctl == NULL) return (EOPNOTSUPP); IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); if (error == 0) getmicrotime(&ifp->if_lastchange); break; case SIOCGIFSTATUS: ifs = (struct ifstat *)data; ifs->ascii[0] = '\0'; case SIOCGIFPSRCADDR: case SIOCGIFPDSTADDR: case SIOCGLIFPHYADDR: case SIOCGIFMEDIA: case SIOCGIFGENERIC: if (ifp->if_ioctl == NULL) return (EOPNOTSUPP); IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); break; case SIOCSIFLLADDR: error = priv_check(td, PRIV_NET_SETLLADDR); if (error) return (error); error = if_setlladdr(ifp, ifr->ifr_addr.sa_data, ifr->ifr_addr.sa_len); break; case SIOCAIFGROUP: { struct ifgroupreq *ifgr = (struct ifgroupreq *)ifr; error = priv_check(td, PRIV_NET_ADDIFGROUP); if (error) return (error); if ((error = if_addgroup(ifp, ifgr->ifgr_group))) return (error); break; } case SIOCGIFGROUP: if ((error = if_getgroup((struct ifgroupreq *)ifr, ifp))) return (error); break; case SIOCDIFGROUP: { struct ifgroupreq *ifgr = (struct ifgroupreq *)ifr; error = priv_check(td, PRIV_NET_DELIFGROUP); if (error) return (error); if ((error = if_delgroup(ifp, ifgr->ifgr_group))) return (error); break; } default: error = ENOIOCTL; break; } return (error); } /* * Interface ioctls. */ int ifioctl(struct socket *so, u_long cmd, caddr_t data, struct thread *td) { struct ifnet *ifp; struct ifreq *ifr; int error; int oif_flags; switch (cmd) { case SIOCGIFCONF: case OSIOCGIFCONF: #ifdef __amd64__ case SIOCGIFCONF32: #endif return (ifconf(cmd, data)); } ifr = (struct ifreq *)data; switch (cmd) { case SIOCIFCREATE: case SIOCIFCREATE2: error = priv_check(td, PRIV_NET_IFCREATE); if (error) return (error); return (if_clone_create(ifr->ifr_name, sizeof(ifr->ifr_name), cmd == SIOCIFCREATE2 ? ifr->ifr_data : NULL)); case SIOCIFDESTROY: error = priv_check(td, PRIV_NET_IFDESTROY); if (error) return (error); return if_clone_destroy(ifr->ifr_name); case SIOCIFGCLONERS: return (if_clone_list((struct if_clonereq *)data)); case SIOCGIFGMEMB: return (if_getgroupmembers((struct ifgroupreq *)data)); } ifp = ifunit(ifr->ifr_name); if (ifp == 0) return (ENXIO); error = ifhwioctl(cmd, ifp, data, td); if (error != ENOIOCTL) return (error); oif_flags = ifp->if_flags; if (so->so_proto == 0) return (EOPNOTSUPP); #ifndef COMPAT_43 error = ((*so->so_proto->pr_usrreqs->pru_control)(so, cmd, data, ifp, td)); #else { int ocmd = cmd; switch (cmd) { case SIOCSIFDSTADDR: case SIOCSIFADDR: case SIOCSIFBRDADDR: case SIOCSIFNETMASK: #if BYTE_ORDER != BIG_ENDIAN if (ifr->ifr_addr.sa_family == 0 && ifr->ifr_addr.sa_len < 16) { ifr->ifr_addr.sa_family = ifr->ifr_addr.sa_len; ifr->ifr_addr.sa_len = 16; } #else if (ifr->ifr_addr.sa_len == 0) ifr->ifr_addr.sa_len = 16; #endif break; case OSIOCGIFADDR: cmd = SIOCGIFADDR; break; case OSIOCGIFDSTADDR: cmd = SIOCGIFDSTADDR; break; case OSIOCGIFBRDADDR: cmd = SIOCGIFBRDADDR; break; case OSIOCGIFNETMASK: cmd = SIOCGIFNETMASK; } error = ((*so->so_proto->pr_usrreqs->pru_control)(so, cmd, data, ifp, td)); switch (ocmd) { case OSIOCGIFADDR: case OSIOCGIFDSTADDR: case OSIOCGIFBRDADDR: case OSIOCGIFNETMASK: *(u_short *)&ifr->ifr_addr = ifr->ifr_addr.sa_family; } } #endif /* COMPAT_43 */ if ((oif_flags ^ ifp->if_flags) & IFF_UP) { #ifdef INET6 DELAY(100);/* XXX: temporary workaround for fxp issue*/ if (ifp->if_flags & IFF_UP) { int s = splimp(); in6_if_up(ifp); splx(s); } #endif } return (error); } /* * The code common to handling reference counted flags, * e.g., in ifpromisc() and if_allmulti(). * The "pflag" argument can specify a permanent mode flag to check, * such as IFF_PPROMISC for promiscuous mode; should be 0 if none. * * Only to be used on stack-owned flags, not driver-owned flags. */ static int if_setflag(struct ifnet *ifp, int flag, int pflag, int *refcount, int onswitch) { struct ifreq ifr; int error; int oldflags, oldcount; /* Sanity checks to catch programming errors */ KASSERT((flag & (IFF_DRV_OACTIVE|IFF_DRV_RUNNING)) == 0, ("%s: setting driver-owned flag %d", __func__, flag)); if (onswitch) KASSERT(*refcount >= 0, ("%s: increment negative refcount %d for flag %d", __func__, *refcount, flag)); else KASSERT(*refcount > 0, ("%s: decrement non-positive refcount %d for flag %d", __func__, *refcount, flag)); /* In case this mode is permanent, just touch refcount */ if (ifp->if_flags & pflag) { *refcount += onswitch ? 1 : -1; return (0); } /* Save ifnet parameters for if_ioctl() may fail */ oldcount = *refcount; oldflags = ifp->if_flags; /* * See if we aren't the only and touching refcount is enough. * Actually toggle interface flag if we are the first or last. */ if (onswitch) { if ((*refcount)++) return (0); ifp->if_flags |= flag; } else { if (--(*refcount)) return (0); ifp->if_flags &= ~flag; } /* Call down the driver since we've changed interface flags */ if (ifp->if_ioctl == NULL) { error = EOPNOTSUPP; goto recover; } ifr.ifr_flags = ifp->if_flags & 0xffff; ifr.ifr_flagshigh = ifp->if_flags >> 16; IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, (caddr_t)&ifr); IFF_UNLOCKGIANT(ifp); if (error) goto recover; /* Notify userland that interface flags have changed */ rt_ifmsg(ifp); return (0); recover: /* Recover after driver error */ *refcount = oldcount; ifp->if_flags = oldflags; return (error); } /* * Set/clear promiscuous mode on interface ifp based on the truth value * of pswitch. The calls are reference counted so that only the first * "on" request actually has an effect, as does the final "off" request. * Results are undefined if the "off" and "on" requests are not matched. */ int ifpromisc(struct ifnet *ifp, int pswitch) { int error; int oldflags = ifp->if_flags; error = if_setflag(ifp, IFF_PROMISC, IFF_PPROMISC, &ifp->if_pcount, pswitch); /* If promiscuous mode status has changed, log a message */ if (error == 0 && ((ifp->if_flags ^ oldflags) & IFF_PROMISC)) log(LOG_INFO, "%s: promiscuous mode %s\n", ifp->if_xname, (ifp->if_flags & IFF_PROMISC) ? "enabled" : "disabled"); return (error); } /* * Return interface configuration * of system. List may be used * in later ioctl's (above) to get * other information. */ /*ARGSUSED*/ static int ifconf(u_long cmd, caddr_t data) { INIT_VNET_NET(curvnet); struct ifconf *ifc = (struct ifconf *)data; #ifdef __amd64__ struct ifconf32 *ifc32 = (struct ifconf32 *)data; struct ifconf ifc_swab; #endif struct ifnet *ifp; struct ifaddr *ifa; struct ifreq ifr; struct sbuf *sb; int error, full = 0, valid_len, max_len; #ifdef __amd64__ if (cmd == SIOCGIFCONF32) { ifc_swab.ifc_len = ifc32->ifc_len; ifc_swab.ifc_buf = (caddr_t)(uintptr_t)ifc32->ifc_buf; ifc = &ifc_swab; } #endif /* Limit initial buffer size to MAXPHYS to avoid DoS from userspace. */ max_len = MAXPHYS - 1; /* Prevent hostile input from being able to crash the system */ if (ifc->ifc_len <= 0) return (EINVAL); again: if (ifc->ifc_len <= max_len) { max_len = ifc->ifc_len; full = 1; } sb = sbuf_new(NULL, NULL, max_len + 1, SBUF_FIXEDLEN); max_len = 0; valid_len = 0; IFNET_RLOCK(); /* could sleep XXX */ TAILQ_FOREACH(ifp, &V_ifnet, if_link) { int addrs; /* * Zero the ifr_name buffer to make sure we don't * disclose the contents of the stack. */ memset(ifr.ifr_name, 0, sizeof(ifr.ifr_name)); if (strlcpy(ifr.ifr_name, ifp->if_xname, sizeof(ifr.ifr_name)) >= sizeof(ifr.ifr_name)) { sbuf_delete(sb); IFNET_RUNLOCK(); return (ENAMETOOLONG); } addrs = 0; TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { struct sockaddr *sa = ifa->ifa_addr; - if (jailed(curthread->td_ucred) && - !prison_if(curthread->td_ucred, sa)) + if (prison_if(curthread->td_ucred, sa) != 0) continue; addrs++; #ifdef COMPAT_43 if (cmd == OSIOCGIFCONF) { struct osockaddr *osa = (struct osockaddr *)&ifr.ifr_addr; ifr.ifr_addr = *sa; osa->sa_family = sa->sa_family; sbuf_bcat(sb, &ifr, sizeof(ifr)); max_len += sizeof(ifr); } else #endif if (sa->sa_len <= sizeof(*sa)) { ifr.ifr_addr = *sa; sbuf_bcat(sb, &ifr, sizeof(ifr)); max_len += sizeof(ifr); } else { sbuf_bcat(sb, &ifr, offsetof(struct ifreq, ifr_addr)); max_len += offsetof(struct ifreq, ifr_addr); sbuf_bcat(sb, sa, sa->sa_len); max_len += sa->sa_len; } if (!sbuf_overflowed(sb)) valid_len = sbuf_len(sb); } if (addrs == 0) { bzero((caddr_t)&ifr.ifr_addr, sizeof(ifr.ifr_addr)); sbuf_bcat(sb, &ifr, sizeof(ifr)); max_len += sizeof(ifr); if (!sbuf_overflowed(sb)) valid_len = sbuf_len(sb); } } IFNET_RUNLOCK(); /* * If we didn't allocate enough space (uncommon), try again. If * we have already allocated as much space as we are allowed, * return what we've got. */ if (valid_len != max_len && !full) { sbuf_delete(sb); goto again; } ifc->ifc_len = valid_len; #ifdef __amd64__ if (cmd == SIOCGIFCONF32) ifc32->ifc_len = valid_len; #endif sbuf_finish(sb); error = copyout(sbuf_data(sb), ifc->ifc_req, ifc->ifc_len); sbuf_delete(sb); return (error); } /* * Just like ifpromisc(), but for all-multicast-reception mode. */ int if_allmulti(struct ifnet *ifp, int onswitch) { return (if_setflag(ifp, IFF_ALLMULTI, 0, &ifp->if_amcount, onswitch)); } struct ifmultiaddr * if_findmulti(struct ifnet *ifp, struct sockaddr *sa) { struct ifmultiaddr *ifma; IF_ADDR_LOCK_ASSERT(ifp); TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) { if (sa->sa_family == AF_LINK) { if (sa_dl_equal(ifma->ifma_addr, sa)) break; } else { if (sa_equal(ifma->ifma_addr, sa)) break; } } return ifma; } /* * Allocate a new ifmultiaddr and initialize based on passed arguments. We * make copies of passed sockaddrs. The ifmultiaddr will not be added to * the ifnet multicast address list here, so the caller must do that and * other setup work (such as notifying the device driver). The reference * count is initialized to 1. */ static struct ifmultiaddr * if_allocmulti(struct ifnet *ifp, struct sockaddr *sa, struct sockaddr *llsa, int mflags) { struct ifmultiaddr *ifma; struct sockaddr *dupsa; ifma = malloc(sizeof *ifma, M_IFMADDR, mflags | M_ZERO); if (ifma == NULL) return (NULL); dupsa = malloc(sa->sa_len, M_IFMADDR, mflags); if (dupsa == NULL) { free(ifma, M_IFMADDR); return (NULL); } bcopy(sa, dupsa, sa->sa_len); ifma->ifma_addr = dupsa; ifma->ifma_ifp = ifp; ifma->ifma_refcount = 1; ifma->ifma_protospec = NULL; if (llsa == NULL) { ifma->ifma_lladdr = NULL; return (ifma); } dupsa = malloc(llsa->sa_len, M_IFMADDR, mflags); if (dupsa == NULL) { free(ifma->ifma_addr, M_IFMADDR); free(ifma, M_IFMADDR); return (NULL); } bcopy(llsa, dupsa, llsa->sa_len); ifma->ifma_lladdr = dupsa; return (ifma); } /* * if_freemulti: free ifmultiaddr structure and possibly attached related * addresses. The caller is responsible for implementing reference * counting, notifying the driver, handling routing messages, and releasing * any dependent link layer state. */ static void if_freemulti(struct ifmultiaddr *ifma) { KASSERT(ifma->ifma_refcount == 0, ("if_freemulti: refcount %d", ifma->ifma_refcount)); KASSERT(ifma->ifma_protospec == NULL, ("if_freemulti: protospec not NULL")); if (ifma->ifma_lladdr != NULL) free(ifma->ifma_lladdr, M_IFMADDR); free(ifma->ifma_addr, M_IFMADDR); free(ifma, M_IFMADDR); } /* * Register an additional multicast address with a network interface. * * - If the address is already present, bump the reference count on the * address and return. * - If the address is not link-layer, look up a link layer address. * - Allocate address structures for one or both addresses, and attach to the * multicast address list on the interface. If automatically adding a link * layer address, the protocol address will own a reference to the link * layer address, to be freed when it is freed. * - Notify the network device driver of an addition to the multicast address * list. * * 'sa' points to caller-owned memory with the desired multicast address. * * 'retifma' will be used to return a pointer to the resulting multicast * address reference, if desired. */ int if_addmulti(struct ifnet *ifp, struct sockaddr *sa, struct ifmultiaddr **retifma) { struct ifmultiaddr *ifma, *ll_ifma; struct sockaddr *llsa; int error; /* * If the address is already present, return a new reference to it; * otherwise, allocate storage and set up a new address. */ IF_ADDR_LOCK(ifp); ifma = if_findmulti(ifp, sa); if (ifma != NULL) { ifma->ifma_refcount++; if (retifma != NULL) *retifma = ifma; IF_ADDR_UNLOCK(ifp); return (0); } /* * The address isn't already present; resolve the protocol address * into a link layer address, and then look that up, bump its * refcount or allocate an ifma for that also. If 'llsa' was * returned, we will need to free it later. */ llsa = NULL; ll_ifma = NULL; if (ifp->if_resolvemulti != NULL) { error = ifp->if_resolvemulti(ifp, &llsa, sa); if (error) goto unlock_out; } /* * Allocate the new address. Don't hook it up yet, as we may also * need to allocate a link layer multicast address. */ ifma = if_allocmulti(ifp, sa, llsa, M_NOWAIT); if (ifma == NULL) { error = ENOMEM; goto free_llsa_out; } /* * If a link layer address is found, we'll need to see if it's * already present in the address list, or allocate is as well. * When this block finishes, the link layer address will be on the * list. */ if (llsa != NULL) { ll_ifma = if_findmulti(ifp, llsa); if (ll_ifma == NULL) { ll_ifma = if_allocmulti(ifp, llsa, NULL, M_NOWAIT); if (ll_ifma == NULL) { --ifma->ifma_refcount; if_freemulti(ifma); error = ENOMEM; goto free_llsa_out; } TAILQ_INSERT_HEAD(&ifp->if_multiaddrs, ll_ifma, ifma_link); } else ll_ifma->ifma_refcount++; ifma->ifma_llifma = ll_ifma; } /* * We now have a new multicast address, ifma, and possibly a new or * referenced link layer address. Add the primary address to the * ifnet address list. */ TAILQ_INSERT_HEAD(&ifp->if_multiaddrs, ifma, ifma_link); if (retifma != NULL) *retifma = ifma; /* * Must generate the message while holding the lock so that 'ifma' * pointer is still valid. */ rt_newmaddrmsg(RTM_NEWMADDR, ifma); IF_ADDR_UNLOCK(ifp); /* * We are certain we have added something, so call down to the * interface to let them know about it. */ if (ifp->if_ioctl != NULL) { IFF_LOCKGIANT(ifp); (void) (*ifp->if_ioctl)(ifp, SIOCADDMULTI, 0); IFF_UNLOCKGIANT(ifp); } if (llsa != NULL) free(llsa, M_IFMADDR); return (0); free_llsa_out: if (llsa != NULL) free(llsa, M_IFMADDR); unlock_out: IF_ADDR_UNLOCK(ifp); return (error); } /* * Delete a multicast group membership by network-layer group address. * * Returns ENOENT if the entry could not be found. If ifp no longer * exists, results are undefined. This entry point should only be used * from subsystems which do appropriate locking to hold ifp for the * duration of the call. * Network-layer protocol domains must use if_delmulti_ifma(). */ int if_delmulti(struct ifnet *ifp, struct sockaddr *sa) { struct ifmultiaddr *ifma; int lastref; #ifdef INVARIANTS struct ifnet *oifp; INIT_VNET_NET(ifp->if_vnet); IFNET_RLOCK(); TAILQ_FOREACH(oifp, &V_ifnet, if_link) if (ifp == oifp) break; if (ifp != oifp) ifp = NULL; IFNET_RUNLOCK(); KASSERT(ifp != NULL, ("%s: ifnet went away", __func__)); #endif if (ifp == NULL) return (ENOENT); IF_ADDR_LOCK(ifp); lastref = 0; ifma = if_findmulti(ifp, sa); if (ifma != NULL) lastref = if_delmulti_locked(ifp, ifma, 0); IF_ADDR_UNLOCK(ifp); if (ifma == NULL) return (ENOENT); if (lastref && ifp->if_ioctl != NULL) { IFF_LOCKGIANT(ifp); (void)(*ifp->if_ioctl)(ifp, SIOCDELMULTI, 0); IFF_UNLOCKGIANT(ifp); } return (0); } /* * Delete a multicast group membership by group membership pointer. * Network-layer protocol domains must use this routine. * * It is safe to call this routine if the ifp disappeared. Callers should * hold IFF_LOCKGIANT() to avoid a LOR in case the hardware needs to be * reconfigured. */ void if_delmulti_ifma(struct ifmultiaddr *ifma) { #ifdef DIAGNOSTIC INIT_VNET_NET(curvnet); #endif struct ifnet *ifp; int lastref; ifp = ifma->ifma_ifp; #ifdef DIAGNOSTIC if (ifp == NULL) { printf("%s: ifma_ifp seems to be detached\n", __func__); } else { struct ifnet *oifp; IFNET_RLOCK(); TAILQ_FOREACH(oifp, &V_ifnet, if_link) if (ifp == oifp) break; if (ifp != oifp) { printf("%s: ifnet %p disappeared\n", __func__, ifp); ifp = NULL; } IFNET_RUNLOCK(); } #endif /* * If and only if the ifnet instance exists: Acquire the address lock. */ if (ifp != NULL) IF_ADDR_LOCK(ifp); lastref = if_delmulti_locked(ifp, ifma, 0); if (ifp != NULL) { /* * If and only if the ifnet instance exists: * Release the address lock. * If the group was left: update the hardware hash filter. */ IF_ADDR_UNLOCK(ifp); if (lastref && ifp->if_ioctl != NULL) { IFF_LOCKGIANT(ifp); (void)(*ifp->if_ioctl)(ifp, SIOCDELMULTI, 0); IFF_UNLOCKGIANT(ifp); } } } /* * Perform deletion of network-layer and/or link-layer multicast address. * * Return 0 if the reference count was decremented. * Return 1 if the final reference was released, indicating that the * hardware hash filter should be reprogrammed. */ static int if_delmulti_locked(struct ifnet *ifp, struct ifmultiaddr *ifma, int detaching) { struct ifmultiaddr *ll_ifma; if (ifp != NULL && ifma->ifma_ifp != NULL) { KASSERT(ifma->ifma_ifp == ifp, ("%s: inconsistent ifp %p", __func__, ifp)); IF_ADDR_LOCK_ASSERT(ifp); } ifp = ifma->ifma_ifp; /* * If the ifnet is detaching, null out references to ifnet, * so that upper protocol layers will notice, and not attempt * to obtain locks for an ifnet which no longer exists. The * routing socket announcement must happen before the ifnet * instance is detached from the system. */ if (detaching) { #ifdef DIAGNOSTIC printf("%s: detaching ifnet instance %p\n", __func__, ifp); #endif /* * ifp may already be nulled out if we are being reentered * to delete the ll_ifma. */ if (ifp != NULL) { rt_newmaddrmsg(RTM_DELMADDR, ifma); ifma->ifma_ifp = NULL; } } if (--ifma->ifma_refcount > 0) return 0; /* * If this ifma is a network-layer ifma, a link-layer ifma may * have been associated with it. Release it first if so. */ ll_ifma = ifma->ifma_llifma; if (ll_ifma != NULL) { KASSERT(ifma->ifma_lladdr != NULL, ("%s: llifma w/o lladdr", __func__)); if (detaching) ll_ifma->ifma_ifp = NULL; /* XXX */ if (--ll_ifma->ifma_refcount == 0) { if (ifp != NULL) { TAILQ_REMOVE(&ifp->if_multiaddrs, ll_ifma, ifma_link); } if_freemulti(ll_ifma); } } if (ifp != NULL) TAILQ_REMOVE(&ifp->if_multiaddrs, ifma, ifma_link); if_freemulti(ifma); /* * The last reference to this instance of struct ifmultiaddr * was released; the hardware should be notified of this change. */ return 1; } /* * Set the link layer address on an interface. * * At this time we only support certain types of interfaces, * and we don't allow the length of the address to change. */ int if_setlladdr(struct ifnet *ifp, const u_char *lladdr, int len) { struct sockaddr_dl *sdl; struct ifaddr *ifa; struct ifreq ifr; ifa = ifp->if_addr; if (ifa == NULL) return (EINVAL); sdl = (struct sockaddr_dl *)ifa->ifa_addr; if (sdl == NULL) return (EINVAL); if (len != sdl->sdl_alen) /* don't allow length to change */ return (EINVAL); switch (ifp->if_type) { case IFT_ETHER: case IFT_FDDI: case IFT_XETHER: case IFT_ISO88025: case IFT_L2VLAN: case IFT_BRIDGE: case IFT_ARCNET: case IFT_IEEE8023ADLAG: bcopy(lladdr, LLADDR(sdl), len); break; default: return (ENODEV); } /* * If the interface is already up, we need * to re-init it in order to reprogram its * address filter. */ if ((ifp->if_flags & IFF_UP) != 0) { if (ifp->if_ioctl) { IFF_LOCKGIANT(ifp); ifp->if_flags &= ~IFF_UP; ifr.ifr_flags = ifp->if_flags & 0xffff; ifr.ifr_flagshigh = ifp->if_flags >> 16; (*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, (caddr_t)&ifr); ifp->if_flags |= IFF_UP; ifr.ifr_flags = ifp->if_flags & 0xffff; ifr.ifr_flagshigh = ifp->if_flags >> 16; (*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, (caddr_t)&ifr); IFF_UNLOCKGIANT(ifp); } #ifdef INET /* * Also send gratuitous ARPs to notify other nodes about * the address change. */ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family == AF_INET) arp_ifinit(ifp, ifa); } #endif } return (0); } /* * The name argument must be a pointer to storage which will last as * long as the interface does. For physical devices, the result of * device_get_name(dev) is a good choice and for pseudo-devices a * static string works well. */ void if_initname(struct ifnet *ifp, const char *name, int unit) { ifp->if_dname = name; ifp->if_dunit = unit; if (unit != IF_DUNIT_NONE) snprintf(ifp->if_xname, IFNAMSIZ, "%s%d", name, unit); else strlcpy(ifp->if_xname, name, IFNAMSIZ); } int if_printf(struct ifnet *ifp, const char * fmt, ...) { va_list ap; int retval; retval = printf("%s: ", ifp->if_xname); va_start(ap, fmt); retval += vprintf(fmt, ap); va_end(ap); return (retval); } /* * When an interface is marked IFF_NEEDSGIANT, its if_start() routine cannot * be called without Giant. However, we often can't acquire the Giant lock * at those points; instead, we run it via a task queue that holds Giant via * if_start_deferred. * * XXXRW: We need to make sure that the ifnet isn't fully detached until any * outstanding if_start_deferred() tasks that will run after the free. This * probably means waiting in if_detach(). */ void if_start(struct ifnet *ifp) { if (ifp->if_flags & IFF_NEEDSGIANT) { if (mtx_owned(&Giant)) (*(ifp)->if_start)(ifp); else taskqueue_enqueue(taskqueue_swi_giant, &ifp->if_starttask); } else (*(ifp)->if_start)(ifp); } static void if_start_deferred(void *context, int pending) { struct ifnet *ifp; GIANT_REQUIRED; ifp = context; (ifp->if_start)(ifp); } /* * Backwards compatibility interface for drivers * that have not implemented it */ static int if_transmit(struct ifnet *ifp, struct mbuf *m) { int error; IFQ_HANDOFF(ifp, m, error); return (error); } int if_handoff(struct ifqueue *ifq, struct mbuf *m, struct ifnet *ifp, int adjust) { int active = 0; IF_LOCK(ifq); if (_IF_QFULL(ifq)) { _IF_DROP(ifq); IF_UNLOCK(ifq); m_freem(m); return (0); } if (ifp != NULL) { ifp->if_obytes += m->m_pkthdr.len + adjust; if (m->m_flags & (M_BCAST|M_MCAST)) ifp->if_omcasts++; active = ifp->if_drv_flags & IFF_DRV_OACTIVE; } _IF_ENQUEUE(ifq, m); IF_UNLOCK(ifq); if (ifp != NULL && !active) if_start(ifp); return (1); } void if_register_com_alloc(u_char type, if_com_alloc_t *a, if_com_free_t *f) { KASSERT(if_com_alloc[type] == NULL, ("if_register_com_alloc: %d already registered", type)); KASSERT(if_com_free[type] == NULL, ("if_register_com_alloc: %d free already registered", type)); if_com_alloc[type] = a; if_com_free[type] = f; } void if_deregister_com_alloc(u_char type) { KASSERT(if_com_alloc[type] != NULL, ("if_deregister_com_alloc: %d not registered", type)); KASSERT(if_com_free[type] != NULL, ("if_deregister_com_alloc: %d free not registered", type)); if_com_alloc[type] = NULL; if_com_free[type] = NULL; } diff --git a/sys/net/rtsock.c b/sys/net/rtsock.c index 12a3cba4b805..8d2d07af4c86 100644 --- a/sys/net/rtsock.c +++ b/sys/net/rtsock.c @@ -1,1511 +1,1512 @@ /*- * Copyright (c) 1988, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)rtsock.c 8.7 (Berkeley) 10/12/95 * $FreeBSD$ */ #include "opt_sctp.h" #include "opt_mpath.h" #include "opt_route.h" #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET6 #include #endif #ifdef SCTP extern void sctp_addr_change(struct ifaddr *ifa, int cmd); #endif /* SCTP */ MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables"); /* NB: these are not modified */ static struct sockaddr route_src = { 2, PF_ROUTE, }; static struct sockaddr sa_zero = { sizeof(sa_zero), AF_INET, }; static struct { int ip_count; /* attached w/ AF_INET */ int ip6_count; /* attached w/ AF_INET6 */ int ipx_count; /* attached w/ AF_IPX */ int any_count; /* total attached */ } route_cb; struct mtx rtsock_mtx; MTX_SYSINIT(rtsock, &rtsock_mtx, "rtsock route_cb lock", MTX_DEF); #define RTSOCK_LOCK() mtx_lock(&rtsock_mtx) #define RTSOCK_UNLOCK() mtx_unlock(&rtsock_mtx) #define RTSOCK_LOCK_ASSERT() mtx_assert(&rtsock_mtx, MA_OWNED) static struct ifqueue rtsintrq; SYSCTL_NODE(_net, OID_AUTO, route, CTLFLAG_RD, 0, ""); SYSCTL_INT(_net_route, OID_AUTO, netisr_maxqlen, CTLFLAG_RW, &rtsintrq.ifq_maxlen, 0, "maximum routing socket dispatch queue length"); struct walkarg { int w_tmemsize; int w_op, w_arg; caddr_t w_tmem; struct sysctl_req *w_req; }; static void rts_input(struct mbuf *m); static struct mbuf *rt_msg1(int type, struct rt_addrinfo *rtinfo); static int rt_msg2(int type, struct rt_addrinfo *rtinfo, caddr_t cp, struct walkarg *w); static int rt_xaddrs(caddr_t cp, caddr_t cplim, struct rt_addrinfo *rtinfo); static int sysctl_dumpentry(struct radix_node *rn, void *vw); static int sysctl_iflist(int af, struct walkarg *w); static int sysctl_ifmalist(int af, struct walkarg *w); static int route_output(struct mbuf *m, struct socket *so); static void rt_setmetrics(u_long which, const struct rt_metrics *in, struct rt_metrics_lite *out); static void rt_getmetrics(const struct rt_metrics_lite *in, struct rt_metrics *out); static void rt_dispatch(struct mbuf *, const struct sockaddr *); static void rts_init(void) { int tmp; rtsintrq.ifq_maxlen = 256; if (TUNABLE_INT_FETCH("net.route.netisr_maxqlen", &tmp)) rtsintrq.ifq_maxlen = tmp; mtx_init(&rtsintrq.ifq_mtx, "rts_inq", NULL, MTX_DEF); netisr_register(NETISR_ROUTE, rts_input, &rtsintrq, 0); } SYSINIT(rtsock, SI_SUB_PROTO_DOMAIN, SI_ORDER_THIRD, rts_init, 0); static void rts_input(struct mbuf *m) { struct sockproto route_proto; unsigned short *family; struct m_tag *tag; route_proto.sp_family = PF_ROUTE; tag = m_tag_find(m, PACKET_TAG_RTSOCKFAM, NULL); if (tag != NULL) { family = (unsigned short *)(tag + 1); route_proto.sp_protocol = *family; m_tag_delete(m, tag); } else route_proto.sp_protocol = 0; raw_input(m, &route_proto, &route_src); } /* * It really doesn't make any sense at all for this code to share much * with raw_usrreq.c, since its functionality is so restricted. XXX */ static void rts_abort(struct socket *so) { raw_usrreqs.pru_abort(so); } static void rts_close(struct socket *so) { raw_usrreqs.pru_close(so); } /* pru_accept is EOPNOTSUPP */ static int rts_attach(struct socket *so, int proto, struct thread *td) { struct rawcb *rp; int s, error; KASSERT(so->so_pcb == NULL, ("rts_attach: so_pcb != NULL")); /* XXX */ rp = malloc(sizeof *rp, M_PCB, M_WAITOK | M_ZERO); if (rp == NULL) return ENOBUFS; /* * The splnet() is necessary to block protocols from sending * error notifications (like RTM_REDIRECT or RTM_LOSING) while * this PCB is extant but incompletely initialized. * Probably we should try to do more of this work beforehand and * eliminate the spl. */ s = splnet(); so->so_pcb = (caddr_t)rp; so->so_fibnum = td->td_proc->p_fibnum; error = raw_attach(so, proto); rp = sotorawcb(so); if (error) { splx(s); so->so_pcb = NULL; free(rp, M_PCB); return error; } RTSOCK_LOCK(); switch(rp->rcb_proto.sp_protocol) { case AF_INET: route_cb.ip_count++; break; case AF_INET6: route_cb.ip6_count++; break; case AF_IPX: route_cb.ipx_count++; break; } route_cb.any_count++; RTSOCK_UNLOCK(); soisconnected(so); so->so_options |= SO_USELOOPBACK; splx(s); return 0; } static int rts_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { return (raw_usrreqs.pru_bind(so, nam, td)); /* xxx just EINVAL */ } static int rts_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { return (raw_usrreqs.pru_connect(so, nam, td)); /* XXX just EINVAL */ } /* pru_connect2 is EOPNOTSUPP */ /* pru_control is EOPNOTSUPP */ static void rts_detach(struct socket *so) { struct rawcb *rp = sotorawcb(so); KASSERT(rp != NULL, ("rts_detach: rp == NULL")); RTSOCK_LOCK(); switch(rp->rcb_proto.sp_protocol) { case AF_INET: route_cb.ip_count--; break; case AF_INET6: route_cb.ip6_count--; break; case AF_IPX: route_cb.ipx_count--; break; } route_cb.any_count--; RTSOCK_UNLOCK(); raw_usrreqs.pru_detach(so); } static int rts_disconnect(struct socket *so) { return (raw_usrreqs.pru_disconnect(so)); } /* pru_listen is EOPNOTSUPP */ static int rts_peeraddr(struct socket *so, struct sockaddr **nam) { return (raw_usrreqs.pru_peeraddr(so, nam)); } /* pru_rcvd is EOPNOTSUPP */ /* pru_rcvoob is EOPNOTSUPP */ static int rts_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, struct mbuf *control, struct thread *td) { return (raw_usrreqs.pru_send(so, flags, m, nam, control, td)); } /* pru_sense is null */ static int rts_shutdown(struct socket *so) { return (raw_usrreqs.pru_shutdown(so)); } static int rts_sockaddr(struct socket *so, struct sockaddr **nam) { return (raw_usrreqs.pru_sockaddr(so, nam)); } static struct pr_usrreqs route_usrreqs = { .pru_abort = rts_abort, .pru_attach = rts_attach, .pru_bind = rts_bind, .pru_connect = rts_connect, .pru_detach = rts_detach, .pru_disconnect = rts_disconnect, .pru_peeraddr = rts_peeraddr, .pru_send = rts_send, .pru_shutdown = rts_shutdown, .pru_sockaddr = rts_sockaddr, .pru_close = rts_close, }; #ifndef _SOCKADDR_UNION_DEFINED #define _SOCKADDR_UNION_DEFINED /* * The union of all possible address formats we handle. */ union sockaddr_union { struct sockaddr sa; struct sockaddr_in sin; struct sockaddr_in6 sin6; }; #endif /* _SOCKADDR_UNION_DEFINED */ static int rtm_get_jailed(struct rt_addrinfo *info, struct ifnet *ifp, struct rtentry *rt, union sockaddr_union *saun, struct ucred *cred) { switch (info->rti_info[RTAX_DST]->sa_family) { #ifdef INET case AF_INET: { struct in_addr ia; /* * 1. Check if the returned address is part of the jail. */ ia = ((struct sockaddr_in *)rt->rt_ifa->ifa_addr)->sin_addr; - if (prison_check_ip4(cred, &ia) != 0) { + if (prison_check_ip4(cred, &ia) == 0) { info->rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr; } else { struct ifaddr *ifa; int found; found = 0; /* * 2. Try to find an address on the given outgoing * interface that belongs to the jail. */ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { struct sockaddr *sa; sa = ifa->ifa_addr; if (sa->sa_family != AF_INET) continue; ia = ((struct sockaddr_in *)sa)->sin_addr; - if (prison_check_ip4(cred, &ia) != 0) { + if (prison_check_ip4(cred, &ia) == 0) { found = 1; break; } } if (!found) { /* * 3. As a last resort return the 'default' * jail address. */ if (prison_get_ip4(cred, &ia) != 0) return (ESRCH); } bzero(&saun->sin, sizeof(struct sockaddr_in)); saun->sin.sin_len = sizeof(struct sockaddr_in); saun->sin.sin_family = AF_INET; saun->sin.sin_addr.s_addr = ia.s_addr; info->rti_info[RTAX_IFA] = (struct sockaddr *)&saun->sin; } break; } #endif #ifdef INET6 case AF_INET6: { struct in6_addr ia6; /* * 1. Check if the returned address is part of the jail. */ bcopy(&((struct sockaddr_in6 *)rt->rt_ifa->ifa_addr)->sin6_addr, &ia6, sizeof(struct in6_addr)); - if (prison_check_ip6(cred, &ia6) != 0) { + if (prison_check_ip6(cred, &ia6) == 0) { info->rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr; } else { struct ifaddr *ifa; int found; found = 0; /* * 2. Try to find an address on the given outgoing * interface that belongs to the jail. */ TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { struct sockaddr *sa; sa = ifa->ifa_addr; if (sa->sa_family != AF_INET6) continue; bcopy(&((struct sockaddr_in6 *)sa)->sin6_addr, &ia6, sizeof(struct in6_addr)); - if (prison_check_ip6(cred, &ia6) != 0) { + if (prison_check_ip6(cred, &ia6) == 0) { found = 1; break; } } if (!found) { /* * 3. As a last resort return the 'default' * jail address. */ if (prison_get_ip6(cred, &ia6) != 0) return (ESRCH); } bzero(&saun->sin6, sizeof(struct sockaddr_in6)); saun->sin6.sin6_len = sizeof(struct sockaddr_in6); saun->sin6.sin6_family = AF_INET6; bcopy(&ia6, &saun->sin6.sin6_addr, sizeof(struct in6_addr)); if (sa6_recoverscope(&saun->sin6) != 0) return (ESRCH); info->rti_info[RTAX_IFA] = (struct sockaddr *)&saun->sin6; } break; } #endif default: return (ESRCH); } return (0); } /*ARGSUSED*/ static int route_output(struct mbuf *m, struct socket *so) { #define sa_equal(a1, a2) (bcmp((a1), (a2), (a1)->sa_len) == 0) INIT_VNET_NET(so->so_vnet); struct rt_msghdr *rtm = NULL; struct rtentry *rt = NULL; struct radix_node_head *rnh; struct rt_addrinfo info; int len, error = 0; struct ifnet *ifp = NULL; union sockaddr_union saun; #define senderr(e) { error = e; goto flush;} if (m == NULL || ((m->m_len < sizeof(long)) && (m = m_pullup(m, sizeof(long))) == NULL)) return (ENOBUFS); if ((m->m_flags & M_PKTHDR) == 0) panic("route_output"); len = m->m_pkthdr.len; if (len < sizeof(*rtm) || len != mtod(m, struct rt_msghdr *)->rtm_msglen) { info.rti_info[RTAX_DST] = NULL; senderr(EINVAL); } R_Malloc(rtm, struct rt_msghdr *, len); if (rtm == NULL) { info.rti_info[RTAX_DST] = NULL; senderr(ENOBUFS); } m_copydata(m, 0, len, (caddr_t)rtm); if (rtm->rtm_version != RTM_VERSION) { info.rti_info[RTAX_DST] = NULL; senderr(EPROTONOSUPPORT); } rtm->rtm_pid = curproc->p_pid; bzero(&info, sizeof(info)); info.rti_addrs = rtm->rtm_addrs; if (rt_xaddrs((caddr_t)(rtm + 1), len + (caddr_t)rtm, &info)) { info.rti_info[RTAX_DST] = NULL; senderr(EINVAL); } info.rti_flags = rtm->rtm_flags; if (info.rti_info[RTAX_DST] == NULL || info.rti_info[RTAX_DST]->sa_family >= AF_MAX || (info.rti_info[RTAX_GATEWAY] != NULL && info.rti_info[RTAX_GATEWAY]->sa_family >= AF_MAX)) senderr(EINVAL); /* * Verify that the caller has the appropriate privilege; RTM_GET * is the only operation the non-superuser is allowed. */ if (rtm->rtm_type != RTM_GET) { error = priv_check(curthread, PRIV_NET_ROUTE); if (error) senderr(error); } switch (rtm->rtm_type) { struct rtentry *saved_nrt; case RTM_ADD: if (info.rti_info[RTAX_GATEWAY] == NULL) senderr(EINVAL); saved_nrt = NULL; /* support for new ARP code */ if (info.rti_info[RTAX_GATEWAY]->sa_family == AF_LINK && (rtm->rtm_flags & RTF_LLDATA) != 0) { error = lla_rt_output(rtm, &info); break; } error = rtrequest1_fib(RTM_ADD, &info, &saved_nrt, so->so_fibnum); if (error == 0 && saved_nrt) { RT_LOCK(saved_nrt); rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &saved_nrt->rt_rmx); rtm->rtm_index = saved_nrt->rt_ifp->if_index; RT_REMREF(saved_nrt); RT_UNLOCK(saved_nrt); } break; case RTM_DELETE: saved_nrt = NULL; /* support for new ARP code */ if (info.rti_info[RTAX_GATEWAY] && (info.rti_info[RTAX_GATEWAY]->sa_family == AF_LINK) && (rtm->rtm_flags & RTF_LLDATA) != 0) { error = lla_rt_output(rtm, &info); break; } error = rtrequest1_fib(RTM_DELETE, &info, &saved_nrt, so->so_fibnum); if (error == 0) { RT_LOCK(saved_nrt); rt = saved_nrt; goto report; } break; case RTM_GET: case RTM_CHANGE: case RTM_LOCK: rnh = V_rt_tables[so->so_fibnum][info.rti_info[RTAX_DST]->sa_family]; if (rnh == NULL) senderr(EAFNOSUPPORT); RADIX_NODE_HEAD_RLOCK(rnh); rt = (struct rtentry *) rnh->rnh_lookup(info.rti_info[RTAX_DST], info.rti_info[RTAX_NETMASK], rnh); if (rt == NULL) { /* XXX looks bogus */ RADIX_NODE_HEAD_RUNLOCK(rnh); senderr(ESRCH); } #ifdef RADIX_MPATH /* * for RTM_CHANGE/LOCK, if we got multipath routes, * we require users to specify a matching RTAX_GATEWAY. * * for RTM_GET, gate is optional even with multipath. * if gate == NULL the first match is returned. * (no need to call rt_mpath_matchgate if gate == NULL) */ if (rn_mpath_capable(rnh) && (rtm->rtm_type != RTM_GET || info.rti_info[RTAX_GATEWAY])) { rt = rt_mpath_matchgate(rt, info.rti_info[RTAX_GATEWAY]); if (!rt) { RADIX_NODE_HEAD_RUNLOCK(rnh); senderr(ESRCH); } } #endif RT_LOCK(rt); RT_ADDREF(rt); RADIX_NODE_HEAD_RUNLOCK(rnh); /* * Fix for PR: 82974 * * RTM_CHANGE/LOCK need a perfect match, rn_lookup() * returns a perfect match in case a netmask is * specified. For host routes only a longest prefix * match is returned so it is necessary to compare the * existence of the netmask. If both have a netmask * rnh_lookup() did a perfect match and if none of them * have a netmask both are host routes which is also a * perfect match. */ if (rtm->rtm_type != RTM_GET && (!rt_mask(rt) != !info.rti_info[RTAX_NETMASK])) { RT_UNLOCK(rt); senderr(ESRCH); } switch(rtm->rtm_type) { case RTM_GET: report: RT_LOCK_ASSERT(rt); - if (jailed(curthread->td_ucred) && - ((rt->rt_flags & RTF_HOST) == 0 || - !prison_if(curthread->td_ucred, rt_key(rt)))) { + if ((rt->rt_flags & RTF_HOST) == 0 + ? jailed(curthread->td_ucred) + : prison_if(curthread->td_ucred, + rt_key(rt)) != 0) { RT_UNLOCK(rt); senderr(ESRCH); } info.rti_info[RTAX_DST] = rt_key(rt); info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; info.rti_info[RTAX_NETMASK] = rt_mask(rt); info.rti_info[RTAX_GENMASK] = 0; if (rtm->rtm_addrs & (RTA_IFP | RTA_IFA)) { ifp = rt->rt_ifp; if (ifp) { info.rti_info[RTAX_IFP] = ifp->if_addr->ifa_addr; if (jailed(curthread->td_ucred)) { error = rtm_get_jailed( &info, ifp, rt, &saun, curthread->td_ucred); if (error != 0) { RT_UNLOCK(rt); senderr(ESRCH); } } else { info.rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr; } if (ifp->if_flags & IFF_POINTOPOINT) info.rti_info[RTAX_BRD] = rt->rt_ifa->ifa_dstaddr; rtm->rtm_index = ifp->if_index; } else { info.rti_info[RTAX_IFP] = NULL; info.rti_info[RTAX_IFA] = NULL; } } else if ((ifp = rt->rt_ifp) != NULL) { rtm->rtm_index = ifp->if_index; } len = rt_msg2(rtm->rtm_type, &info, NULL, NULL); if (len > rtm->rtm_msglen) { struct rt_msghdr *new_rtm; R_Malloc(new_rtm, struct rt_msghdr *, len); if (new_rtm == NULL) { RT_UNLOCK(rt); senderr(ENOBUFS); } bcopy(rtm, new_rtm, rtm->rtm_msglen); Free(rtm); rtm = new_rtm; } (void)rt_msg2(rtm->rtm_type, &info, (caddr_t)rtm, NULL); rtm->rtm_flags = rt->rt_flags; rtm->rtm_use = 0; rt_getmetrics(&rt->rt_rmx, &rtm->rtm_rmx); rtm->rtm_addrs = info.rti_addrs; break; case RTM_CHANGE: /* * New gateway could require new ifaddr, ifp; * flags may also be different; ifp may be specified * by ll sockaddr when protocol address is ambiguous */ if (((rt->rt_flags & RTF_GATEWAY) && info.rti_info[RTAX_GATEWAY] != NULL) || info.rti_info[RTAX_IFP] != NULL || (info.rti_info[RTAX_IFA] != NULL && !sa_equal(info.rti_info[RTAX_IFA], rt->rt_ifa->ifa_addr))) { RT_UNLOCK(rt); RADIX_NODE_HEAD_LOCK(rnh); error = rt_getifa_fib(&info, rt->rt_fibnum); RADIX_NODE_HEAD_UNLOCK(rnh); if (error != 0) senderr(error); RT_LOCK(rt); } if (info.rti_ifa != NULL && info.rti_ifa != rt->rt_ifa && rt->rt_ifa != NULL && rt->rt_ifa->ifa_rtrequest != NULL) { rt->rt_ifa->ifa_rtrequest(RTM_DELETE, rt, &info); IFAFREE(rt->rt_ifa); } if (info.rti_info[RTAX_GATEWAY] != NULL) { RT_UNLOCK(rt); RADIX_NODE_HEAD_LOCK(rnh); RT_LOCK(rt); error = rt_setgate(rt, rt_key(rt), info.rti_info[RTAX_GATEWAY]); RADIX_NODE_HEAD_UNLOCK(rnh); if (error != 0) { RT_UNLOCK(rt); senderr(error); } rt->rt_flags |= RTF_GATEWAY; } if (info.rti_ifa != NULL && info.rti_ifa != rt->rt_ifa) { IFAREF(info.rti_ifa); rt->rt_ifa = info.rti_ifa; rt->rt_ifp = info.rti_ifp; } /* Allow some flags to be toggled on change. */ if (rtm->rtm_fmask & RTF_FMASK) rt->rt_flags = (rt->rt_flags & ~rtm->rtm_fmask) | (rtm->rtm_flags & rtm->rtm_fmask); rt_setmetrics(rtm->rtm_inits, &rtm->rtm_rmx, &rt->rt_rmx); rtm->rtm_index = rt->rt_ifp->if_index; if (rt->rt_ifa && rt->rt_ifa->ifa_rtrequest) rt->rt_ifa->ifa_rtrequest(RTM_ADD, rt, &info); /* FALLTHROUGH */ case RTM_LOCK: /* We don't support locks anymore */ break; } RT_UNLOCK(rt); break; default: senderr(EOPNOTSUPP); } flush: if (rtm) { if (error) rtm->rtm_errno = error; else rtm->rtm_flags |= RTF_DONE; } if (rt) /* XXX can this be true? */ RTFREE(rt); { struct rawcb *rp = NULL; /* * Check to see if we don't want our own messages. */ if ((so->so_options & SO_USELOOPBACK) == 0) { if (route_cb.any_count <= 1) { if (rtm) Free(rtm); m_freem(m); return (error); } /* There is another listener, so construct message */ rp = sotorawcb(so); } if (rtm) { m_copyback(m, 0, rtm->rtm_msglen, (caddr_t)rtm); if (m->m_pkthdr.len < rtm->rtm_msglen) { m_freem(m); m = NULL; } else if (m->m_pkthdr.len > rtm->rtm_msglen) m_adj(m, rtm->rtm_msglen - m->m_pkthdr.len); Free(rtm); } if (m) { if (rp) { /* * XXX insure we don't get a copy by * invalidating our protocol */ unsigned short family = rp->rcb_proto.sp_family; rp->rcb_proto.sp_family = 0; rt_dispatch(m, info.rti_info[RTAX_DST]); rp->rcb_proto.sp_family = family; } else rt_dispatch(m, info.rti_info[RTAX_DST]); } } return (error); #undef sa_equal } static void rt_setmetrics(u_long which, const struct rt_metrics *in, struct rt_metrics_lite *out) { #define metric(f, e) if (which & (f)) out->e = in->e; /* * Only these are stored in the routing entry since introduction * of tcp hostcache. The rest is ignored. */ metric(RTV_MTU, rmx_mtu); /* Userland -> kernel timebase conversion. */ if (which & RTV_EXPIRE) out->rmx_expire = in->rmx_expire ? in->rmx_expire - time_second + time_uptime : 0; #undef metric } static void rt_getmetrics(const struct rt_metrics_lite *in, struct rt_metrics *out) { #define metric(e) out->e = in->e; bzero(out, sizeof(*out)); metric(rmx_mtu); /* Kernel -> userland timebase conversion. */ out->rmx_expire = in->rmx_expire ? in->rmx_expire - time_uptime + time_second : 0; #undef metric } /* * Extract the addresses of the passed sockaddrs. * Do a little sanity checking so as to avoid bad memory references. * This data is derived straight from userland. */ static int rt_xaddrs(caddr_t cp, caddr_t cplim, struct rt_addrinfo *rtinfo) { struct sockaddr *sa; int i; for (i = 0; i < RTAX_MAX && cp < cplim; i++) { if ((rtinfo->rti_addrs & (1 << i)) == 0) continue; sa = (struct sockaddr *)cp; /* * It won't fit. */ if (cp + sa->sa_len > cplim) return (EINVAL); /* * there are no more.. quit now * If there are more bits, they are in error. * I've seen this. route(1) can evidently generate these. * This causes kernel to core dump. * for compatibility, If we see this, point to a safe address. */ if (sa->sa_len == 0) { rtinfo->rti_info[i] = &sa_zero; return (0); /* should be EINVAL but for compat */ } /* accept it */ rtinfo->rti_info[i] = sa; cp += SA_SIZE(sa); } return (0); } static struct mbuf * rt_msg1(int type, struct rt_addrinfo *rtinfo) { struct rt_msghdr *rtm; struct mbuf *m; int i; struct sockaddr *sa; int len, dlen; switch (type) { case RTM_DELADDR: case RTM_NEWADDR: len = sizeof(struct ifa_msghdr); break; case RTM_DELMADDR: case RTM_NEWMADDR: len = sizeof(struct ifma_msghdr); break; case RTM_IFINFO: len = sizeof(struct if_msghdr); break; case RTM_IFANNOUNCE: case RTM_IEEE80211: len = sizeof(struct if_announcemsghdr); break; default: len = sizeof(struct rt_msghdr); } if (len > MCLBYTES) panic("rt_msg1"); m = m_gethdr(M_DONTWAIT, MT_DATA); if (m && len > MHLEN) { MCLGET(m, M_DONTWAIT); if ((m->m_flags & M_EXT) == 0) { m_free(m); m = NULL; } } if (m == NULL) return (m); m->m_pkthdr.len = m->m_len = len; m->m_pkthdr.rcvif = NULL; rtm = mtod(m, struct rt_msghdr *); bzero((caddr_t)rtm, len); for (i = 0; i < RTAX_MAX; i++) { if ((sa = rtinfo->rti_info[i]) == NULL) continue; rtinfo->rti_addrs |= (1 << i); dlen = SA_SIZE(sa); m_copyback(m, len, dlen, (caddr_t)sa); len += dlen; } if (m->m_pkthdr.len != len) { m_freem(m); return (NULL); } rtm->rtm_msglen = len; rtm->rtm_version = RTM_VERSION; rtm->rtm_type = type; return (m); } static int rt_msg2(int type, struct rt_addrinfo *rtinfo, caddr_t cp, struct walkarg *w) { int i; int len, dlen, second_time = 0; caddr_t cp0; rtinfo->rti_addrs = 0; again: switch (type) { case RTM_DELADDR: case RTM_NEWADDR: len = sizeof(struct ifa_msghdr); break; case RTM_IFINFO: len = sizeof(struct if_msghdr); break; case RTM_NEWMADDR: len = sizeof(struct ifma_msghdr); break; default: len = sizeof(struct rt_msghdr); } cp0 = cp; if (cp0) cp += len; for (i = 0; i < RTAX_MAX; i++) { struct sockaddr *sa; if ((sa = rtinfo->rti_info[i]) == NULL) continue; rtinfo->rti_addrs |= (1 << i); dlen = SA_SIZE(sa); if (cp) { bcopy((caddr_t)sa, cp, (unsigned)dlen); cp += dlen; } len += dlen; } len = ALIGN(len); if (cp == NULL && w != NULL && !second_time) { struct walkarg *rw = w; if (rw->w_req) { if (rw->w_tmemsize < len) { if (rw->w_tmem) free(rw->w_tmem, M_RTABLE); rw->w_tmem = (caddr_t) malloc(len, M_RTABLE, M_NOWAIT); if (rw->w_tmem) rw->w_tmemsize = len; } if (rw->w_tmem) { cp = rw->w_tmem; second_time = 1; goto again; } } } if (cp) { struct rt_msghdr *rtm = (struct rt_msghdr *)cp0; rtm->rtm_version = RTM_VERSION; rtm->rtm_type = type; rtm->rtm_msglen = len; } return (len); } /* * This routine is called to generate a message from the routing * socket indicating that a redirect has occured, a routing lookup * has failed, or that a protocol has detected timeouts to a particular * destination. */ void rt_missmsg(int type, struct rt_addrinfo *rtinfo, int flags, int error) { struct rt_msghdr *rtm; struct mbuf *m; struct sockaddr *sa = rtinfo->rti_info[RTAX_DST]; if (route_cb.any_count == 0) return; m = rt_msg1(type, rtinfo); if (m == NULL) return; rtm = mtod(m, struct rt_msghdr *); rtm->rtm_flags = RTF_DONE | flags; rtm->rtm_errno = error; rtm->rtm_addrs = rtinfo->rti_addrs; rt_dispatch(m, sa); } /* * This routine is called to generate a message from the routing * socket indicating that the status of a network interface has changed. */ void rt_ifmsg(struct ifnet *ifp) { struct if_msghdr *ifm; struct mbuf *m; struct rt_addrinfo info; if (route_cb.any_count == 0) return; bzero((caddr_t)&info, sizeof(info)); m = rt_msg1(RTM_IFINFO, &info); if (m == NULL) return; ifm = mtod(m, struct if_msghdr *); ifm->ifm_index = ifp->if_index; ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags; ifm->ifm_data = ifp->if_data; ifm->ifm_addrs = 0; rt_dispatch(m, NULL); } /* * This is called to generate messages from the routing socket * indicating a network interface has had addresses associated with it. * if we ever reverse the logic and replace messages TO the routing * socket indicate a request to configure interfaces, then it will * be unnecessary as the routing socket will automatically generate * copies of it. */ void rt_newaddrmsg(int cmd, struct ifaddr *ifa, int error, struct rtentry *rt) { struct rt_addrinfo info; struct sockaddr *sa = NULL; int pass; struct mbuf *m = NULL; struct ifnet *ifp = ifa->ifa_ifp; KASSERT(cmd == RTM_ADD || cmd == RTM_DELETE, ("unexpected cmd %u", cmd)); #ifdef SCTP /* * notify the SCTP stack * this will only get called when an address is added/deleted * XXX pass the ifaddr struct instead if ifa->ifa_addr... */ sctp_addr_change(ifa, cmd); #endif /* SCTP */ if (route_cb.any_count == 0) return; for (pass = 1; pass < 3; pass++) { bzero((caddr_t)&info, sizeof(info)); if ((cmd == RTM_ADD && pass == 1) || (cmd == RTM_DELETE && pass == 2)) { struct ifa_msghdr *ifam; int ncmd = cmd == RTM_ADD ? RTM_NEWADDR : RTM_DELADDR; info.rti_info[RTAX_IFA] = sa = ifa->ifa_addr; info.rti_info[RTAX_IFP] = ifp->if_addr->ifa_addr; info.rti_info[RTAX_NETMASK] = ifa->ifa_netmask; info.rti_info[RTAX_BRD] = ifa->ifa_dstaddr; if ((m = rt_msg1(ncmd, &info)) == NULL) continue; ifam = mtod(m, struct ifa_msghdr *); ifam->ifam_index = ifp->if_index; ifam->ifam_metric = ifa->ifa_metric; ifam->ifam_flags = ifa->ifa_flags; ifam->ifam_addrs = info.rti_addrs; } if ((cmd == RTM_ADD && pass == 2) || (cmd == RTM_DELETE && pass == 1)) { struct rt_msghdr *rtm; if (rt == NULL) continue; info.rti_info[RTAX_NETMASK] = rt_mask(rt); info.rti_info[RTAX_DST] = sa = rt_key(rt); info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; if ((m = rt_msg1(cmd, &info)) == NULL) continue; rtm = mtod(m, struct rt_msghdr *); rtm->rtm_index = ifp->if_index; rtm->rtm_flags |= rt->rt_flags; rtm->rtm_errno = error; rtm->rtm_addrs = info.rti_addrs; } rt_dispatch(m, sa); } } /* * This is the analogue to the rt_newaddrmsg which performs the same * function but for multicast group memberhips. This is easier since * there is no route state to worry about. */ void rt_newmaddrmsg(int cmd, struct ifmultiaddr *ifma) { struct rt_addrinfo info; struct mbuf *m = NULL; struct ifnet *ifp = ifma->ifma_ifp; struct ifma_msghdr *ifmam; if (route_cb.any_count == 0) return; bzero((caddr_t)&info, sizeof(info)); info.rti_info[RTAX_IFA] = ifma->ifma_addr; info.rti_info[RTAX_IFP] = ifp ? ifp->if_addr->ifa_addr : NULL; /* * If a link-layer address is present, present it as a ``gateway'' * (similarly to how ARP entries, e.g., are presented). */ info.rti_info[RTAX_GATEWAY] = ifma->ifma_lladdr; m = rt_msg1(cmd, &info); if (m == NULL) return; ifmam = mtod(m, struct ifma_msghdr *); KASSERT(ifp != NULL, ("%s: link-layer multicast address w/o ifp\n", __func__)); ifmam->ifmam_index = ifp->if_index; ifmam->ifmam_addrs = info.rti_addrs; rt_dispatch(m, ifma->ifma_addr); } static struct mbuf * rt_makeifannouncemsg(struct ifnet *ifp, int type, int what, struct rt_addrinfo *info) { struct if_announcemsghdr *ifan; struct mbuf *m; if (route_cb.any_count == 0) return NULL; bzero((caddr_t)info, sizeof(*info)); m = rt_msg1(type, info); if (m != NULL) { ifan = mtod(m, struct if_announcemsghdr *); ifan->ifan_index = ifp->if_index; strlcpy(ifan->ifan_name, ifp->if_xname, sizeof(ifan->ifan_name)); ifan->ifan_what = what; } return m; } /* * This is called to generate routing socket messages indicating * IEEE80211 wireless events. * XXX we piggyback on the RTM_IFANNOUNCE msg format in a clumsy way. */ void rt_ieee80211msg(struct ifnet *ifp, int what, void *data, size_t data_len) { struct mbuf *m; struct rt_addrinfo info; m = rt_makeifannouncemsg(ifp, RTM_IEEE80211, what, &info); if (m != NULL) { /* * Append the ieee80211 data. Try to stick it in the * mbuf containing the ifannounce msg; otherwise allocate * a new mbuf and append. * * NB: we assume m is a single mbuf. */ if (data_len > M_TRAILINGSPACE(m)) { struct mbuf *n = m_get(M_NOWAIT, MT_DATA); if (n == NULL) { m_freem(m); return; } bcopy(data, mtod(n, void *), data_len); n->m_len = data_len; m->m_next = n; } else if (data_len > 0) { bcopy(data, mtod(m, u_int8_t *) + m->m_len, data_len); m->m_len += data_len; } if (m->m_flags & M_PKTHDR) m->m_pkthdr.len += data_len; mtod(m, struct if_announcemsghdr *)->ifan_msglen += data_len; rt_dispatch(m, NULL); } } /* * This is called to generate routing socket messages indicating * network interface arrival and departure. */ void rt_ifannouncemsg(struct ifnet *ifp, int what) { struct mbuf *m; struct rt_addrinfo info; m = rt_makeifannouncemsg(ifp, RTM_IFANNOUNCE, what, &info); if (m != NULL) rt_dispatch(m, NULL); } static void rt_dispatch(struct mbuf *m, const struct sockaddr *sa) { INIT_VNET_NET(curvnet); struct m_tag *tag; /* * Preserve the family from the sockaddr, if any, in an m_tag for * use when injecting the mbuf into the routing socket buffer from * the netisr. */ if (sa != NULL) { tag = m_tag_get(PACKET_TAG_RTSOCKFAM, sizeof(unsigned short), M_NOWAIT); if (tag == NULL) { m_freem(m); return; } *(unsigned short *)(tag + 1) = sa->sa_family; m_tag_prepend(m, tag); } netisr_queue(NETISR_ROUTE, m); /* mbuf is free'd on failure. */ } /* * This is used in dumping the kernel table via sysctl(). */ static int sysctl_dumpentry(struct radix_node *rn, void *vw) { struct walkarg *w = vw; struct rtentry *rt = (struct rtentry *)rn; int error = 0, size; struct rt_addrinfo info; if (w->w_op == NET_RT_FLAGS && !(rt->rt_flags & w->w_arg)) return 0; - if (jailed(w->w_req->td->td_ucred) && - ((rt->rt_flags & RTF_HOST) == 0 || - !prison_if(w->w_req->td->td_ucred, rt_key(rt)))) + if ((rt->rt_flags & RTF_HOST) == 0 + ? jailed(w->w_req->td->td_ucred) + : prison_if(w->w_req->td->td_ucred, rt_key(rt)) != 0) return (0); bzero((caddr_t)&info, sizeof(info)); info.rti_info[RTAX_DST] = rt_key(rt); info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; info.rti_info[RTAX_NETMASK] = rt_mask(rt); info.rti_info[RTAX_GENMASK] = 0; if (rt->rt_ifp) { info.rti_info[RTAX_IFP] = rt->rt_ifp->if_addr->ifa_addr; info.rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr; if (rt->rt_ifp->if_flags & IFF_POINTOPOINT) info.rti_info[RTAX_BRD] = rt->rt_ifa->ifa_dstaddr; } size = rt_msg2(RTM_GET, &info, NULL, w); if (w->w_req && w->w_tmem) { struct rt_msghdr *rtm = (struct rt_msghdr *)w->w_tmem; rtm->rtm_flags = rt->rt_flags; rtm->rtm_use = rt->rt_rmx.rmx_pksent; rt_getmetrics(&rt->rt_rmx, &rtm->rtm_rmx); rtm->rtm_index = rt->rt_ifp->if_index; rtm->rtm_errno = rtm->rtm_pid = rtm->rtm_seq = 0; rtm->rtm_addrs = info.rti_addrs; error = SYSCTL_OUT(w->w_req, (caddr_t)rtm, size); return (error); } return (error); } static int sysctl_iflist(int af, struct walkarg *w) { INIT_VNET_NET(curvnet); struct ifnet *ifp; struct ifaddr *ifa; struct rt_addrinfo info; int len, error = 0; bzero((caddr_t)&info, sizeof(info)); IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { if (w->w_arg && w->w_arg != ifp->if_index) continue; ifa = ifp->if_addr; info.rti_info[RTAX_IFP] = ifa->ifa_addr; len = rt_msg2(RTM_IFINFO, &info, NULL, w); info.rti_info[RTAX_IFP] = NULL; if (w->w_req && w->w_tmem) { struct if_msghdr *ifm; ifm = (struct if_msghdr *)w->w_tmem; ifm->ifm_index = ifp->if_index; ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags; ifm->ifm_data = ifp->if_data; ifm->ifm_addrs = info.rti_addrs; error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len); if (error) goto done; } while ((ifa = TAILQ_NEXT(ifa, ifa_link)) != NULL) { if (af && af != ifa->ifa_addr->sa_family) continue; - if (jailed(w->w_req->td->td_ucred) && - !prison_if(w->w_req->td->td_ucred, ifa->ifa_addr)) + if (prison_if(w->w_req->td->td_ucred, + ifa->ifa_addr) != 0) continue; info.rti_info[RTAX_IFA] = ifa->ifa_addr; info.rti_info[RTAX_NETMASK] = ifa->ifa_netmask; info.rti_info[RTAX_BRD] = ifa->ifa_dstaddr; len = rt_msg2(RTM_NEWADDR, &info, NULL, w); if (w->w_req && w->w_tmem) { struct ifa_msghdr *ifam; ifam = (struct ifa_msghdr *)w->w_tmem; ifam->ifam_index = ifa->ifa_ifp->if_index; ifam->ifam_flags = ifa->ifa_flags; ifam->ifam_metric = ifa->ifa_metric; ifam->ifam_addrs = info.rti_addrs; error = SYSCTL_OUT(w->w_req, w->w_tmem, len); if (error) goto done; } } info.rti_info[RTAX_IFA] = info.rti_info[RTAX_NETMASK] = info.rti_info[RTAX_BRD] = NULL; } done: IFNET_RUNLOCK(); return (error); } static int sysctl_ifmalist(int af, struct walkarg *w) { INIT_VNET_NET(curvnet); struct ifnet *ifp; struct ifmultiaddr *ifma; struct rt_addrinfo info; int len, error = 0; struct ifaddr *ifa; bzero((caddr_t)&info, sizeof(info)); IFNET_RLOCK(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { if (w->w_arg && w->w_arg != ifp->if_index) continue; ifa = ifp->if_addr; info.rti_info[RTAX_IFP] = ifa ? ifa->ifa_addr : NULL; IF_ADDR_LOCK(ifp); TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) { if (af && af != ifma->ifma_addr->sa_family) continue; - if (jailed(w->w_req->td->td_ucred) && - !prison_if(w->w_req->td->td_ucred, ifma->ifma_addr)) + if (prison_if(w->w_req->td->td_ucred, + ifma->ifma_addr) != 0) continue; info.rti_info[RTAX_IFA] = ifma->ifma_addr; info.rti_info[RTAX_GATEWAY] = (ifma->ifma_addr->sa_family != AF_LINK) ? ifma->ifma_lladdr : NULL; len = rt_msg2(RTM_NEWMADDR, &info, NULL, w); if (w->w_req && w->w_tmem) { struct ifma_msghdr *ifmam; ifmam = (struct ifma_msghdr *)w->w_tmem; ifmam->ifmam_index = ifma->ifma_ifp->if_index; ifmam->ifmam_flags = 0; ifmam->ifmam_addrs = info.rti_addrs; error = SYSCTL_OUT(w->w_req, w->w_tmem, len); if (error) { IF_ADDR_UNLOCK(ifp); goto done; } } } IF_ADDR_UNLOCK(ifp); } done: IFNET_RUNLOCK(); return (error); } static int sysctl_rtsock(SYSCTL_HANDLER_ARGS) { INIT_VNET_NET(curvnet); int *name = (int *)arg1; u_int namelen = arg2; struct radix_node_head *rnh; int i, lim, error = EINVAL; u_char af; struct walkarg w; name ++; namelen--; if (req->newptr) return (EPERM); if (namelen != 3) return ((namelen < 3) ? EISDIR : ENOTDIR); af = name[0]; if (af > AF_MAX) return (EINVAL); bzero(&w, sizeof(w)); w.w_op = name[1]; w.w_arg = name[2]; w.w_req = req; error = sysctl_wire_old_buffer(req, 0); if (error) return (error); switch (w.w_op) { case NET_RT_DUMP: case NET_RT_FLAGS: if (af == 0) { /* dump all tables */ i = 1; lim = AF_MAX; } else /* dump only one table */ i = lim = af; /* * take care of llinfo entries, the caller must * specify an AF */ if (w.w_op == NET_RT_FLAGS && (w.w_arg == 0 || w.w_arg & RTF_LLINFO)) { if (af != 0) error = lltable_sysctl_dumparp(af, w.w_req); else error = EINVAL; break; } /* * take care of routing entries */ for (error = 0; error == 0 && i <= lim; i++) if ((rnh = V_rt_tables[req->td->td_proc->p_fibnum][i]) != NULL) { RADIX_NODE_HEAD_LOCK(rnh); error = rnh->rnh_walktree(rnh, sysctl_dumpentry, &w); RADIX_NODE_HEAD_UNLOCK(rnh); } else if (af != 0) error = EAFNOSUPPORT; break; case NET_RT_IFLIST: error = sysctl_iflist(af, &w); break; case NET_RT_IFMALIST: error = sysctl_ifmalist(af, &w); break; } if (w.w_tmem) free(w.w_tmem, M_RTABLE); return (error); } SYSCTL_NODE(_net, PF_ROUTE, routetable, CTLFLAG_RD, sysctl_rtsock, ""); /* * Definitions of protocols supported in the ROUTE domain. */ static struct domain routedomain; /* or at least forward */ static struct protosw routesw[] = { { .pr_type = SOCK_RAW, .pr_domain = &routedomain, .pr_flags = PR_ATOMIC|PR_ADDR, .pr_output = route_output, .pr_ctlinput = raw_ctlinput, .pr_init = raw_init, .pr_usrreqs = &route_usrreqs } }; static struct domain routedomain = { .dom_family = PF_ROUTE, .dom_name = "route", .dom_protosw = routesw, .dom_protoswNPROTOSW = &routesw[sizeof(routesw)/sizeof(routesw[0])] }; DOMAIN_SET(route); diff --git a/sys/netinet/in.c b/sys/netinet/in.c index 03c72697c123..bf1ebae81e85 100644 --- a/sys/netinet/in.c +++ b/sys/netinet/in.c @@ -1,1273 +1,1272 @@ /*- * Copyright (c) 1982, 1986, 1991, 1993 * The Regents of the University of California. All rights reserved. * Copyright (C) 2001 WIDE Project. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)in.c 8.4 (Berkeley) 1/9/95 */ #include __FBSDID("$FreeBSD$"); #include "opt_carp.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static int in_mask2len(struct in_addr *); static void in_len2mask(struct in_addr *, int); static int in_lifaddr_ioctl(struct socket *, u_long, caddr_t, struct ifnet *, struct thread *); static int in_addprefix(struct in_ifaddr *, int); static int in_scrubprefix(struct in_ifaddr *); static void in_socktrim(struct sockaddr_in *); static int in_ifinit(struct ifnet *, struct in_ifaddr *, struct sockaddr_in *, int); static void in_purgemaddrs(struct ifnet *); #ifdef VIMAGE_GLOBALS static int subnetsarelocal; static int sameprefixcarponly; extern struct inpcbinfo ripcbinfo; #endif SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip, OID_AUTO, subnets_are_local, CTLFLAG_RW, subnetsarelocal, 0, "Treat all subnets as directly connected"); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip, OID_AUTO, same_prefix_carp_only, CTLFLAG_RW, sameprefixcarponly, 0, "Refuse to create same prefixes on different interfaces"); /* * Return 1 if an internet address is for a ``local'' host * (one to which we have a connection). If subnetsarelocal * is true, this includes other subnets of the local net. * Otherwise, it includes only the directly-connected (sub)nets. */ int in_localaddr(struct in_addr in) { INIT_VNET_INET(curvnet); register u_long i = ntohl(in.s_addr); register struct in_ifaddr *ia; if (V_subnetsarelocal) { TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) if ((i & ia->ia_netmask) == ia->ia_net) return (1); } else { TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) if ((i & ia->ia_subnetmask) == ia->ia_subnet) return (1); } return (0); } /* * Return 1 if an internet address is for the local host and configured * on one of its interfaces. */ int in_localip(struct in_addr in) { INIT_VNET_INET(curvnet); struct in_ifaddr *ia; LIST_FOREACH(ia, INADDR_HASH(in.s_addr), ia_hash) { if (IA_SIN(ia)->sin_addr.s_addr == in.s_addr) return (1); } return (0); } /* * Determine whether an IP address is in a reserved set of addresses * that may not be forwarded, or whether datagrams to that destination * may be forwarded. */ int in_canforward(struct in_addr in) { register u_long i = ntohl(in.s_addr); register u_long net; if (IN_EXPERIMENTAL(i) || IN_MULTICAST(i) || IN_LINKLOCAL(i)) return (0); if (IN_CLASSA(i)) { net = i & IN_CLASSA_NET; if (net == 0 || net == (IN_LOOPBACKNET << IN_CLASSA_NSHIFT)) return (0); } return (1); } /* * Trim a mask in a sockaddr */ static void in_socktrim(struct sockaddr_in *ap) { register char *cplim = (char *) &ap->sin_addr; register char *cp = (char *) (&ap->sin_addr + 1); ap->sin_len = 0; while (--cp >= cplim) if (*cp) { (ap)->sin_len = cp - (char *) (ap) + 1; break; } } static int in_mask2len(mask) struct in_addr *mask; { int x, y; u_char *p; p = (u_char *)mask; for (x = 0; x < sizeof(*mask); x++) { if (p[x] != 0xff) break; } y = 0; if (x < sizeof(*mask)) { for (y = 0; y < 8; y++) { if ((p[x] & (0x80 >> y)) == 0) break; } } return (x * 8 + y); } static void in_len2mask(struct in_addr *mask, int len) { int i; u_char *p; p = (u_char *)mask; bzero(mask, sizeof(*mask)); for (i = 0; i < len / 8; i++) p[i] = 0xff; if (len % 8) p[i] = (0xff00 >> (len % 8)) & 0xff; } /* * Generic internet control operations (ioctl's). * Ifp is 0 if not an interface-specific ioctl. */ /* ARGSUSED */ int in_control(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp, struct thread *td) { INIT_VNET_INET(curvnet); /* both so and ifp can be NULL here! */ register struct ifreq *ifr = (struct ifreq *)data; register struct in_ifaddr *ia, *iap; register struct ifaddr *ifa; struct in_addr allhosts_addr; struct in_addr dst; struct in_ifaddr *oia; struct in_aliasreq *ifra = (struct in_aliasreq *)data; struct sockaddr_in oldaddr; int error, hostIsNew, iaIsNew, maskIsNew, s; int iaIsFirst; ia = NULL; iaIsFirst = 0; iaIsNew = 0; allhosts_addr.s_addr = htonl(INADDR_ALLHOSTS_GROUP); switch (cmd) { case SIOCALIFADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_ADDIFADDR); if (error) return (error); } if (ifp == NULL) return (EINVAL); return in_lifaddr_ioctl(so, cmd, data, ifp, td); case SIOCDLIFADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_DELIFADDR); if (error) return (error); } if (ifp == NULL) return (EINVAL); return in_lifaddr_ioctl(so, cmd, data, ifp, td); case SIOCGLIFADDR: if (ifp == NULL) return (EINVAL); return in_lifaddr_ioctl(so, cmd, data, ifp, td); } /* * Find address for this interface, if it exists. * * If an alias address was specified, find that one instead of * the first one on the interface, if possible. */ if (ifp != NULL) { dst = ((struct sockaddr_in *)&ifr->ifr_addr)->sin_addr; LIST_FOREACH(iap, INADDR_HASH(dst.s_addr), ia_hash) if (iap->ia_ifp == ifp && iap->ia_addr.sin_addr.s_addr == dst.s_addr) { if (td == NULL || prison_check_ip4( - td->td_ucred, &dst)) + td->td_ucred, &dst) == 0) ia = iap; break; } if (ia == NULL) TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { iap = ifatoia(ifa); if (iap->ia_addr.sin_family == AF_INET) { if (td != NULL && - !prison_check_ip4(td->td_ucred, - &iap->ia_addr.sin_addr)) + prison_check_ip4(td->td_ucred, + &iap->ia_addr.sin_addr) != 0) continue; ia = iap; break; } } if (ia == NULL) iaIsFirst = 1; } switch (cmd) { case SIOCAIFADDR: case SIOCDIFADDR: if (ifp == NULL) return (EADDRNOTAVAIL); if (ifra->ifra_addr.sin_family == AF_INET) { for (oia = ia; ia; ia = TAILQ_NEXT(ia, ia_link)) { if (ia->ia_ifp == ifp && ia->ia_addr.sin_addr.s_addr == ifra->ifra_addr.sin_addr.s_addr) break; } if ((ifp->if_flags & IFF_POINTOPOINT) && (cmd == SIOCAIFADDR) && (ifra->ifra_dstaddr.sin_addr.s_addr == INADDR_ANY)) { return (EDESTADDRREQ); } } if (cmd == SIOCDIFADDR && ia == NULL) return (EADDRNOTAVAIL); /* FALLTHROUGH */ case SIOCSIFADDR: case SIOCSIFNETMASK: case SIOCSIFDSTADDR: if (td != NULL) { error = priv_check(td, (cmd == SIOCDIFADDR) ? PRIV_NET_DELIFADDR : PRIV_NET_ADDIFADDR); if (error) return (error); } if (ifp == NULL) return (EADDRNOTAVAIL); if (ia == NULL) { ia = (struct in_ifaddr *) malloc(sizeof *ia, M_IFADDR, M_WAITOK | M_ZERO); if (ia == NULL) return (ENOBUFS); /* * Protect from ipintr() traversing address list * while we're modifying it. */ s = splnet(); ifa = &ia->ia_ifa; IFA_LOCK_INIT(ifa); ifa->ifa_addr = (struct sockaddr *)&ia->ia_addr; ifa->ifa_dstaddr = (struct sockaddr *)&ia->ia_dstaddr; ifa->ifa_netmask = (struct sockaddr *)&ia->ia_sockmask; ifa->ifa_refcnt = 1; TAILQ_INSERT_TAIL(&ifp->if_addrhead, ifa, ifa_link); ia->ia_sockmask.sin_len = 8; ia->ia_sockmask.sin_family = AF_INET; if (ifp->if_flags & IFF_BROADCAST) { ia->ia_broadaddr.sin_len = sizeof(ia->ia_addr); ia->ia_broadaddr.sin_family = AF_INET; } ia->ia_ifp = ifp; TAILQ_INSERT_TAIL(&V_in_ifaddrhead, ia, ia_link); splx(s); iaIsNew = 1; } break; case SIOCSIFBRDADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_ADDIFADDR); if (error) return (error); } /* FALLTHROUGH */ case SIOCGIFADDR: case SIOCGIFNETMASK: case SIOCGIFDSTADDR: case SIOCGIFBRDADDR: if (ia == NULL) return (EADDRNOTAVAIL); break; } switch (cmd) { case SIOCGIFADDR: *((struct sockaddr_in *)&ifr->ifr_addr) = ia->ia_addr; return (0); case SIOCGIFBRDADDR: if ((ifp->if_flags & IFF_BROADCAST) == 0) return (EINVAL); *((struct sockaddr_in *)&ifr->ifr_dstaddr) = ia->ia_broadaddr; return (0); case SIOCGIFDSTADDR: if ((ifp->if_flags & IFF_POINTOPOINT) == 0) return (EINVAL); *((struct sockaddr_in *)&ifr->ifr_dstaddr) = ia->ia_dstaddr; return (0); case SIOCGIFNETMASK: *((struct sockaddr_in *)&ifr->ifr_addr) = ia->ia_sockmask; return (0); case SIOCSIFDSTADDR: if ((ifp->if_flags & IFF_POINTOPOINT) == 0) return (EINVAL); oldaddr = ia->ia_dstaddr; ia->ia_dstaddr = *(struct sockaddr_in *)&ifr->ifr_dstaddr; if (ifp->if_ioctl != NULL) { IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, SIOCSIFDSTADDR, (caddr_t)ia); IFF_UNLOCKGIANT(ifp); if (error) { ia->ia_dstaddr = oldaddr; return (error); } } if (ia->ia_flags & IFA_ROUTE) { ia->ia_ifa.ifa_dstaddr = (struct sockaddr *)&oldaddr; rtinit(&(ia->ia_ifa), (int)RTM_DELETE, RTF_HOST); ia->ia_ifa.ifa_dstaddr = (struct sockaddr *)&ia->ia_dstaddr; rtinit(&(ia->ia_ifa), (int)RTM_ADD, RTF_HOST|RTF_UP); } return (0); case SIOCSIFBRDADDR: if ((ifp->if_flags & IFF_BROADCAST) == 0) return (EINVAL); ia->ia_broadaddr = *(struct sockaddr_in *)&ifr->ifr_broadaddr; return (0); case SIOCSIFADDR: error = in_ifinit(ifp, ia, (struct sockaddr_in *) &ifr->ifr_addr, 1); if (error != 0 && iaIsNew) break; if (error == 0) { if (iaIsFirst && (ifp->if_flags & IFF_MULTICAST) != 0) in_addmulti(&allhosts_addr, ifp); EVENTHANDLER_INVOKE(ifaddr_event, ifp); } return (0); case SIOCSIFNETMASK: ia->ia_sockmask.sin_addr = ifra->ifra_addr.sin_addr; ia->ia_subnetmask = ntohl(ia->ia_sockmask.sin_addr.s_addr); return (0); case SIOCAIFADDR: maskIsNew = 0; hostIsNew = 1; error = 0; if (ia->ia_addr.sin_family == AF_INET) { if (ifra->ifra_addr.sin_len == 0) { ifra->ifra_addr = ia->ia_addr; hostIsNew = 0; } else if (ifra->ifra_addr.sin_addr.s_addr == ia->ia_addr.sin_addr.s_addr) hostIsNew = 0; } if (ifra->ifra_mask.sin_len) { in_ifscrub(ifp, ia); ia->ia_sockmask = ifra->ifra_mask; ia->ia_sockmask.sin_family = AF_INET; ia->ia_subnetmask = ntohl(ia->ia_sockmask.sin_addr.s_addr); maskIsNew = 1; } if ((ifp->if_flags & IFF_POINTOPOINT) && (ifra->ifra_dstaddr.sin_family == AF_INET)) { in_ifscrub(ifp, ia); ia->ia_dstaddr = ifra->ifra_dstaddr; maskIsNew = 1; /* We lie; but the effect's the same */ } if (ifra->ifra_addr.sin_family == AF_INET && (hostIsNew || maskIsNew)) error = in_ifinit(ifp, ia, &ifra->ifra_addr, 0); if (error != 0 && iaIsNew) break; if ((ifp->if_flags & IFF_BROADCAST) && (ifra->ifra_broadaddr.sin_family == AF_INET)) ia->ia_broadaddr = ifra->ifra_broadaddr; if (error == 0) { if (iaIsFirst && (ifp->if_flags & IFF_MULTICAST) != 0) in_addmulti(&allhosts_addr, ifp); EVENTHANDLER_INVOKE(ifaddr_event, ifp); } return (error); case SIOCDIFADDR: /* * in_ifscrub kills the interface route. */ in_ifscrub(ifp, ia); /* * in_ifadown gets rid of all the rest of * the routes. This is not quite the right * thing to do, but at least if we are running * a routing process they will come back. */ in_ifadown(&ia->ia_ifa, 1); EVENTHANDLER_INVOKE(ifaddr_event, ifp); error = 0; break; default: if (ifp == NULL || ifp->if_ioctl == NULL) return (EOPNOTSUPP); IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, cmd, data); IFF_UNLOCKGIANT(ifp); return (error); } /* * Protect from ipintr() traversing address list while we're modifying * it. */ s = splnet(); TAILQ_REMOVE(&ifp->if_addrhead, &ia->ia_ifa, ifa_link); TAILQ_REMOVE(&V_in_ifaddrhead, ia, ia_link); if (ia->ia_addr.sin_family == AF_INET) { LIST_REMOVE(ia, ia_hash); /* * If this is the last IPv4 address configured on this * interface, leave the all-hosts group. * XXX: This is quite ugly because of locking and structure. */ oia = NULL; IFP_TO_IA(ifp, oia); if (oia == NULL) { struct in_multi *inm; IFF_LOCKGIANT(ifp); IN_MULTI_LOCK(); IN_LOOKUP_MULTI(allhosts_addr, ifp, inm); if (inm != NULL) in_delmulti_locked(inm); IN_MULTI_UNLOCK(); IFF_UNLOCKGIANT(ifp); } } IFAFREE(&ia->ia_ifa); splx(s); return (error); } /* * SIOC[GAD]LIFADDR. * SIOCGLIFADDR: get first address. (?!?) * SIOCGLIFADDR with IFLR_PREFIX: * get first address that matches the specified prefix. * SIOCALIFADDR: add the specified address. * SIOCALIFADDR with IFLR_PREFIX: * EINVAL since we can't deduce hostid part of the address. * SIOCDLIFADDR: delete the specified address. * SIOCDLIFADDR with IFLR_PREFIX: * delete the first address that matches the specified prefix. * return values: * EINVAL on invalid parameters * EADDRNOTAVAIL on prefix match failed/specified address not found * other values may be returned from in_ioctl() */ static int in_lifaddr_ioctl(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp, struct thread *td) { struct if_laddrreq *iflr = (struct if_laddrreq *)data; struct ifaddr *ifa; /* sanity checks */ if (data == NULL || ifp == NULL) { panic("invalid argument to in_lifaddr_ioctl"); /*NOTRECHED*/ } switch (cmd) { case SIOCGLIFADDR: /* address must be specified on GET with IFLR_PREFIX */ if ((iflr->flags & IFLR_PREFIX) == 0) break; /*FALLTHROUGH*/ case SIOCALIFADDR: case SIOCDLIFADDR: /* address must be specified on ADD and DELETE */ if (iflr->addr.ss_family != AF_INET) return (EINVAL); if (iflr->addr.ss_len != sizeof(struct sockaddr_in)) return (EINVAL); /* XXX need improvement */ if (iflr->dstaddr.ss_family && iflr->dstaddr.ss_family != AF_INET) return (EINVAL); if (iflr->dstaddr.ss_family && iflr->dstaddr.ss_len != sizeof(struct sockaddr_in)) return (EINVAL); break; default: /*shouldn't happen*/ return (EOPNOTSUPP); } if (sizeof(struct in_addr) * 8 < iflr->prefixlen) return (EINVAL); switch (cmd) { case SIOCALIFADDR: { struct in_aliasreq ifra; if (iflr->flags & IFLR_PREFIX) return (EINVAL); /* copy args to in_aliasreq, perform ioctl(SIOCAIFADDR_IN6). */ bzero(&ifra, sizeof(ifra)); bcopy(iflr->iflr_name, ifra.ifra_name, sizeof(ifra.ifra_name)); bcopy(&iflr->addr, &ifra.ifra_addr, iflr->addr.ss_len); if (iflr->dstaddr.ss_family) { /*XXX*/ bcopy(&iflr->dstaddr, &ifra.ifra_dstaddr, iflr->dstaddr.ss_len); } ifra.ifra_mask.sin_family = AF_INET; ifra.ifra_mask.sin_len = sizeof(struct sockaddr_in); in_len2mask(&ifra.ifra_mask.sin_addr, iflr->prefixlen); return (in_control(so, SIOCAIFADDR, (caddr_t)&ifra, ifp, td)); } case SIOCGLIFADDR: case SIOCDLIFADDR: { struct in_ifaddr *ia; struct in_addr mask, candidate, match; struct sockaddr_in *sin; bzero(&mask, sizeof(mask)); bzero(&match, sizeof(match)); if (iflr->flags & IFLR_PREFIX) { /* lookup a prefix rather than address. */ in_len2mask(&mask, iflr->prefixlen); sin = (struct sockaddr_in *)&iflr->addr; match.s_addr = sin->sin_addr.s_addr; match.s_addr &= mask.s_addr; /* if you set extra bits, that's wrong */ if (match.s_addr != sin->sin_addr.s_addr) return (EINVAL); } else { /* on getting an address, take the 1st match */ /* on deleting an address, do exact match */ if (cmd != SIOCGLIFADDR) { in_len2mask(&mask, 32); sin = (struct sockaddr_in *)&iflr->addr; match.s_addr = sin->sin_addr.s_addr; } } TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (match.s_addr == 0) break; candidate.s_addr = ((struct sockaddr_in *)&ifa->ifa_addr)->sin_addr.s_addr; candidate.s_addr &= mask.s_addr; if (candidate.s_addr == match.s_addr) break; } if (ifa == NULL) return (EADDRNOTAVAIL); ia = (struct in_ifaddr *)ifa; if (cmd == SIOCGLIFADDR) { /* fill in the if_laddrreq structure */ bcopy(&ia->ia_addr, &iflr->addr, ia->ia_addr.sin_len); if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { bcopy(&ia->ia_dstaddr, &iflr->dstaddr, ia->ia_dstaddr.sin_len); } else bzero(&iflr->dstaddr, sizeof(iflr->dstaddr)); iflr->prefixlen = in_mask2len(&ia->ia_sockmask.sin_addr); iflr->flags = 0; /*XXX*/ return (0); } else { struct in_aliasreq ifra; /* fill in_aliasreq and do ioctl(SIOCDIFADDR_IN6) */ bzero(&ifra, sizeof(ifra)); bcopy(iflr->iflr_name, ifra.ifra_name, sizeof(ifra.ifra_name)); bcopy(&ia->ia_addr, &ifra.ifra_addr, ia->ia_addr.sin_len); if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { bcopy(&ia->ia_dstaddr, &ifra.ifra_dstaddr, ia->ia_dstaddr.sin_len); } bcopy(&ia->ia_sockmask, &ifra.ifra_dstaddr, ia->ia_sockmask.sin_len); return (in_control(so, SIOCDIFADDR, (caddr_t)&ifra, ifp, td)); } } } return (EOPNOTSUPP); /*just for safety*/ } /* * Delete any existing route for an interface. */ void in_ifscrub(struct ifnet *ifp, struct in_ifaddr *ia) { in_scrubprefix(ia); } /* * Initialize an interface's internet address * and routing table entry. */ static int in_ifinit(struct ifnet *ifp, struct in_ifaddr *ia, struct sockaddr_in *sin, int scrub) { INIT_VNET_INET(ifp->if_vnet); register u_long i = ntohl(sin->sin_addr.s_addr); struct sockaddr_in oldaddr; int s = splimp(), flags = RTF_UP, error = 0; oldaddr = ia->ia_addr; if (oldaddr.sin_family == AF_INET) LIST_REMOVE(ia, ia_hash); ia->ia_addr = *sin; if (ia->ia_addr.sin_family == AF_INET) LIST_INSERT_HEAD(INADDR_HASH(ia->ia_addr.sin_addr.s_addr), ia, ia_hash); /* * Give the interface a chance to initialize * if this is its first address, * and to validate the address if necessary. */ if (ifp->if_ioctl != NULL) { IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, SIOCSIFADDR, (caddr_t)ia); IFF_UNLOCKGIANT(ifp); if (error) { splx(s); /* LIST_REMOVE(ia, ia_hash) is done in in_control */ ia->ia_addr = oldaddr; if (ia->ia_addr.sin_family == AF_INET) LIST_INSERT_HEAD(INADDR_HASH( ia->ia_addr.sin_addr.s_addr), ia, ia_hash); else /* * If oldaddr family is not AF_INET (e.g. * interface has been just created) in_control * does not call LIST_REMOVE, and we end up * with bogus ia entries in hash */ LIST_REMOVE(ia, ia_hash); return (error); } } splx(s); if (scrub) { ia->ia_ifa.ifa_addr = (struct sockaddr *)&oldaddr; in_ifscrub(ifp, ia); ia->ia_ifa.ifa_addr = (struct sockaddr *)&ia->ia_addr; } if (IN_CLASSA(i)) ia->ia_netmask = IN_CLASSA_NET; else if (IN_CLASSB(i)) ia->ia_netmask = IN_CLASSB_NET; else ia->ia_netmask = IN_CLASSC_NET; /* * The subnet mask usually includes at least the standard network part, * but may may be smaller in the case of supernetting. * If it is set, we believe it. */ if (ia->ia_subnetmask == 0) { ia->ia_subnetmask = ia->ia_netmask; ia->ia_sockmask.sin_addr.s_addr = htonl(ia->ia_subnetmask); } else ia->ia_netmask &= ia->ia_subnetmask; ia->ia_net = i & ia->ia_netmask; ia->ia_subnet = i & ia->ia_subnetmask; in_socktrim(&ia->ia_sockmask); #ifdef DEV_CARP /* * XXX: carp(4) does not have interface route */ if (ifp->if_type == IFT_CARP) return (0); #endif /* * Add route for the network. */ ia->ia_ifa.ifa_metric = ifp->if_metric; if (ifp->if_flags & IFF_BROADCAST) { ia->ia_broadaddr.sin_addr.s_addr = htonl(ia->ia_subnet | ~ia->ia_subnetmask); ia->ia_netbroadcast.s_addr = htonl(ia->ia_net | ~ ia->ia_netmask); } else if (ifp->if_flags & IFF_LOOPBACK) { ia->ia_dstaddr = ia->ia_addr; flags |= RTF_HOST; } else if (ifp->if_flags & IFF_POINTOPOINT) { if (ia->ia_dstaddr.sin_family != AF_INET) return (0); flags |= RTF_HOST; } if ((error = in_addprefix(ia, flags)) != 0) return (error); return (error); } #define rtinitflags(x) \ ((((x)->ia_ifp->if_flags & (IFF_LOOPBACK | IFF_POINTOPOINT)) != 0) \ ? RTF_HOST : 0) /* * Check if we have a route for the given prefix already or add one accordingly. */ static int in_addprefix(struct in_ifaddr *target, int flags) { INIT_VNET_INET(curvnet); struct in_ifaddr *ia; struct in_addr prefix, mask, p, m; int error; if ((flags & RTF_HOST) != 0) { prefix = target->ia_dstaddr.sin_addr; mask.s_addr = 0; } else { prefix = target->ia_addr.sin_addr; mask = target->ia_sockmask.sin_addr; prefix.s_addr &= mask.s_addr; } TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) { if (rtinitflags(ia)) { p = ia->ia_addr.sin_addr; if (prefix.s_addr != p.s_addr) continue; } else { p = ia->ia_addr.sin_addr; m = ia->ia_sockmask.sin_addr; p.s_addr &= m.s_addr; if (prefix.s_addr != p.s_addr || mask.s_addr != m.s_addr) continue; } /* * If we got a matching prefix route inserted by other * interface address, we are done here. */ if (ia->ia_flags & IFA_ROUTE) { if (V_sameprefixcarponly && target->ia_ifp->if_type != IFT_CARP && ia->ia_ifp->if_type != IFT_CARP) return (EEXIST); else return (0); } } /* * No-one seem to have this prefix route, so we try to insert it. */ error = rtinit(&target->ia_ifa, (int)RTM_ADD, flags); if (!error) target->ia_flags |= IFA_ROUTE; return (error); } extern void arp_ifscrub(struct ifnet *ifp, uint32_t addr); /* * If there is no other address in the system that can serve a route to the * same prefix, remove the route. Hand over the route to the new address * otherwise. */ static int in_scrubprefix(struct in_ifaddr *target) { INIT_VNET_INET(curvnet); struct in_ifaddr *ia; struct in_addr prefix, mask, p; int error; if ((target->ia_flags & IFA_ROUTE) == 0) return (0); if (rtinitflags(target)) prefix = target->ia_dstaddr.sin_addr; else { prefix = target->ia_addr.sin_addr; mask = target->ia_sockmask.sin_addr; prefix.s_addr &= mask.s_addr; /* remove arp cache */ arp_ifscrub(target->ia_ifp, IA_SIN(target)->sin_addr.s_addr); } TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) { if (rtinitflags(ia)) p = ia->ia_dstaddr.sin_addr; else { p = ia->ia_addr.sin_addr; p.s_addr &= ia->ia_sockmask.sin_addr.s_addr; } if (prefix.s_addr != p.s_addr) continue; /* * If we got a matching prefix address, move IFA_ROUTE and * the route itself to it. Make sure that routing daemons * get a heads-up. * * XXX: a special case for carp(4) interface */ if ((ia->ia_flags & IFA_ROUTE) == 0 #ifdef DEV_CARP && (ia->ia_ifp->if_type != IFT_CARP) #endif ) { rtinit(&(target->ia_ifa), (int)RTM_DELETE, rtinitflags(target)); target->ia_flags &= ~IFA_ROUTE; error = rtinit(&ia->ia_ifa, (int)RTM_ADD, rtinitflags(ia) | RTF_UP); if (error == 0) ia->ia_flags |= IFA_ROUTE; return (error); } } /* * As no-one seem to have this prefix, we can remove the route. */ rtinit(&(target->ia_ifa), (int)RTM_DELETE, rtinitflags(target)); target->ia_flags &= ~IFA_ROUTE; return (0); } #undef rtinitflags /* * Return 1 if the address might be a local broadcast address. */ int in_broadcast(struct in_addr in, struct ifnet *ifp) { register struct ifaddr *ifa; u_long t; if (in.s_addr == INADDR_BROADCAST || in.s_addr == INADDR_ANY) return (1); if ((ifp->if_flags & IFF_BROADCAST) == 0) return (0); t = ntohl(in.s_addr); /* * Look through the list of addresses for a match * with a broadcast address. */ #define ia ((struct in_ifaddr *)ifa) TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) if (ifa->ifa_addr->sa_family == AF_INET && (in.s_addr == ia->ia_broadaddr.sin_addr.s_addr || in.s_addr == ia->ia_netbroadcast.s_addr || /* * Check for old-style (host 0) broadcast. */ t == ia->ia_subnet || t == ia->ia_net) && /* * Check for an all one subnetmask. These * only exist when an interface gets a secondary * address. */ ia->ia_subnetmask != (u_long)0xffffffff) return (1); return (0); #undef ia } /* * Delete all IPv4 multicast address records, and associated link-layer * multicast address records, associated with ifp. */ static void in_purgemaddrs(struct ifnet *ifp) { INIT_VNET_INET(ifp->if_vnet); struct in_multi *inm; struct in_multi *oinm; IFF_LOCKGIANT(ifp); IN_MULTI_LOCK(); LIST_FOREACH_SAFE(inm, &V_in_multihead, inm_link, oinm) { if (inm->inm_ifp == ifp) in_delmulti_locked(inm); } IN_MULTI_UNLOCK(); IFF_UNLOCKGIANT(ifp); } /* * On interface removal, clean up IPv4 data structures hung off of the ifnet. */ void in_ifdetach(struct ifnet *ifp) { INIT_VNET_INET(ifp->if_vnet); in_pcbpurgeif0(&V_ripcbinfo, ifp); in_pcbpurgeif0(&V_udbinfo, ifp); in_purgemaddrs(ifp); } #include #include #include struct in_llentry { struct llentry base; struct sockaddr_in l3_addr4; }; static struct llentry * in_lltable_new(const struct sockaddr *l3addr, u_int flags) { struct in_llentry *lle; lle = malloc(sizeof(struct in_llentry), M_LLTABLE, M_DONTWAIT | M_ZERO); if (lle == NULL) /* NB: caller generates msg */ return NULL; callout_init(&lle->base.la_timer, CALLOUT_MPSAFE); /* * For IPv4 this will trigger "arpresolve" to generate * an ARP request. */ lle->base.la_expire = time_second; /* mark expired */ lle->l3_addr4 = *(const struct sockaddr_in *)l3addr; lle->base.lle_refcnt = 1; LLE_LOCK_INIT(&lle->base); return &lle->base; } /* * Deletes an address from the address table. * This function is called by the timer functions * such as arptimer() and nd6_llinfo_timer(), and * the caller does the locking. */ static void in_lltable_free(struct lltable *llt, struct llentry *lle) { LLE_WUNLOCK(lle); LLE_LOCK_DESTROY(lle); free(lle, M_LLTABLE); } static int in_lltable_rtcheck(struct ifnet *ifp, const struct sockaddr *l3addr) { struct rtentry *rt; KASSERT(l3addr->sa_family == AF_INET, ("sin_family %d", l3addr->sa_family)); /* XXX rtalloc1 should take a const param */ rt = rtalloc1(__DECONST(struct sockaddr *, l3addr), 0, 0); if (rt == NULL || (rt->rt_flags & RTF_GATEWAY) || rt->rt_ifp != ifp) { log(LOG_INFO, "IPv4 address: \"%s\" is not on the network\n", inet_ntoa(((const struct sockaddr_in *)l3addr)->sin_addr)); if (rt != NULL) RTFREE_LOCKED(rt); return (EINVAL); } RTFREE_LOCKED(rt); return 0; } /* * Return NULL if not found or marked for deletion. * If found return lle read locked. */ static struct llentry * in_lltable_lookup(struct lltable *llt, u_int flags, const struct sockaddr *l3addr) { const struct sockaddr_in *sin = (const struct sockaddr_in *)l3addr; struct ifnet *ifp = llt->llt_ifp; struct llentry *lle; struct llentries *lleh; u_int hashkey; IF_AFDATA_LOCK_ASSERT(ifp); KASSERT(l3addr->sa_family == AF_INET, ("sin_family %d", l3addr->sa_family)); hashkey = sin->sin_addr.s_addr; lleh = &llt->lle_head[LLATBL_HASH(hashkey, LLTBL_HASHMASK)]; LIST_FOREACH(lle, lleh, lle_next) { struct sockaddr_in *sa2 = (struct sockaddr_in *)L3_ADDR(lle); if (lle->la_flags & LLE_DELETED) continue; if (sa2->sin_addr.s_addr == sin->sin_addr.s_addr) break; } if (lle == NULL) { #ifdef DIAGNOSTICS if (flags & LLE_DELETE) log(LOG_INFO, "interface address is missing from cache = %p in delete\n", lle); #endif if (!(flags & LLE_CREATE)) return (NULL); /* * A route that covers the given address must have * been installed 1st because we are doing a resolution, * verify this. */ if (!(flags & LLE_IFADDR) && in_lltable_rtcheck(ifp, l3addr) != 0) goto done; lle = in_lltable_new(l3addr, flags); if (lle == NULL) { log(LOG_INFO, "lla_lookup: new lle malloc failed\n"); goto done; } lle->la_flags = flags & ~LLE_CREATE; if ((flags & (LLE_CREATE | LLE_IFADDR)) == (LLE_CREATE | LLE_IFADDR)) { bcopy(IF_LLADDR(ifp), &lle->ll_addr, ifp->if_addrlen); lle->la_flags |= (LLE_VALID | LLE_STATIC); } lle->lle_tbl = llt; lle->lle_head = lleh; LIST_INSERT_HEAD(lleh, lle, lle_next); } else if (flags & LLE_DELETE) { if (!(lle->la_flags & LLE_IFADDR) || (flags & LLE_IFADDR)) { LLE_WLOCK(lle); lle->la_flags = LLE_DELETED; LLE_WUNLOCK(lle); #ifdef DIAGNOSTICS log(LOG_INFO, "ifaddr cache = %p is deleted\n", lle); #endif } lle = (void *)-1; } if (LLE_IS_VALID(lle)) { if (flags & LLE_EXCLUSIVE) LLE_WLOCK(lle); else LLE_RLOCK(lle); } done: return (lle); } static int in_lltable_dump(struct lltable *llt, struct sysctl_req *wr) { #define SIN(lle) ((struct sockaddr_in *) L3_ADDR(lle)) struct ifnet *ifp = llt->llt_ifp; struct llentry *lle; /* XXX stack use */ struct { struct rt_msghdr rtm; struct sockaddr_inarp sin; struct sockaddr_dl sdl; } arpc; int error, i; /* XXXXX * current IFNET_RLOCK() is mapped to IFNET_WLOCK() * so it is okay to use this ASSERT, change it when * IFNET lock is finalized */ IFNET_WLOCK_ASSERT(); error = 0; for (i = 0; i < LLTBL_HASHTBL_SIZE; i++) { LIST_FOREACH(lle, &llt->lle_head[i], lle_next) { struct sockaddr_dl *sdl; /* skip deleted entries */ if ((lle->la_flags & (LLE_DELETED|LLE_VALID)) != LLE_VALID) continue; /* Skip if jailed and not a valid IP of the prison. */ - if (jailed(wr->td->td_ucred) && - !prison_if(wr->td->td_ucred, L3_ADDR(lle))) + if (prison_if(wr->td->td_ucred, L3_ADDR(lle)) != 0) continue; /* * produce a msg made of: * struct rt_msghdr; * struct sockaddr_inarp; (IPv4) * struct sockaddr_dl; */ bzero(&arpc, sizeof(arpc)); arpc.rtm.rtm_msglen = sizeof(arpc); arpc.rtm.rtm_version = RTM_VERSION; arpc.rtm.rtm_type = RTM_GET; arpc.rtm.rtm_flags = RTF_UP; arpc.rtm.rtm_addrs = RTA_DST | RTA_GATEWAY; arpc.sin.sin_family = AF_INET; arpc.sin.sin_len = sizeof(arpc.sin); arpc.sin.sin_addr.s_addr = SIN(lle)->sin_addr.s_addr; /* publish */ if (lle->la_flags & LLE_PUB) { arpc.rtm.rtm_flags |= RTF_ANNOUNCE; /* proxy only */ if (lle->la_flags & LLE_PROXY) arpc.sin.sin_other = SIN_PROXY; } sdl = &arpc.sdl; sdl->sdl_family = AF_LINK; sdl->sdl_len = sizeof(*sdl); sdl->sdl_alen = ifp->if_addrlen; sdl->sdl_index = ifp->if_index; sdl->sdl_type = ifp->if_type; bcopy(&lle->ll_addr, LLADDR(sdl), ifp->if_addrlen); arpc.rtm.rtm_rmx.rmx_expire = lle->la_flags & LLE_STATIC ? 0 : lle->la_expire; arpc.rtm.rtm_flags |= (RTF_HOST | RTF_LLDATA); if (lle->la_flags & LLE_STATIC) arpc.rtm.rtm_flags |= RTF_STATIC; arpc.rtm.rtm_index = ifp->if_index; error = SYSCTL_OUT(wr, &arpc, sizeof(arpc)); if (error) break; } } return error; #undef SIN } void * in_domifattach(struct ifnet *ifp) { struct lltable *llt = lltable_init(ifp, AF_INET); if (llt != NULL) { llt->llt_new = in_lltable_new; llt->llt_free = in_lltable_free; llt->llt_rtcheck = in_lltable_rtcheck; llt->llt_lookup = in_lltable_lookup; llt->llt_dump = in_lltable_dump; } return (llt); } void in_domifdetach(struct ifnet *ifp __unused, void *aux) { struct lltable *llt = (struct lltable *)aux; lltable_free(llt); } diff --git a/sys/netinet/in_pcb.c b/sys/netinet/in_pcb.c index 4a7e4f92a183..25494cfbfb9f 100644 --- a/sys/netinet/in_pcb.c +++ b/sys/netinet/in_pcb.c @@ -1,1937 +1,1936 @@ /*- * Copyright (c) 1982, 1986, 1991, 1993, 1995 * The Regents of the University of California. * Copyright (c) 2007-2008 Robert N. M. Watson * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)in_pcb.c 8.4 (Berkeley) 5/24/95 */ #include __FBSDID("$FreeBSD$"); #include "opt_ddb.h" #include "opt_inet.h" #include "opt_ipsec.h" #include "opt_inet6.h" #include "opt_mac.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef DDB #include #endif #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET6 #include #include #include #endif /* INET6 */ #ifdef IPSEC #include #include #endif /* IPSEC */ #include #ifdef VIMAGE_GLOBALS /* * These configure the range of local port addresses assigned to * "unspecified" outgoing connections/packets/whatever. */ int ipport_lowfirstauto; int ipport_lowlastauto; int ipport_firstauto; int ipport_lastauto; int ipport_hifirstauto; int ipport_hilastauto; /* * Reserved ports accessible only to root. There are significant * security considerations that must be accounted for when changing these, * but the security benefits can be great. Please be careful. */ int ipport_reservedhigh; int ipport_reservedlow; /* Variables dealing with random ephemeral port allocation. */ int ipport_randomized; int ipport_randomcps; int ipport_randomtime; int ipport_stoprandom; int ipport_tcpallocs; int ipport_tcplastcount; #endif #define RANGECHK(var, min, max) \ if ((var) < (min)) { (var) = (min); } \ else if ((var) > (max)) { (var) = (max); } static int sysctl_net_ipport_check(SYSCTL_HANDLER_ARGS) { INIT_VNET_INET(curvnet); int error; error = sysctl_handle_int(oidp, oidp->oid_arg1, oidp->oid_arg2, req); if (error == 0) { RANGECHK(V_ipport_lowfirstauto, 1, IPPORT_RESERVED - 1); RANGECHK(V_ipport_lowlastauto, 1, IPPORT_RESERVED - 1); RANGECHK(V_ipport_firstauto, IPPORT_RESERVED, IPPORT_MAX); RANGECHK(V_ipport_lastauto, IPPORT_RESERVED, IPPORT_MAX); RANGECHK(V_ipport_hifirstauto, IPPORT_RESERVED, IPPORT_MAX); RANGECHK(V_ipport_hilastauto, IPPORT_RESERVED, IPPORT_MAX); } return (error); } #undef RANGECHK SYSCTL_NODE(_net_inet_ip, IPPROTO_IP, portrange, CTLFLAG_RW, 0, "IP Ports"); SYSCTL_V_PROC(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, lowfirst, CTLTYPE_INT|CTLFLAG_RW, ipport_lowfirstauto, 0, &sysctl_net_ipport_check, "I", ""); SYSCTL_V_PROC(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, lowlast, CTLTYPE_INT|CTLFLAG_RW, ipport_lowlastauto, 0, &sysctl_net_ipport_check, "I", ""); SYSCTL_V_PROC(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, first, CTLTYPE_INT|CTLFLAG_RW, ipport_firstauto, 0, &sysctl_net_ipport_check, "I", ""); SYSCTL_V_PROC(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, last, CTLTYPE_INT|CTLFLAG_RW, ipport_lastauto, 0, &sysctl_net_ipport_check, "I", ""); SYSCTL_V_PROC(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, hifirst, CTLTYPE_INT|CTLFLAG_RW, ipport_hifirstauto, 0, &sysctl_net_ipport_check, "I", ""); SYSCTL_V_PROC(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, hilast, CTLTYPE_INT|CTLFLAG_RW, ipport_hilastauto, 0, &sysctl_net_ipport_check, "I", ""); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, reservedhigh, CTLFLAG_RW|CTLFLAG_SECURE, ipport_reservedhigh, 0, ""); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, reservedlow, CTLFLAG_RW|CTLFLAG_SECURE, ipport_reservedlow, 0, ""); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, randomized, CTLFLAG_RW, ipport_randomized, 0, "Enable random port allocation"); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, randomcps, CTLFLAG_RW, ipport_randomcps, 0, "Maximum number of random port " "allocations before switching to a sequental one"); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_ip_portrange, OID_AUTO, randomtime, CTLFLAG_RW, ipport_randomtime, 0, "Minimum time to keep sequental port " "allocation before switching to a random one"); /* * in_pcb.c: manage the Protocol Control Blocks. * * NOTE: It is assumed that most of these functions will be called with * the pcbinfo lock held, and often, the inpcb lock held, as these utility * functions often modify hash chains or addresses in pcbs. */ /* * Allocate a PCB and associate it with the socket. * On success return with the PCB locked. */ int in_pcballoc(struct socket *so, struct inpcbinfo *pcbinfo) { #ifdef INET6 INIT_VNET_INET6(curvnet); #endif struct inpcb *inp; int error; INP_INFO_WLOCK_ASSERT(pcbinfo); error = 0; inp = uma_zalloc(pcbinfo->ipi_zone, M_NOWAIT); if (inp == NULL) return (ENOBUFS); bzero(inp, inp_zero_size); inp->inp_pcbinfo = pcbinfo; inp->inp_socket = so; inp->inp_cred = crhold(so->so_cred); inp->inp_inc.inc_fibnum = so->so_fibnum; #ifdef MAC error = mac_inpcb_init(inp, M_NOWAIT); if (error != 0) goto out; SOCK_LOCK(so); mac_inpcb_create(so, inp); SOCK_UNLOCK(so); #endif #ifdef IPSEC error = ipsec_init_policy(so, &inp->inp_sp); if (error != 0) { #ifdef MAC mac_inpcb_destroy(inp); #endif goto out; } #endif /*IPSEC*/ #ifdef INET6 if (INP_SOCKAF(so) == AF_INET6) { inp->inp_vflag |= INP_IPV6PROTO; if (V_ip6_v6only) inp->inp_flags |= IN6P_IPV6_V6ONLY; } #endif LIST_INSERT_HEAD(pcbinfo->ipi_listhead, inp, inp_list); pcbinfo->ipi_count++; so->so_pcb = (caddr_t)inp; #ifdef INET6 if (V_ip6_auto_flowlabel) inp->inp_flags |= IN6P_AUTOFLOWLABEL; #endif INP_WLOCK(inp); inp->inp_gencnt = ++pcbinfo->ipi_gencnt; inp->inp_refcount = 1; /* Reference from the inpcbinfo */ #if defined(IPSEC) || defined(MAC) out: if (error != 0) { crfree(inp->inp_cred); uma_zfree(pcbinfo->ipi_zone, inp); } #endif return (error); } int in_pcbbind(struct inpcb *inp, struct sockaddr *nam, struct ucred *cred) { int anonport, error; INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); if (inp->inp_lport != 0 || inp->inp_laddr.s_addr != INADDR_ANY) return (EINVAL); anonport = inp->inp_lport == 0 && (nam == NULL || ((struct sockaddr_in *)nam)->sin_port == 0); error = in_pcbbind_setup(inp, nam, &inp->inp_laddr.s_addr, &inp->inp_lport, cred); if (error) return (error); if (in_pcbinshash(inp) != 0) { inp->inp_laddr.s_addr = INADDR_ANY; inp->inp_lport = 0; return (EAGAIN); } if (anonport) inp->inp_flags |= INP_ANONPORT; return (0); } /* * Set up a bind operation on a PCB, performing port allocation * as required, but do not actually modify the PCB. Callers can * either complete the bind by setting inp_laddr/inp_lport and * calling in_pcbinshash(), or they can just use the resulting * port and address to authorise the sending of a once-off packet. * * On error, the values of *laddrp and *lportp are not changed. */ int in_pcbbind_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp, u_short *lportp, struct ucred *cred) { INIT_VNET_INET(inp->inp_vnet); struct socket *so = inp->inp_socket; unsigned short *lastport; struct sockaddr_in *sin; struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; struct in_addr laddr; u_short lport = 0; int wild = 0, reuseport = (so->so_options & SO_REUSEPORT); int error; int dorandom; /* * Because no actual state changes occur here, a global write lock on * the pcbinfo isn't required. */ INP_INFO_LOCK_ASSERT(pcbinfo); INP_LOCK_ASSERT(inp); if (TAILQ_EMPTY(&V_in_ifaddrhead)) /* XXX broken! */ return (EADDRNOTAVAIL); laddr.s_addr = *laddrp; if (nam != NULL && laddr.s_addr != INADDR_ANY) return (EINVAL); if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) wild = INPLOOKUP_WILDCARD; if (nam) { sin = (struct sockaddr_in *)nam; if (nam->sa_len != sizeof (*sin)) return (EINVAL); #ifdef notdef /* * We should check the family, but old programs * incorrectly fail to initialize it. */ if (sin->sin_family != AF_INET) return (EAFNOSUPPORT); #endif - if (prison_local_ip4(cred, &sin->sin_addr)) - return (EINVAL); + error = prison_local_ip4(cred, &sin->sin_addr); + if (error) + return (error); if (sin->sin_port != *lportp) { /* Don't allow the port to change. */ if (*lportp != 0) return (EINVAL); lport = sin->sin_port; } /* NB: lport is left as 0 if the port isn't being changed. */ if (IN_MULTICAST(ntohl(sin->sin_addr.s_addr))) { /* * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; * allow complete duplication of binding if * SO_REUSEPORT is set, or if SO_REUSEADDR is set * and a multicast address is bound on both * new and duplicated sockets. */ if (so->so_options & SO_REUSEADDR) reuseport = SO_REUSEADDR|SO_REUSEPORT; } else if (sin->sin_addr.s_addr != INADDR_ANY) { sin->sin_port = 0; /* yech... */ bzero(&sin->sin_zero, sizeof(sin->sin_zero)); /* * Is the address a local IP address? * If INP_NONLOCALOK is set, then the socket may be bound * to any endpoint address, local or not. */ if ( #if defined(IP_NONLOCALBIND) ((inp->inp_flags & INP_NONLOCALOK) == 0) && #endif (ifa_ifwithaddr((struct sockaddr *)sin) == 0)) return (EADDRNOTAVAIL); } laddr = sin->sin_addr; if (lport) { struct inpcb *t; struct tcptw *tw; /* GROSS */ if (ntohs(lport) <= V_ipport_reservedhigh && ntohs(lport) >= V_ipport_reservedlow && priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT, 0)) return (EACCES); if (!IN_MULTICAST(ntohl(sin->sin_addr.s_addr)) && priv_check_cred(inp->inp_cred, PRIV_NETINET_REUSEPORT, 0) != 0) { t = in_pcblookup_local(pcbinfo, sin->sin_addr, lport, INPLOOKUP_WILDCARD, cred); /* * XXX * This entire block sorely needs a rewrite. */ if (t && ((t->inp_vflag & INP_TIMEWAIT) == 0) && (so->so_type != SOCK_STREAM || ntohl(t->inp_faddr.s_addr) == INADDR_ANY) && (ntohl(sin->sin_addr.s_addr) != INADDR_ANY || ntohl(t->inp_laddr.s_addr) != INADDR_ANY || (t->inp_socket->so_options & SO_REUSEPORT) == 0) && (inp->inp_cred->cr_uid != t->inp_cred->cr_uid)) return (EADDRINUSE); } - if (prison_local_ip4(cred, &sin->sin_addr)) - return (EADDRNOTAVAIL); + error = prison_local_ip4(cred, &sin->sin_addr); + if (error) + return (error); t = in_pcblookup_local(pcbinfo, sin->sin_addr, lport, wild, cred); if (t && (t->inp_vflag & INP_TIMEWAIT)) { /* * XXXRW: If an incpb has had its timewait * state recycled, we treat the address as * being in use (for now). This is better * than a panic, but not desirable. */ tw = intotw(inp); if (tw == NULL || (reuseport & tw->tw_so_options) == 0) return (EADDRINUSE); } else if (t && (reuseport & t->inp_socket->so_options) == 0) { #ifdef INET6 if (ntohl(sin->sin_addr.s_addr) != INADDR_ANY || ntohl(t->inp_laddr.s_addr) != INADDR_ANY || INP_SOCKAF(so) == INP_SOCKAF(t->inp_socket)) #endif return (EADDRINUSE); } } } if (*lportp != 0) lport = *lportp; if (lport == 0) { u_short first, last, aux; int count; - if (prison_local_ip4(cred, &laddr)) - return (EINVAL); + error = prison_local_ip4(cred, &laddr); + if (error) + return (error); if (inp->inp_flags & INP_HIGHPORT) { first = V_ipport_hifirstauto; /* sysctl */ last = V_ipport_hilastauto; lastport = &pcbinfo->ipi_lasthi; } else if (inp->inp_flags & INP_LOWPORT) { error = priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT, 0); if (error) return error; first = V_ipport_lowfirstauto; /* 1023 */ last = V_ipport_lowlastauto; /* 600 */ lastport = &pcbinfo->ipi_lastlow; } else { first = V_ipport_firstauto; /* sysctl */ last = V_ipport_lastauto; lastport = &pcbinfo->ipi_lastport; } /* * For UDP, use random port allocation as long as the user * allows it. For TCP (and as of yet unknown) connections, * use random port allocation only if the user allows it AND * ipport_tick() allows it. */ if (V_ipport_randomized && (!V_ipport_stoprandom || pcbinfo == &V_udbinfo)) dorandom = 1; else dorandom = 0; /* * It makes no sense to do random port allocation if * we have the only port available. */ if (first == last) dorandom = 0; /* Make sure to not include UDP packets in the count. */ if (pcbinfo != &V_udbinfo) V_ipport_tcpallocs++; /* * Instead of having two loops further down counting up or down * make sure that first is always <= last and go with only one * code path implementing all logic. */ if (first > last) { aux = first; first = last; last = aux; } if (dorandom) *lastport = first + (arc4random() % (last - first)); count = last - first; do { if (count-- < 0) /* completely used? */ return (EADDRNOTAVAIL); ++*lastport; if (*lastport < first || *lastport > last) *lastport = first; lport = htons(*lastport); } while (in_pcblookup_local(pcbinfo, laddr, lport, wild, cred)); } - if (prison_local_ip4(cred, &laddr)) - return (EINVAL); + error = prison_local_ip4(cred, &laddr); + if (error) + return (error); *laddrp = laddr.s_addr; *lportp = lport; return (0); } /* * Connect from a socket to a specified address. * Both address and port must be specified in argument sin. * If don't have a local address for this socket yet, * then pick one. */ int in_pcbconnect(struct inpcb *inp, struct sockaddr *nam, struct ucred *cred) { u_short lport, fport; in_addr_t laddr, faddr; int anonport, error; INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); lport = inp->inp_lport; laddr = inp->inp_laddr.s_addr; anonport = (lport == 0); error = in_pcbconnect_setup(inp, nam, &laddr, &lport, &faddr, &fport, NULL, cred); if (error) return (error); /* Do the initial binding of the local address if required. */ if (inp->inp_laddr.s_addr == INADDR_ANY && inp->inp_lport == 0) { inp->inp_lport = lport; inp->inp_laddr.s_addr = laddr; if (in_pcbinshash(inp) != 0) { inp->inp_laddr.s_addr = INADDR_ANY; inp->inp_lport = 0; return (EAGAIN); } } /* Commit the remaining changes. */ inp->inp_lport = lport; inp->inp_laddr.s_addr = laddr; inp->inp_faddr.s_addr = faddr; inp->inp_fport = fport; in_pcbrehash(inp); if (anonport) inp->inp_flags |= INP_ANONPORT; return (0); } /* * Do proper source address selection on an unbound socket in case * of connect. Take jails into account as well. */ static int in_pcbladdr(struct inpcb *inp, struct in_addr *faddr, struct in_addr *laddr, struct ucred *cred) { struct in_ifaddr *ia; struct ifaddr *ifa; struct sockaddr *sa; struct sockaddr_in *sin; struct route sro; int error; KASSERT(laddr != NULL, ("%s: laddr NULL", __func__)); error = 0; ia = NULL; bzero(&sro, sizeof(sro)); sin = (struct sockaddr_in *)&sro.ro_dst; sin->sin_family = AF_INET; sin->sin_len = sizeof(struct sockaddr_in); sin->sin_addr.s_addr = faddr->s_addr; /* * If route is known our src addr is taken from the i/f, * else punt. * * Find out route to destination. */ if ((inp->inp_socket->so_options & SO_DONTROUTE) == 0) in_rtalloc_ign(&sro, 0, inp->inp_inc.inc_fibnum); /* * If we found a route, use the address corresponding to * the outgoing interface. * * Otherwise assume faddr is reachable on a directly connected * network and try to find a corresponding interface to take * the source address from. */ if (sro.ro_rt == NULL || sro.ro_rt->rt_ifp == NULL) { struct ifnet *ifp; ia = ifatoia(ifa_ifwithdstaddr((struct sockaddr *)sin)); if (ia == NULL) ia = ifatoia(ifa_ifwithnet((struct sockaddr *)sin)); if (ia == NULL) { error = ENETUNREACH; goto done; } if (cred == NULL || !jailed(cred)) { laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } ifp = ia->ia_ifp; ia = NULL; TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { sa = ifa->ifa_addr; if (sa->sa_family != AF_INET) continue; sin = (struct sockaddr_in *)sa; - if (prison_check_ip4(cred, &sin->sin_addr)) { + if (prison_check_ip4(cred, &sin->sin_addr) == 0) { ia = (struct in_ifaddr *)ifa; break; } } if (ia != NULL) { laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } /* 3. As a last resort return the 'default' jail address. */ - if (prison_get_ip4(cred, laddr) != 0) - error = EADDRNOTAVAIL; + error = prison_get_ip4(cred, laddr); goto done; } /* * If the outgoing interface on the route found is not * a loopback interface, use the address from that interface. * In case of jails do those three steps: * 1. check if the interface address belongs to the jail. If so use it. * 2. check if we have any address on the outgoing interface * belonging to this jail. If so use it. * 3. as a last resort return the 'default' jail address. */ if ((sro.ro_rt->rt_ifp->if_flags & IFF_LOOPBACK) == 0) { /* If not jailed, use the default returned. */ if (cred == NULL || !jailed(cred)) { ia = (struct in_ifaddr *)sro.ro_rt->rt_ifa; laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } /* Jailed. */ /* 1. Check if the iface address belongs to the jail. */ sin = (struct sockaddr_in *)sro.ro_rt->rt_ifa->ifa_addr; - if (prison_check_ip4(cred, &sin->sin_addr)) { + if (prison_check_ip4(cred, &sin->sin_addr) == 0) { ia = (struct in_ifaddr *)sro.ro_rt->rt_ifa; laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } /* * 2. Check if we have any address on the outgoing interface * belonging to this jail. */ TAILQ_FOREACH(ifa, &sro.ro_rt->rt_ifp->if_addrhead, ifa_link) { sa = ifa->ifa_addr; if (sa->sa_family != AF_INET) continue; sin = (struct sockaddr_in *)sa; - if (prison_check_ip4(cred, &sin->sin_addr)) { + if (prison_check_ip4(cred, &sin->sin_addr) == 0) { ia = (struct in_ifaddr *)ifa; break; } } if (ia != NULL) { laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } /* 3. As a last resort return the 'default' jail address. */ - if (prison_get_ip4(cred, laddr) != 0) - error = EADDRNOTAVAIL; + error = prison_get_ip4(cred, laddr); goto done; } /* * The outgoing interface is marked with 'loopback net', so a route * to ourselves is here. * Try to find the interface of the destination address and then * take the address from there. That interface is not necessarily * a loopback interface. * In case of jails, check that it is an address of the jail * and if we cannot find, fall back to the 'default' jail address. */ if ((sro.ro_rt->rt_ifp->if_flags & IFF_LOOPBACK) != 0) { struct sockaddr_in sain; bzero(&sain, sizeof(struct sockaddr_in)); sain.sin_family = AF_INET; sain.sin_len = sizeof(struct sockaddr_in); sain.sin_addr.s_addr = faddr->s_addr; ia = ifatoia(ifa_ifwithdstaddr(sintosa(&sain))); if (ia == NULL) ia = ifatoia(ifa_ifwithnet(sintosa(&sain))); if (cred == NULL || !jailed(cred)) { #if __FreeBSD_version < 800000 if (ia == NULL) ia = (struct in_ifaddr *)sro.ro_rt->rt_ifa; #endif if (ia == NULL) { error = ENETUNREACH; goto done; } laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } /* Jailed. */ if (ia != NULL) { struct ifnet *ifp; ifp = ia->ia_ifp; ia = NULL; TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { sa = ifa->ifa_addr; if (sa->sa_family != AF_INET) continue; sin = (struct sockaddr_in *)sa; - if (prison_check_ip4(cred, &sin->sin_addr)) { + if (prison_check_ip4(cred, + &sin->sin_addr) == 0) { ia = (struct in_ifaddr *)ifa; break; } } if (ia != NULL) { laddr->s_addr = ia->ia_addr.sin_addr.s_addr; goto done; } } /* 3. As a last resort return the 'default' jail address. */ - if (prison_get_ip4(cred, laddr) != 0) - error = EADDRNOTAVAIL; + error = prison_get_ip4(cred, laddr); goto done; } done: if (sro.ro_rt != NULL) RTFREE(sro.ro_rt); return (error); } /* * Set up for a connect from a socket to the specified address. * On entry, *laddrp and *lportp should contain the current local * address and port for the PCB; these are updated to the values * that should be placed in inp_laddr and inp_lport to complete * the connect. * * On success, *faddrp and *fportp will be set to the remote address * and port. These are not updated in the error case. * * If the operation fails because the connection already exists, * *oinpp will be set to the PCB of that connection so that the * caller can decide to override it. In all other cases, *oinpp * is set to NULL. */ int in_pcbconnect_setup(struct inpcb *inp, struct sockaddr *nam, in_addr_t *laddrp, u_short *lportp, in_addr_t *faddrp, u_short *fportp, struct inpcb **oinpp, struct ucred *cred) { INIT_VNET_INET(inp->inp_vnet); struct sockaddr_in *sin = (struct sockaddr_in *)nam; struct in_ifaddr *ia; struct inpcb *oinp; - struct in_addr laddr, faddr, jailia; + struct in_addr laddr, faddr; u_short lport, fport; int error; /* * Because a global state change doesn't actually occur here, a read * lock is sufficient. */ INP_INFO_LOCK_ASSERT(inp->inp_pcbinfo); INP_LOCK_ASSERT(inp); if (oinpp != NULL) *oinpp = NULL; if (nam->sa_len != sizeof (*sin)) return (EINVAL); if (sin->sin_family != AF_INET) return (EAFNOSUPPORT); if (sin->sin_port == 0) return (EADDRNOTAVAIL); laddr.s_addr = *laddrp; lport = *lportp; faddr = sin->sin_addr; fport = sin->sin_port; if (!TAILQ_EMPTY(&V_in_ifaddrhead)) { /* * If the destination address is INADDR_ANY, * use the primary local address. * If the supplied address is INADDR_BROADCAST, * and the primary interface supports broadcast, * choose the broadcast address for that interface. */ if (faddr.s_addr == INADDR_ANY) { - if (cred != NULL && jailed(cred)) { - if (prison_get_ip4(cred, &jailia) != 0) - return (EADDRNOTAVAIL); - faddr.s_addr = jailia.s_addr; - } else { - faddr = - IA_SIN(TAILQ_FIRST(&V_in_ifaddrhead))-> - sin_addr; - } + faddr = + IA_SIN(TAILQ_FIRST(&V_in_ifaddrhead))->sin_addr; + if (cred != NULL && + (error = prison_get_ip4(cred, &faddr)) != 0) + return (error); } else if (faddr.s_addr == (u_long)INADDR_BROADCAST && (TAILQ_FIRST(&V_in_ifaddrhead)->ia_ifp->if_flags & IFF_BROADCAST)) faddr = satosin(&TAILQ_FIRST( &V_in_ifaddrhead)->ia_broadaddr)->sin_addr; } if (laddr.s_addr == INADDR_ANY) { error = in_pcbladdr(inp, &faddr, &laddr, cred); if (error) return (error); /* * If the destination address is multicast and an outgoing * interface has been set as a multicast option, use the * address of that interface as our source address. */ if (IN_MULTICAST(ntohl(faddr.s_addr)) && inp->inp_moptions != NULL) { struct ip_moptions *imo; struct ifnet *ifp; imo = inp->inp_moptions; if (imo->imo_multicast_ifp != NULL) { ifp = imo->imo_multicast_ifp; TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) if (ia->ia_ifp == ifp) break; if (ia == NULL) return (EADDRNOTAVAIL); laddr = ia->ia_addr.sin_addr; } } } oinp = in_pcblookup_hash(inp->inp_pcbinfo, faddr, fport, laddr, lport, 0, NULL); if (oinp != NULL) { if (oinpp != NULL) *oinpp = oinp; return (EADDRINUSE); } if (lport == 0) { error = in_pcbbind_setup(inp, NULL, &laddr.s_addr, &lport, cred); if (error) return (error); } *laddrp = laddr.s_addr; *lportp = lport; *faddrp = faddr.s_addr; *fportp = fport; return (0); } void in_pcbdisconnect(struct inpcb *inp) { INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); inp->inp_faddr.s_addr = INADDR_ANY; inp->inp_fport = 0; in_pcbrehash(inp); } /* * in_pcbdetach() is responsibe for disassociating a socket from an inpcb. * For most protocols, this will be invoked immediately prior to calling * in_pcbfree(). However, with TCP the inpcb may significantly outlive the * socket, in which case in_pcbfree() is deferred. */ void in_pcbdetach(struct inpcb *inp) { KASSERT(inp->inp_socket != NULL, ("%s: inp_socket == NULL", __func__)); inp->inp_socket->so_pcb = NULL; inp->inp_socket = NULL; } /* * in_pcbfree_internal() frees an inpcb that has been detached from its * socket, and whose reference count has reached 0. It will also remove the * inpcb from any global lists it might remain on. */ static void in_pcbfree_internal(struct inpcb *inp) { struct inpcbinfo *ipi = inp->inp_pcbinfo; KASSERT(inp->inp_socket == NULL, ("%s: inp_socket != NULL", __func__)); KASSERT(inp->inp_refcount == 0, ("%s: refcount !0", __func__)); INP_INFO_WLOCK_ASSERT(ipi); INP_WLOCK_ASSERT(inp); #ifdef IPSEC if (inp->inp_sp != NULL) ipsec_delete_pcbpolicy(inp); #endif /* IPSEC */ inp->inp_gencnt = ++ipi->ipi_gencnt; in_pcbremlists(inp); #ifdef INET6 if (inp->inp_vflag & INP_IPV6PROTO) { ip6_freepcbopts(inp->in6p_outputopts); ip6_freemoptions(inp->in6p_moptions); } #endif if (inp->inp_options) (void)m_free(inp->inp_options); if (inp->inp_moptions != NULL) inp_freemoptions(inp->inp_moptions); inp->inp_vflag = 0; crfree(inp->inp_cred); #ifdef MAC mac_inpcb_destroy(inp); #endif INP_WUNLOCK(inp); uma_zfree(ipi->ipi_zone, inp); } /* * in_pcbref() bumps the reference count on an inpcb in order to maintain * stability of an inpcb pointer despite the inpcb lock being released. This * is used in TCP when the inpcbinfo lock needs to be acquired or upgraded, * but where the inpcb lock is already held. * * While the inpcb will not be freed, releasing the inpcb lock means that the * connection's state may change, so the caller should be careful to * revalidate any cached state on reacquiring the lock. Drop the reference * using in_pcbrele(). */ void in_pcbref(struct inpcb *inp) { INP_WLOCK_ASSERT(inp); KASSERT(inp->inp_refcount > 0, ("%s: refcount 0", __func__)); inp->inp_refcount++; } /* * Drop a refcount on an inpcb elevated using in_pcbref(); because a call to * in_pcbfree() may have been made between in_pcbref() and in_pcbrele(), we * return a flag indicating whether or not the inpcb remains valid. If it is * valid, we return with the inpcb lock held. */ int in_pcbrele(struct inpcb *inp) { #ifdef INVARIANTS struct inpcbinfo *ipi = inp->inp_pcbinfo; #endif KASSERT(inp->inp_refcount > 0, ("%s: refcount 0", __func__)); INP_INFO_WLOCK_ASSERT(ipi); INP_WLOCK_ASSERT(inp); inp->inp_refcount--; if (inp->inp_refcount > 0) return (0); in_pcbfree_internal(inp); return (1); } /* * Unconditionally schedule an inpcb to be freed by decrementing its * reference count, which should occur only after the inpcb has been detached * from its socket. If another thread holds a temporary reference (acquired * using in_pcbref()) then the free is deferred until that reference is * released using in_pcbrele(), but the inpcb is still unlocked. */ void in_pcbfree(struct inpcb *inp) { #ifdef INVARIANTS struct inpcbinfo *ipi = inp->inp_pcbinfo; #endif KASSERT(inp->inp_socket == NULL, ("%s: inp_socket != NULL", __func__)); INP_INFO_WLOCK_ASSERT(ipi); INP_WLOCK_ASSERT(inp); if (!in_pcbrele(inp)) INP_WUNLOCK(inp); } /* * in_pcbdrop() removes an inpcb from hashed lists, releasing its address and * port reservation, and preventing it from being returned by inpcb lookups. * * It is used by TCP to mark an inpcb as unused and avoid future packet * delivery or event notification when a socket remains open but TCP has * closed. This might occur as a result of a shutdown()-initiated TCP close * or a RST on the wire, and allows the port binding to be reused while still * maintaining the invariant that so_pcb always points to a valid inpcb until * in_pcbdetach(). * * XXXRW: An inp_lport of 0 is used to indicate that the inpcb is not on hash * lists, but can lead to confusing netstat output, as open sockets with * closed TCP connections will no longer appear to have their bound port * number. An explicit flag would be better, as it would allow us to leave * the port number intact after the connection is dropped. * * XXXRW: Possibly in_pcbdrop() should also prevent future notifications by * in_pcbnotifyall() and in_pcbpurgeif0()? */ void in_pcbdrop(struct inpcb *inp) { INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); inp->inp_vflag |= INP_DROPPED; if (inp->inp_lport) { struct inpcbport *phd = inp->inp_phd; LIST_REMOVE(inp, inp_hash); LIST_REMOVE(inp, inp_portlist); if (LIST_FIRST(&phd->phd_pcblist) == NULL) { LIST_REMOVE(phd, phd_hash); free(phd, M_PCB); } inp->inp_lport = 0; } } /* * Common routines to return the socket addresses associated with inpcbs. */ struct sockaddr * in_sockaddr(in_port_t port, struct in_addr *addr_p) { struct sockaddr_in *sin; sin = malloc(sizeof *sin, M_SONAME, M_WAITOK | M_ZERO); sin->sin_family = AF_INET; sin->sin_len = sizeof(*sin); sin->sin_addr = *addr_p; sin->sin_port = port; return (struct sockaddr *)sin; } int in_getsockaddr(struct socket *so, struct sockaddr **nam) { struct inpcb *inp; struct in_addr addr; in_port_t port; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in_getsockaddr: inp == NULL")); INP_RLOCK(inp); port = inp->inp_lport; addr = inp->inp_laddr; INP_RUNLOCK(inp); *nam = in_sockaddr(port, &addr); return 0; } int in_getpeeraddr(struct socket *so, struct sockaddr **nam) { struct inpcb *inp; struct in_addr addr; in_port_t port; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in_getpeeraddr: inp == NULL")); INP_RLOCK(inp); port = inp->inp_fport; addr = inp->inp_faddr; INP_RUNLOCK(inp); *nam = in_sockaddr(port, &addr); return 0; } void in_pcbnotifyall(struct inpcbinfo *pcbinfo, struct in_addr faddr, int errno, struct inpcb *(*notify)(struct inpcb *, int)) { struct inpcb *inp, *inp_temp; INP_INFO_WLOCK(pcbinfo); LIST_FOREACH_SAFE(inp, pcbinfo->ipi_listhead, inp_list, inp_temp) { INP_WLOCK(inp); #ifdef INET6 if ((inp->inp_vflag & INP_IPV4) == 0) { INP_WUNLOCK(inp); continue; } #endif if (inp->inp_faddr.s_addr != faddr.s_addr || inp->inp_socket == NULL) { INP_WUNLOCK(inp); continue; } if ((*notify)(inp, errno)) INP_WUNLOCK(inp); } INP_INFO_WUNLOCK(pcbinfo); } void in_pcbpurgeif0(struct inpcbinfo *pcbinfo, struct ifnet *ifp) { struct inpcb *inp; struct ip_moptions *imo; int i, gap; INP_INFO_RLOCK(pcbinfo); LIST_FOREACH(inp, pcbinfo->ipi_listhead, inp_list) { INP_WLOCK(inp); imo = inp->inp_moptions; if ((inp->inp_vflag & INP_IPV4) && imo != NULL) { /* * Unselect the outgoing interface if it is being * detached. */ if (imo->imo_multicast_ifp == ifp) imo->imo_multicast_ifp = NULL; /* * Drop multicast group membership if we joined * through the interface being detached. */ for (i = 0, gap = 0; i < imo->imo_num_memberships; i++) { if (imo->imo_membership[i]->inm_ifp == ifp) { in_delmulti(imo->imo_membership[i]); gap++; } else if (gap != 0) imo->imo_membership[i - gap] = imo->imo_membership[i]; } imo->imo_num_memberships -= gap; } INP_WUNLOCK(inp); } INP_INFO_RUNLOCK(pcbinfo); } /* * Lookup a PCB based on the local address and port. */ #define INP_LOOKUP_MAPPED_PCB_COST 3 struct inpcb * in_pcblookup_local(struct inpcbinfo *pcbinfo, struct in_addr laddr, u_short lport, int wild_okay, struct ucred *cred) { struct inpcb *inp; #ifdef INET6 int matchwild = 3 + INP_LOOKUP_MAPPED_PCB_COST; #else int matchwild = 3; #endif int wildcard; INP_INFO_LOCK_ASSERT(pcbinfo); if (!wild_okay) { struct inpcbhead *head; /* * Look for an unconnected (wildcard foreign addr) PCB that * matches the local address and port we're looking for. */ head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport, 0, pcbinfo->ipi_hashmask)]; LIST_FOREACH(inp, head, inp_hash) { #ifdef INET6 /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV4) == 0) continue; #endif if (inp->inp_faddr.s_addr == INADDR_ANY && inp->inp_laddr.s_addr == laddr.s_addr && inp->inp_lport == lport) { /* * Found? */ if (cred == NULL || inp->inp_cred->cr_prison == cred->cr_prison) return (inp); } } /* * Not found. */ return (NULL); } else { struct inpcbporthead *porthash; struct inpcbport *phd; struct inpcb *match = NULL; /* * Best fit PCB lookup. * * First see if this local port is in use by looking on the * port hash list. */ porthash = &pcbinfo->ipi_porthashbase[INP_PCBPORTHASH(lport, pcbinfo->ipi_porthashmask)]; LIST_FOREACH(phd, porthash, phd_hash) { if (phd->phd_port == lport) break; } if (phd != NULL) { /* * Port is in use by one or more PCBs. Look for best * fit. */ LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) { wildcard = 0; if (cred != NULL && inp->inp_cred->cr_prison != cred->cr_prison) continue; #ifdef INET6 /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV4) == 0) continue; /* * We never select the PCB that has * INP_IPV6 flag and is bound to :: if * we have another PCB which is bound * to 0.0.0.0. If a PCB has the * INP_IPV6 flag, then we set its cost * higher than IPv4 only PCBs. * * Note that the case only happens * when a socket is bound to ::, under * the condition that the use of the * mapped address is allowed. */ if ((inp->inp_vflag & INP_IPV6) != 0) wildcard += INP_LOOKUP_MAPPED_PCB_COST; #endif if (inp->inp_faddr.s_addr != INADDR_ANY) wildcard++; if (inp->inp_laddr.s_addr != INADDR_ANY) { if (laddr.s_addr == INADDR_ANY) wildcard++; else if (inp->inp_laddr.s_addr != laddr.s_addr) continue; } else { if (laddr.s_addr != INADDR_ANY) wildcard++; } if (wildcard < matchwild) { match = inp; matchwild = wildcard; if (matchwild == 0) break; } } } return (match); } } #undef INP_LOOKUP_MAPPED_PCB_COST /* * Lookup PCB in hash list. */ struct inpcb * in_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in_addr faddr, u_int fport_arg, struct in_addr laddr, u_int lport_arg, int wildcard, struct ifnet *ifp) { struct inpcbhead *head; struct inpcb *inp, *tmpinp; u_short fport = fport_arg, lport = lport_arg; INP_INFO_LOCK_ASSERT(pcbinfo); /* * First look for an exact match. */ tmpinp = NULL; head = &pcbinfo->ipi_hashbase[INP_PCBHASH(faddr.s_addr, lport, fport, pcbinfo->ipi_hashmask)]; LIST_FOREACH(inp, head, inp_hash) { #ifdef INET6 /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV4) == 0) continue; #endif if (inp->inp_faddr.s_addr == faddr.s_addr && inp->inp_laddr.s_addr == laddr.s_addr && inp->inp_fport == fport && inp->inp_lport == lport) { /* * XXX We should be able to directly return * the inp here, without any checks. * Well unless both bound with SO_REUSEPORT? */ if (jailed(inp->inp_cred)) return (inp); if (tmpinp == NULL) tmpinp = inp; } } if (tmpinp != NULL) return (tmpinp); /* * Then look for a wildcard match, if requested. */ if (wildcard == INPLOOKUP_WILDCARD) { struct inpcb *local_wild = NULL, *local_exact = NULL; #ifdef INET6 struct inpcb *local_wild_mapped = NULL; #endif struct inpcb *jail_wild = NULL; int injail; /* * Order of socket selection - we always prefer jails. * 1. jailed, non-wild. * 2. jailed, wild. * 3. non-jailed, non-wild. * 4. non-jailed, wild. */ head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport, 0, pcbinfo->ipi_hashmask)]; LIST_FOREACH(inp, head, inp_hash) { #ifdef INET6 /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV4) == 0) continue; #endif if (inp->inp_faddr.s_addr != INADDR_ANY || inp->inp_lport != lport) continue; /* XXX inp locking */ if (ifp && ifp->if_type == IFT_FAITH && (inp->inp_flags & INP_FAITH) == 0) continue; injail = jailed(inp->inp_cred); if (injail) { - if (!prison_check_ip4(inp->inp_cred, &laddr)) + if (prison_check_ip4(inp->inp_cred, + &laddr) != 0) continue; } else { if (local_exact != NULL) continue; } if (inp->inp_laddr.s_addr == laddr.s_addr) { if (injail) return (inp); else local_exact = inp; } else if (inp->inp_laddr.s_addr == INADDR_ANY) { #ifdef INET6 /* XXX inp locking, NULL check */ if (inp->inp_vflag & INP_IPV6PROTO) local_wild_mapped = inp; else #endif /* INET6 */ if (injail) jail_wild = inp; else local_wild = inp; } } /* LIST_FOREACH */ if (jail_wild != NULL) return (jail_wild); if (local_exact != NULL) return (local_exact); if (local_wild != NULL) return (local_wild); #ifdef INET6 if (local_wild_mapped != NULL) return (local_wild_mapped); #endif /* defined(INET6) */ } /* if (wildcard == INPLOOKUP_WILDCARD) */ return (NULL); } /* * Insert PCB onto various hash lists. */ int in_pcbinshash(struct inpcb *inp) { struct inpcbhead *pcbhash; struct inpcbporthead *pcbporthash; struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; struct inpcbport *phd; u_int32_t hashkey_faddr; INP_INFO_WLOCK_ASSERT(pcbinfo); INP_WLOCK_ASSERT(inp); #ifdef INET6 if (inp->inp_vflag & INP_IPV6) hashkey_faddr = inp->in6p_faddr.s6_addr32[3] /* XXX */; else #endif /* INET6 */ hashkey_faddr = inp->inp_faddr.s_addr; pcbhash = &pcbinfo->ipi_hashbase[INP_PCBHASH(hashkey_faddr, inp->inp_lport, inp->inp_fport, pcbinfo->ipi_hashmask)]; pcbporthash = &pcbinfo->ipi_porthashbase[ INP_PCBPORTHASH(inp->inp_lport, pcbinfo->ipi_porthashmask)]; /* * Go through port list and look for a head for this lport. */ LIST_FOREACH(phd, pcbporthash, phd_hash) { if (phd->phd_port == inp->inp_lport) break; } /* * If none exists, malloc one and tack it on. */ if (phd == NULL) { phd = malloc(sizeof(struct inpcbport), M_PCB, M_NOWAIT); if (phd == NULL) { return (ENOBUFS); /* XXX */ } phd->phd_port = inp->inp_lport; LIST_INIT(&phd->phd_pcblist); LIST_INSERT_HEAD(pcbporthash, phd, phd_hash); } inp->inp_phd = phd; LIST_INSERT_HEAD(&phd->phd_pcblist, inp, inp_portlist); LIST_INSERT_HEAD(pcbhash, inp, inp_hash); return (0); } /* * Move PCB to the proper hash bucket when { faddr, fport } have been * changed. NOTE: This does not handle the case of the lport changing (the * hashed port list would have to be updated as well), so the lport must * not change after in_pcbinshash() has been called. */ void in_pcbrehash(struct inpcb *inp) { struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; struct inpcbhead *head; u_int32_t hashkey_faddr; INP_INFO_WLOCK_ASSERT(pcbinfo); INP_WLOCK_ASSERT(inp); #ifdef INET6 if (inp->inp_vflag & INP_IPV6) hashkey_faddr = inp->in6p_faddr.s6_addr32[3] /* XXX */; else #endif /* INET6 */ hashkey_faddr = inp->inp_faddr.s_addr; head = &pcbinfo->ipi_hashbase[INP_PCBHASH(hashkey_faddr, inp->inp_lport, inp->inp_fport, pcbinfo->ipi_hashmask)]; LIST_REMOVE(inp, inp_hash); LIST_INSERT_HEAD(head, inp, inp_hash); } /* * Remove PCB from various lists. */ void in_pcbremlists(struct inpcb *inp) { struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; INP_INFO_WLOCK_ASSERT(pcbinfo); INP_WLOCK_ASSERT(inp); inp->inp_gencnt = ++pcbinfo->ipi_gencnt; if (inp->inp_lport) { struct inpcbport *phd = inp->inp_phd; LIST_REMOVE(inp, inp_hash); LIST_REMOVE(inp, inp_portlist); if (LIST_FIRST(&phd->phd_pcblist) == NULL) { LIST_REMOVE(phd, phd_hash); free(phd, M_PCB); } } LIST_REMOVE(inp, inp_list); pcbinfo->ipi_count--; } /* * A set label operation has occurred at the socket layer, propagate the * label change into the in_pcb for the socket. */ void in_pcbsosetlabel(struct socket *so) { #ifdef MAC struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in_pcbsosetlabel: so->so_pcb == NULL")); INP_WLOCK(inp); SOCK_LOCK(so); mac_inpcb_sosetlabel(so, inp); SOCK_UNLOCK(so); INP_WUNLOCK(inp); #endif } /* * ipport_tick runs once per second, determining if random port allocation * should be continued. If more than ipport_randomcps ports have been * allocated in the last second, then we return to sequential port * allocation. We return to random allocation only once we drop below * ipport_randomcps for at least ipport_randomtime seconds. */ void ipport_tick(void *xtp) { VNET_ITERATOR_DECL(vnet_iter); VNET_LIST_RLOCK(); VNET_FOREACH(vnet_iter) { CURVNET_SET(vnet_iter); /* XXX appease INVARIANTS here */ INIT_VNET_INET(vnet_iter); if (V_ipport_tcpallocs <= V_ipport_tcplastcount + V_ipport_randomcps) { if (V_ipport_stoprandom > 0) V_ipport_stoprandom--; } else V_ipport_stoprandom = V_ipport_randomtime; V_ipport_tcplastcount = V_ipport_tcpallocs; CURVNET_RESTORE(); } VNET_LIST_RUNLOCK(); callout_reset(&ipport_tick_callout, hz, ipport_tick, NULL); } void inp_wlock(struct inpcb *inp) { INP_WLOCK(inp); } void inp_wunlock(struct inpcb *inp) { INP_WUNLOCK(inp); } void inp_rlock(struct inpcb *inp) { INP_RLOCK(inp); } void inp_runlock(struct inpcb *inp) { INP_RUNLOCK(inp); } #ifdef INVARIANTS void inp_lock_assert(struct inpcb *inp) { INP_WLOCK_ASSERT(inp); } void inp_unlock_assert(struct inpcb *inp) { INP_UNLOCK_ASSERT(inp); } #endif void inp_apply_all(void (*func)(struct inpcb *, void *), void *arg) { INIT_VNET_INET(curvnet); struct inpcb *inp; INP_INFO_RLOCK(&V_tcbinfo); LIST_FOREACH(inp, V_tcbinfo.ipi_listhead, inp_list) { INP_WLOCK(inp); func(inp, arg); INP_WUNLOCK(inp); } INP_INFO_RUNLOCK(&V_tcbinfo); } struct socket * inp_inpcbtosocket(struct inpcb *inp) { INP_WLOCK_ASSERT(inp); return (inp->inp_socket); } struct tcpcb * inp_inpcbtotcpcb(struct inpcb *inp) { INP_WLOCK_ASSERT(inp); return ((struct tcpcb *)inp->inp_ppcb); } int inp_ip_tos_get(const struct inpcb *inp) { return (inp->inp_ip_tos); } void inp_ip_tos_set(struct inpcb *inp, int val) { inp->inp_ip_tos = val; } void inp_4tuple_get(struct inpcb *inp, uint32_t *laddr, uint16_t *lp, uint32_t *faddr, uint16_t *fp) { INP_LOCK_ASSERT(inp); *laddr = inp->inp_laddr.s_addr; *faddr = inp->inp_faddr.s_addr; *lp = inp->inp_lport; *fp = inp->inp_fport; } struct inpcb * so_sotoinpcb(struct socket *so) { return (sotoinpcb(so)); } struct tcpcb * so_sototcpcb(struct socket *so) { return (sototcpcb(so)); } #ifdef DDB static void db_print_indent(int indent) { int i; for (i = 0; i < indent; i++) db_printf(" "); } static void db_print_inconninfo(struct in_conninfo *inc, const char *name, int indent) { char faddr_str[48], laddr_str[48]; db_print_indent(indent); db_printf("%s at %p\n", name, inc); indent += 2; #ifdef INET6 if (inc->inc_flags & INC_ISIPV6) { /* IPv6. */ ip6_sprintf(laddr_str, &inc->inc6_laddr); ip6_sprintf(faddr_str, &inc->inc6_faddr); } else { #endif /* IPv4. */ inet_ntoa_r(inc->inc_laddr, laddr_str); inet_ntoa_r(inc->inc_faddr, faddr_str); #ifdef INET6 } #endif db_print_indent(indent); db_printf("inc_laddr %s inc_lport %u\n", laddr_str, ntohs(inc->inc_lport)); db_print_indent(indent); db_printf("inc_faddr %s inc_fport %u\n", faddr_str, ntohs(inc->inc_fport)); } static void db_print_inpflags(int inp_flags) { int comma; comma = 0; if (inp_flags & INP_RECVOPTS) { db_printf("%sINP_RECVOPTS", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_RECVRETOPTS) { db_printf("%sINP_RECVRETOPTS", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_RECVDSTADDR) { db_printf("%sINP_RECVDSTADDR", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_HDRINCL) { db_printf("%sINP_HDRINCL", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_HIGHPORT) { db_printf("%sINP_HIGHPORT", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_LOWPORT) { db_printf("%sINP_LOWPORT", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_ANONPORT) { db_printf("%sINP_ANONPORT", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_RECVIF) { db_printf("%sINP_RECVIF", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_MTUDISC) { db_printf("%sINP_MTUDISC", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_FAITH) { db_printf("%sINP_FAITH", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_RECVTTL) { db_printf("%sINP_RECVTTL", comma ? ", " : ""); comma = 1; } if (inp_flags & INP_DONTFRAG) { db_printf("%sINP_DONTFRAG", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_IPV6_V6ONLY) { db_printf("%sIN6P_IPV6_V6ONLY", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_PKTINFO) { db_printf("%sIN6P_PKTINFO", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_HOPLIMIT) { db_printf("%sIN6P_HOPLIMIT", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_HOPOPTS) { db_printf("%sIN6P_HOPOPTS", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_DSTOPTS) { db_printf("%sIN6P_DSTOPTS", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_RTHDR) { db_printf("%sIN6P_RTHDR", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_RTHDRDSTOPTS) { db_printf("%sIN6P_RTHDRDSTOPTS", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_TCLASS) { db_printf("%sIN6P_TCLASS", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_AUTOFLOWLABEL) { db_printf("%sIN6P_AUTOFLOWLABEL", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_RFC2292) { db_printf("%sIN6P_RFC2292", comma ? ", " : ""); comma = 1; } if (inp_flags & IN6P_MTU) { db_printf("IN6P_MTU%s", comma ? ", " : ""); comma = 1; } } static void db_print_inpvflag(u_char inp_vflag) { int comma; comma = 0; if (inp_vflag & INP_IPV4) { db_printf("%sINP_IPV4", comma ? ", " : ""); comma = 1; } if (inp_vflag & INP_IPV6) { db_printf("%sINP_IPV6", comma ? ", " : ""); comma = 1; } if (inp_vflag & INP_IPV6PROTO) { db_printf("%sINP_IPV6PROTO", comma ? ", " : ""); comma = 1; } if (inp_vflag & INP_TIMEWAIT) { db_printf("%sINP_TIMEWAIT", comma ? ", " : ""); comma = 1; } if (inp_vflag & INP_ONESBCAST) { db_printf("%sINP_ONESBCAST", comma ? ", " : ""); comma = 1; } if (inp_vflag & INP_DROPPED) { db_printf("%sINP_DROPPED", comma ? ", " : ""); comma = 1; } if (inp_vflag & INP_SOCKREF) { db_printf("%sINP_SOCKREF", comma ? ", " : ""); comma = 1; } } void db_print_inpcb(struct inpcb *inp, const char *name, int indent) { db_print_indent(indent); db_printf("%s at %p\n", name, inp); indent += 2; db_print_indent(indent); db_printf("inp_flow: 0x%x\n", inp->inp_flow); db_print_inconninfo(&inp->inp_inc, "inp_conninfo", indent); db_print_indent(indent); db_printf("inp_ppcb: %p inp_pcbinfo: %p inp_socket: %p\n", inp->inp_ppcb, inp->inp_pcbinfo, inp->inp_socket); db_print_indent(indent); db_printf("inp_label: %p inp_flags: 0x%x (", inp->inp_label, inp->inp_flags); db_print_inpflags(inp->inp_flags); db_printf(")\n"); db_print_indent(indent); db_printf("inp_sp: %p inp_vflag: 0x%x (", inp->inp_sp, inp->inp_vflag); db_print_inpvflag(inp->inp_vflag); db_printf(")\n"); db_print_indent(indent); db_printf("inp_ip_ttl: %d inp_ip_p: %d inp_ip_minttl: %d\n", inp->inp_ip_ttl, inp->inp_ip_p, inp->inp_ip_minttl); db_print_indent(indent); #ifdef INET6 if (inp->inp_vflag & INP_IPV6) { db_printf("in6p_options: %p in6p_outputopts: %p " "in6p_moptions: %p\n", inp->in6p_options, inp->in6p_outputopts, inp->in6p_moptions); db_printf("in6p_icmp6filt: %p in6p_cksum %d " "in6p_hops %u\n", inp->in6p_icmp6filt, inp->in6p_cksum, inp->in6p_hops); } else #endif { db_printf("inp_ip_tos: %d inp_ip_options: %p " "inp_ip_moptions: %p\n", inp->inp_ip_tos, inp->inp_options, inp->inp_moptions); } db_print_indent(indent); db_printf("inp_phd: %p inp_gencnt: %ju\n", inp->inp_phd, (uintmax_t)inp->inp_gencnt); } DB_SHOW_COMMAND(inpcb, db_show_inpcb) { struct inpcb *inp; if (!have_addr) { db_printf("usage: show inpcb \n"); return; } inp = (struct inpcb *)addr; db_print_inpcb(inp, "inpcb", 0); } #endif diff --git a/sys/netinet/raw_ip.c b/sys/netinet/raw_ip.c index 8f6d8ecd4a55..27f4b9ca179e 100644 --- a/sys/netinet/raw_ip.c +++ b/sys/netinet/raw_ip.c @@ -1,1014 +1,1011 @@ /*- * Copyright (c) 1982, 1986, 1988, 1993 * The Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95 */ #include __FBSDID("$FreeBSD$"); #include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_mac.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef IPSEC #include #endif /*IPSEC*/ #include #ifdef VIMAGE_GLOBALS struct inpcbhead ripcb; struct inpcbinfo ripcbinfo; #endif /* control hooks for ipfw and dummynet */ ip_fw_ctl_t *ip_fw_ctl_ptr = NULL; ip_dn_ctl_t *ip_dn_ctl_ptr = NULL; /* * Hooks for multicast routing. They all default to NULL, so leave them not * initialized and rely on BSS being set to 0. */ /* * The socket used to communicate with the multicast routing daemon. */ #ifdef VIMAGE_GLOBALS struct socket *ip_mrouter; #endif /* * The various mrouter and rsvp functions. */ int (*ip_mrouter_set)(struct socket *, struct sockopt *); int (*ip_mrouter_get)(struct socket *, struct sockopt *); int (*ip_mrouter_done)(void); int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *, struct ip_moptions *); int (*mrt_ioctl)(int, caddr_t, int); int (*legal_vif_num)(int); u_long (*ip_mcast_src)(int); void (*rsvp_input_p)(struct mbuf *m, int off); int (*ip_rsvp_vif)(struct socket *, struct sockopt *); void (*ip_rsvp_force_done)(struct socket *); /* * Hash functions */ #define INP_PCBHASH_RAW_SIZE 256 #define INP_PCBHASH_RAW(proto, laddr, faddr, mask) \ (((proto) + (laddr) + (faddr)) % (mask) + 1) static void rip_inshash(struct inpcb *inp) { struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; struct inpcbhead *pcbhash; int hash; INP_INFO_WLOCK_ASSERT(pcbinfo); INP_WLOCK_ASSERT(inp); if (inp->inp_ip_p != 0 && inp->inp_laddr.s_addr != INADDR_ANY && inp->inp_faddr.s_addr != INADDR_ANY) { hash = INP_PCBHASH_RAW(inp->inp_ip_p, inp->inp_laddr.s_addr, inp->inp_faddr.s_addr, pcbinfo->ipi_hashmask); } else hash = 0; pcbhash = &pcbinfo->ipi_hashbase[hash]; LIST_INSERT_HEAD(pcbhash, inp, inp_hash); } static void rip_delhash(struct inpcb *inp) { INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); LIST_REMOVE(inp, inp_hash); } /* * Raw interface to IP protocol. */ /* * Initialize raw connection block q. */ static void rip_zone_change(void *tag) { INIT_VNET_INET(curvnet); uma_zone_set_max(V_ripcbinfo.ipi_zone, maxsockets); } static int rip_inpcb_init(void *mem, int size, int flags) { struct inpcb *inp = mem; INP_LOCK_INIT(inp, "inp", "rawinp"); return (0); } void rip_init(void) { INIT_VNET_INET(curvnet); INP_INFO_LOCK_INIT(&V_ripcbinfo, "rip"); LIST_INIT(&V_ripcb); V_ripcbinfo.ipi_listhead = &V_ripcb; V_ripcbinfo.ipi_hashbase = hashinit(INP_PCBHASH_RAW_SIZE, M_PCB, &V_ripcbinfo.ipi_hashmask); V_ripcbinfo.ipi_porthashbase = hashinit(1, M_PCB, &V_ripcbinfo.ipi_porthashmask); V_ripcbinfo.ipi_zone = uma_zcreate("ripcb", sizeof(struct inpcb), NULL, NULL, rip_inpcb_init, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE); uma_zone_set_max(V_ripcbinfo.ipi_zone, maxsockets); EVENTHANDLER_REGISTER(maxsockets_change, rip_zone_change, NULL, EVENTHANDLER_PRI_ANY); } static int rip_append(struct inpcb *last, struct ip *ip, struct mbuf *n, struct sockaddr_in *ripsrc) { int policyfail = 0; INP_RLOCK_ASSERT(last); #ifdef IPSEC /* check AH/ESP integrity. */ if (ipsec4_in_reject(n, last)) { policyfail = 1; } #endif /* IPSEC */ #ifdef MAC if (!policyfail && mac_inpcb_check_deliver(last, n) != 0) policyfail = 1; #endif /* Check the minimum TTL for socket. */ if (last->inp_ip_minttl && last->inp_ip_minttl > ip->ip_ttl) policyfail = 1; if (!policyfail) { struct mbuf *opts = NULL; struct socket *so; so = last->inp_socket; if ((last->inp_flags & INP_CONTROLOPTS) || (so->so_options & (SO_TIMESTAMP | SO_BINTIME))) ip_savecontrol(last, &opts, ip, n); SOCKBUF_LOCK(&so->so_rcv); if (sbappendaddr_locked(&so->so_rcv, (struct sockaddr *)ripsrc, n, opts) == 0) { /* should notify about lost packet */ m_freem(n); if (opts) m_freem(opts); SOCKBUF_UNLOCK(&so->so_rcv); } else sorwakeup_locked(so); } else m_freem(n); return (policyfail); } /* * Setup generic address and protocol structures for raw_input routine, then * pass them along with mbuf chain. */ void rip_input(struct mbuf *m, int off) { INIT_VNET_INET(curvnet); struct ip *ip = mtod(m, struct ip *); int proto = ip->ip_p; struct inpcb *inp, *last; struct sockaddr_in ripsrc; int hash; bzero(&ripsrc, sizeof(ripsrc)); ripsrc.sin_len = sizeof(ripsrc); ripsrc.sin_family = AF_INET; ripsrc.sin_addr = ip->ip_src; last = NULL; hash = INP_PCBHASH_RAW(proto, ip->ip_src.s_addr, ip->ip_dst.s_addr, V_ripcbinfo.ipi_hashmask); INP_INFO_RLOCK(&V_ripcbinfo); LIST_FOREACH(inp, &V_ripcbinfo.ipi_hashbase[hash], inp_hash) { if (inp->inp_ip_p != proto) continue; #ifdef INET6 /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV4) == 0) continue; #endif if (inp->inp_laddr.s_addr != ip->ip_dst.s_addr) continue; if (inp->inp_faddr.s_addr != ip->ip_src.s_addr) continue; - if (jailed(inp->inp_cred)) { - if (!prison_check_ip4(inp->inp_cred, &ip->ip_dst)) - continue; - } + if (prison_check_ip4(inp->inp_cred, &ip->ip_dst) != 0) + continue; if (last != NULL) { struct mbuf *n; n = m_copy(m, 0, (int)M_COPYALL); if (n != NULL) (void) rip_append(last, ip, n, &ripsrc); /* XXX count dropped packet */ INP_RUNLOCK(last); } INP_RLOCK(inp); last = inp; } LIST_FOREACH(inp, &V_ripcbinfo.ipi_hashbase[0], inp_hash) { if (inp->inp_ip_p && inp->inp_ip_p != proto) continue; #ifdef INET6 /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV4) == 0) continue; #endif if (inp->inp_laddr.s_addr && inp->inp_laddr.s_addr != ip->ip_dst.s_addr) continue; if (inp->inp_faddr.s_addr && inp->inp_faddr.s_addr != ip->ip_src.s_addr) continue; - if (jailed(inp->inp_cred)) { - if (!prison_check_ip4(inp->inp_cred, &ip->ip_dst)) - continue; - } + if (prison_check_ip4(inp->inp_cred, &ip->ip_dst) != 0) + continue; if (last != NULL) { struct mbuf *n; n = m_copy(m, 0, (int)M_COPYALL); if (n != NULL) (void) rip_append(last, ip, n, &ripsrc); /* XXX count dropped packet */ INP_RUNLOCK(last); } INP_RLOCK(inp); last = inp; } INP_INFO_RUNLOCK(&V_ripcbinfo); if (last != NULL) { if (rip_append(last, ip, m, &ripsrc) != 0) V_ipstat.ips_delivered--; INP_RUNLOCK(last); } else { m_freem(m); V_ipstat.ips_noproto++; V_ipstat.ips_delivered--; } } /* * Generate IP header and pass packet to ip_output. Tack on options user may * have setup with control call. */ int rip_output(struct mbuf *m, struct socket *so, u_long dst) { INIT_VNET_INET(so->so_vnet); struct ip *ip; int error; struct inpcb *inp = sotoinpcb(so); int flags = ((so->so_options & SO_DONTROUTE) ? IP_ROUTETOIF : 0) | IP_ALLOWBROADCAST; /* * If the user handed us a complete IP packet, use it. Otherwise, * allocate an mbuf for a header and fill it in. */ if ((inp->inp_flags & INP_HDRINCL) == 0) { if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) { m_freem(m); return(EMSGSIZE); } M_PREPEND(m, sizeof(struct ip), M_DONTWAIT); if (m == NULL) return(ENOBUFS); INP_RLOCK(inp); ip = mtod(m, struct ip *); ip->ip_tos = inp->inp_ip_tos; if (inp->inp_flags & INP_DONTFRAG) ip->ip_off = IP_DF; else ip->ip_off = 0; ip->ip_p = inp->inp_ip_p; ip->ip_len = m->m_pkthdr.len; - if (jailed(inp->inp_cred)) { - if (prison_get_ip4(inp->inp_cred, &ip->ip_src) != 0) { - INP_RUNLOCK(inp); - m_freem(m); - return (EPERM); - } - } else { - ip->ip_src = inp->inp_laddr; + ip->ip_src = inp->inp_laddr; + error = prison_get_ip4(inp->inp_cred, &ip->ip_src); + if (error != 0) { + INP_RUNLOCK(inp); + m_freem(m); + return (error); } ip->ip_dst.s_addr = dst; ip->ip_ttl = inp->inp_ip_ttl; } else { if (m->m_pkthdr.len > IP_MAXPACKET) { m_freem(m); return(EMSGSIZE); } INP_RLOCK(inp); ip = mtod(m, struct ip *); - if (!prison_check_ip4(inp->inp_cred, &ip->ip_src)) { + error = prison_check_ip4(inp->inp_cred, &ip->ip_src); + if (error != 0) { INP_RUNLOCK(inp); m_freem(m); - return (EPERM); + return (error); } /* * Don't allow both user specified and setsockopt options, * and don't allow packet length sizes that will crash. */ if (((ip->ip_hl != (sizeof (*ip) >> 2)) && inp->inp_options) || (ip->ip_len > m->m_pkthdr.len) || (ip->ip_len < (ip->ip_hl << 2))) { INP_RUNLOCK(inp); m_freem(m); return (EINVAL); } if (ip->ip_id == 0) ip->ip_id = ip_newid(); /* * XXX prevent ip_output from overwriting header fields. */ flags |= IP_RAWOUTPUT; V_ipstat.ips_rawout++; } if (inp->inp_flags & INP_ONESBCAST) flags |= IP_SENDONES; #ifdef MAC mac_inpcb_create_mbuf(inp, m); #endif error = ip_output(m, inp->inp_options, NULL, flags, inp->inp_moptions, inp); INP_RUNLOCK(inp); return (error); } /* * Raw IP socket option processing. * * IMPORTANT NOTE regarding access control: Traditionally, raw sockets could * only be created by a privileged process, and as such, socket option * operations to manage system properties on any raw socket were allowed to * take place without explicit additional access control checks. However, * raw sockets can now also be created in jail(), and therefore explicit * checks are now required. Likewise, raw sockets can be used by a process * after it gives up privilege, so some caution is required. For options * passed down to the IP layer via ip_ctloutput(), checks are assumed to be * performed in ip_ctloutput() and therefore no check occurs here. * Unilaterally checking priv_check() here breaks normal IP socket option * operations on raw sockets. * * When adding new socket options here, make sure to add access control * checks here as necessary. */ int rip_ctloutput(struct socket *so, struct sockopt *sopt) { struct inpcb *inp = sotoinpcb(so); int error, optval; if (sopt->sopt_level != IPPROTO_IP) { if ((sopt->sopt_level == SOL_SOCKET) && (sopt->sopt_name == SO_SETFIB)) { inp->inp_inc.inc_fibnum = so->so_fibnum; return (0); } return (EINVAL); } error = 0; switch (sopt->sopt_dir) { case SOPT_GET: switch (sopt->sopt_name) { case IP_HDRINCL: optval = inp->inp_flags & INP_HDRINCL; error = sooptcopyout(sopt, &optval, sizeof optval); break; case IP_FW_ADD: /* ADD actually returns the body... */ case IP_FW_GET: case IP_FW_TABLE_GETSIZE: case IP_FW_TABLE_LIST: case IP_FW_NAT_GET_CONFIG: case IP_FW_NAT_GET_LOG: if (ip_fw_ctl_ptr != NULL) error = ip_fw_ctl_ptr(sopt); else error = ENOPROTOOPT; break; case IP_DUMMYNET_GET: if (ip_dn_ctl_ptr != NULL) error = ip_dn_ctl_ptr(sopt); else error = ENOPROTOOPT; break ; case MRT_INIT: case MRT_DONE: case MRT_ADD_VIF: case MRT_DEL_VIF: case MRT_ADD_MFC: case MRT_DEL_MFC: case MRT_VERSION: case MRT_ASSERT: case MRT_API_SUPPORT: case MRT_API_CONFIG: case MRT_ADD_BW_UPCALL: case MRT_DEL_BW_UPCALL: error = priv_check(curthread, PRIV_NETINET_MROUTE); if (error != 0) return (error); error = ip_mrouter_get ? ip_mrouter_get(so, sopt) : EOPNOTSUPP; break; default: error = ip_ctloutput(so, sopt); break; } break; case SOPT_SET: switch (sopt->sopt_name) { case IP_HDRINCL: error = sooptcopyin(sopt, &optval, sizeof optval, sizeof optval); if (error) break; if (optval) inp->inp_flags |= INP_HDRINCL; else inp->inp_flags &= ~INP_HDRINCL; break; case IP_FW_ADD: case IP_FW_DEL: case IP_FW_FLUSH: case IP_FW_ZERO: case IP_FW_RESETLOG: case IP_FW_TABLE_ADD: case IP_FW_TABLE_DEL: case IP_FW_TABLE_FLUSH: case IP_FW_NAT_CFG: case IP_FW_NAT_DEL: if (ip_fw_ctl_ptr != NULL) error = ip_fw_ctl_ptr(sopt); else error = ENOPROTOOPT; break; case IP_DUMMYNET_CONFIGURE: case IP_DUMMYNET_DEL: case IP_DUMMYNET_FLUSH: if (ip_dn_ctl_ptr != NULL) error = ip_dn_ctl_ptr(sopt); else error = ENOPROTOOPT ; break ; case IP_RSVP_ON: error = priv_check(curthread, PRIV_NETINET_MROUTE); if (error != 0) return (error); error = ip_rsvp_init(so); break; case IP_RSVP_OFF: error = priv_check(curthread, PRIV_NETINET_MROUTE); if (error != 0) return (error); error = ip_rsvp_done(); break; case IP_RSVP_VIF_ON: case IP_RSVP_VIF_OFF: error = priv_check(curthread, PRIV_NETINET_MROUTE); if (error != 0) return (error); error = ip_rsvp_vif ? ip_rsvp_vif(so, sopt) : EINVAL; break; case MRT_INIT: case MRT_DONE: case MRT_ADD_VIF: case MRT_DEL_VIF: case MRT_ADD_MFC: case MRT_DEL_MFC: case MRT_VERSION: case MRT_ASSERT: case MRT_API_SUPPORT: case MRT_API_CONFIG: case MRT_ADD_BW_UPCALL: case MRT_DEL_BW_UPCALL: error = priv_check(curthread, PRIV_NETINET_MROUTE); if (error != 0) return (error); error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : EOPNOTSUPP; break; default: error = ip_ctloutput(so, sopt); break; } break; } return (error); } /* * This function exists solely to receive the PRC_IFDOWN messages which are * sent by if_down(). It looks for an ifaddr whose ifa_addr is sa, and calls * in_ifadown() to remove all routes corresponding to that address. It also * receives the PRC_IFUP messages from if_up() and reinstalls the interface * routes. */ void rip_ctlinput(int cmd, struct sockaddr *sa, void *vip) { INIT_VNET_INET(curvnet); struct in_ifaddr *ia; struct ifnet *ifp; int err; int flags; switch (cmd) { case PRC_IFDOWN: TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) { if (ia->ia_ifa.ifa_addr == sa && (ia->ia_flags & IFA_ROUTE)) { /* * in_ifscrub kills the interface route. */ in_ifscrub(ia->ia_ifp, ia); /* * in_ifadown gets rid of all the rest of the * routes. This is not quite the right thing * to do, but at least if we are running a * routing process they will come back. */ in_ifadown(&ia->ia_ifa, 0); break; } } break; case PRC_IFUP: TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) { if (ia->ia_ifa.ifa_addr == sa) break; } if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) return; flags = RTF_UP; ifp = ia->ia_ifa.ifa_ifp; if ((ifp->if_flags & IFF_LOOPBACK) || (ifp->if_flags & IFF_POINTOPOINT)) flags |= RTF_HOST; err = rtinit(&ia->ia_ifa, RTM_ADD, flags); if (err == 0) ia->ia_flags |= IFA_ROUTE; break; } } u_long rip_sendspace = 9216; u_long rip_recvspace = 9216; SYSCTL_ULONG(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); SYSCTL_ULONG(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, &rip_recvspace, 0, "Maximum space for incoming raw IP datagrams"); static int rip_attach(struct socket *so, int proto, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp == NULL, ("rip_attach: inp != NULL")); error = priv_check(td, PRIV_NETINET_RAW); if (error) return (error); if (proto >= IPPROTO_MAX || proto < 0) return EPROTONOSUPPORT; error = soreserve(so, rip_sendspace, rip_recvspace); if (error) return (error); INP_INFO_WLOCK(&V_ripcbinfo); error = in_pcballoc(so, &V_ripcbinfo); if (error) { INP_INFO_WUNLOCK(&V_ripcbinfo); return (error); } inp = (struct inpcb *)so->so_pcb; inp->inp_vflag |= INP_IPV4; inp->inp_ip_p = proto; inp->inp_ip_ttl = V_ip_defttl; rip_inshash(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); INP_WUNLOCK(inp); return (0); } static void rip_detach(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_detach: inp == NULL")); KASSERT(inp->inp_faddr.s_addr == INADDR_ANY, ("rip_detach: not closed")); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); rip_delhash(inp); if (so == V_ip_mrouter && ip_mrouter_done) ip_mrouter_done(); if (ip_rsvp_force_done) ip_rsvp_force_done(so); if (so == V_ip_rsvpd) ip_rsvp_done(); in_pcbdetach(inp); in_pcbfree(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); } static void rip_dodisconnect(struct socket *so, struct inpcb *inp) { INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); rip_delhash(inp); inp->inp_faddr.s_addr = INADDR_ANY; rip_inshash(inp); SOCK_LOCK(so); so->so_state &= ~SS_ISCONNECTED; SOCK_UNLOCK(so); } static void rip_abort(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_abort: inp == NULL")); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); rip_dodisconnect(so, inp); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); } static void rip_close(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_close: inp == NULL")); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); rip_dodisconnect(so, inp); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); } static int rip_disconnect(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; if ((so->so_state & SS_ISCONNECTED) == 0) return (ENOTCONN); inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_disconnect: inp == NULL")); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); rip_dodisconnect(so, inp); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (0); } static int rip_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_NET(so->so_vnet); INIT_VNET_INET(so->so_vnet); struct sockaddr_in *addr = (struct sockaddr_in *)nam; struct inpcb *inp; + int error; if (nam->sa_len != sizeof(*addr)) return (EINVAL); - if (!prison_check_ip4(td->td_ucred, &addr->sin_addr)) - return (EADDRNOTAVAIL); + error = prison_check_ip4(td->td_ucred, &addr->sin_addr); + if (error != 0) + return (error); if (TAILQ_EMPTY(&V_ifnet) || (addr->sin_family != AF_INET && addr->sin_family != AF_IMPLINK) || (addr->sin_addr.s_addr && ifa_ifwithaddr((struct sockaddr *)addr) == 0)) return (EADDRNOTAVAIL); inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_bind: inp == NULL")); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); rip_delhash(inp); inp->inp_laddr = addr->sin_addr; rip_inshash(inp); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (0); } static int rip_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_NET(so->so_vnet); INIT_VNET_INET(so->so_vnet); struct sockaddr_in *addr = (struct sockaddr_in *)nam; struct inpcb *inp; if (nam->sa_len != sizeof(*addr)) return (EINVAL); if (TAILQ_EMPTY(&V_ifnet)) return (EADDRNOTAVAIL); if (addr->sin_family != AF_INET && addr->sin_family != AF_IMPLINK) return (EAFNOSUPPORT); inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_connect: inp == NULL")); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); rip_delhash(inp); inp->inp_faddr = addr->sin_addr; rip_inshash(inp); soisconnected(so); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (0); } static int rip_shutdown(struct socket *so) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_shutdown: inp == NULL")); INP_WLOCK(inp); socantsendmore(so); INP_WUNLOCK(inp); return (0); } static int rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, struct mbuf *control, struct thread *td) { struct inpcb *inp; u_long dst; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip_send: inp == NULL")); /* * Note: 'dst' reads below are unlocked. */ if (so->so_state & SS_ISCONNECTED) { if (nam) { m_freem(m); return (EISCONN); } dst = inp->inp_faddr.s_addr; /* Unlocked read. */ } else { if (nam == NULL) { m_freem(m); return (ENOTCONN); } dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; } return (rip_output(m, so, dst)); } static int rip_pcblist(SYSCTL_HANDLER_ARGS) { INIT_VNET_INET(curvnet); int error, i, n; struct inpcb *inp, **inp_list; inp_gen_t gencnt; struct xinpgen xig; /* * The process of preparing the TCB list is too time-consuming and * resource-intensive to repeat twice on every request. */ if (req->oldptr == 0) { n = V_ripcbinfo.ipi_count; req->oldidx = 2 * (sizeof xig) + (n + n/8) * sizeof(struct xinpcb); return (0); } if (req->newptr != 0) return (EPERM); /* * OK, now we're committed to doing something. */ INP_INFO_RLOCK(&V_ripcbinfo); gencnt = V_ripcbinfo.ipi_gencnt; n = V_ripcbinfo.ipi_count; INP_INFO_RUNLOCK(&V_ripcbinfo); xig.xig_len = sizeof xig; xig.xig_count = n; xig.xig_gen = gencnt; xig.xig_sogen = so_gencnt; error = SYSCTL_OUT(req, &xig, sizeof xig); if (error) return (error); inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK); if (inp_list == 0) return (ENOMEM); INP_INFO_RLOCK(&V_ripcbinfo); for (inp = LIST_FIRST(V_ripcbinfo.ipi_listhead), i = 0; inp && i < n; inp = LIST_NEXT(inp, inp_list)) { INP_RLOCK(inp); if (inp->inp_gencnt <= gencnt && cr_canseeinpcb(req->td->td_ucred, inp) == 0) { /* XXX held references? */ inp_list[i++] = inp; } INP_RUNLOCK(inp); } INP_INFO_RUNLOCK(&V_ripcbinfo); n = i; error = 0; for (i = 0; i < n; i++) { inp = inp_list[i]; INP_RLOCK(inp); if (inp->inp_gencnt <= gencnt) { struct xinpcb xi; bzero(&xi, sizeof(xi)); xi.xi_len = sizeof xi; /* XXX should avoid extra copy */ bcopy(inp, &xi.xi_inp, sizeof *inp); if (inp->inp_socket) sotoxsocket(inp->inp_socket, &xi.xi_socket); INP_RUNLOCK(inp); error = SYSCTL_OUT(req, &xi, sizeof xi); } else INP_RUNLOCK(inp); } if (!error) { /* * Give the user an updated idea of our state. If the * generation differs from what we told her before, she knows * that something happened while we were processing this * request, and it might be necessary to retry. */ INP_INFO_RLOCK(&V_ripcbinfo); xig.xig_gen = V_ripcbinfo.ipi_gencnt; xig.xig_sogen = so_gencnt; xig.xig_count = V_ripcbinfo.ipi_count; INP_INFO_RUNLOCK(&V_ripcbinfo); error = SYSCTL_OUT(req, &xig, sizeof xig); } free(inp_list, M_TEMP); return (error); } SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, 0, 0, rip_pcblist, "S,xinpcb", "List of active raw IP sockets"); struct pr_usrreqs rip_usrreqs = { .pru_abort = rip_abort, .pru_attach = rip_attach, .pru_bind = rip_bind, .pru_connect = rip_connect, .pru_control = in_control, .pru_detach = rip_detach, .pru_disconnect = rip_disconnect, .pru_peeraddr = in_getpeeraddr, .pru_send = rip_send, .pru_shutdown = rip_shutdown, .pru_sockaddr = in_getsockaddr, .pru_sosetlabel = in_pcbsosetlabel, .pru_close = rip_close, }; diff --git a/sys/netinet/tcp_usrreq.c b/sys/netinet/tcp_usrreq.c index f6dcae1a20b0..89987c4c7e42 100644 --- a/sys/netinet/tcp_usrreq.c +++ b/sys/netinet/tcp_usrreq.c @@ -1,1898 +1,1895 @@ /*- * Copyright (c) 1982, 1986, 1988, 1993 * The Regents of the University of California. * Copyright (c) 2006-2007 Robert N. M. Watson * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * From: @(#)tcp_usrreq.c 8.2 (Berkeley) 1/3/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_ddb.h" #include "opt_inet.h" #include "opt_inet6.h" #include "opt_tcpdebug.h" #include #include #include #include #include #include #ifdef INET6 #include #endif /* INET6 */ #include #include #include #include #include #include #ifdef DDB #include #endif #include #include #include #include #ifdef INET6 #include #endif #include #ifdef INET6 #include #endif #include #include #ifdef INET6 #include #include #endif #include #include #include #include #include #include #ifdef TCPDEBUG #include #endif #include #include /* * TCP protocol interface to socket abstraction. */ static int tcp_attach(struct socket *); static int tcp_connect(struct tcpcb *, struct sockaddr *, struct thread *td); #ifdef INET6 static int tcp6_connect(struct tcpcb *, struct sockaddr *, struct thread *td); #endif /* INET6 */ static void tcp_disconnect(struct tcpcb *); static void tcp_usrclosed(struct tcpcb *); static void tcp_fill_info(struct tcpcb *, struct tcp_info *); #ifdef TCPDEBUG #define TCPDEBUG0 int ostate = 0 #define TCPDEBUG1() ostate = tp ? tp->t_state : 0 #define TCPDEBUG2(req) if (tp && (so->so_options & SO_DEBUG)) \ tcp_trace(TA_USER, ostate, tp, 0, 0, req) #else #define TCPDEBUG0 #define TCPDEBUG1() #define TCPDEBUG2(req) #endif /* * TCP attaches to socket via pru_attach(), reserving space, * and an internet control block. */ static int tcp_usr_attach(struct socket *so, int proto, struct thread *td) { struct inpcb *inp; struct tcpcb *tp = NULL; int error; TCPDEBUG0; inp = sotoinpcb(so); KASSERT(inp == NULL, ("tcp_usr_attach: inp != NULL")); TCPDEBUG1(); error = tcp_attach(so); if (error) goto out; if ((so->so_options & SO_LINGER) && so->so_linger == 0) so->so_linger = TCP_LINGERTIME; inp = sotoinpcb(so); tp = intotcpcb(inp); out: TCPDEBUG2(PRU_ATTACH); return error; } /* * tcp_detach is called when the socket layer loses its final reference * to the socket, be it a file descriptor reference, a reference from TCP, * etc. At this point, there is only one case in which we will keep around * inpcb state: time wait. * * This function can probably be re-absorbed back into tcp_usr_detach() now * that there is a single detach path. */ static void tcp_detach(struct socket *so, struct inpcb *inp) { struct tcpcb *tp; #ifdef INVARIANTS INIT_VNET_INET(so->so_vnet); #endif INP_INFO_WLOCK_ASSERT(&V_tcbinfo); INP_WLOCK_ASSERT(inp); KASSERT(so->so_pcb == inp, ("tcp_detach: so_pcb != inp")); KASSERT(inp->inp_socket == so, ("tcp_detach: inp_socket != so")); tp = intotcpcb(inp); if (inp->inp_vflag & INP_TIMEWAIT) { /* * There are two cases to handle: one in which the time wait * state is being discarded (INP_DROPPED), and one in which * this connection will remain in timewait. In the former, * it is time to discard all state (except tcptw, which has * already been discarded by the timewait close code, which * should be further up the call stack somewhere). In the * latter case, we detach from the socket, but leave the pcb * present until timewait ends. * * XXXRW: Would it be cleaner to free the tcptw here? */ if (inp->inp_vflag & INP_DROPPED) { KASSERT(tp == NULL, ("tcp_detach: INP_TIMEWAIT && " "INP_DROPPED && tp != NULL")); in_pcbdetach(inp); in_pcbfree(inp); } else { in_pcbdetach(inp); INP_WUNLOCK(inp); } } else { /* * If the connection is not in timewait, we consider two * two conditions: one in which no further processing is * necessary (dropped || embryonic), and one in which TCP is * not yet done, but no longer requires the socket, so the * pcb will persist for the time being. * * XXXRW: Does the second case still occur? */ if (inp->inp_vflag & INP_DROPPED || tp->t_state < TCPS_SYN_SENT) { tcp_discardcb(tp); in_pcbdetach(inp); in_pcbfree(inp); } else in_pcbdetach(inp); } } /* * pru_detach() detaches the TCP protocol from the socket. * If the protocol state is non-embryonic, then can't * do this directly: have to initiate a pru_disconnect(), * which may finish later; embryonic TCB's can just * be discarded here. */ static void tcp_usr_detach(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_detach: inp == NULL")); INP_INFO_WLOCK(&V_tcbinfo); INP_WLOCK(inp); KASSERT(inp->inp_socket != NULL, ("tcp_usr_detach: inp_socket == NULL")); tcp_detach(so, inp); INP_INFO_WUNLOCK(&V_tcbinfo); } /* * Give the socket an address. */ static int tcp_usr_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; struct sockaddr_in *sinp; sinp = (struct sockaddr_in *)nam; if (nam->sa_len != sizeof (*sinp)) return (EINVAL); /* * Must check for multicast addresses and disallow binding * to them. */ if (sinp->sin_family == AF_INET && IN_MULTICAST(ntohl(sinp->sin_addr.s_addr))) return (EAFNOSUPPORT); TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_bind: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = EINVAL; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); error = in_pcbbind(inp, nam, td->td_ucred); out: TCPDEBUG2(PRU_BIND); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } #ifdef INET6 static int tcp6_usr_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; struct sockaddr_in6 *sin6p; sin6p = (struct sockaddr_in6 *)nam; if (nam->sa_len != sizeof (*sin6p)) return (EINVAL); /* * Must check for multicast addresses and disallow binding * to them. */ if (sin6p->sin6_family == AF_INET6 && IN6_IS_ADDR_MULTICAST(&sin6p->sin6_addr)) return (EAFNOSUPPORT); TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp6_usr_bind: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = EINVAL; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); inp->inp_vflag &= ~INP_IPV4; inp->inp_vflag |= INP_IPV6; if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) { if (IN6_IS_ADDR_UNSPECIFIED(&sin6p->sin6_addr)) inp->inp_vflag |= INP_IPV4; else if (IN6_IS_ADDR_V4MAPPED(&sin6p->sin6_addr)) { struct sockaddr_in sin; in6_sin6_2_sin(&sin, sin6p); inp->inp_vflag |= INP_IPV4; inp->inp_vflag &= ~INP_IPV6; error = in_pcbbind(inp, (struct sockaddr *)&sin, td->td_ucred); goto out; } } error = in6_pcbbind(inp, nam, td->td_ucred); out: TCPDEBUG2(PRU_BIND); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } #endif /* INET6 */ /* * Prepare to accept connections. */ static int tcp_usr_listen(struct socket *so, int backlog, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_listen: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = EINVAL; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); SOCK_LOCK(so); error = solisten_proto_check(so); if (error == 0 && inp->inp_lport == 0) error = in_pcbbind(inp, (struct sockaddr *)0, td->td_ucred); if (error == 0) { tp->t_state = TCPS_LISTEN; solisten_proto(so, backlog); tcp_offload_listen_open(tp); } SOCK_UNLOCK(so); out: TCPDEBUG2(PRU_LISTEN); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } #ifdef INET6 static int tcp6_usr_listen(struct socket *so, int backlog, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp6_usr_listen: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = EINVAL; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); SOCK_LOCK(so); error = solisten_proto_check(so); if (error == 0 && inp->inp_lport == 0) { inp->inp_vflag &= ~INP_IPV4; if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) inp->inp_vflag |= INP_IPV4; error = in6_pcbbind(inp, (struct sockaddr *)0, td->td_ucred); } if (error == 0) { tp->t_state = TCPS_LISTEN; solisten_proto(so, backlog); } SOCK_UNLOCK(so); out: TCPDEBUG2(PRU_LISTEN); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } #endif /* INET6 */ /* * Initiate connection to peer. * Create a template for use in transmissions on this connection. * Enter SYN_SENT state, and mark socket as connecting. * Start keep-alive timer, and seed output sequence space. * Send initial segment on connection. */ static int tcp_usr_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; struct sockaddr_in *sinp; sinp = (struct sockaddr_in *)nam; if (nam->sa_len != sizeof (*sinp)) return (EINVAL); /* * Must disallow TCP ``connections'' to multicast addresses. */ if (sinp->sin_family == AF_INET && IN_MULTICAST(ntohl(sinp->sin_addr.s_addr))) return (EAFNOSUPPORT); - if (prison_remote_ip4(td->td_ucred, &sinp->sin_addr) != 0) - return (EINVAL); + if ((error = prison_remote_ip4(td->td_ucred, &sinp->sin_addr)) != 0) + return (error); TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_connect: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = EINVAL; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); if ((error = tcp_connect(tp, nam, td)) != 0) goto out; error = tcp_output_connect(so, nam); out: TCPDEBUG2(PRU_CONNECT); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } #ifdef INET6 static int tcp6_usr_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; struct sockaddr_in6 *sin6p; TCPDEBUG0; sin6p = (struct sockaddr_in6 *)nam; if (nam->sa_len != sizeof (*sin6p)) return (EINVAL); /* * Must disallow TCP ``connections'' to multicast addresses. */ if (sin6p->sin6_family == AF_INET6 && IN6_IS_ADDR_MULTICAST(&sin6p->sin6_addr)) return (EAFNOSUPPORT); INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp6_usr_connect: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = EINVAL; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); if (IN6_IS_ADDR_V4MAPPED(&sin6p->sin6_addr)) { struct sockaddr_in sin; if ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0) { error = EINVAL; goto out; } in6_sin6_2_sin(&sin, sin6p); inp->inp_vflag |= INP_IPV4; inp->inp_vflag &= ~INP_IPV6; - if (prison_remote_ip4(td->td_ucred, &sin.sin_addr) != 0) { - error = EINVAL; + if ((error = prison_remote_ip4(td->td_ucred, + &sin.sin_addr)) != 0) goto out; - } if ((error = tcp_connect(tp, (struct sockaddr *)&sin, td)) != 0) goto out; error = tcp_output_connect(so, nam); goto out; } inp->inp_vflag &= ~INP_IPV4; inp->inp_vflag |= INP_IPV6; inp->inp_inc.inc_flags |= INC_ISIPV6; - if (prison_remote_ip6(td->td_ucred, &sin6p->sin6_addr) != 0) { - error = EINVAL; + if ((error = prison_remote_ip6(td->td_ucred, &sin6p->sin6_addr)) != 0) goto out; - } if ((error = tcp6_connect(tp, nam, td)) != 0) goto out; error = tcp_output_connect(so, nam); out: TCPDEBUG2(PRU_CONNECT); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } #endif /* INET6 */ /* * Initiate disconnect from peer. * If connection never passed embryonic stage, just drop; * else if don't need to let data drain, then can just drop anyways, * else have to begin TCP shutdown process: mark socket disconnecting, * drain unread data, state switch to reflect user close, and * send segment (e.g. FIN) to peer. Socket will be really disconnected * when peer sends FIN and acks ours. * * SHOULD IMPLEMENT LATER PRU_CONNECT VIA REALLOC TCPCB. */ static int tcp_usr_disconnect(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; struct tcpcb *tp = NULL; int error = 0; TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_disconnect: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNRESET; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); tcp_disconnect(tp); out: TCPDEBUG2(PRU_DISCONNECT); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } /* * Accept a connection. Essentially all the work is * done at higher levels; just return the address * of the peer, storing through addr. */ static int tcp_usr_accept(struct socket *so, struct sockaddr **nam) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp = NULL; struct tcpcb *tp = NULL; struct in_addr addr; in_port_t port = 0; TCPDEBUG0; if (so->so_state & SS_ISDISCONNECTED) return (ECONNABORTED); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_accept: inp == NULL")); INP_INFO_RLOCK(&V_tcbinfo); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNABORTED; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); /* * We inline in_getpeeraddr and COMMON_END here, so that we can * copy the data of interest and defer the malloc until after we * release the lock. */ port = inp->inp_fport; addr = inp->inp_faddr; out: TCPDEBUG2(PRU_ACCEPT); INP_WUNLOCK(inp); INP_INFO_RUNLOCK(&V_tcbinfo); if (error == 0) *nam = in_sockaddr(port, &addr); return error; } #ifdef INET6 static int tcp6_usr_accept(struct socket *so, struct sockaddr **nam) { struct inpcb *inp = NULL; int error = 0; struct tcpcb *tp = NULL; struct in_addr addr; struct in6_addr addr6; in_port_t port = 0; int v4 = 0; TCPDEBUG0; if (so->so_state & SS_ISDISCONNECTED) return (ECONNABORTED); inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp6_usr_accept: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNABORTED; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); /* * We inline in6_mapped_peeraddr and COMMON_END here, so that we can * copy the data of interest and defer the malloc until after we * release the lock. */ if (inp->inp_vflag & INP_IPV4) { v4 = 1; port = inp->inp_fport; addr = inp->inp_faddr; } else { port = inp->inp_fport; addr6 = inp->in6p_faddr; } out: TCPDEBUG2(PRU_ACCEPT); INP_WUNLOCK(inp); if (error == 0) { if (v4) *nam = in6_v4mapsin6_sockaddr(port, &addr); else *nam = in6_sockaddr(port, &addr6); } return error; } #endif /* INET6 */ /* * Mark the connection as being incapable of further output. */ static int tcp_usr_shutdown(struct socket *so) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; TCPDEBUG0; INP_INFO_WLOCK(&V_tcbinfo); inp = sotoinpcb(so); KASSERT(inp != NULL, ("inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNRESET; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); socantsendmore(so); tcp_usrclosed(tp); error = tcp_output_disconnect(tp); out: TCPDEBUG2(PRU_SHUTDOWN); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } /* * After a receive, possibly send window update to peer. */ static int tcp_usr_rcvd(struct socket *so, int flags) { struct inpcb *inp; struct tcpcb *tp = NULL; int error = 0; TCPDEBUG0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_rcvd: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNRESET; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); tcp_output_rcvd(tp); out: TCPDEBUG2(PRU_RCVD); INP_WUNLOCK(inp); return (error); } /* * Do a send by putting data in output queue and updating urgent * marker if URG set. Possibly send more data. Unlike the other * pru_*() routines, the mbuf chains are our responsibility. We * must either enqueue them or free them. The other pru_* routines * generally are caller-frees. */ static int tcp_usr_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, struct mbuf *control, struct thread *td) { INIT_VNET_INET(so->so_vnet); int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; int headlocked = 0; #ifdef INET6 int isipv6; #endif TCPDEBUG0; /* * We require the pcbinfo lock in two cases: * * (1) An implied connect is taking place, which can result in * binding IPs and ports and hence modification of the pcb hash * chains. * * (2) PRUS_EOF is set, resulting in explicit close on the send. */ if ((nam != NULL) || (flags & PRUS_EOF)) { INP_INFO_WLOCK(&V_tcbinfo); headlocked = 1; } inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_send: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { if (control) m_freem(control); if (m) m_freem(m); error = ECONNRESET; goto out; } #ifdef INET6 isipv6 = nam && nam->sa_family == AF_INET6; #endif /* INET6 */ tp = intotcpcb(inp); TCPDEBUG1(); if (control) { /* TCP doesn't do control messages (rights, creds, etc) */ if (control->m_len) { m_freem(control); if (m) m_freem(m); error = EINVAL; goto out; } m_freem(control); /* empty control, just free it */ } if (!(flags & PRUS_OOB)) { sbappendstream(&so->so_snd, m); if (nam && tp->t_state < TCPS_SYN_SENT) { /* * Do implied connect if not yet connected, * initialize window to default value, and * initialize maxseg/maxopd using peer's cached * MSS. */ INP_INFO_WLOCK_ASSERT(&V_tcbinfo); #ifdef INET6 if (isipv6) error = tcp6_connect(tp, nam, td); else #endif /* INET6 */ error = tcp_connect(tp, nam, td); if (error) goto out; tp->snd_wnd = TTCP_CLIENT_SND_WND; tcp_mss(tp, -1); } if (flags & PRUS_EOF) { /* * Close the send side of the connection after * the data is sent. */ INP_INFO_WLOCK_ASSERT(&V_tcbinfo); socantsendmore(so); tcp_usrclosed(tp); } if (headlocked) { INP_INFO_WUNLOCK(&V_tcbinfo); headlocked = 0; } if (tp != NULL) { if (flags & PRUS_MORETOCOME) tp->t_flags |= TF_MORETOCOME; error = tcp_output_send(tp); if (flags & PRUS_MORETOCOME) tp->t_flags &= ~TF_MORETOCOME; } } else { /* * XXXRW: PRUS_EOF not implemented with PRUS_OOB? */ SOCKBUF_LOCK(&so->so_snd); if (sbspace(&so->so_snd) < -512) { SOCKBUF_UNLOCK(&so->so_snd); m_freem(m); error = ENOBUFS; goto out; } /* * According to RFC961 (Assigned Protocols), * the urgent pointer points to the last octet * of urgent data. We continue, however, * to consider it to indicate the first octet * of data past the urgent section. * Otherwise, snd_up should be one lower. */ sbappendstream_locked(&so->so_snd, m); SOCKBUF_UNLOCK(&so->so_snd); if (nam && tp->t_state < TCPS_SYN_SENT) { /* * Do implied connect if not yet connected, * initialize window to default value, and * initialize maxseg/maxopd using peer's cached * MSS. */ INP_INFO_WLOCK_ASSERT(&V_tcbinfo); #ifdef INET6 if (isipv6) error = tcp6_connect(tp, nam, td); else #endif /* INET6 */ error = tcp_connect(tp, nam, td); if (error) goto out; tp->snd_wnd = TTCP_CLIENT_SND_WND; tcp_mss(tp, -1); INP_INFO_WUNLOCK(&V_tcbinfo); headlocked = 0; } else if (nam) { INP_INFO_WUNLOCK(&V_tcbinfo); headlocked = 0; } tp->snd_up = tp->snd_una + so->so_snd.sb_cc; tp->t_flags |= TF_FORCEDATA; error = tcp_output_send(tp); tp->t_flags &= ~TF_FORCEDATA; } out: TCPDEBUG2((flags & PRUS_OOB) ? PRU_SENDOOB : ((flags & PRUS_EOF) ? PRU_SEND_EOF : PRU_SEND)); INP_WUNLOCK(inp); if (headlocked) INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } /* * Abort the TCP. Drop the connection abruptly. */ static void tcp_usr_abort(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; struct tcpcb *tp = NULL; TCPDEBUG0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_abort: inp == NULL")); INP_INFO_WLOCK(&V_tcbinfo); INP_WLOCK(inp); KASSERT(inp->inp_socket != NULL, ("tcp_usr_abort: inp_socket == NULL")); /* * If we still have full TCP state, and we're not dropped, drop. */ if (!(inp->inp_vflag & INP_TIMEWAIT) && !(inp->inp_vflag & INP_DROPPED)) { tp = intotcpcb(inp); TCPDEBUG1(); tcp_drop(tp, ECONNABORTED); TCPDEBUG2(PRU_ABORT); } if (!(inp->inp_vflag & INP_DROPPED)) { SOCK_LOCK(so); so->so_state |= SS_PROTOREF; SOCK_UNLOCK(so); inp->inp_vflag |= INP_SOCKREF; } INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); } /* * TCP socket is closed. Start friendly disconnect. */ static void tcp_usr_close(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; struct tcpcb *tp = NULL; TCPDEBUG0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_close: inp == NULL")); INP_INFO_WLOCK(&V_tcbinfo); INP_WLOCK(inp); KASSERT(inp->inp_socket != NULL, ("tcp_usr_close: inp_socket == NULL")); /* * If we still have full TCP state, and we're not dropped, initiate * a disconnect. */ if (!(inp->inp_vflag & INP_TIMEWAIT) && !(inp->inp_vflag & INP_DROPPED)) { tp = intotcpcb(inp); TCPDEBUG1(); tcp_disconnect(tp); TCPDEBUG2(PRU_CLOSE); } if (!(inp->inp_vflag & INP_DROPPED)) { SOCK_LOCK(so); so->so_state |= SS_PROTOREF; SOCK_UNLOCK(so); inp->inp_vflag |= INP_SOCKREF; } INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); } /* * Receive out-of-band data. */ static int tcp_usr_rcvoob(struct socket *so, struct mbuf *m, int flags) { int error = 0; struct inpcb *inp; struct tcpcb *tp = NULL; TCPDEBUG0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_rcvoob: inp == NULL")); INP_WLOCK(inp); if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNRESET; goto out; } tp = intotcpcb(inp); TCPDEBUG1(); if ((so->so_oobmark == 0 && (so->so_rcv.sb_state & SBS_RCVATMARK) == 0) || so->so_options & SO_OOBINLINE || tp->t_oobflags & TCPOOB_HADDATA) { error = EINVAL; goto out; } if ((tp->t_oobflags & TCPOOB_HAVEDATA) == 0) { error = EWOULDBLOCK; goto out; } m->m_len = 1; *mtod(m, caddr_t) = tp->t_iobc; if ((flags & MSG_PEEK) == 0) tp->t_oobflags ^= (TCPOOB_HAVEDATA | TCPOOB_HADDATA); out: TCPDEBUG2(PRU_RCVOOB); INP_WUNLOCK(inp); return (error); } struct pr_usrreqs tcp_usrreqs = { .pru_abort = tcp_usr_abort, .pru_accept = tcp_usr_accept, .pru_attach = tcp_usr_attach, .pru_bind = tcp_usr_bind, .pru_connect = tcp_usr_connect, .pru_control = in_control, .pru_detach = tcp_usr_detach, .pru_disconnect = tcp_usr_disconnect, .pru_listen = tcp_usr_listen, .pru_peeraddr = in_getpeeraddr, .pru_rcvd = tcp_usr_rcvd, .pru_rcvoob = tcp_usr_rcvoob, .pru_send = tcp_usr_send, .pru_shutdown = tcp_usr_shutdown, .pru_sockaddr = in_getsockaddr, .pru_sosetlabel = in_pcbsosetlabel, .pru_close = tcp_usr_close, }; #ifdef INET6 struct pr_usrreqs tcp6_usrreqs = { .pru_abort = tcp_usr_abort, .pru_accept = tcp6_usr_accept, .pru_attach = tcp_usr_attach, .pru_bind = tcp6_usr_bind, .pru_connect = tcp6_usr_connect, .pru_control = in6_control, .pru_detach = tcp_usr_detach, .pru_disconnect = tcp_usr_disconnect, .pru_listen = tcp6_usr_listen, .pru_peeraddr = in6_mapped_peeraddr, .pru_rcvd = tcp_usr_rcvd, .pru_rcvoob = tcp_usr_rcvoob, .pru_send = tcp_usr_send, .pru_shutdown = tcp_usr_shutdown, .pru_sockaddr = in6_mapped_sockaddr, .pru_sosetlabel = in_pcbsosetlabel, .pru_close = tcp_usr_close, }; #endif /* INET6 */ /* * Common subroutine to open a TCP connection to remote host specified * by struct sockaddr_in in mbuf *nam. Call in_pcbbind to assign a local * port number if needed. Call in_pcbconnect_setup to do the routing and * to choose a local host address (interface). If there is an existing * incarnation of the same connection in TIME-WAIT state and if the remote * host was sending CC options and if the connection duration was < MSL, then * truncate the previous TIME-WAIT state and proceed. * Initialize connection parameters and enter SYN-SENT state. */ static int tcp_connect(struct tcpcb *tp, struct sockaddr *nam, struct thread *td) { struct inpcb *inp = tp->t_inpcb, *oinp; struct socket *so = inp->inp_socket; INIT_VNET_INET(so->so_vnet); struct in_addr laddr; u_short lport; int error; INP_INFO_WLOCK_ASSERT(&V_tcbinfo); INP_WLOCK_ASSERT(inp); if (inp->inp_lport == 0) { error = in_pcbbind(inp, (struct sockaddr *)0, td->td_ucred); if (error) return error; } /* * Cannot simply call in_pcbconnect, because there might be an * earlier incarnation of this same connection still in * TIME_WAIT state, creating an ADDRINUSE error. */ laddr = inp->inp_laddr; lport = inp->inp_lport; error = in_pcbconnect_setup(inp, nam, &laddr.s_addr, &lport, &inp->inp_faddr.s_addr, &inp->inp_fport, &oinp, td->td_ucred); if (error && oinp == NULL) return error; if (oinp) return EADDRINUSE; inp->inp_laddr = laddr; in_pcbrehash(inp); /* * Compute window scaling to request: * Scale to fit into sweet spot. See tcp_syncache.c. * XXX: This should move to tcp_output(). */ while (tp->request_r_scale < TCP_MAX_WINSHIFT && (TCP_MAXWIN << tp->request_r_scale) < sb_max) tp->request_r_scale++; soisconnecting(so); V_tcpstat.tcps_connattempt++; tp->t_state = TCPS_SYN_SENT; tcp_timer_activate(tp, TT_KEEP, tcp_keepinit); tp->iss = tcp_new_isn(tp); tp->t_bw_rtseq = tp->iss; tcp_sendseqinit(tp); return 0; } #ifdef INET6 static int tcp6_connect(struct tcpcb *tp, struct sockaddr *nam, struct thread *td) { struct inpcb *inp = tp->t_inpcb, *oinp; struct socket *so = inp->inp_socket; INIT_VNET_INET(so->so_vnet); struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam; struct in6_addr *addr6; int error; INP_INFO_WLOCK_ASSERT(&V_tcbinfo); INP_WLOCK_ASSERT(inp); if (inp->inp_lport == 0) { error = in6_pcbbind(inp, (struct sockaddr *)0, td->td_ucred); if (error) return error; } /* * Cannot simply call in_pcbconnect, because there might be an * earlier incarnation of this same connection still in * TIME_WAIT state, creating an ADDRINUSE error. * in6_pcbladdr() also handles scope zone IDs. */ error = in6_pcbladdr(inp, nam, &addr6); if (error) return error; oinp = in6_pcblookup_hash(inp->inp_pcbinfo, &sin6->sin6_addr, sin6->sin6_port, IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr) ? addr6 : &inp->in6p_laddr, inp->inp_lport, 0, NULL); if (oinp) return EADDRINUSE; if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) inp->in6p_laddr = *addr6; inp->in6p_faddr = sin6->sin6_addr; inp->inp_fport = sin6->sin6_port; /* update flowinfo - draft-itojun-ipv6-flowlabel-api-00 */ inp->inp_flow &= ~IPV6_FLOWLABEL_MASK; if (inp->inp_flags & IN6P_AUTOFLOWLABEL) inp->inp_flow |= (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); in_pcbrehash(inp); /* Compute window scaling to request. */ while (tp->request_r_scale < TCP_MAX_WINSHIFT && (TCP_MAXWIN << tp->request_r_scale) < so->so_rcv.sb_hiwat) tp->request_r_scale++; soisconnecting(so); V_tcpstat.tcps_connattempt++; tp->t_state = TCPS_SYN_SENT; tcp_timer_activate(tp, TT_KEEP, tcp_keepinit); tp->iss = tcp_new_isn(tp); tp->t_bw_rtseq = tp->iss; tcp_sendseqinit(tp); return 0; } #endif /* INET6 */ /* * Export TCP internal state information via a struct tcp_info, based on the * Linux 2.6 API. Not ABI compatible as our constants are mapped differently * (TCP state machine, etc). We export all information using FreeBSD-native * constants -- for example, the numeric values for tcpi_state will differ * from Linux. */ static void tcp_fill_info(struct tcpcb *tp, struct tcp_info *ti) { INP_WLOCK_ASSERT(tp->t_inpcb); bzero(ti, sizeof(*ti)); ti->tcpi_state = tp->t_state; if ((tp->t_flags & TF_REQ_TSTMP) && (tp->t_flags & TF_RCVD_TSTMP)) ti->tcpi_options |= TCPI_OPT_TIMESTAMPS; if (tp->t_flags & TF_SACK_PERMIT) ti->tcpi_options |= TCPI_OPT_SACK; if ((tp->t_flags & TF_REQ_SCALE) && (tp->t_flags & TF_RCVD_SCALE)) { ti->tcpi_options |= TCPI_OPT_WSCALE; ti->tcpi_snd_wscale = tp->snd_scale; ti->tcpi_rcv_wscale = tp->rcv_scale; } ti->tcpi_rtt = ((u_int64_t)tp->t_srtt * tick) >> TCP_RTT_SHIFT; ti->tcpi_rttvar = ((u_int64_t)tp->t_rttvar * tick) >> TCP_RTTVAR_SHIFT; ti->tcpi_snd_ssthresh = tp->snd_ssthresh; ti->tcpi_snd_cwnd = tp->snd_cwnd; /* * FreeBSD-specific extension fields for tcp_info. */ ti->tcpi_rcv_space = tp->rcv_wnd; ti->tcpi_rcv_nxt = tp->rcv_nxt; ti->tcpi_snd_wnd = tp->snd_wnd; ti->tcpi_snd_bwnd = tp->snd_bwnd; ti->tcpi_snd_nxt = tp->snd_nxt; ti->__tcpi_snd_mss = tp->t_maxseg; ti->__tcpi_rcv_mss = tp->t_maxseg; if (tp->t_flags & TF_TOE) ti->tcpi_options |= TCPI_OPT_TOE; } /* * tcp_ctloutput() must drop the inpcb lock before performing copyin on * socket option arguments. When it re-acquires the lock after the copy, it * has to revalidate that the connection is still valid for the socket * option. */ #define INP_WLOCK_RECHECK(inp) do { \ INP_WLOCK(inp); \ if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { \ INP_WUNLOCK(inp); \ return (ECONNRESET); \ } \ tp = intotcpcb(inp); \ } while(0) int tcp_ctloutput(struct socket *so, struct sockopt *sopt) { INIT_VNET_INET(so->so_vnet); int error, opt, optval; struct inpcb *inp; struct tcpcb *tp; struct tcp_info ti; error = 0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_ctloutput: inp == NULL")); INP_WLOCK(inp); if (sopt->sopt_level != IPPROTO_TCP) { #ifdef INET6 if (inp->inp_vflag & INP_IPV6PROTO) { INP_WUNLOCK(inp); error = ip6_ctloutput(so, sopt); } else { #endif /* INET6 */ INP_WUNLOCK(inp); error = ip_ctloutput(so, sopt); #ifdef INET6 } #endif return (error); } if (inp->inp_vflag & (INP_TIMEWAIT | INP_DROPPED)) { INP_WUNLOCK(inp); return (ECONNRESET); } switch (sopt->sopt_dir) { case SOPT_SET: switch (sopt->sopt_name) { #ifdef TCP_SIGNATURE case TCP_MD5SIG: INP_WUNLOCK(inp); error = sooptcopyin(sopt, &optval, sizeof optval, sizeof optval); if (error) return (error); INP_WLOCK_RECHECK(inp); if (optval > 0) tp->t_flags |= TF_SIGNATURE; else tp->t_flags &= ~TF_SIGNATURE; INP_WUNLOCK(inp); break; #endif /* TCP_SIGNATURE */ case TCP_NODELAY: case TCP_NOOPT: INP_WUNLOCK(inp); error = sooptcopyin(sopt, &optval, sizeof optval, sizeof optval); if (error) return (error); INP_WLOCK_RECHECK(inp); switch (sopt->sopt_name) { case TCP_NODELAY: opt = TF_NODELAY; break; case TCP_NOOPT: opt = TF_NOOPT; break; default: opt = 0; /* dead code to fool gcc */ break; } if (optval) tp->t_flags |= opt; else tp->t_flags &= ~opt; INP_WUNLOCK(inp); break; case TCP_NOPUSH: INP_WUNLOCK(inp); error = sooptcopyin(sopt, &optval, sizeof optval, sizeof optval); if (error) return (error); INP_WLOCK_RECHECK(inp); if (optval) tp->t_flags |= TF_NOPUSH; else { tp->t_flags &= ~TF_NOPUSH; error = tcp_output(tp); } INP_WUNLOCK(inp); break; case TCP_MAXSEG: INP_WUNLOCK(inp); error = sooptcopyin(sopt, &optval, sizeof optval, sizeof optval); if (error) return (error); INP_WLOCK_RECHECK(inp); if (optval > 0 && optval <= tp->t_maxseg && optval + 40 >= V_tcp_minmss) tp->t_maxseg = optval; else error = EINVAL; INP_WUNLOCK(inp); break; case TCP_INFO: INP_WUNLOCK(inp); error = EINVAL; break; default: INP_WUNLOCK(inp); error = ENOPROTOOPT; break; } break; case SOPT_GET: tp = intotcpcb(inp); switch (sopt->sopt_name) { #ifdef TCP_SIGNATURE case TCP_MD5SIG: optval = (tp->t_flags & TF_SIGNATURE) ? 1 : 0; INP_WUNLOCK(inp); error = sooptcopyout(sopt, &optval, sizeof optval); break; #endif case TCP_NODELAY: optval = tp->t_flags & TF_NODELAY; INP_WUNLOCK(inp); error = sooptcopyout(sopt, &optval, sizeof optval); break; case TCP_MAXSEG: optval = tp->t_maxseg; INP_WUNLOCK(inp); error = sooptcopyout(sopt, &optval, sizeof optval); break; case TCP_NOOPT: optval = tp->t_flags & TF_NOOPT; INP_WUNLOCK(inp); error = sooptcopyout(sopt, &optval, sizeof optval); break; case TCP_NOPUSH: optval = tp->t_flags & TF_NOPUSH; INP_WUNLOCK(inp); error = sooptcopyout(sopt, &optval, sizeof optval); break; case TCP_INFO: tcp_fill_info(tp, &ti); INP_WUNLOCK(inp); error = sooptcopyout(sopt, &ti, sizeof ti); break; default: INP_WUNLOCK(inp); error = ENOPROTOOPT; break; } break; } return (error); } #undef INP_WLOCK_RECHECK /* * tcp_sendspace and tcp_recvspace are the default send and receive window * sizes, respectively. These are obsolescent (this information should * be set by the route). */ u_long tcp_sendspace = 1024*32; SYSCTL_ULONG(_net_inet_tcp, TCPCTL_SENDSPACE, sendspace, CTLFLAG_RW, &tcp_sendspace , 0, "Maximum outgoing TCP datagram size"); u_long tcp_recvspace = 1024*64; SYSCTL_ULONG(_net_inet_tcp, TCPCTL_RECVSPACE, recvspace, CTLFLAG_RW, &tcp_recvspace , 0, "Maximum incoming TCP datagram size"); /* * Attach TCP protocol to socket, allocating * internet protocol control block, tcp control block, * bufer space, and entering LISTEN state if to accept connections. */ static int tcp_attach(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct tcpcb *tp; struct inpcb *inp; int error; if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { error = soreserve(so, tcp_sendspace, tcp_recvspace); if (error) return (error); } so->so_rcv.sb_flags |= SB_AUTOSIZE; so->so_snd.sb_flags |= SB_AUTOSIZE; INP_INFO_WLOCK(&V_tcbinfo); error = in_pcballoc(so, &V_tcbinfo); if (error) { INP_INFO_WUNLOCK(&V_tcbinfo); return (error); } inp = sotoinpcb(so); #ifdef INET6 if (inp->inp_vflag & INP_IPV6PROTO) { inp->inp_vflag |= INP_IPV6; inp->in6p_hops = -1; /* use kernel default */ } else #endif inp->inp_vflag |= INP_IPV4; tp = tcp_newtcpcb(inp); if (tp == NULL) { in_pcbdetach(inp); in_pcbfree(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (ENOBUFS); } tp->t_state = TCPS_CLOSED; INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_tcbinfo); return (0); } /* * Initiate (or continue) disconnect. * If embryonic state, just send reset (once). * If in ``let data drain'' option and linger null, just drop. * Otherwise (hard), mark socket disconnecting and drop * current input data; switch states based on user close, and * send segment to peer (with FIN). */ static void tcp_disconnect(struct tcpcb *tp) { struct inpcb *inp = tp->t_inpcb; struct socket *so = inp->inp_socket; #ifdef INVARIANTS INIT_VNET_INET(so->so_vnet); #endif INP_INFO_WLOCK_ASSERT(&V_tcbinfo); INP_WLOCK_ASSERT(inp); /* * Neither tcp_close() nor tcp_drop() should return NULL, as the * socket is still open. */ if (tp->t_state < TCPS_ESTABLISHED) { tp = tcp_close(tp); KASSERT(tp != NULL, ("tcp_disconnect: tcp_close() returned NULL")); } else if ((so->so_options & SO_LINGER) && so->so_linger == 0) { tp = tcp_drop(tp, 0); KASSERT(tp != NULL, ("tcp_disconnect: tcp_drop() returned NULL")); } else { soisdisconnecting(so); sbflush(&so->so_rcv); tcp_usrclosed(tp); if (!(inp->inp_vflag & INP_DROPPED)) tcp_output_disconnect(tp); } } /* * User issued close, and wish to trail through shutdown states: * if never received SYN, just forget it. If got a SYN from peer, * but haven't sent FIN, then go to FIN_WAIT_1 state to send peer a FIN. * If already got a FIN from peer, then almost done; go to LAST_ACK * state. In all other cases, have already sent FIN to peer (e.g. * after PRU_SHUTDOWN), and just have to play tedious game waiting * for peer to send FIN or not respond to keep-alives, etc. * We can let the user exit from the close as soon as the FIN is acked. */ static void tcp_usrclosed(struct tcpcb *tp) { #ifdef INVARIANTS INIT_VNET_INET(tp->t_inpcb->inp_vnet); #endif INP_INFO_WLOCK_ASSERT(&V_tcbinfo); INP_WLOCK_ASSERT(tp->t_inpcb); switch (tp->t_state) { case TCPS_LISTEN: tcp_offload_listen_close(tp); /* FALLTHROUGH */ case TCPS_CLOSED: tp->t_state = TCPS_CLOSED; tp = tcp_close(tp); /* * tcp_close() should never return NULL here as the socket is * still open. */ KASSERT(tp != NULL, ("tcp_usrclosed: tcp_close() returned NULL")); break; case TCPS_SYN_SENT: case TCPS_SYN_RECEIVED: tp->t_flags |= TF_NEEDFIN; break; case TCPS_ESTABLISHED: tp->t_state = TCPS_FIN_WAIT_1; break; case TCPS_CLOSE_WAIT: tp->t_state = TCPS_LAST_ACK; break; } if (tp->t_state >= TCPS_FIN_WAIT_2) { soisdisconnected(tp->t_inpcb->inp_socket); /* Prevent the connection hanging in FIN_WAIT_2 forever. */ if (tp->t_state == TCPS_FIN_WAIT_2) { int timeout; timeout = (tcp_fast_finwait2_recycle) ? tcp_finwait2_timeout : tcp_maxidle; tcp_timer_activate(tp, TT_2MSL, timeout); } } } #ifdef DDB static void db_print_indent(int indent) { int i; for (i = 0; i < indent; i++) db_printf(" "); } static void db_print_tstate(int t_state) { switch (t_state) { case TCPS_CLOSED: db_printf("TCPS_CLOSED"); return; case TCPS_LISTEN: db_printf("TCPS_LISTEN"); return; case TCPS_SYN_SENT: db_printf("TCPS_SYN_SENT"); return; case TCPS_SYN_RECEIVED: db_printf("TCPS_SYN_RECEIVED"); return; case TCPS_ESTABLISHED: db_printf("TCPS_ESTABLISHED"); return; case TCPS_CLOSE_WAIT: db_printf("TCPS_CLOSE_WAIT"); return; case TCPS_FIN_WAIT_1: db_printf("TCPS_FIN_WAIT_1"); return; case TCPS_CLOSING: db_printf("TCPS_CLOSING"); return; case TCPS_LAST_ACK: db_printf("TCPS_LAST_ACK"); return; case TCPS_FIN_WAIT_2: db_printf("TCPS_FIN_WAIT_2"); return; case TCPS_TIME_WAIT: db_printf("TCPS_TIME_WAIT"); return; default: db_printf("unknown"); return; } } static void db_print_tflags(u_int t_flags) { int comma; comma = 0; if (t_flags & TF_ACKNOW) { db_printf("%sTF_ACKNOW", comma ? ", " : ""); comma = 1; } if (t_flags & TF_DELACK) { db_printf("%sTF_DELACK", comma ? ", " : ""); comma = 1; } if (t_flags & TF_NODELAY) { db_printf("%sTF_NODELAY", comma ? ", " : ""); comma = 1; } if (t_flags & TF_NOOPT) { db_printf("%sTF_NOOPT", comma ? ", " : ""); comma = 1; } if (t_flags & TF_SENTFIN) { db_printf("%sTF_SENTFIN", comma ? ", " : ""); comma = 1; } if (t_flags & TF_REQ_SCALE) { db_printf("%sTF_REQ_SCALE", comma ? ", " : ""); comma = 1; } if (t_flags & TF_RCVD_SCALE) { db_printf("%sTF_RECVD_SCALE", comma ? ", " : ""); comma = 1; } if (t_flags & TF_REQ_TSTMP) { db_printf("%sTF_REQ_TSTMP", comma ? ", " : ""); comma = 1; } if (t_flags & TF_RCVD_TSTMP) { db_printf("%sTF_RCVD_TSTMP", comma ? ", " : ""); comma = 1; } if (t_flags & TF_SACK_PERMIT) { db_printf("%sTF_SACK_PERMIT", comma ? ", " : ""); comma = 1; } if (t_flags & TF_NEEDSYN) { db_printf("%sTF_NEEDSYN", comma ? ", " : ""); comma = 1; } if (t_flags & TF_NEEDFIN) { db_printf("%sTF_NEEDFIN", comma ? ", " : ""); comma = 1; } if (t_flags & TF_NOPUSH) { db_printf("%sTF_NOPUSH", comma ? ", " : ""); comma = 1; } if (t_flags & TF_NOPUSH) { db_printf("%sTF_NOPUSH", comma ? ", " : ""); comma = 1; } if (t_flags & TF_MORETOCOME) { db_printf("%sTF_MORETOCOME", comma ? ", " : ""); comma = 1; } if (t_flags & TF_LQ_OVERFLOW) { db_printf("%sTF_LQ_OVERFLOW", comma ? ", " : ""); comma = 1; } if (t_flags & TF_LASTIDLE) { db_printf("%sTF_LASTIDLE", comma ? ", " : ""); comma = 1; } if (t_flags & TF_RXWIN0SENT) { db_printf("%sTF_RXWIN0SENT", comma ? ", " : ""); comma = 1; } if (t_flags & TF_FASTRECOVERY) { db_printf("%sTF_FASTRECOVERY", comma ? ", " : ""); comma = 1; } if (t_flags & TF_WASFRECOVERY) { db_printf("%sTF_WASFRECOVERY", comma ? ", " : ""); comma = 1; } if (t_flags & TF_SIGNATURE) { db_printf("%sTF_SIGNATURE", comma ? ", " : ""); comma = 1; } if (t_flags & TF_FORCEDATA) { db_printf("%sTF_FORCEDATA", comma ? ", " : ""); comma = 1; } if (t_flags & TF_TSO) { db_printf("%sTF_TSO", comma ? ", " : ""); comma = 1; } if (t_flags & TF_ECN_PERMIT) { db_printf("%sTF_ECN_PERMIT", comma ? ", " : ""); comma = 1; } } static void db_print_toobflags(char t_oobflags) { int comma; comma = 0; if (t_oobflags & TCPOOB_HAVEDATA) { db_printf("%sTCPOOB_HAVEDATA", comma ? ", " : ""); comma = 1; } if (t_oobflags & TCPOOB_HADDATA) { db_printf("%sTCPOOB_HADDATA", comma ? ", " : ""); comma = 1; } } static void db_print_tcpcb(struct tcpcb *tp, const char *name, int indent) { db_print_indent(indent); db_printf("%s at %p\n", name, tp); indent += 2; db_print_indent(indent); db_printf("t_segq first: %p t_segqlen: %d t_dupacks: %d\n", LIST_FIRST(&tp->t_segq), tp->t_segqlen, tp->t_dupacks); db_print_indent(indent); db_printf("tt_rexmt: %p tt_persist: %p tt_keep: %p\n", &tp->t_timers->tt_rexmt, &tp->t_timers->tt_persist, &tp->t_timers->tt_keep); db_print_indent(indent); db_printf("tt_2msl: %p tt_delack: %p t_inpcb: %p\n", &tp->t_timers->tt_2msl, &tp->t_timers->tt_delack, tp->t_inpcb); db_print_indent(indent); db_printf("t_state: %d (", tp->t_state); db_print_tstate(tp->t_state); db_printf(")\n"); db_print_indent(indent); db_printf("t_flags: 0x%x (", tp->t_flags); db_print_tflags(tp->t_flags); db_printf(")\n"); db_print_indent(indent); db_printf("snd_una: 0x%08x snd_max: 0x%08x snd_nxt: x0%08x\n", tp->snd_una, tp->snd_max, tp->snd_nxt); db_print_indent(indent); db_printf("snd_up: 0x%08x snd_wl1: 0x%08x snd_wl2: 0x%08x\n", tp->snd_up, tp->snd_wl1, tp->snd_wl2); db_print_indent(indent); db_printf("iss: 0x%08x irs: 0x%08x rcv_nxt: 0x%08x\n", tp->iss, tp->irs, tp->rcv_nxt); db_print_indent(indent); db_printf("rcv_adv: 0x%08x rcv_wnd: %lu rcv_up: 0x%08x\n", tp->rcv_adv, tp->rcv_wnd, tp->rcv_up); db_print_indent(indent); db_printf("snd_wnd: %lu snd_cwnd: %lu snd_bwnd: %lu\n", tp->snd_wnd, tp->snd_cwnd, tp->snd_bwnd); db_print_indent(indent); db_printf("snd_ssthresh: %lu snd_bandwidth: %lu snd_recover: " "0x%08x\n", tp->snd_ssthresh, tp->snd_bandwidth, tp->snd_recover); db_print_indent(indent); db_printf("t_maxopd: %u t_rcvtime: %lu t_startime: %lu\n", tp->t_maxopd, tp->t_rcvtime, tp->t_starttime); db_print_indent(indent); db_printf("t_rttime: %d t_rtsq: 0x%08x t_bw_rtttime: %d\n", tp->t_rtttime, tp->t_rtseq, tp->t_bw_rtttime); db_print_indent(indent); db_printf("t_bw_rtseq: 0x%08x t_rxtcur: %d t_maxseg: %u " "t_srtt: %d\n", tp->t_bw_rtseq, tp->t_rxtcur, tp->t_maxseg, tp->t_srtt); db_print_indent(indent); db_printf("t_rttvar: %d t_rxtshift: %d t_rttmin: %u " "t_rttbest: %u\n", tp->t_rttvar, tp->t_rxtshift, tp->t_rttmin, tp->t_rttbest); db_print_indent(indent); db_printf("t_rttupdated: %lu max_sndwnd: %lu t_softerror: %d\n", tp->t_rttupdated, tp->max_sndwnd, tp->t_softerror); db_print_indent(indent); db_printf("t_oobflags: 0x%x (", tp->t_oobflags); db_print_toobflags(tp->t_oobflags); db_printf(") t_iobc: 0x%02x\n", tp->t_iobc); db_print_indent(indent); db_printf("snd_scale: %u rcv_scale: %u request_r_scale: %u\n", tp->snd_scale, tp->rcv_scale, tp->request_r_scale); db_print_indent(indent); db_printf("ts_recent: %u ts_recent_age: %lu\n", tp->ts_recent, tp->ts_recent_age); db_print_indent(indent); db_printf("ts_offset: %u last_ack_sent: 0x%08x snd_cwnd_prev: " "%lu\n", tp->ts_offset, tp->last_ack_sent, tp->snd_cwnd_prev); db_print_indent(indent); db_printf("snd_ssthresh_prev: %lu snd_recover_prev: 0x%08x " "t_badrxtwin: %lu\n", tp->snd_ssthresh_prev, tp->snd_recover_prev, tp->t_badrxtwin); db_print_indent(indent); db_printf("snd_numholes: %d snd_holes first: %p\n", tp->snd_numholes, TAILQ_FIRST(&tp->snd_holes)); db_print_indent(indent); db_printf("snd_fack: 0x%08x rcv_numsacks: %d sack_newdata: " "0x%08x\n", tp->snd_fack, tp->rcv_numsacks, tp->sack_newdata); /* Skip sackblks, sackhint. */ db_print_indent(indent); db_printf("t_rttlow: %d rfbuf_ts: %u rfbuf_cnt: %d\n", tp->t_rttlow, tp->rfbuf_ts, tp->rfbuf_cnt); } DB_SHOW_COMMAND(tcpcb, db_show_tcpcb) { struct tcpcb *tp; if (!have_addr) { db_printf("usage: show tcpcb \n"); return; } tp = (struct tcpcb *)addr; db_print_tcpcb(tp, "tcpcb", 0); } #endif diff --git a/sys/netinet/udp_usrreq.c b/sys/netinet/udp_usrreq.c index 7eaa0e365467..804f5fe78d40 100644 --- a/sys/netinet/udp_usrreq.c +++ b/sys/netinet/udp_usrreq.c @@ -1,1370 +1,1370 @@ /*- * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 * The Regents of the University of California. * Copyright (c) 2008 Robert N. M. Watson * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)udp_usrreq.c 8.6 (Berkeley) 5/23/95 */ #include __FBSDID("$FreeBSD$"); #include "opt_ipfw.h" #include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_mac.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET6 #include #endif #include #include #include #include #ifdef INET6 #include #endif #include #include #include #ifdef IPSEC #include #endif #include #include /* * UDP protocol implementation. * Per RFC 768, August, 1980. */ #ifdef VIMAGE_GLOBALS int udp_blackhole; #endif /* * BSD 4.2 defaulted the udp checksum to be off. Turning off udp checksums * removes the only data integrity mechanism for packets and malformed * packets that would otherwise be discarded due to bad checksums, and may * cause problems (especially for NFS data blocks). */ static int udp_cksum = 1; SYSCTL_INT(_net_inet_udp, UDPCTL_CHECKSUM, checksum, CTLFLAG_RW, &udp_cksum, 0, "compute udp checksum"); int udp_log_in_vain = 0; SYSCTL_INT(_net_inet_udp, OID_AUTO, log_in_vain, CTLFLAG_RW, &udp_log_in_vain, 0, "Log all incoming UDP packets"); SYSCTL_V_INT(V_NET, vnet_inet, _net_inet_udp, OID_AUTO, blackhole, CTLFLAG_RW, udp_blackhole, 0, "Do not send port unreachables for refused connects"); u_long udp_sendspace = 9216; /* really max datagram size */ /* 40 1K datagrams */ SYSCTL_ULONG(_net_inet_udp, UDPCTL_MAXDGRAM, maxdgram, CTLFLAG_RW, &udp_sendspace, 0, "Maximum outgoing UDP datagram size"); u_long udp_recvspace = 40 * (1024 + #ifdef INET6 sizeof(struct sockaddr_in6) #else sizeof(struct sockaddr_in) #endif ); SYSCTL_ULONG(_net_inet_udp, UDPCTL_RECVSPACE, recvspace, CTLFLAG_RW, &udp_recvspace, 0, "Maximum space for incoming UDP datagrams"); #ifdef VIMAGE_GLOBALS struct inpcbhead udb; /* from udp_var.h */ struct inpcbinfo udbinfo; struct udpstat udpstat; /* from udp_var.h */ #endif #ifndef UDBHASHSIZE #define UDBHASHSIZE 128 #endif SYSCTL_V_STRUCT(V_NET, vnet_inet, _net_inet_udp, UDPCTL_STATS, stats, CTLFLAG_RW, udpstat, udpstat, "UDP statistics (struct udpstat, netinet/udp_var.h)"); static void udp_detach(struct socket *so); static int udp_output(struct inpcb *, struct mbuf *, struct sockaddr *, struct mbuf *, struct thread *); static void udp_zone_change(void *tag) { uma_zone_set_max(V_udbinfo.ipi_zone, maxsockets); } static int udp_inpcb_init(void *mem, int size, int flags) { struct inpcb *inp; inp = mem; INP_LOCK_INIT(inp, "inp", "udpinp"); return (0); } void udp_init(void) { INIT_VNET_INET(curvnet); V_udp_blackhole = 0; INP_INFO_LOCK_INIT(&V_udbinfo, "udp"); LIST_INIT(&V_udb); V_udbinfo.ipi_listhead = &V_udb; V_udbinfo.ipi_hashbase = hashinit(UDBHASHSIZE, M_PCB, &V_udbinfo.ipi_hashmask); V_udbinfo.ipi_porthashbase = hashinit(UDBHASHSIZE, M_PCB, &V_udbinfo.ipi_porthashmask); V_udbinfo.ipi_zone = uma_zcreate("udpcb", sizeof(struct inpcb), NULL, NULL, udp_inpcb_init, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE); uma_zone_set_max(V_udbinfo.ipi_zone, maxsockets); EVENTHANDLER_REGISTER(maxsockets_change, udp_zone_change, NULL, EVENTHANDLER_PRI_ANY); } /* * Subroutine of udp_input(), which appends the provided mbuf chain to the * passed pcb/socket. The caller must provide a sockaddr_in via udp_in that * contains the source address. If the socket ends up being an IPv6 socket, * udp_append() will convert to a sockaddr_in6 before passing the address * into the socket code. */ static void udp_append(struct inpcb *inp, struct ip *ip, struct mbuf *n, int off, struct sockaddr_in *udp_in) { struct sockaddr *append_sa; struct socket *so; struct mbuf *opts = 0; #ifdef INET6 struct sockaddr_in6 udp_in6; #endif INP_RLOCK_ASSERT(inp); #ifdef IPSEC /* Check AH/ESP integrity. */ if (ipsec4_in_reject(n, inp)) { INIT_VNET_IPSEC(curvnet); m_freem(n); V_ipsec4stat.in_polvio++; return; } #endif /* IPSEC */ #ifdef MAC if (mac_inpcb_check_deliver(inp, n) != 0) { m_freem(n); return; } #endif if (inp->inp_flags & INP_CONTROLOPTS || inp->inp_socket->so_options & (SO_TIMESTAMP | SO_BINTIME)) { #ifdef INET6 if (inp->inp_vflag & INP_IPV6) (void)ip6_savecontrol_v4(inp, n, &opts, NULL); else #endif ip_savecontrol(inp, &opts, ip, n); } #ifdef INET6 if (inp->inp_vflag & INP_IPV6) { bzero(&udp_in6, sizeof(udp_in6)); udp_in6.sin6_len = sizeof(udp_in6); udp_in6.sin6_family = AF_INET6; in6_sin_2_v4mapsin6(udp_in, &udp_in6); append_sa = (struct sockaddr *)&udp_in6; } else #endif append_sa = (struct sockaddr *)udp_in; m_adj(n, off); so = inp->inp_socket; SOCKBUF_LOCK(&so->so_rcv); if (sbappendaddr_locked(&so->so_rcv, append_sa, n, opts) == 0) { INIT_VNET_INET(so->so_vnet); SOCKBUF_UNLOCK(&so->so_rcv); m_freem(n); if (opts) m_freem(opts); V_udpstat.udps_fullsock++; } else sorwakeup_locked(so); } void udp_input(struct mbuf *m, int off) { INIT_VNET_INET(curvnet); int iphlen = off; struct ip *ip; struct udphdr *uh; struct ifnet *ifp; struct inpcb *inp; int len; struct ip save_ip; struct sockaddr_in udp_in; #ifdef IPFIREWALL_FORWARD struct m_tag *fwd_tag; #endif ifp = m->m_pkthdr.rcvif; V_udpstat.udps_ipackets++; /* * Strip IP options, if any; should skip this, make available to * user, and use on returned packets, but we don't yet have a way to * check the checksum with options still present. */ if (iphlen > sizeof (struct ip)) { ip_stripoptions(m, (struct mbuf *)0); iphlen = sizeof(struct ip); } /* * Get IP and UDP header together in first mbuf. */ ip = mtod(m, struct ip *); if (m->m_len < iphlen + sizeof(struct udphdr)) { if ((m = m_pullup(m, iphlen + sizeof(struct udphdr))) == 0) { V_udpstat.udps_hdrops++; return; } ip = mtod(m, struct ip *); } uh = (struct udphdr *)((caddr_t)ip + iphlen); /* * Destination port of 0 is illegal, based on RFC768. */ if (uh->uh_dport == 0) goto badunlocked; /* * Construct sockaddr format source address. Stuff source address * and datagram in user buffer. */ bzero(&udp_in, sizeof(udp_in)); udp_in.sin_len = sizeof(udp_in); udp_in.sin_family = AF_INET; udp_in.sin_port = uh->uh_sport; udp_in.sin_addr = ip->ip_src; /* * Make mbuf data length reflect UDP length. If not enough data to * reflect UDP length, drop. */ len = ntohs((u_short)uh->uh_ulen); if (ip->ip_len != len) { if (len > ip->ip_len || len < sizeof(struct udphdr)) { V_udpstat.udps_badlen++; goto badunlocked; } m_adj(m, len - ip->ip_len); /* ip->ip_len = len; */ } /* * Save a copy of the IP header in case we want restore it for * sending an ICMP error message in response. */ if (!V_udp_blackhole) save_ip = *ip; else memset(&save_ip, 0, sizeof(save_ip)); /* * Checksum extended UDP header and data. */ if (uh->uh_sum) { u_short uh_sum; if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) { if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR) uh_sum = m->m_pkthdr.csum_data; else uh_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr, htonl((u_short)len + m->m_pkthdr.csum_data + IPPROTO_UDP)); uh_sum ^= 0xffff; } else { char b[9]; bcopy(((struct ipovly *)ip)->ih_x1, b, 9); bzero(((struct ipovly *)ip)->ih_x1, 9); ((struct ipovly *)ip)->ih_len = uh->uh_ulen; uh_sum = in_cksum(m, len + sizeof (struct ip)); bcopy(b, ((struct ipovly *)ip)->ih_x1, 9); } if (uh_sum) { V_udpstat.udps_badsum++; m_freem(m); return; } } else V_udpstat.udps_nosum++; #ifdef IPFIREWALL_FORWARD /* * Grab info from PACKET_TAG_IPFORWARD tag prepended to the chain. */ fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL); if (fwd_tag != NULL) { struct sockaddr_in *next_hop; /* * Do the hack. */ next_hop = (struct sockaddr_in *)(fwd_tag + 1); ip->ip_dst = next_hop->sin_addr; uh->uh_dport = ntohs(next_hop->sin_port); /* * Remove the tag from the packet. We don't need it anymore. */ m_tag_delete(m, fwd_tag); } #endif INP_INFO_RLOCK(&V_udbinfo); if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) || in_broadcast(ip->ip_dst, ifp)) { struct inpcb *last; struct ip_moptions *imo; last = NULL; LIST_FOREACH(inp, &V_udb, inp_list) { if (inp->inp_lport != uh->uh_dport) continue; #ifdef INET6 if ((inp->inp_vflag & INP_IPV4) == 0) continue; #endif if (inp->inp_laddr.s_addr != INADDR_ANY && inp->inp_laddr.s_addr != ip->ip_dst.s_addr) continue; if (inp->inp_faddr.s_addr != INADDR_ANY && inp->inp_faddr.s_addr != ip->ip_src.s_addr) continue; /* * XXX: Do not check source port of incoming datagram * unless inp_connect() has been called to bind the * fport part of the 4-tuple; the source could be * trying to talk to us with an ephemeral port. */ if (inp->inp_fport != 0 && inp->inp_fport != uh->uh_sport) continue; INP_RLOCK(inp); /* * Handle socket delivery policy for any-source * and source-specific multicast. [RFC3678] */ imo = inp->inp_moptions; if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) && imo != NULL) { struct sockaddr_in sin; struct in_msource *ims; int blocked, mode; size_t idx; bzero(&sin, sizeof(struct sockaddr_in)); sin.sin_len = sizeof(struct sockaddr_in); sin.sin_family = AF_INET; sin.sin_addr = ip->ip_dst; blocked = 0; idx = imo_match_group(imo, ifp, (struct sockaddr *)&sin); if (idx == -1) { /* * No group membership for this socket. * Do not bump udps_noportbcast, as * this will happen further down. */ blocked++; } else { /* * Check for a multicast source filter * entry on this socket for this group. * MCAST_EXCLUDE is the default * behaviour. It means default accept; * entries, if present, denote sources * to be excluded from delivery. */ ims = imo_match_source(imo, idx, (struct sockaddr *)&udp_in); mode = imo->imo_mfilters[idx].imf_fmode; if ((ims != NULL && mode == MCAST_EXCLUDE) || (ims == NULL && mode == MCAST_INCLUDE)) { #ifdef DIAGNOSTIC if (bootverbose) { printf("%s: blocked by" " source filter\n", __func__); } #endif V_udpstat.udps_filtermcast++; blocked++; } } if (blocked != 0) { INP_RUNLOCK(inp); continue; } } if (last != NULL) { struct mbuf *n; n = m_copy(m, 0, M_COPYALL); if (last->inp_ppcb == NULL) { if (n != NULL) udp_append(last, ip, n, iphlen + sizeof(struct udphdr), &udp_in); INP_RUNLOCK(last); } else { /* * Engage the tunneling protocol we * will have to leave the info_lock * up, since we are hunting through * multiple UDP's. * */ udp_tun_func_t tunnel_func; tunnel_func = (udp_tun_func_t)last->inp_ppcb; tunnel_func(n, iphlen, last); INP_RUNLOCK(last); } } last = inp; /* * Don't look for additional matches if this one does * not have either the SO_REUSEPORT or SO_REUSEADDR * socket options set. This heuristic avoids * searching through all pcbs in the common case of a * non-shared port. It assumes that an application * will never clear these options after setting them. */ if ((last->inp_socket->so_options & (SO_REUSEPORT|SO_REUSEADDR)) == 0) break; } if (last == NULL) { /* * No matching pcb found; discard datagram. (No need * to send an ICMP Port Unreachable for a broadcast * or multicast datgram.) */ V_udpstat.udps_noportbcast++; goto badheadlocked; } if (last->inp_ppcb == NULL) { udp_append(last, ip, m, iphlen + sizeof(struct udphdr), &udp_in); INP_RUNLOCK(last); INP_INFO_RUNLOCK(&V_udbinfo); } else { /* * Engage the tunneling protocol. */ udp_tun_func_t tunnel_func; tunnel_func = (udp_tun_func_t)last->inp_ppcb; tunnel_func(m, iphlen, last); INP_RUNLOCK(last); INP_INFO_RUNLOCK(&V_udbinfo); } return; } /* * Locate pcb for datagram. */ inp = in_pcblookup_hash(&V_udbinfo, ip->ip_src, uh->uh_sport, ip->ip_dst, uh->uh_dport, 1, ifp); if (inp == NULL) { if (udp_log_in_vain) { char buf[4*sizeof "123"]; strcpy(buf, inet_ntoa(ip->ip_dst)); log(LOG_INFO, "Connection attempt to UDP %s:%d from %s:%d\n", buf, ntohs(uh->uh_dport), inet_ntoa(ip->ip_src), ntohs(uh->uh_sport)); } V_udpstat.udps_noport++; if (m->m_flags & (M_BCAST | M_MCAST)) { V_udpstat.udps_noportbcast++; goto badheadlocked; } if (V_udp_blackhole) goto badheadlocked; if (badport_bandlim(BANDLIM_ICMP_UNREACH) < 0) goto badheadlocked; *ip = save_ip; ip->ip_len += iphlen; icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PORT, 0, 0); INP_INFO_RUNLOCK(&V_udbinfo); return; } /* * Check the minimum TTL for socket. */ INP_RLOCK(inp); INP_INFO_RUNLOCK(&V_udbinfo); if (inp->inp_ip_minttl && inp->inp_ip_minttl > ip->ip_ttl) { INP_RUNLOCK(inp); goto badunlocked; } if (inp->inp_ppcb != NULL) { /* * Engage the tunneling protocol. */ udp_tun_func_t tunnel_func; tunnel_func = (udp_tun_func_t)inp->inp_ppcb; tunnel_func(m, iphlen, inp); INP_RUNLOCK(inp); return; } udp_append(inp, ip, m, iphlen + sizeof(struct udphdr), &udp_in); INP_RUNLOCK(inp); return; badheadlocked: if (inp) INP_RUNLOCK(inp); INP_INFO_RUNLOCK(&V_udbinfo); badunlocked: m_freem(m); } /* * Notify a udp user of an asynchronous error; just wake up so that they can * collect error status. */ struct inpcb * udp_notify(struct inpcb *inp, int errno) { /* * While udp_ctlinput() always calls udp_notify() with a read lock * when invoking it directly, in_pcbnotifyall() currently uses write * locks due to sharing code with TCP. For now, accept either a read * or a write lock, but a read lock is sufficient. */ INP_LOCK_ASSERT(inp); inp->inp_socket->so_error = errno; sorwakeup(inp->inp_socket); sowwakeup(inp->inp_socket); return (inp); } void udp_ctlinput(int cmd, struct sockaddr *sa, void *vip) { INIT_VNET_INET(curvnet); struct ip *ip = vip; struct udphdr *uh; struct in_addr faddr; struct inpcb *inp; faddr = ((struct sockaddr_in *)sa)->sin_addr; if (sa->sa_family != AF_INET || faddr.s_addr == INADDR_ANY) return; /* * Redirects don't need to be handled up here. */ if (PRC_IS_REDIRECT(cmd)) return; /* * Hostdead is ugly because it goes linearly through all PCBs. * * XXX: We never get this from ICMP, otherwise it makes an excellent * DoS attack on machines with many connections. */ if (cmd == PRC_HOSTDEAD) ip = NULL; else if ((unsigned)cmd >= PRC_NCMDS || inetctlerrmap[cmd] == 0) return; if (ip != NULL) { uh = (struct udphdr *)((caddr_t)ip + (ip->ip_hl << 2)); INP_INFO_RLOCK(&V_udbinfo); inp = in_pcblookup_hash(&V_udbinfo, faddr, uh->uh_dport, ip->ip_src, uh->uh_sport, 0, NULL); if (inp != NULL) { INP_RLOCK(inp); if (inp->inp_socket != NULL) { udp_notify(inp, inetctlerrmap[cmd]); } INP_RUNLOCK(inp); } INP_INFO_RUNLOCK(&V_udbinfo); } else in_pcbnotifyall(&V_udbinfo, faddr, inetctlerrmap[cmd], udp_notify); } static int udp_pcblist(SYSCTL_HANDLER_ARGS) { INIT_VNET_INET(curvnet); int error, i, n; struct inpcb *inp, **inp_list; inp_gen_t gencnt; struct xinpgen xig; /* * The process of preparing the PCB list is too time-consuming and * resource-intensive to repeat twice on every request. */ if (req->oldptr == 0) { n = V_udbinfo.ipi_count; req->oldidx = 2 * (sizeof xig) + (n + n/8) * sizeof(struct xinpcb); return (0); } if (req->newptr != 0) return (EPERM); /* * OK, now we're committed to doing something. */ INP_INFO_RLOCK(&V_udbinfo); gencnt = V_udbinfo.ipi_gencnt; n = V_udbinfo.ipi_count; INP_INFO_RUNLOCK(&V_udbinfo); error = sysctl_wire_old_buffer(req, 2 * (sizeof xig) + n * sizeof(struct xinpcb)); if (error != 0) return (error); xig.xig_len = sizeof xig; xig.xig_count = n; xig.xig_gen = gencnt; xig.xig_sogen = so_gencnt; error = SYSCTL_OUT(req, &xig, sizeof xig); if (error) return (error); inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK); if (inp_list == 0) return (ENOMEM); INP_INFO_RLOCK(&V_udbinfo); for (inp = LIST_FIRST(V_udbinfo.ipi_listhead), i = 0; inp && i < n; inp = LIST_NEXT(inp, inp_list)) { INP_RLOCK(inp); if (inp->inp_gencnt <= gencnt && cr_canseeinpcb(req->td->td_ucred, inp) == 0) inp_list[i++] = inp; INP_RUNLOCK(inp); } INP_INFO_RUNLOCK(&V_udbinfo); n = i; error = 0; for (i = 0; i < n; i++) { inp = inp_list[i]; INP_RLOCK(inp); if (inp->inp_gencnt <= gencnt) { struct xinpcb xi; bzero(&xi, sizeof(xi)); xi.xi_len = sizeof xi; /* XXX should avoid extra copy */ bcopy(inp, &xi.xi_inp, sizeof *inp); if (inp->inp_socket) sotoxsocket(inp->inp_socket, &xi.xi_socket); xi.xi_inp.inp_gencnt = inp->inp_gencnt; INP_RUNLOCK(inp); error = SYSCTL_OUT(req, &xi, sizeof xi); } else INP_RUNLOCK(inp); } if (!error) { /* * Give the user an updated idea of our state. If the * generation differs from what we told her before, she knows * that something happened while we were processing this * request, and it might be necessary to retry. */ INP_INFO_RLOCK(&V_udbinfo); xig.xig_gen = V_udbinfo.ipi_gencnt; xig.xig_sogen = so_gencnt; xig.xig_count = V_udbinfo.ipi_count; INP_INFO_RUNLOCK(&V_udbinfo); error = SYSCTL_OUT(req, &xig, sizeof xig); } free(inp_list, M_TEMP); return (error); } SYSCTL_PROC(_net_inet_udp, UDPCTL_PCBLIST, pcblist, CTLFLAG_RD, 0, 0, udp_pcblist, "S,xinpcb", "List of active UDP sockets"); static int udp_getcred(SYSCTL_HANDLER_ARGS) { INIT_VNET_INET(curvnet); struct xucred xuc; struct sockaddr_in addrs[2]; struct inpcb *inp; int error; error = priv_check(req->td, PRIV_NETINET_GETCRED); if (error) return (error); error = SYSCTL_IN(req, addrs, sizeof(addrs)); if (error) return (error); INP_INFO_RLOCK(&V_udbinfo); inp = in_pcblookup_hash(&V_udbinfo, addrs[1].sin_addr, addrs[1].sin_port, addrs[0].sin_addr, addrs[0].sin_port, 1, NULL); if (inp != NULL) { INP_RLOCK(inp); INP_INFO_RUNLOCK(&V_udbinfo); if (inp->inp_socket == NULL) error = ENOENT; if (error == 0) error = cr_canseeinpcb(req->td->td_ucred, inp); if (error == 0) cru2x(inp->inp_cred, &xuc); INP_RUNLOCK(inp); } else { INP_INFO_RUNLOCK(&V_udbinfo); error = ENOENT; } if (error == 0) error = SYSCTL_OUT(req, &xuc, sizeof(struct xucred)); return (error); } SYSCTL_PROC(_net_inet_udp, OID_AUTO, getcred, CTLTYPE_OPAQUE|CTLFLAG_RW|CTLFLAG_PRISON, 0, 0, udp_getcred, "S,xucred", "Get the xucred of a UDP connection"); static int udp_output(struct inpcb *inp, struct mbuf *m, struct sockaddr *addr, struct mbuf *control, struct thread *td) { INIT_VNET_INET(inp->inp_vnet); struct udpiphdr *ui; int len = m->m_pkthdr.len; struct in_addr faddr, laddr; struct cmsghdr *cm; struct sockaddr_in *sin, src; int error = 0; int ipflags; u_short fport, lport; int unlock_udbinfo; /* * udp_output() may need to temporarily bind or connect the current * inpcb. As such, we don't know up front whether we will need the * pcbinfo lock or not. Do any work to decide what is needed up * front before acquiring any locks. */ if (len + sizeof(struct udpiphdr) > IP_MAXPACKET) { if (control) m_freem(control); m_freem(m); return (EMSGSIZE); } src.sin_family = 0; if (control != NULL) { /* * XXX: Currently, we assume all the optional information is * stored in a single mbuf. */ if (control->m_next) { m_freem(control); m_freem(m); return (EINVAL); } for (; control->m_len > 0; control->m_data += CMSG_ALIGN(cm->cmsg_len), control->m_len -= CMSG_ALIGN(cm->cmsg_len)) { cm = mtod(control, struct cmsghdr *); if (control->m_len < sizeof(*cm) || cm->cmsg_len == 0 || cm->cmsg_len > control->m_len) { error = EINVAL; break; } if (cm->cmsg_level != IPPROTO_IP) continue; switch (cm->cmsg_type) { case IP_SENDSRCADDR: if (cm->cmsg_len != CMSG_LEN(sizeof(struct in_addr))) { error = EINVAL; break; } bzero(&src, sizeof(src)); src.sin_family = AF_INET; src.sin_len = sizeof(src); src.sin_port = inp->inp_lport; src.sin_addr = *(struct in_addr *)CMSG_DATA(cm); break; default: error = ENOPROTOOPT; break; } if (error) break; } m_freem(control); } if (error) { m_freem(m); return (error); } /* * Depending on whether or not the application has bound or connected * the socket, we may have to do varying levels of work. The optimal * case is for a connected UDP socket, as a global lock isn't * required at all. * * In order to decide which we need, we require stability of the * inpcb binding, which we ensure by acquiring a read lock on the * inpcb. This doesn't strictly follow the lock order, so we play * the trylock and retry game; note that we may end up with more * conservative locks than required the second time around, so later * assertions have to accept that. Further analysis of the number of * misses under contention is required. */ sin = (struct sockaddr_in *)addr; INP_RLOCK(inp); if (sin != NULL && (inp->inp_laddr.s_addr == INADDR_ANY && inp->inp_lport == 0)) { INP_RUNLOCK(inp); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); unlock_udbinfo = 2; } else if ((sin != NULL && ( (sin->sin_addr.s_addr == INADDR_ANY) || (sin->sin_addr.s_addr == INADDR_BROADCAST) || (inp->inp_laddr.s_addr == INADDR_ANY) || (inp->inp_lport == 0))) || (src.sin_family == AF_INET)) { if (!INP_INFO_TRY_RLOCK(&V_udbinfo)) { INP_RUNLOCK(inp); INP_INFO_RLOCK(&V_udbinfo); INP_RLOCK(inp); } unlock_udbinfo = 1; } else unlock_udbinfo = 0; /* * If the IP_SENDSRCADDR control message was specified, override the * source address for this datagram. Its use is invalidated if the * address thus specified is incomplete or clobbers other inpcbs. */ laddr = inp->inp_laddr; lport = inp->inp_lport; if (src.sin_family == AF_INET) { INP_INFO_LOCK_ASSERT(&V_udbinfo); if ((lport == 0) || (laddr.s_addr == INADDR_ANY && src.sin_addr.s_addr == INADDR_ANY)) { error = EINVAL; goto release; } error = in_pcbbind_setup(inp, (struct sockaddr *)&src, &laddr.s_addr, &lport, td->td_ucred); if (error) goto release; } /* * If a UDP socket has been connected, then a local address/port will * have been selected and bound. * * If a UDP socket has not been connected to, then an explicit * destination address must be used, in which case a local * address/port may not have been selected and bound. */ if (sin != NULL) { INP_LOCK_ASSERT(inp); if (inp->inp_faddr.s_addr != INADDR_ANY) { error = EISCONN; goto release; } /* * Jail may rewrite the destination address, so let it do * that before we use it. */ - if (prison_remote_ip4(td->td_ucred, &sin->sin_addr) != 0) { - error = EINVAL; + error = prison_remote_ip4(td->td_ucred, &sin->sin_addr); + if (error) goto release; - } /* * If a local address or port hasn't yet been selected, or if * the destination address needs to be rewritten due to using * a special INADDR_ constant, invoke in_pcbconnect_setup() * to do the heavy lifting. Once a port is selected, we * commit the binding back to the socket; we also commit the * binding of the address if in jail. * * If we already have a valid binding and we're not * requesting a destination address rewrite, use a fast path. */ if (inp->inp_laddr.s_addr == INADDR_ANY || inp->inp_lport == 0 || sin->sin_addr.s_addr == INADDR_ANY || sin->sin_addr.s_addr == INADDR_BROADCAST) { INP_INFO_LOCK_ASSERT(&V_udbinfo); error = in_pcbconnect_setup(inp, addr, &laddr.s_addr, &lport, &faddr.s_addr, &fport, NULL, td->td_ucred); if (error) goto release; /* * XXXRW: Why not commit the port if the address is * !INADDR_ANY? */ /* Commit the local port if newly assigned. */ if (inp->inp_laddr.s_addr == INADDR_ANY && inp->inp_lport == 0) { INP_INFO_WLOCK_ASSERT(&V_udbinfo); INP_WLOCK_ASSERT(inp); /* * Remember addr if jailed, to prevent * rebinding. */ if (jailed(td->td_ucred)) inp->inp_laddr = laddr; inp->inp_lport = lport; if (in_pcbinshash(inp) != 0) { inp->inp_lport = 0; error = EAGAIN; goto release; } inp->inp_flags |= INP_ANONPORT; } } else { faddr = sin->sin_addr; fport = sin->sin_port; } } else { INP_LOCK_ASSERT(inp); faddr = inp->inp_faddr; fport = inp->inp_fport; if (faddr.s_addr == INADDR_ANY) { error = ENOTCONN; goto release; } } /* * Calculate data length and get a mbuf for UDP, IP, and possible * link-layer headers. Immediate slide the data pointer back forward * since we won't use that space at this layer. */ M_PREPEND(m, sizeof(struct udpiphdr) + max_linkhdr, M_DONTWAIT); if (m == NULL) { error = ENOBUFS; goto release; } m->m_data += max_linkhdr; m->m_len -= max_linkhdr; m->m_pkthdr.len -= max_linkhdr; /* * Fill in mbuf with extended UDP header and addresses and length put * into network format. */ ui = mtod(m, struct udpiphdr *); bzero(ui->ui_x1, sizeof(ui->ui_x1)); /* XXX still needed? */ ui->ui_pr = IPPROTO_UDP; ui->ui_src = laddr; ui->ui_dst = faddr; ui->ui_sport = lport; ui->ui_dport = fport; ui->ui_ulen = htons((u_short)len + sizeof(struct udphdr)); /* * Set the Don't Fragment bit in the IP header. */ if (inp->inp_flags & INP_DONTFRAG) { struct ip *ip; ip = (struct ip *)&ui->ui_i; ip->ip_off |= IP_DF; } ipflags = 0; if (inp->inp_socket->so_options & SO_DONTROUTE) ipflags |= IP_ROUTETOIF; if (inp->inp_socket->so_options & SO_BROADCAST) ipflags |= IP_ALLOWBROADCAST; if (inp->inp_flags & INP_ONESBCAST) ipflags |= IP_SENDONES; #ifdef MAC mac_inpcb_create_mbuf(inp, m); #endif /* * Set up checksum and output datagram. */ if (udp_cksum) { if (inp->inp_flags & INP_ONESBCAST) faddr.s_addr = INADDR_BROADCAST; ui->ui_sum = in_pseudo(ui->ui_src.s_addr, faddr.s_addr, htons((u_short)len + sizeof(struct udphdr) + IPPROTO_UDP)); m->m_pkthdr.csum_flags = CSUM_UDP; m->m_pkthdr.csum_data = offsetof(struct udphdr, uh_sum); } else ui->ui_sum = 0; ((struct ip *)ui)->ip_len = sizeof (struct udpiphdr) + len; ((struct ip *)ui)->ip_ttl = inp->inp_ip_ttl; /* XXX */ ((struct ip *)ui)->ip_tos = inp->inp_ip_tos; /* XXX */ V_udpstat.udps_opackets++; if (unlock_udbinfo == 2) INP_INFO_WUNLOCK(&V_udbinfo); else if (unlock_udbinfo == 1) INP_INFO_RUNLOCK(&V_udbinfo); error = ip_output(m, inp->inp_options, NULL, ipflags, inp->inp_moptions, inp); if (unlock_udbinfo == 2) INP_WUNLOCK(inp); else INP_RUNLOCK(inp); return (error); release: if (unlock_udbinfo == 2) { INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); } else if (unlock_udbinfo == 1) { INP_RUNLOCK(inp); INP_INFO_RUNLOCK(&V_udbinfo); } else INP_RUNLOCK(inp); m_freem(m); return (error); } static void udp_abort(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_abort: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (inp->inp_faddr.s_addr != INADDR_ANY) { in_pcbdisconnect(inp); inp->inp_laddr.s_addr = INADDR_ANY; soisdisconnected(so); } INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); } static int udp_attach(struct socket *so, int proto, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp == NULL, ("udp_attach: inp != NULL")); error = soreserve(so, udp_sendspace, udp_recvspace); if (error) return (error); INP_INFO_WLOCK(&V_udbinfo); error = in_pcballoc(so, &V_udbinfo); if (error) { INP_INFO_WUNLOCK(&V_udbinfo); return (error); } inp = (struct inpcb *)so->so_pcb; INP_INFO_WUNLOCK(&V_udbinfo); inp->inp_vflag |= INP_IPV4; inp->inp_ip_ttl = V_ip_defttl; /* * UDP does not have a per-protocol pcb (inp->inp_ppcb). * We use this pointer for kernel tunneling pointer. * If we ever need to have a protocol block we will * need to move this function pointer there. Null * in this pointer means "do the normal thing". */ inp->inp_ppcb = NULL; INP_WUNLOCK(inp); return (0); } int udp_set_kernel_tunneling(struct socket *so, udp_tun_func_t f) { struct inpcb *inp; inp = (struct inpcb *)so->so_pcb; KASSERT(so->so_type == SOCK_DGRAM, ("udp_set_kernel_tunneling: !dgram")); KASSERT(so->so_pcb != NULL, ("udp_set_kernel_tunneling: NULL inp")); if (so->so_type != SOCK_DGRAM) { /* Not UDP socket... sorry! */ return (ENOTSUP); } if (inp == NULL) { /* NULL INP? */ return (EINVAL); } INP_WLOCK(inp); if (inp->inp_ppcb != NULL) { INP_WUNLOCK(inp); return (EBUSY); } inp->inp_ppcb = f; INP_WUNLOCK(inp); return (0); } static int udp_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_bind: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); error = in_pcbbind(inp, nam, td->td_ucred); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (error); } static void udp_close(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_close: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (inp->inp_faddr.s_addr != INADDR_ANY) { in_pcbdisconnect(inp); inp->inp_laddr.s_addr = INADDR_ANY; soisdisconnected(so); } INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); } static int udp_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; struct sockaddr_in *sin; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_connect: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (inp->inp_faddr.s_addr != INADDR_ANY) { INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (EISCONN); } sin = (struct sockaddr_in *)nam; - if (prison_remote_ip4(td->td_ucred, &sin->sin_addr) != 0) { + error = prison_remote_ip4(td->td_ucred, &sin->sin_addr); + if (error != 0) { INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); - return (EAFNOSUPPORT); + return (error); } error = in_pcbconnect(inp, nam, td->td_ucred); if (error == 0) soisconnected(so); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (error); } static void udp_detach(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_detach: inp == NULL")); KASSERT(inp->inp_faddr.s_addr == INADDR_ANY, ("udp_detach: not disconnected")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); in_pcbdetach(inp); in_pcbfree(inp); INP_INFO_WUNLOCK(&V_udbinfo); } static int udp_disconnect(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_disconnect: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (inp->inp_faddr.s_addr == INADDR_ANY) { INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (ENOTCONN); } in_pcbdisconnect(inp); inp->inp_laddr.s_addr = INADDR_ANY; SOCK_LOCK(so); so->so_state &= ~SS_ISCONNECTED; /* XXX */ SOCK_UNLOCK(so); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (0); } static int udp_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *addr, struct mbuf *control, struct thread *td) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_send: inp == NULL")); return (udp_output(inp, m, addr, control, td)); } int udp_shutdown(struct socket *so) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp_shutdown: inp == NULL")); INP_WLOCK(inp); socantsendmore(so); INP_WUNLOCK(inp); return (0); } struct pr_usrreqs udp_usrreqs = { .pru_abort = udp_abort, .pru_attach = udp_attach, .pru_bind = udp_bind, .pru_connect = udp_connect, .pru_control = in_control, .pru_detach = udp_detach, .pru_disconnect = udp_disconnect, .pru_peeraddr = in_getpeeraddr, .pru_send = udp_send, .pru_soreceive = soreceive_dgram, .pru_sosend = sosend_dgram, .pru_shutdown = udp_shutdown, .pru_sockaddr = in_getsockaddr, .pru_sosetlabel = in_pcbsosetlabel, .pru_close = udp_close, }; diff --git a/sys/netinet6/in6.c b/sys/netinet6/in6.c index 693a2f979d38..81a37f845b79 100644 --- a/sys/netinet6/in6.c +++ b/sys/netinet6/in6.c @@ -1,2389 +1,2388 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 3. Neither the name of the project 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 PROJECT 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 PROJECT 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: in6.c,v 1.259 2002/01/21 11:37:50 keiichi Exp $ */ /*- * Copyright (c) 1982, 1986, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)in.c 8.2 (Berkeley) 11/15/93 */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include MALLOC_DEFINE(M_IP6MADDR, "in6_multi", "internet multicast address"); /* * Definitions of some costant IP6 addresses. */ const struct in6_addr in6addr_any = IN6ADDR_ANY_INIT; const struct in6_addr in6addr_loopback = IN6ADDR_LOOPBACK_INIT; const struct in6_addr in6addr_nodelocal_allnodes = IN6ADDR_NODELOCAL_ALLNODES_INIT; const struct in6_addr in6addr_linklocal_allnodes = IN6ADDR_LINKLOCAL_ALLNODES_INIT; const struct in6_addr in6addr_linklocal_allrouters = IN6ADDR_LINKLOCAL_ALLROUTERS_INIT; const struct in6_addr in6mask0 = IN6MASK0; const struct in6_addr in6mask32 = IN6MASK32; const struct in6_addr in6mask64 = IN6MASK64; const struct in6_addr in6mask96 = IN6MASK96; const struct in6_addr in6mask128 = IN6MASK128; const struct sockaddr_in6 sa6_any = { sizeof(sa6_any), AF_INET6, 0, 0, IN6ADDR_ANY_INIT, 0 }; static int in6_lifaddr_ioctl __P((struct socket *, u_long, caddr_t, struct ifnet *, struct thread *)); static int in6_ifinit __P((struct ifnet *, struct in6_ifaddr *, struct sockaddr_in6 *, int)); static void in6_unlink_ifa(struct in6_ifaddr *, struct ifnet *); struct in6_multihead in6_multihead; /* XXX BSS initialization */ int (*faithprefix_p)(struct in6_addr *); int in6_mask2len(struct in6_addr *mask, u_char *lim0) { int x = 0, y; u_char *lim = lim0, *p; /* ignore the scope_id part */ if (lim0 == NULL || lim0 - (u_char *)mask > sizeof(*mask)) lim = (u_char *)mask + sizeof(*mask); for (p = (u_char *)mask; p < lim; x++, p++) { if (*p != 0xff) break; } y = 0; if (p < lim) { for (y = 0; y < 8; y++) { if ((*p & (0x80 >> y)) == 0) break; } } /* * when the limit pointer is given, do a stricter check on the * remaining bits. */ if (p < lim) { if (y != 0 && (*p & (0x00ff >> y)) != 0) return (-1); for (p = p + 1; p < lim; p++) if (*p != 0) return (-1); } return x * 8 + y; } #define ifa2ia6(ifa) ((struct in6_ifaddr *)(ifa)) #define ia62ifa(ia6) (&((ia6)->ia_ifa)) int in6_control(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp, struct thread *td) { INIT_VNET_INET6(curvnet); struct in6_ifreq *ifr = (struct in6_ifreq *)data; struct in6_ifaddr *ia = NULL; struct in6_aliasreq *ifra = (struct in6_aliasreq *)data; struct sockaddr_in6 *sa6; int error; switch (cmd) { case SIOCGETSGCNT_IN6: case SIOCGETMIFCNT_IN6: return (mrt6_ioctl ? mrt6_ioctl(cmd, data) : EOPNOTSUPP); } switch(cmd) { case SIOCAADDRCTL_POLICY: case SIOCDADDRCTL_POLICY: if (td != NULL) { error = priv_check(td, PRIV_NETINET_ADDRCTRL6); if (error) return (error); } return (in6_src_ioctl(cmd, data)); } if (ifp == NULL) return (EOPNOTSUPP); switch (cmd) { case SIOCSNDFLUSH_IN6: case SIOCSPFXFLUSH_IN6: case SIOCSRTRFLUSH_IN6: case SIOCSDEFIFACE_IN6: case SIOCSIFINFO_FLAGS: if (td != NULL) { error = priv_check(td, PRIV_NETINET_ND6); if (error) return (error); } /* FALLTHROUGH */ case OSIOCGIFINFO_IN6: case SIOCGIFINFO_IN6: case SIOCSIFINFO_IN6: case SIOCGDRLST_IN6: case SIOCGPRLST_IN6: case SIOCGNBRINFO_IN6: case SIOCGDEFIFACE_IN6: return (nd6_ioctl(cmd, data, ifp)); } switch (cmd) { case SIOCSIFPREFIX_IN6: case SIOCDIFPREFIX_IN6: case SIOCAIFPREFIX_IN6: case SIOCCIFPREFIX_IN6: case SIOCSGIFPREFIX_IN6: case SIOCGIFPREFIX_IN6: log(LOG_NOTICE, "prefix ioctls are now invalidated. " "please use ifconfig.\n"); return (EOPNOTSUPP); } switch (cmd) { case SIOCSSCOPE6: if (td != NULL) { error = priv_check(td, PRIV_NETINET_SCOPE6); if (error) return (error); } return (scope6_set(ifp, (struct scope6_id *)ifr->ifr_ifru.ifru_scope_id)); case SIOCGSCOPE6: return (scope6_get(ifp, (struct scope6_id *)ifr->ifr_ifru.ifru_scope_id)); case SIOCGSCOPE6DEF: return (scope6_get_default((struct scope6_id *) ifr->ifr_ifru.ifru_scope_id)); } switch (cmd) { case SIOCALIFADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_ADDIFADDR); if (error) return (error); } return in6_lifaddr_ioctl(so, cmd, data, ifp, td); case SIOCDLIFADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_DELIFADDR); if (error) return (error); } /* FALLTHROUGH */ case SIOCGLIFADDR: return in6_lifaddr_ioctl(so, cmd, data, ifp, td); } /* * Find address for this interface, if it exists. * * In netinet code, we have checked ifra_addr in SIOCSIF*ADDR operation * only, and used the first interface address as the target of other * operations (without checking ifra_addr). This was because netinet * code/API assumed at most 1 interface address per interface. * Since IPv6 allows a node to assign multiple addresses * on a single interface, we almost always look and check the * presence of ifra_addr, and reject invalid ones here. * It also decreases duplicated code among SIOC*_IN6 operations. */ switch (cmd) { case SIOCAIFADDR_IN6: case SIOCSIFPHYADDR_IN6: sa6 = &ifra->ifra_addr; break; case SIOCSIFADDR_IN6: case SIOCGIFADDR_IN6: case SIOCSIFDSTADDR_IN6: case SIOCSIFNETMASK_IN6: case SIOCGIFDSTADDR_IN6: case SIOCGIFNETMASK_IN6: case SIOCDIFADDR_IN6: case SIOCGIFPSRCADDR_IN6: case SIOCGIFPDSTADDR_IN6: case SIOCGIFAFLAG_IN6: case SIOCSNDFLUSH_IN6: case SIOCSPFXFLUSH_IN6: case SIOCSRTRFLUSH_IN6: case SIOCGIFALIFETIME_IN6: case SIOCSIFALIFETIME_IN6: case SIOCGIFSTAT_IN6: case SIOCGIFSTAT_ICMP6: sa6 = &ifr->ifr_addr; break; default: sa6 = NULL; break; } if (sa6 && sa6->sin6_family == AF_INET6) { int error = 0; if (sa6->sin6_scope_id != 0) error = sa6_embedscope(sa6, 0); else error = in6_setscope(&sa6->sin6_addr, ifp, NULL); if (error != 0) return (error); - if (td != NULL && !prison_check_ip6(td->td_ucred, - &sa6->sin6_addr)) - return (EADDRNOTAVAIL); + if (td != NULL && (error = prison_check_ip6(td->td_ucred, + &sa6->sin6_addr)) != 0) + return (error); ia = in6ifa_ifpwithaddr(ifp, &sa6->sin6_addr); } else ia = NULL; switch (cmd) { case SIOCSIFADDR_IN6: case SIOCSIFDSTADDR_IN6: case SIOCSIFNETMASK_IN6: /* * Since IPv6 allows a node to assign multiple addresses * on a single interface, SIOCSIFxxx ioctls are deprecated. */ /* we decided to obsolete this command (20000704) */ return (EINVAL); case SIOCDIFADDR_IN6: /* * for IPv4, we look for existing in_ifaddr here to allow * "ifconfig if0 delete" to remove the first IPv4 address on * the interface. For IPv6, as the spec allows multiple * interface address from the day one, we consider "remove the * first one" semantics to be not preferable. */ if (ia == NULL) return (EADDRNOTAVAIL); /* FALLTHROUGH */ case SIOCAIFADDR_IN6: /* * We always require users to specify a valid IPv6 address for * the corresponding operation. */ if (ifra->ifra_addr.sin6_family != AF_INET6 || ifra->ifra_addr.sin6_len != sizeof(struct sockaddr_in6)) return (EAFNOSUPPORT); if (td != NULL) { error = priv_check(td, (cmd == SIOCDIFADDR_IN6) ? PRIV_NET_DELIFADDR : PRIV_NET_ADDIFADDR); if (error) return (error); } break; case SIOCGIFADDR_IN6: /* This interface is basically deprecated. use SIOCGIFCONF. */ /* FALLTHROUGH */ case SIOCGIFAFLAG_IN6: case SIOCGIFNETMASK_IN6: case SIOCGIFDSTADDR_IN6: case SIOCGIFALIFETIME_IN6: /* must think again about its semantics */ if (ia == NULL) return (EADDRNOTAVAIL); break; case SIOCSIFALIFETIME_IN6: { struct in6_addrlifetime *lt; if (td != NULL) { error = priv_check(td, PRIV_NETINET_ALIFETIME6); if (error) return (error); } if (ia == NULL) return (EADDRNOTAVAIL); /* sanity for overflow - beware unsigned */ lt = &ifr->ifr_ifru.ifru_lifetime; if (lt->ia6t_vltime != ND6_INFINITE_LIFETIME && lt->ia6t_vltime + time_second < time_second) { return EINVAL; } if (lt->ia6t_pltime != ND6_INFINITE_LIFETIME && lt->ia6t_pltime + time_second < time_second) { return EINVAL; } break; } } switch (cmd) { case SIOCGIFADDR_IN6: ifr->ifr_addr = ia->ia_addr; if ((error = sa6_recoverscope(&ifr->ifr_addr)) != 0) return (error); break; case SIOCGIFDSTADDR_IN6: if ((ifp->if_flags & IFF_POINTOPOINT) == 0) return (EINVAL); /* * XXX: should we check if ifa_dstaddr is NULL and return * an error? */ ifr->ifr_dstaddr = ia->ia_dstaddr; if ((error = sa6_recoverscope(&ifr->ifr_dstaddr)) != 0) return (error); break; case SIOCGIFNETMASK_IN6: ifr->ifr_addr = ia->ia_prefixmask; break; case SIOCGIFAFLAG_IN6: ifr->ifr_ifru.ifru_flags6 = ia->ia6_flags; break; case SIOCGIFSTAT_IN6: if (ifp == NULL) return EINVAL; bzero(&ifr->ifr_ifru.ifru_stat, sizeof(ifr->ifr_ifru.ifru_stat)); ifr->ifr_ifru.ifru_stat = *((struct in6_ifextra *)ifp->if_afdata[AF_INET6])->in6_ifstat; break; case SIOCGIFSTAT_ICMP6: if (ifp == NULL) return EINVAL; bzero(&ifr->ifr_ifru.ifru_icmp6stat, sizeof(ifr->ifr_ifru.ifru_icmp6stat)); ifr->ifr_ifru.ifru_icmp6stat = *((struct in6_ifextra *)ifp->if_afdata[AF_INET6])->icmp6_ifstat; break; case SIOCGIFALIFETIME_IN6: ifr->ifr_ifru.ifru_lifetime = ia->ia6_lifetime; if (ia->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { time_t maxexpire; struct in6_addrlifetime *retlt = &ifr->ifr_ifru.ifru_lifetime; /* * XXX: adjust expiration time assuming time_t is * signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (ia->ia6_lifetime.ia6t_vltime < maxexpire - ia->ia6_updatetime) { retlt->ia6t_expire = ia->ia6_updatetime + ia->ia6_lifetime.ia6t_vltime; } else retlt->ia6t_expire = maxexpire; } if (ia->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { time_t maxexpire; struct in6_addrlifetime *retlt = &ifr->ifr_ifru.ifru_lifetime; /* * XXX: adjust expiration time assuming time_t is * signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (ia->ia6_lifetime.ia6t_pltime < maxexpire - ia->ia6_updatetime) { retlt->ia6t_preferred = ia->ia6_updatetime + ia->ia6_lifetime.ia6t_pltime; } else retlt->ia6t_preferred = maxexpire; } break; case SIOCSIFALIFETIME_IN6: ia->ia6_lifetime = ifr->ifr_ifru.ifru_lifetime; /* for sanity */ if (ia->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_expire = time_second + ia->ia6_lifetime.ia6t_vltime; } else ia->ia6_lifetime.ia6t_expire = 0; if (ia->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_preferred = time_second + ia->ia6_lifetime.ia6t_pltime; } else ia->ia6_lifetime.ia6t_preferred = 0; break; case SIOCAIFADDR_IN6: { int i, error = 0; struct nd_prefixctl pr0; struct nd_prefix *pr; /* * first, make or update the interface address structure, * and link it to the list. */ if ((error = in6_update_ifa(ifp, ifra, ia, 0)) != 0) return (error); if ((ia = in6ifa_ifpwithaddr(ifp, &ifra->ifra_addr.sin6_addr)) == NULL) { /* * this can happen when the user specify the 0 valid * lifetime. */ break; } /* * then, make the prefix on-link on the interface. * XXX: we'd rather create the prefix before the address, but * we need at least one address to install the corresponding * interface route, so we configure the address first. */ /* * convert mask to prefix length (prefixmask has already * been validated in in6_update_ifa(). */ bzero(&pr0, sizeof(pr0)); pr0.ndpr_ifp = ifp; pr0.ndpr_plen = in6_mask2len(&ifra->ifra_prefixmask.sin6_addr, NULL); if (pr0.ndpr_plen == 128) { break; /* we don't need to install a host route. */ } pr0.ndpr_prefix = ifra->ifra_addr; /* apply the mask for safety. */ for (i = 0; i < 4; i++) { pr0.ndpr_prefix.sin6_addr.s6_addr32[i] &= ifra->ifra_prefixmask.sin6_addr.s6_addr32[i]; } /* * XXX: since we don't have an API to set prefix (not address) * lifetimes, we just use the same lifetimes as addresses. * The (temporarily) installed lifetimes can be overridden by * later advertised RAs (when accept_rtadv is non 0), which is * an intended behavior. */ pr0.ndpr_raf_onlink = 1; /* should be configurable? */ pr0.ndpr_raf_auto = ((ifra->ifra_flags & IN6_IFF_AUTOCONF) != 0); pr0.ndpr_vltime = ifra->ifra_lifetime.ia6t_vltime; pr0.ndpr_pltime = ifra->ifra_lifetime.ia6t_pltime; /* add the prefix if not yet. */ if ((pr = nd6_prefix_lookup(&pr0)) == NULL) { /* * nd6_prelist_add will install the corresponding * interface route. */ if ((error = nd6_prelist_add(&pr0, NULL, &pr)) != 0) return (error); if (pr == NULL) { log(LOG_ERR, "nd6_prelist_add succeeded but " "no prefix\n"); return (EINVAL); /* XXX panic here? */ } } /* relate the address to the prefix */ if (ia->ia6_ndpr == NULL) { ia->ia6_ndpr = pr; pr->ndpr_refcnt++; /* * If this is the first autoconf address from the * prefix, create a temporary address as well * (when required). */ if ((ia->ia6_flags & IN6_IFF_AUTOCONF) && V_ip6_use_tempaddr && pr->ndpr_refcnt == 1) { int e; if ((e = in6_tmpifadd(ia, 1, 0)) != 0) { log(LOG_NOTICE, "in6_control: failed " "to create a temporary address, " "errno=%d\n", e); } } } /* * this might affect the status of autoconfigured addresses, * that is, this address might make other addresses detached. */ pfxlist_onlink_check(); if (error == 0 && ia) EVENTHANDLER_INVOKE(ifaddr_event, ifp); break; } case SIOCDIFADDR_IN6: { struct nd_prefix *pr; /* * If the address being deleted is the only one that owns * the corresponding prefix, expire the prefix as well. * XXX: theoretically, we don't have to worry about such * relationship, since we separate the address management * and the prefix management. We do this, however, to provide * as much backward compatibility as possible in terms of * the ioctl operation. * Note that in6_purgeaddr() will decrement ndpr_refcnt. */ pr = ia->ia6_ndpr; in6_purgeaddr(&ia->ia_ifa); if (pr && pr->ndpr_refcnt == 0) prelist_remove(pr); EVENTHANDLER_INVOKE(ifaddr_event, ifp); break; } default: if (ifp == NULL || ifp->if_ioctl == 0) return (EOPNOTSUPP); return ((*ifp->if_ioctl)(ifp, cmd, data)); } return (0); } /* * Update parameters of an IPv6 interface address. * If necessary, a new entry is created and linked into address chains. * This function is separated from in6_control(). * XXX: should this be performed under splnet()? */ int in6_update_ifa(struct ifnet *ifp, struct in6_aliasreq *ifra, struct in6_ifaddr *ia, int flags) { INIT_VNET_INET6(ifp->if_vnet); INIT_VPROCG(TD_TO_VPROCG(curthread)); /* XXX V_hostname needs this */ int error = 0, hostIsNew = 0, plen = -1; struct in6_ifaddr *oia; struct sockaddr_in6 dst6; struct in6_addrlifetime *lt; struct in6_multi_mship *imm; struct in6_multi *in6m_sol; struct rtentry *rt; int delay; char ip6buf[INET6_ADDRSTRLEN]; /* Validate parameters */ if (ifp == NULL || ifra == NULL) /* this maybe redundant */ return (EINVAL); /* * The destination address for a p2p link must have a family * of AF_UNSPEC or AF_INET6. */ if ((ifp->if_flags & IFF_POINTOPOINT) != 0 && ifra->ifra_dstaddr.sin6_family != AF_INET6 && ifra->ifra_dstaddr.sin6_family != AF_UNSPEC) return (EAFNOSUPPORT); /* * validate ifra_prefixmask. don't check sin6_family, netmask * does not carry fields other than sin6_len. */ if (ifra->ifra_prefixmask.sin6_len > sizeof(struct sockaddr_in6)) return (EINVAL); /* * Because the IPv6 address architecture is classless, we require * users to specify a (non 0) prefix length (mask) for a new address. * We also require the prefix (when specified) mask is valid, and thus * reject a non-consecutive mask. */ if (ia == NULL && ifra->ifra_prefixmask.sin6_len == 0) return (EINVAL); if (ifra->ifra_prefixmask.sin6_len != 0) { plen = in6_mask2len(&ifra->ifra_prefixmask.sin6_addr, (u_char *)&ifra->ifra_prefixmask + ifra->ifra_prefixmask.sin6_len); if (plen <= 0) return (EINVAL); } else { /* * In this case, ia must not be NULL. We just use its prefix * length. */ plen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); } /* * If the destination address on a p2p interface is specified, * and the address is a scoped one, validate/set the scope * zone identifier. */ dst6 = ifra->ifra_dstaddr; if ((ifp->if_flags & (IFF_POINTOPOINT|IFF_LOOPBACK)) != 0 && (dst6.sin6_family == AF_INET6)) { struct in6_addr in6_tmp; u_int32_t zoneid; in6_tmp = dst6.sin6_addr; if (in6_setscope(&in6_tmp, ifp, &zoneid)) return (EINVAL); /* XXX: should be impossible */ if (dst6.sin6_scope_id != 0) { if (dst6.sin6_scope_id != zoneid) return (EINVAL); } else /* user omit to specify the ID. */ dst6.sin6_scope_id = zoneid; /* convert into the internal form */ if (sa6_embedscope(&dst6, 0)) return (EINVAL); /* XXX: should be impossible */ } /* * The destination address can be specified only for a p2p or a * loopback interface. If specified, the corresponding prefix length * must be 128. */ if (ifra->ifra_dstaddr.sin6_family == AF_INET6) { if ((ifp->if_flags & (IFF_POINTOPOINT|IFF_LOOPBACK)) == 0) { /* XXX: noisy message */ nd6log((LOG_INFO, "in6_update_ifa: a destination can " "be specified for a p2p or a loopback IF only\n")); return (EINVAL); } if (plen != 128) { nd6log((LOG_INFO, "in6_update_ifa: prefixlen should " "be 128 when dstaddr is specified\n")); return (EINVAL); } } /* lifetime consistency check */ lt = &ifra->ifra_lifetime; if (lt->ia6t_pltime > lt->ia6t_vltime) return (EINVAL); if (lt->ia6t_vltime == 0) { /* * the following log might be noisy, but this is a typical * configuration mistake or a tool's bug. */ nd6log((LOG_INFO, "in6_update_ifa: valid lifetime is 0 for %s\n", ip6_sprintf(ip6buf, &ifra->ifra_addr.sin6_addr))); if (ia == NULL) return (0); /* there's nothing to do */ } /* * If this is a new address, allocate a new ifaddr and link it * into chains. */ if (ia == NULL) { hostIsNew = 1; /* * When in6_update_ifa() is called in a process of a received * RA, it is called under an interrupt context. So, we should * call malloc with M_NOWAIT. */ ia = (struct in6_ifaddr *) malloc(sizeof(*ia), M_IFADDR, M_NOWAIT); if (ia == NULL) return (ENOBUFS); bzero((caddr_t)ia, sizeof(*ia)); LIST_INIT(&ia->ia6_memberships); /* Initialize the address and masks, and put time stamp */ IFA_LOCK_INIT(&ia->ia_ifa); ia->ia_ifa.ifa_addr = (struct sockaddr *)&ia->ia_addr; ia->ia_addr.sin6_family = AF_INET6; ia->ia_addr.sin6_len = sizeof(ia->ia_addr); ia->ia6_createtime = time_second; if ((ifp->if_flags & (IFF_POINTOPOINT | IFF_LOOPBACK)) != 0) { /* * XXX: some functions expect that ifa_dstaddr is not * NULL for p2p interfaces. */ ia->ia_ifa.ifa_dstaddr = (struct sockaddr *)&ia->ia_dstaddr; } else { ia->ia_ifa.ifa_dstaddr = NULL; } ia->ia_ifa.ifa_netmask = (struct sockaddr *)&ia->ia_prefixmask; ia->ia_ifp = ifp; if ((oia = V_in6_ifaddr) != NULL) { for ( ; oia->ia_next; oia = oia->ia_next) continue; oia->ia_next = ia; } else V_in6_ifaddr = ia; ia->ia_ifa.ifa_refcnt = 1; TAILQ_INSERT_TAIL(&ifp->if_addrlist, &ia->ia_ifa, ifa_list); } /* update timestamp */ ia->ia6_updatetime = time_second; /* set prefix mask */ if (ifra->ifra_prefixmask.sin6_len) { /* * We prohibit changing the prefix length of an existing * address, because * + such an operation should be rare in IPv6, and * + the operation would confuse prefix management. */ if (ia->ia_prefixmask.sin6_len && in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL) != plen) { nd6log((LOG_INFO, "in6_update_ifa: the prefix length of an" " existing (%s) address should not be changed\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr))); error = EINVAL; goto unlink; } ia->ia_prefixmask = ifra->ifra_prefixmask; } /* * If a new destination address is specified, scrub the old one and * install the new destination. Note that the interface must be * p2p or loopback (see the check above.) */ if (dst6.sin6_family == AF_INET6 && !IN6_ARE_ADDR_EQUAL(&dst6.sin6_addr, &ia->ia_dstaddr.sin6_addr)) { int e; if ((ia->ia_flags & IFA_ROUTE) != 0 && (e = rtinit(&(ia->ia_ifa), (int)RTM_DELETE, RTF_HOST)) != 0) { nd6log((LOG_ERR, "in6_update_ifa: failed to remove " "a route to the old destination: %s\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr))); /* proceed anyway... */ } else ia->ia_flags &= ~IFA_ROUTE; ia->ia_dstaddr = dst6; } /* * Set lifetimes. We do not refer to ia6t_expire and ia6t_preferred * to see if the address is deprecated or invalidated, but initialize * these members for applications. */ ia->ia6_lifetime = ifra->ifra_lifetime; if (ia->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_expire = time_second + ia->ia6_lifetime.ia6t_vltime; } else ia->ia6_lifetime.ia6t_expire = 0; if (ia->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_preferred = time_second + ia->ia6_lifetime.ia6t_pltime; } else ia->ia6_lifetime.ia6t_preferred = 0; /* reset the interface and routing table appropriately. */ if ((error = in6_ifinit(ifp, ia, &ifra->ifra_addr, hostIsNew)) != 0) goto unlink; /* * configure address flags. */ ia->ia6_flags = ifra->ifra_flags; /* * backward compatibility - if IN6_IFF_DEPRECATED is set from the * userland, make it deprecated. */ if ((ifra->ifra_flags & IN6_IFF_DEPRECATED) != 0) { ia->ia6_lifetime.ia6t_pltime = 0; ia->ia6_lifetime.ia6t_preferred = time_second; } /* * Make the address tentative before joining multicast addresses, * so that corresponding MLD responses would not have a tentative * source address. */ ia->ia6_flags &= ~IN6_IFF_DUPLICATED; /* safety */ if (hostIsNew && in6if_do_dad(ifp)) ia->ia6_flags |= IN6_IFF_TENTATIVE; /* * We are done if we have simply modified an existing address. */ if (!hostIsNew) return (error); /* * Beyond this point, we should call in6_purgeaddr upon an error, * not just go to unlink. */ /* Join necessary multicast groups */ in6m_sol = NULL; if ((ifp->if_flags & IFF_MULTICAST) != 0) { struct sockaddr_in6 mltaddr, mltmask; struct in6_addr llsol; /* join solicited multicast addr for new host id */ bzero(&llsol, sizeof(struct in6_addr)); llsol.s6_addr32[0] = IPV6_ADDR_INT32_MLL; llsol.s6_addr32[1] = 0; llsol.s6_addr32[2] = htonl(1); llsol.s6_addr32[3] = ifra->ifra_addr.sin6_addr.s6_addr32[3]; llsol.s6_addr8[12] = 0xff; if ((error = in6_setscope(&llsol, ifp, NULL)) != 0) { /* XXX: should not happen */ log(LOG_ERR, "in6_update_ifa: " "in6_setscope failed\n"); goto cleanup; } delay = 0; if ((flags & IN6_IFAUPDATE_DADDELAY)) { /* * We need a random delay for DAD on the address * being configured. It also means delaying * transmission of the corresponding MLD report to * avoid report collision. * [draft-ietf-ipv6-rfc2462bis-02.txt] */ delay = arc4random() % (MAX_RTR_SOLICITATION_DELAY * hz); } imm = in6_joingroup(ifp, &llsol, &error, delay); if (imm == NULL) { nd6log((LOG_WARNING, "in6_update_ifa: addmulti failed for " "%s on %s (errno=%d)\n", ip6_sprintf(ip6buf, &llsol), if_name(ifp), error)); in6_purgeaddr((struct ifaddr *)ia); return (error); } LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); in6m_sol = imm->i6mm_maddr; bzero(&mltmask, sizeof(mltmask)); mltmask.sin6_len = sizeof(struct sockaddr_in6); mltmask.sin6_family = AF_INET6; mltmask.sin6_addr = in6mask32; #define MLTMASK_LEN 4 /* mltmask's masklen (=32bit=4octet) */ /* * join link-local all-nodes address */ bzero(&mltaddr, sizeof(mltaddr)); mltaddr.sin6_len = sizeof(struct sockaddr_in6); mltaddr.sin6_family = AF_INET6; mltaddr.sin6_addr = in6addr_linklocal_allnodes; if ((error = in6_setscope(&mltaddr.sin6_addr, ifp, NULL)) != 0) goto cleanup; /* XXX: should not fail */ /* * XXX: do we really need this automatic routes? * We should probably reconsider this stuff. Most applications * actually do not need the routes, since they usually specify * the outgoing interface. */ rt = rtalloc1((struct sockaddr *)&mltaddr, 0, 0UL); if (rt) { /* XXX: only works in !SCOPEDROUTING case. */ if (memcmp(&mltaddr.sin6_addr, &((struct sockaddr_in6 *)rt_key(rt))->sin6_addr, MLTMASK_LEN)) { RTFREE_LOCKED(rt); rt = NULL; } } if (!rt) { error = rtrequest(RTM_ADD, (struct sockaddr *)&mltaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&mltmask, RTF_UP, (struct rtentry **)0); if (error) goto cleanup; } else { RTFREE_LOCKED(rt); } imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, 0); if (!imm) { nd6log((LOG_WARNING, "in6_update_ifa: addmulti failed for " "%s on %s (errno=%d)\n", ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); goto cleanup; } LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); /* * join node information group address */ #define hostnamelen strlen(V_hostname) delay = 0; if ((flags & IN6_IFAUPDATE_DADDELAY)) { /* * The spec doesn't say anything about delay for this * group, but the same logic should apply. */ delay = arc4random() % (MAX_RTR_SOLICITATION_DELAY * hz); } mtx_lock(&hostname_mtx); if (in6_nigroup(ifp, V_hostname, hostnamelen, &mltaddr.sin6_addr) == 0) { mtx_unlock(&hostname_mtx); imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, delay); /* XXX jinmei */ if (!imm) { nd6log((LOG_WARNING, "in6_update_ifa: " "addmulti failed for %s on %s " "(errno=%d)\n", ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); /* XXX not very fatal, go on... */ } else { LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); } } else mtx_unlock(&hostname_mtx); #undef hostnamelen /* * join interface-local all-nodes address. * (ff01::1%ifN, and ff01::%ifN/32) */ mltaddr.sin6_addr = in6addr_nodelocal_allnodes; if ((error = in6_setscope(&mltaddr.sin6_addr, ifp, NULL)) != 0) goto cleanup; /* XXX: should not fail */ /* XXX: again, do we really need the route? */ rt = rtalloc1((struct sockaddr *)&mltaddr, 0, 0UL); if (rt) { if (memcmp(&mltaddr.sin6_addr, &((struct sockaddr_in6 *)rt_key(rt))->sin6_addr, MLTMASK_LEN)) { RTFREE_LOCKED(rt); rt = NULL; } } if (!rt) { error = rtrequest(RTM_ADD, (struct sockaddr *)&mltaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&mltmask, RTF_UP, (struct rtentry **)0); if (error) goto cleanup; } else RTFREE_LOCKED(rt); imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, 0); if (!imm) { nd6log((LOG_WARNING, "in6_update_ifa: " "addmulti failed for %s on %s " "(errno=%d)\n", ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); goto cleanup; } LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); #undef MLTMASK_LEN } /* * Perform DAD, if needed. * XXX It may be of use, if we can administratively * disable DAD. */ if (hostIsNew && in6if_do_dad(ifp) && ((ifra->ifra_flags & IN6_IFF_NODAD) == 0) && (ia->ia6_flags & IN6_IFF_TENTATIVE)) { int mindelay, maxdelay; delay = 0; if ((flags & IN6_IFAUPDATE_DADDELAY)) { /* * We need to impose a delay before sending an NS * for DAD. Check if we also needed a delay for the * corresponding MLD message. If we did, the delay * should be larger than the MLD delay (this could be * relaxed a bit, but this simple logic is at least * safe). */ mindelay = 0; if (in6m_sol != NULL && in6m_sol->in6m_state == MLD_REPORTPENDING) { mindelay = in6m_sol->in6m_timer; } maxdelay = MAX_RTR_SOLICITATION_DELAY * hz; if (maxdelay - mindelay == 0) delay = 0; else { delay = (arc4random() % (maxdelay - mindelay)) + mindelay; } } nd6_dad_start((struct ifaddr *)ia, delay); } return (error); unlink: /* * XXX: if a change of an existing address failed, keep the entry * anyway. */ if (hostIsNew) in6_unlink_ifa(ia, ifp); return (error); cleanup: in6_purgeaddr(&ia->ia_ifa); return error; } void in6_purgeaddr(struct ifaddr *ifa) { struct ifnet *ifp = ifa->ifa_ifp; struct in6_ifaddr *ia = (struct in6_ifaddr *) ifa; struct in6_multi_mship *imm; /* stop DAD processing */ nd6_dad_stop(ifa); IF_AFDATA_LOCK(ifp); lla_lookup(LLTABLE6(ifp), (LLE_DELETE | LLE_IFADDR), (struct sockaddr *)&ia->ia_addr); IF_AFDATA_UNLOCK(ifp); /* * leave from multicast groups we have joined for the interface */ while ((imm = ia->ia6_memberships.lh_first) != NULL) { LIST_REMOVE(imm, i6mm_chain); in6_leavegroup(imm); } in6_unlink_ifa(ia, ifp); } static void in6_unlink_ifa(struct in6_ifaddr *ia, struct ifnet *ifp) { INIT_VNET_INET6(ifp->if_vnet); struct in6_ifaddr *oia; int s = splnet(); TAILQ_REMOVE(&ifp->if_addrlist, &ia->ia_ifa, ifa_list); oia = ia; if (oia == (ia = V_in6_ifaddr)) V_in6_ifaddr = ia->ia_next; else { while (ia->ia_next && (ia->ia_next != oia)) ia = ia->ia_next; if (ia->ia_next) ia->ia_next = oia->ia_next; else { /* search failed */ printf("Couldn't unlink in6_ifaddr from in6_ifaddr\n"); } } /* * Release the reference to the base prefix. There should be a * positive reference. */ if (oia->ia6_ndpr == NULL) { nd6log((LOG_NOTICE, "in6_unlink_ifa: autoconf'ed address " "%p has no prefix\n", oia)); } else { oia->ia6_ndpr->ndpr_refcnt--; oia->ia6_ndpr = NULL; } /* * Also, if the address being removed is autoconf'ed, call * pfxlist_onlink_check() since the release might affect the status of * other (detached) addresses. */ if ((oia->ia6_flags & IN6_IFF_AUTOCONF)) { pfxlist_onlink_check(); } /* * release another refcnt for the link from in6_ifaddr. * Note that we should decrement the refcnt at least once for all *BSD. */ IFAFREE(&oia->ia_ifa); splx(s); } void in6_purgeif(struct ifnet *ifp) { struct ifaddr *ifa, *nifa; for (ifa = TAILQ_FIRST(&ifp->if_addrlist); ifa != NULL; ifa = nifa) { nifa = TAILQ_NEXT(ifa, ifa_list); if (ifa->ifa_addr->sa_family != AF_INET6) continue; in6_purgeaddr(ifa); } in6_ifdetach(ifp); } /* * SIOC[GAD]LIFADDR. * SIOCGLIFADDR: get first address. (?) * SIOCGLIFADDR with IFLR_PREFIX: * get first address that matches the specified prefix. * SIOCALIFADDR: add the specified address. * SIOCALIFADDR with IFLR_PREFIX: * add the specified prefix, filling hostid part from * the first link-local address. prefixlen must be <= 64. * SIOCDLIFADDR: delete the specified address. * SIOCDLIFADDR with IFLR_PREFIX: * delete the first address that matches the specified prefix. * return values: * EINVAL on invalid parameters * EADDRNOTAVAIL on prefix match failed/specified address not found * other values may be returned from in6_ioctl() * * NOTE: SIOCALIFADDR(with IFLR_PREFIX set) allows prefixlen less than 64. * this is to accomodate address naming scheme other than RFC2374, * in the future. * RFC2373 defines interface id to be 64bit, but it allows non-RFC2374 * address encoding scheme. (see figure on page 8) */ static int in6_lifaddr_ioctl(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp, struct thread *td) { struct if_laddrreq *iflr = (struct if_laddrreq *)data; struct ifaddr *ifa; struct sockaddr *sa; /* sanity checks */ if (!data || !ifp) { panic("invalid argument to in6_lifaddr_ioctl"); /* NOTREACHED */ } switch (cmd) { case SIOCGLIFADDR: /* address must be specified on GET with IFLR_PREFIX */ if ((iflr->flags & IFLR_PREFIX) == 0) break; /* FALLTHROUGH */ case SIOCALIFADDR: case SIOCDLIFADDR: /* address must be specified on ADD and DELETE */ sa = (struct sockaddr *)&iflr->addr; if (sa->sa_family != AF_INET6) return EINVAL; if (sa->sa_len != sizeof(struct sockaddr_in6)) return EINVAL; /* XXX need improvement */ sa = (struct sockaddr *)&iflr->dstaddr; if (sa->sa_family && sa->sa_family != AF_INET6) return EINVAL; if (sa->sa_len && sa->sa_len != sizeof(struct sockaddr_in6)) return EINVAL; break; default: /* shouldn't happen */ #if 0 panic("invalid cmd to in6_lifaddr_ioctl"); /* NOTREACHED */ #else return EOPNOTSUPP; #endif } if (sizeof(struct in6_addr) * 8 < iflr->prefixlen) return EINVAL; switch (cmd) { case SIOCALIFADDR: { struct in6_aliasreq ifra; struct in6_addr *hostid = NULL; int prefixlen; if ((iflr->flags & IFLR_PREFIX) != 0) { struct sockaddr_in6 *sin6; /* * hostid is to fill in the hostid part of the * address. hostid points to the first link-local * address attached to the interface. */ ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, 0); if (!ifa) return EADDRNOTAVAIL; hostid = IFA_IN6(ifa); /* prefixlen must be <= 64. */ if (64 < iflr->prefixlen) return EINVAL; prefixlen = iflr->prefixlen; /* hostid part must be zero. */ sin6 = (struct sockaddr_in6 *)&iflr->addr; if (sin6->sin6_addr.s6_addr32[2] != 0 || sin6->sin6_addr.s6_addr32[3] != 0) { return EINVAL; } } else prefixlen = iflr->prefixlen; /* copy args to in6_aliasreq, perform ioctl(SIOCAIFADDR_IN6). */ bzero(&ifra, sizeof(ifra)); bcopy(iflr->iflr_name, ifra.ifra_name, sizeof(ifra.ifra_name)); bcopy(&iflr->addr, &ifra.ifra_addr, ((struct sockaddr *)&iflr->addr)->sa_len); if (hostid) { /* fill in hostid part */ ifra.ifra_addr.sin6_addr.s6_addr32[2] = hostid->s6_addr32[2]; ifra.ifra_addr.sin6_addr.s6_addr32[3] = hostid->s6_addr32[3]; } if (((struct sockaddr *)&iflr->dstaddr)->sa_family) { /* XXX */ bcopy(&iflr->dstaddr, &ifra.ifra_dstaddr, ((struct sockaddr *)&iflr->dstaddr)->sa_len); if (hostid) { ifra.ifra_dstaddr.sin6_addr.s6_addr32[2] = hostid->s6_addr32[2]; ifra.ifra_dstaddr.sin6_addr.s6_addr32[3] = hostid->s6_addr32[3]; } } ifra.ifra_prefixmask.sin6_len = sizeof(struct sockaddr_in6); in6_prefixlen2mask(&ifra.ifra_prefixmask.sin6_addr, prefixlen); ifra.ifra_flags = iflr->flags & ~IFLR_PREFIX; return in6_control(so, SIOCAIFADDR_IN6, (caddr_t)&ifra, ifp, td); } case SIOCGLIFADDR: case SIOCDLIFADDR: { struct in6_ifaddr *ia; struct in6_addr mask, candidate, match; struct sockaddr_in6 *sin6; int cmp; bzero(&mask, sizeof(mask)); if (iflr->flags & IFLR_PREFIX) { /* lookup a prefix rather than address. */ in6_prefixlen2mask(&mask, iflr->prefixlen); sin6 = (struct sockaddr_in6 *)&iflr->addr; bcopy(&sin6->sin6_addr, &match, sizeof(match)); match.s6_addr32[0] &= mask.s6_addr32[0]; match.s6_addr32[1] &= mask.s6_addr32[1]; match.s6_addr32[2] &= mask.s6_addr32[2]; match.s6_addr32[3] &= mask.s6_addr32[3]; /* if you set extra bits, that's wrong */ if (bcmp(&match, &sin6->sin6_addr, sizeof(match))) return EINVAL; cmp = 1; } else { if (cmd == SIOCGLIFADDR) { /* on getting an address, take the 1st match */ cmp = 0; /* XXX */ } else { /* on deleting an address, do exact match */ in6_prefixlen2mask(&mask, 128); sin6 = (struct sockaddr_in6 *)&iflr->addr; bcopy(&sin6->sin6_addr, &match, sizeof(match)); cmp = 1; } } TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (!cmp) break; /* * XXX: this is adhoc, but is necessary to allow * a user to specify fe80::/64 (not /10) for a * link-local address. */ bcopy(IFA_IN6(ifa), &candidate, sizeof(candidate)); in6_clearscope(&candidate); candidate.s6_addr32[0] &= mask.s6_addr32[0]; candidate.s6_addr32[1] &= mask.s6_addr32[1]; candidate.s6_addr32[2] &= mask.s6_addr32[2]; candidate.s6_addr32[3] &= mask.s6_addr32[3]; if (IN6_ARE_ADDR_EQUAL(&candidate, &match)) break; } if (!ifa) return EADDRNOTAVAIL; ia = ifa2ia6(ifa); if (cmd == SIOCGLIFADDR) { int error; /* fill in the if_laddrreq structure */ bcopy(&ia->ia_addr, &iflr->addr, ia->ia_addr.sin6_len); error = sa6_recoverscope( (struct sockaddr_in6 *)&iflr->addr); if (error != 0) return (error); if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { bcopy(&ia->ia_dstaddr, &iflr->dstaddr, ia->ia_dstaddr.sin6_len); error = sa6_recoverscope( (struct sockaddr_in6 *)&iflr->dstaddr); if (error != 0) return (error); } else bzero(&iflr->dstaddr, sizeof(iflr->dstaddr)); iflr->prefixlen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); iflr->flags = ia->ia6_flags; /* XXX */ return 0; } else { struct in6_aliasreq ifra; /* fill in6_aliasreq and do ioctl(SIOCDIFADDR_IN6) */ bzero(&ifra, sizeof(ifra)); bcopy(iflr->iflr_name, ifra.ifra_name, sizeof(ifra.ifra_name)); bcopy(&ia->ia_addr, &ifra.ifra_addr, ia->ia_addr.sin6_len); if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { bcopy(&ia->ia_dstaddr, &ifra.ifra_dstaddr, ia->ia_dstaddr.sin6_len); } else { bzero(&ifra.ifra_dstaddr, sizeof(ifra.ifra_dstaddr)); } bcopy(&ia->ia_prefixmask, &ifra.ifra_dstaddr, ia->ia_prefixmask.sin6_len); ifra.ifra_flags = ia->ia6_flags; return in6_control(so, SIOCDIFADDR_IN6, (caddr_t)&ifra, ifp, td); } } } return EOPNOTSUPP; /* just for safety */ } /* * Initialize an interface's intetnet6 address * and routing table entry. */ static int in6_ifinit(struct ifnet *ifp, struct in6_ifaddr *ia, struct sockaddr_in6 *sin6, int newhost) { int error = 0, plen, ifacount = 0; int s = splimp(); struct ifaddr *ifa; /* * Give the interface a chance to initialize * if this is its first address, * and to validate the address if necessary. */ TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ifacount++; } ia->ia_addr = *sin6; if (ifacount <= 1 && ifp->if_ioctl) { IFF_LOCKGIANT(ifp); error = (*ifp->if_ioctl)(ifp, SIOCSIFADDR, (caddr_t)ia); IFF_UNLOCKGIANT(ifp); if (error) { splx(s); return (error); } } splx(s); ia->ia_ifa.ifa_metric = ifp->if_metric; /* we could do in(6)_socktrim here, but just omit it at this moment. */ /* * Special case: * If a new destination address is specified for a point-to-point * interface, install a route to the destination as an interface * direct route. * XXX: the logic below rejects assigning multiple addresses on a p2p * interface that share the same destination. */ plen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); /* XXX */ if (!(ia->ia_flags & IFA_ROUTE) && plen == 128) { struct sockaddr *dstaddr; int rtflags = RTF_UP | RTF_HOST; /* * use the interface address if configuring an * interface address with a /128 prefix len */ if (ia->ia_dstaddr.sin6_family == AF_INET6) dstaddr = (struct sockaddr *)&ia->ia_dstaddr; else dstaddr = (struct sockaddr *)&ia->ia_addr; error = rtrequest(RTM_ADD, (struct sockaddr *)dstaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&ia->ia_prefixmask, ia->ia_flags | rtflags, NULL); if (error != 0) return (error); ia->ia_flags |= IFA_ROUTE; } /* Add ownaddr as loopback rtentry, if necessary (ex. on p2p link). */ if (newhost) { struct llentry *ln; IF_AFDATA_LOCK(ifp); ia->ia_ifa.ifa_rtrequest = NULL; /* XXX QL * we need to report rt_newaddrmsg */ ln = lla_lookup(LLTABLE6(ifp), (LLE_CREATE | LLE_IFADDR | LLE_EXCLUSIVE), (struct sockaddr *)&ia->ia_addr); IF_AFDATA_UNLOCK(ifp); if (ln != NULL) { ln->la_expire = 0; /* for IPv6 this means permanent */ ln->ln_state = ND6_LLINFO_REACHABLE; LLE_WUNLOCK(ln); } } return (error); } struct in6_multi_mship * in6_joingroup(struct ifnet *ifp, struct in6_addr *addr, int *errorp, int delay) { struct in6_multi_mship *imm; imm = malloc(sizeof(*imm), M_IP6MADDR, M_NOWAIT); if (!imm) { *errorp = ENOBUFS; return NULL; } imm->i6mm_maddr = in6_addmulti(addr, ifp, errorp, delay); if (!imm->i6mm_maddr) { /* *errorp is alrady set */ free(imm, M_IP6MADDR); return NULL; } return imm; } int in6_leavegroup(struct in6_multi_mship *imm) { if (imm->i6mm_maddr) in6_delmulti(imm->i6mm_maddr); free(imm, M_IP6MADDR); return 0; } /* * Find an IPv6 interface link-local address specific to an interface. */ struct in6_ifaddr * in6ifa_ifpforlinklocal(struct ifnet *ifp, int ignoreflags) { struct ifaddr *ifa; TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (IN6_IS_ADDR_LINKLOCAL(IFA_IN6(ifa))) { if ((((struct in6_ifaddr *)ifa)->ia6_flags & ignoreflags) != 0) continue; break; } } return ((struct in6_ifaddr *)ifa); } /* * find the internet address corresponding to a given interface and address. */ struct in6_ifaddr * in6ifa_ifpwithaddr(struct ifnet *ifp, struct in6_addr *addr) { struct ifaddr *ifa; TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (IN6_ARE_ADDR_EQUAL(addr, IFA_IN6(ifa))) break; } return ((struct in6_ifaddr *)ifa); } /* * Convert IP6 address to printable (loggable) representation. Caller * has to make sure that ip6buf is at least INET6_ADDRSTRLEN long. */ static char digits[] = "0123456789abcdef"; char * ip6_sprintf(char *ip6buf, const struct in6_addr *addr) { int i; char *cp; const u_int16_t *a = (const u_int16_t *)addr; const u_int8_t *d; int dcolon = 0, zero = 0; cp = ip6buf; for (i = 0; i < 8; i++) { if (dcolon == 1) { if (*a == 0) { if (i == 7) *cp++ = ':'; a++; continue; } else dcolon = 2; } if (*a == 0) { if (dcolon == 0 && *(a + 1) == 0) { if (i == 0) *cp++ = ':'; *cp++ = ':'; dcolon = 1; } else { *cp++ = '0'; *cp++ = ':'; } a++; continue; } d = (const u_char *)a; /* Try to eliminate leading zeros in printout like in :0001. */ zero = 1; *cp = digits[*d >> 4]; if (*cp != '0') { zero = 0; cp++; } *cp = digits[*d++ & 0xf]; if (zero == 0 || (*cp != '0')) { zero = 0; cp++; } *cp = digits[*d >> 4]; if (zero == 0 || (*cp != '0')) { zero = 0; cp++; } *cp++ = digits[*d & 0xf]; *cp++ = ':'; a++; } *--cp = '\0'; return (ip6buf); } int in6_localaddr(struct in6_addr *in6) { INIT_VNET_INET6(curvnet); struct in6_ifaddr *ia; if (IN6_IS_ADDR_LOOPBACK(in6) || IN6_IS_ADDR_LINKLOCAL(in6)) return 1; for (ia = V_in6_ifaddr; ia; ia = ia->ia_next) { if (IN6_ARE_MASKED_ADDR_EQUAL(in6, &ia->ia_addr.sin6_addr, &ia->ia_prefixmask.sin6_addr)) { return 1; } } return (0); } int in6_is_addr_deprecated(struct sockaddr_in6 *sa6) { INIT_VNET_INET6(curvnet); struct in6_ifaddr *ia; for (ia = V_in6_ifaddr; ia; ia = ia->ia_next) { if (IN6_ARE_ADDR_EQUAL(&ia->ia_addr.sin6_addr, &sa6->sin6_addr) && (ia->ia6_flags & IN6_IFF_DEPRECATED) != 0) return (1); /* true */ /* XXX: do we still have to go thru the rest of the list? */ } return (0); /* false */ } /* * return length of part which dst and src are equal * hard coding... */ int in6_matchlen(struct in6_addr *src, struct in6_addr *dst) { int match = 0; u_char *s = (u_char *)src, *d = (u_char *)dst; u_char *lim = s + 16, r; while (s < lim) if ((r = (*d++ ^ *s++)) != 0) { while (r < 128) { match++; r <<= 1; } break; } else match += 8; return match; } /* XXX: to be scope conscious */ int in6_are_prefix_equal(struct in6_addr *p1, struct in6_addr *p2, int len) { int bytelen, bitlen; /* sanity check */ if (0 > len || len > 128) { log(LOG_ERR, "in6_are_prefix_equal: invalid prefix length(%d)\n", len); return (0); } bytelen = len / 8; bitlen = len % 8; if (bcmp(&p1->s6_addr, &p2->s6_addr, bytelen)) return (0); if (bitlen != 0 && p1->s6_addr[bytelen] >> (8 - bitlen) != p2->s6_addr[bytelen] >> (8 - bitlen)) return (0); return (1); } void in6_prefixlen2mask(struct in6_addr *maskp, int len) { u_char maskarray[8] = {0x80, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc, 0xfe, 0xff}; int bytelen, bitlen, i; /* sanity check */ if (0 > len || len > 128) { log(LOG_ERR, "in6_prefixlen2mask: invalid prefix length(%d)\n", len); return; } bzero(maskp, sizeof(*maskp)); bytelen = len / 8; bitlen = len % 8; for (i = 0; i < bytelen; i++) maskp->s6_addr[i] = 0xff; if (bitlen) maskp->s6_addr[bytelen] = maskarray[bitlen - 1]; } /* * return the best address out of the same scope. if no address was * found, return the first valid address from designated IF. */ struct in6_ifaddr * in6_ifawithifp(struct ifnet *ifp, struct in6_addr *dst) { INIT_VNET_INET6(curvnet); int dst_scope = in6_addrscope(dst), blen = -1, tlen; struct ifaddr *ifa; struct in6_ifaddr *besta = 0; struct in6_ifaddr *dep[2]; /* last-resort: deprecated */ dep[0] = dep[1] = NULL; /* * We first look for addresses in the same scope. * If there is one, return it. * If two or more, return one which matches the dst longest. * If none, return one of global addresses assigned other ifs. */ TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST) continue; /* XXX: is there any case to allow anycast? */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_NOTREADY) continue; /* don't use this interface */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DETACHED) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DEPRECATED) { if (V_ip6_use_deprecated) dep[0] = (struct in6_ifaddr *)ifa; continue; } if (dst_scope == in6_addrscope(IFA_IN6(ifa))) { /* * call in6_matchlen() as few as possible */ if (besta) { if (blen == -1) blen = in6_matchlen(&besta->ia_addr.sin6_addr, dst); tlen = in6_matchlen(IFA_IN6(ifa), dst); if (tlen > blen) { blen = tlen; besta = (struct in6_ifaddr *)ifa; } } else besta = (struct in6_ifaddr *)ifa; } } if (besta) return (besta); TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST) continue; /* XXX: is there any case to allow anycast? */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_NOTREADY) continue; /* don't use this interface */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DETACHED) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DEPRECATED) { if (V_ip6_use_deprecated) dep[1] = (struct in6_ifaddr *)ifa; continue; } return (struct in6_ifaddr *)ifa; } /* use the last-resort values, that are, deprecated addresses */ if (dep[0]) return dep[0]; if (dep[1]) return dep[1]; return NULL; } /* * perform DAD when interface becomes IFF_UP. */ void in6_if_up(struct ifnet *ifp) { struct ifaddr *ifa; struct in6_ifaddr *ia; TAILQ_FOREACH(ifa, &ifp->if_addrlist, ifa_list) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; if (ia->ia6_flags & IN6_IFF_TENTATIVE) { /* * The TENTATIVE flag was likely set by hand * beforehand, implicitly indicating the need for DAD. * We may be able to skip the random delay in this * case, but we impose delays just in case. */ nd6_dad_start(ifa, arc4random() % (MAX_RTR_SOLICITATION_DELAY * hz)); } } /* * special cases, like 6to4, are handled in in6_ifattach */ in6_ifattach(ifp, NULL); } int in6if_do_dad(struct ifnet *ifp) { if ((ifp->if_flags & IFF_LOOPBACK) != 0) return (0); switch (ifp->if_type) { #ifdef IFT_DUMMY case IFT_DUMMY: #endif case IFT_FAITH: /* * These interfaces do not have the IFF_LOOPBACK flag, * but loop packets back. We do not have to do DAD on such * interfaces. We should even omit it, because loop-backed * NS would confuse the DAD procedure. */ return (0); default: /* * Our DAD routine requires the interface up and running. * However, some interfaces can be up before the RUNNING * status. Additionaly, users may try to assign addresses * before the interface becomes up (or running). * We simply skip DAD in such a case as a work around. * XXX: we should rather mark "tentative" on such addresses, * and do DAD after the interface becomes ready. */ if (!((ifp->if_flags & IFF_UP) && (ifp->if_drv_flags & IFF_DRV_RUNNING))) return (0); return (1); } } /* * Calculate max IPv6 MTU through all the interfaces and store it * to in6_maxmtu. */ void in6_setmaxmtu(void) { INIT_VNET_NET(curvnet); INIT_VNET_INET6(curvnet); unsigned long maxmtu = 0; struct ifnet *ifp; IFNET_RLOCK(); for (ifp = TAILQ_FIRST(&V_ifnet); ifp; ifp = TAILQ_NEXT(ifp, if_list)) { /* this function can be called during ifnet initialization */ if (!ifp->if_afdata[AF_INET6]) continue; if ((ifp->if_flags & IFF_LOOPBACK) == 0 && IN6_LINKMTU(ifp) > maxmtu) maxmtu = IN6_LINKMTU(ifp); } IFNET_RUNLOCK(); if (maxmtu) /* update only when maxmtu is positive */ V_in6_maxmtu = maxmtu; } /* * Provide the length of interface identifiers to be used for the link attached * to the given interface. The length should be defined in "IPv6 over * xxx-link" document. Note that address architecture might also define * the length for a particular set of address prefixes, regardless of the * link type. As clarified in rfc2462bis, those two definitions should be * consistent, and those really are as of August 2004. */ int in6_if2idlen(struct ifnet *ifp) { switch (ifp->if_type) { case IFT_ETHER: /* RFC2464 */ #ifdef IFT_PROPVIRTUAL case IFT_PROPVIRTUAL: /* XXX: no RFC. treat it as ether */ #endif #ifdef IFT_L2VLAN case IFT_L2VLAN: /* ditto */ #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: /* ditto */ #endif #ifdef IFT_MIP case IFT_MIP: /* ditto */ #endif return (64); case IFT_FDDI: /* RFC2467 */ return (64); case IFT_ISO88025: /* RFC2470 (IPv6 over Token Ring) */ return (64); case IFT_PPP: /* RFC2472 */ return (64); case IFT_ARCNET: /* RFC2497 */ return (64); case IFT_FRELAY: /* RFC2590 */ return (64); case IFT_IEEE1394: /* RFC3146 */ return (64); case IFT_GIF: return (64); /* draft-ietf-v6ops-mech-v2-07 */ case IFT_LOOP: return (64); /* XXX: is this really correct? */ default: /* * Unknown link type: * It might be controversial to use the today's common constant * of 64 for these cases unconditionally. For full compliance, * we should return an error in this case. On the other hand, * if we simply miss the standard for the link type or a new * standard is defined for a new link type, the IFID length * is very likely to be the common constant. As a compromise, * we always use the constant, but make an explicit notice * indicating the "unknown" case. */ printf("in6_if2idlen: unknown link type (%d)\n", ifp->if_type); return (64); } } #include struct in6_llentry { struct llentry base; struct sockaddr_in6 l3_addr6; }; static struct llentry * in6_lltable_new(const struct sockaddr *l3addr, u_int flags) { struct in6_llentry *lle; lle = malloc(sizeof(struct in6_llentry), M_LLTABLE, M_DONTWAIT | M_ZERO); if (lle == NULL) /* NB: caller generates msg */ return NULL; callout_init(&lle->base.ln_timer_ch, CALLOUT_MPSAFE); lle->l3_addr6 = *(const struct sockaddr_in6 *)l3addr; lle->base.lle_refcnt = 1; LLE_LOCK_INIT(&lle->base); return &lle->base; } /* * Deletes an address from the address table. * This function is called by the timer functions * such as arptimer() and nd6_llinfo_timer(), and * the caller does the locking. */ static void in6_lltable_free(struct lltable *llt, struct llentry *lle) { LLE_WUNLOCK(lle); LLE_LOCK_DESTROY(lle); free(lle, M_LLTABLE); } static int in6_lltable_rtcheck(struct ifnet *ifp, const struct sockaddr *l3addr) { struct rtentry *rt; char ip6buf[INET6_ADDRSTRLEN]; KASSERT(l3addr->sa_family == AF_INET6, ("sin_family %d", l3addr->sa_family)); /* XXX rtalloc1 should take a const param */ rt = rtalloc1(__DECONST(struct sockaddr *, l3addr), 0, 0); if (rt == NULL || (rt->rt_flags & RTF_GATEWAY) || rt->rt_ifp != ifp) { struct ifaddr *ifa; /* * Create an ND6 cache for an IPv6 neighbor * that is not covered by our own prefix. */ /* XXX ifaof_ifpforaddr should take a const param */ ifa = ifaof_ifpforaddr(__DECONST(struct sockaddr *, l3addr), ifp); if (ifa != NULL) { if (rt != NULL) RTFREE_LOCKED(rt); return 0; } log(LOG_INFO, "IPv6 address: \"%s\" is not on the network\n", ip6_sprintf(ip6buf, &((const struct sockaddr_in6 *)l3addr)->sin6_addr)); if (rt != NULL) RTFREE_LOCKED(rt); return EINVAL; } RTFREE_LOCKED(rt); return 0; } static struct llentry * in6_lltable_lookup(struct lltable *llt, u_int flags, const struct sockaddr *l3addr) { const struct sockaddr_in6 *sin6 = (const struct sockaddr_in6 *)l3addr; struct ifnet *ifp = llt->llt_ifp; struct llentry *lle; struct llentries *lleh; u_int hashkey; IF_AFDATA_LOCK_ASSERT(ifp); KASSERT(l3addr->sa_family == AF_INET6, ("sin_family %d", l3addr->sa_family)); hashkey = sin6->sin6_addr.s6_addr32[3]; lleh = &llt->lle_head[LLATBL_HASH(hashkey, LLTBL_HASHMASK)]; LIST_FOREACH(lle, lleh, lle_next) { struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)L3_ADDR(lle); if (lle->la_flags & LLE_DELETED) continue; if (bcmp(&sa6->sin6_addr, &sin6->sin6_addr, sizeof(struct in6_addr)) == 0) break; } if (lle == NULL) { if (!(flags & LLE_CREATE)) return (NULL); /* * A route that covers the given address must have * been installed 1st because we are doing a resolution, * verify this. */ if (!(flags & LLE_IFADDR) && in6_lltable_rtcheck(ifp, l3addr) != 0) return NULL; lle = in6_lltable_new(l3addr, flags); if (lle == NULL) { log(LOG_INFO, "lla_lookup: new lle malloc failed\n"); return NULL; } lle->la_flags = flags & ~LLE_CREATE; if ((flags & (LLE_CREATE | LLE_IFADDR)) == (LLE_CREATE | LLE_IFADDR)) { bcopy(IF_LLADDR(ifp), &lle->ll_addr, ifp->if_addrlen); lle->la_flags |= (LLE_VALID | LLE_STATIC); } lle->lle_tbl = llt; lle->lle_head = lleh; LIST_INSERT_HEAD(lleh, lle, lle_next); } else if (flags & LLE_DELETE) { if (!(lle->la_flags & LLE_IFADDR) || (flags & LLE_IFADDR)) { LLE_WLOCK(lle); lle->la_flags = LLE_DELETED; LLE_WUNLOCK(lle); #ifdef DIAGNOSTICS log(LOG_INFO, "ifaddr cache = %p is deleted\n", lle); #endif } lle = (void *)-1; } if (LLE_IS_VALID(lle)) { if (flags & LLE_EXCLUSIVE) LLE_WLOCK(lle); else LLE_RLOCK(lle); } return (lle); } static int in6_lltable_dump(struct lltable *llt, struct sysctl_req *wr) { struct ifnet *ifp = llt->llt_ifp; struct llentry *lle; /* XXX stack use */ struct { struct rt_msghdr rtm; struct sockaddr_in6 sin6; /* * ndp.c assumes that sdl is word aligned */ #ifdef __LP64__ uint32_t pad; #endif struct sockaddr_dl sdl; } ndpc; int i, error; /* XXXXX * current IFNET_RLOCK() is mapped to IFNET_WLOCK() * so it is okay to use this ASSERT, change it when * IFNET lock is finalized */ IFNET_WLOCK_ASSERT(); error = 0; for (i = 0; i < LLTBL_HASHTBL_SIZE; i++) { LIST_FOREACH(lle, &llt->lle_head[i], lle_next) { struct sockaddr_dl *sdl; /* skip deleted or invalid entries */ if ((lle->la_flags & (LLE_DELETED|LLE_VALID)) != LLE_VALID) continue; /* Skip if jailed and not a valid IP of the prison. */ - if (jailed(wr->td->td_ucred) && - !prison_if(wr->td->td_ucred, L3_ADDR(lle))) + if (prison_if(wr->td->td_ucred, L3_ADDR(lle)) != 0) continue; /* * produce a msg made of: * struct rt_msghdr; * struct sockaddr_in6 (IPv6) * struct sockaddr_dl; */ bzero(&ndpc, sizeof(ndpc)); ndpc.rtm.rtm_msglen = sizeof(ndpc); ndpc.rtm.rtm_version = RTM_VERSION; ndpc.rtm.rtm_type = RTM_GET; ndpc.rtm.rtm_flags = RTF_UP; ndpc.rtm.rtm_addrs = RTA_DST | RTA_GATEWAY; ndpc.sin6.sin6_family = AF_INET6; ndpc.sin6.sin6_len = sizeof(ndpc.sin6); bcopy(L3_ADDR(lle), &ndpc.sin6, L3_ADDR_LEN(lle)); /* publish */ if (lle->la_flags & LLE_PUB) ndpc.rtm.rtm_flags |= RTF_ANNOUNCE; sdl = &ndpc.sdl; sdl->sdl_family = AF_LINK; sdl->sdl_len = sizeof(*sdl); sdl->sdl_alen = ifp->if_addrlen; sdl->sdl_index = ifp->if_index; sdl->sdl_type = ifp->if_type; bcopy(&lle->ll_addr, LLADDR(sdl), ifp->if_addrlen); ndpc.rtm.rtm_rmx.rmx_expire = lle->la_flags & LLE_STATIC ? 0 : lle->la_expire; ndpc.rtm.rtm_flags |= (RTF_HOST | RTF_LLDATA); if (lle->la_flags & LLE_STATIC) ndpc.rtm.rtm_flags |= RTF_STATIC; ndpc.rtm.rtm_index = ifp->if_index; error = SYSCTL_OUT(wr, &ndpc, sizeof(ndpc)); if (error) break; } } return error; } void * in6_domifattach(struct ifnet *ifp) { struct in6_ifextra *ext; ext = (struct in6_ifextra *)malloc(sizeof(*ext), M_IFADDR, M_WAITOK); bzero(ext, sizeof(*ext)); ext->in6_ifstat = (struct in6_ifstat *)malloc(sizeof(struct in6_ifstat), M_IFADDR, M_WAITOK); bzero(ext->in6_ifstat, sizeof(*ext->in6_ifstat)); ext->icmp6_ifstat = (struct icmp6_ifstat *)malloc(sizeof(struct icmp6_ifstat), M_IFADDR, M_WAITOK); bzero(ext->icmp6_ifstat, sizeof(*ext->icmp6_ifstat)); ext->nd_ifinfo = nd6_ifattach(ifp); ext->scope6_id = scope6_ifattach(ifp); ext->lltable = lltable_init(ifp, AF_INET6); if (ext->lltable != NULL) { ext->lltable->llt_new = in6_lltable_new; ext->lltable->llt_free = in6_lltable_free; ext->lltable->llt_rtcheck = in6_lltable_rtcheck; ext->lltable->llt_lookup = in6_lltable_lookup; ext->lltable->llt_dump = in6_lltable_dump; } return ext; } void in6_domifdetach(struct ifnet *ifp, void *aux) { struct in6_ifextra *ext = (struct in6_ifextra *)aux; scope6_ifdetach(ext->scope6_id); nd6_ifdetach(ext->nd_ifinfo); lltable_free(ext->lltable); free(ext->in6_ifstat, M_IFADDR); free(ext->icmp6_ifstat, M_IFADDR); free(ext, M_IFADDR); } /* * Convert sockaddr_in6 to sockaddr_in. Original sockaddr_in6 must be * v4 mapped addr or v4 compat addr */ void in6_sin6_2_sin(struct sockaddr_in *sin, struct sockaddr_in6 *sin6) { bzero(sin, sizeof(*sin)); sin->sin_len = sizeof(struct sockaddr_in); sin->sin_family = AF_INET; sin->sin_port = sin6->sin6_port; sin->sin_addr.s_addr = sin6->sin6_addr.s6_addr32[3]; } /* Convert sockaddr_in to sockaddr_in6 in v4 mapped addr format. */ void in6_sin_2_v4mapsin6(struct sockaddr_in *sin, struct sockaddr_in6 *sin6) { bzero(sin6, sizeof(*sin6)); sin6->sin6_len = sizeof(struct sockaddr_in6); sin6->sin6_family = AF_INET6; sin6->sin6_port = sin->sin_port; sin6->sin6_addr.s6_addr32[0] = 0; sin6->sin6_addr.s6_addr32[1] = 0; sin6->sin6_addr.s6_addr32[2] = IPV6_ADDR_INT32_SMP; sin6->sin6_addr.s6_addr32[3] = sin->sin_addr.s_addr; } /* Convert sockaddr_in6 into sockaddr_in. */ void in6_sin6_2_sin_in_sock(struct sockaddr *nam) { struct sockaddr_in *sin_p; struct sockaddr_in6 sin6; /* * Save original sockaddr_in6 addr and convert it * to sockaddr_in. */ sin6 = *(struct sockaddr_in6 *)nam; sin_p = (struct sockaddr_in *)nam; in6_sin6_2_sin(sin_p, &sin6); } /* Convert sockaddr_in into sockaddr_in6 in v4 mapped addr format. */ void in6_sin_2_v4mapsin6_in_sock(struct sockaddr **nam) { struct sockaddr_in *sin_p; struct sockaddr_in6 *sin6_p; sin6_p = malloc(sizeof *sin6_p, M_SONAME, M_WAITOK); sin_p = (struct sockaddr_in *)*nam; in6_sin_2_v4mapsin6(sin_p, sin6_p); free(*nam, M_SONAME); *nam = (struct sockaddr *)sin6_p; } diff --git a/sys/netinet6/in6_pcb.c b/sys/netinet6/in6_pcb.c index 753f45df0e0c..13c55cdeb837 100644 --- a/sys/netinet6/in6_pcb.c +++ b/sys/netinet6/in6_pcb.c @@ -1,936 +1,934 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 3. Neither the name of the project 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 PROJECT 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 PROJECT 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: in6_pcb.c,v 1.31 2001/05/21 05:45:10 jinmei Exp $ */ /*- * Copyright (c) 1982, 1986, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_mac.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include struct in6_addr zeroin6_addr; int in6_pcbbind(register struct inpcb *inp, struct sockaddr *nam, struct ucred *cred) { INIT_VNET_INET6(inp->inp_vnet); INIT_VNET_INET(inp->inp_vnet); struct socket *so = inp->inp_socket; struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)NULL; struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; u_short lport = 0; - int wild = 0, reuseport = (so->so_options & SO_REUSEPORT); + int error, wild = 0, reuseport = (so->so_options & SO_REUSEPORT); INP_INFO_WLOCK_ASSERT(pcbinfo); INP_WLOCK_ASSERT(inp); if (!V_in6_ifaddr) /* XXX broken! */ return (EADDRNOTAVAIL); if (inp->inp_lport || !IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) return (EINVAL); if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) wild = INPLOOKUP_WILDCARD; if (nam) { - int error; - sin6 = (struct sockaddr_in6 *)nam; if (nam->sa_len != sizeof(*sin6)) return (EINVAL); /* * family check. */ if (nam->sa_family != AF_INET6) return (EAFNOSUPPORT); if ((error = sa6_embedscope(sin6, V_ip6_use_defzone)) != 0) return(error); - if (prison_local_ip6(cred, &sin6->sin6_addr, - ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) - return (EINVAL); + if ((error = prison_local_ip6(cred, &sin6->sin6_addr, + ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0) + return (error); lport = sin6->sin6_port; if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) { /* * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; * allow compepte duplication of binding if * SO_REUSEPORT is set, or if SO_REUSEADDR is set * and a multicast address is bound on both * new and duplicated sockets. */ if (so->so_options & SO_REUSEADDR) reuseport = SO_REUSEADDR|SO_REUSEPORT; } else if (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) { struct ifaddr *ia = NULL; sin6->sin6_port = 0; /* yech... */ if ((ia = ifa_ifwithaddr((struct sockaddr *)sin6)) == 0) return (EADDRNOTAVAIL); /* * XXX: bind to an anycast address might accidentally * cause sending a packet with anycast source address. * We should allow to bind to a deprecated address, since * the application dares to use it. */ if (ia && ((struct in6_ifaddr *)ia)->ia6_flags & (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY|IN6_IFF_DETACHED)) { return (EADDRNOTAVAIL); } } if (lport) { struct inpcb *t; /* GROSS */ if (ntohs(lport) <= V_ipport_reservedhigh && ntohs(lport) >= V_ipport_reservedlow && priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT, 0)) return (EACCES); if (!IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr) && priv_check_cred(inp->inp_cred, PRIV_NETINET_REUSEPORT, 0) != 0) { t = in6_pcblookup_local(pcbinfo, &sin6->sin6_addr, lport, INPLOOKUP_WILDCARD, cred); if (t && ((t->inp_vflag & INP_TIMEWAIT) == 0) && (so->so_type != SOCK_STREAM || IN6_IS_ADDR_UNSPECIFIED(&t->in6p_faddr)) && (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr) || !IN6_IS_ADDR_UNSPECIFIED(&t->in6p_laddr) || (t->inp_socket->so_options & SO_REUSEPORT) == 0) && (inp->inp_cred->cr_uid != t->inp_cred->cr_uid)) return (EADDRINUSE); if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0 && IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) { struct sockaddr_in sin; in6_sin6_2_sin(&sin, sin6); t = in_pcblookup_local(pcbinfo, sin.sin_addr, lport, INPLOOKUP_WILDCARD, cred); if (t && ((t->inp_vflag & INP_TIMEWAIT) == 0) && (so->so_type != SOCK_STREAM || ntohl(t->inp_faddr.s_addr) == INADDR_ANY) && (inp->inp_cred->cr_uid != t->inp_cred->cr_uid)) return (EADDRINUSE); } } - if (prison_local_ip6(cred, &sin6->sin6_addr, - ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) - return (EADDRNOTAVAIL); + if ((error = prison_local_ip6(cred, &sin6->sin6_addr, + ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0) + return (error); t = in6_pcblookup_local(pcbinfo, &sin6->sin6_addr, lport, wild, cred); if (t && (reuseport & ((t->inp_vflag & INP_TIMEWAIT) ? intotw(t)->tw_so_options : t->inp_socket->so_options)) == 0) return (EADDRINUSE); if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0 && IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) { struct sockaddr_in sin; in6_sin6_2_sin(&sin, sin6); t = in_pcblookup_local(pcbinfo, sin.sin_addr, lport, wild, cred); if (t && t->inp_vflag & INP_TIMEWAIT) { if ((reuseport & intotw(t)->tw_so_options) == 0 && (ntohl(t->inp_laddr.s_addr) != INADDR_ANY || ((inp->inp_vflag & INP_IPV6PROTO) == (t->inp_vflag & INP_IPV6PROTO)))) return (EADDRINUSE); } else if (t && (reuseport & t->inp_socket->so_options) == 0 && (ntohl(t->inp_laddr.s_addr) != INADDR_ANY || INP_SOCKAF(so) == INP_SOCKAF(t->inp_socket))) return (EADDRINUSE); } } inp->in6p_laddr = sin6->sin6_addr; } - if (prison_local_ip6(cred, &inp->in6p_laddr, - ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) - return (EINVAL); + if ((error = prison_local_ip6(cred, &inp->in6p_laddr, + ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0) + return (error); if (lport == 0) { - int e; - if ((e = in6_pcbsetport(&inp->in6p_laddr, inp, cred)) != 0) - return (e); + if ((error = in6_pcbsetport(&inp->in6p_laddr, inp, cred)) != 0) + return (error); } else { inp->inp_lport = lport; if (in_pcbinshash(inp) != 0) { inp->in6p_laddr = in6addr_any; inp->inp_lport = 0; return (EAGAIN); } } return (0); } /* * Transform old in6_pcbconnect() into an inner subroutine for new * in6_pcbconnect(): Do some validity-checking on the remote * address (in mbuf 'nam') and then determine local host address * (i.e., which interface) to use to access that remote host. * * This preserves definition of in6_pcbconnect(), while supporting a * slightly different version for T/TCP. (This is more than * a bit of a kludge, but cleaning up the internal interfaces would * have forced minor changes in every protocol). */ int in6_pcbladdr(register struct inpcb *inp, struct sockaddr *nam, struct in6_addr **plocal_addr6) { INIT_VNET_INET6(inp->inp_vnet); register struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam; int error = 0; struct ifnet *ifp = NULL; int scope_ambiguous = 0; INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); if (nam->sa_len != sizeof (*sin6)) return (EINVAL); if (sin6->sin6_family != AF_INET6) return (EAFNOSUPPORT); if (sin6->sin6_port == 0) return (EADDRNOTAVAIL); if (sin6->sin6_scope_id == 0 && !V_ip6_use_defzone) scope_ambiguous = 1; if ((error = sa6_embedscope(sin6, V_ip6_use_defzone)) != 0) return(error); if (V_in6_ifaddr) { /* * If the destination address is UNSPECIFIED addr, * use the loopback addr, e.g ::1. */ if (IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) sin6->sin6_addr = in6addr_loopback; } - if (prison_remote_ip6(inp->inp_cred, &sin6->sin6_addr) != 0) - return (EADDRNOTAVAIL); + if ((error = prison_remote_ip6(inp->inp_cred, &sin6->sin6_addr)) != 0) + return (error); /* * XXX: in6_selectsrc might replace the bound local address * with the address specified by setsockopt(IPV6_PKTINFO). * Is it the intended behavior? */ *plocal_addr6 = in6_selectsrc(sin6, inp->in6p_outputopts, inp, NULL, inp->inp_cred, &ifp, &error); if (ifp && scope_ambiguous && (error = in6_setscope(&sin6->sin6_addr, ifp, NULL)) != 0) { return(error); } if (*plocal_addr6 == NULL) { if (error == 0) error = EADDRNOTAVAIL; return (error); } /* * Don't do pcblookup call here; return interface in * plocal_addr6 * and exit to caller, that will do the lookup. */ return (0); } /* * Outer subroutine: * Connect from a socket to a specified address. * Both address and port must be specified in argument sin. * If don't have a local address for this socket yet, * then pick one. */ int in6_pcbconnect(register struct inpcb *inp, struct sockaddr *nam, struct ucred *cred) { struct in6_addr *addr6; register struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam; int error; INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); /* * Call inner routine, to assign local interface address. * in6_pcbladdr() may automatically fill in sin6_scope_id. */ if ((error = in6_pcbladdr(inp, nam, &addr6)) != 0) return (error); if (in6_pcblookup_hash(inp->inp_pcbinfo, &sin6->sin6_addr, sin6->sin6_port, IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr) ? addr6 : &inp->in6p_laddr, inp->inp_lport, 0, NULL) != NULL) { return (EADDRINUSE); } if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) { if (inp->inp_lport == 0) { error = in6_pcbbind(inp, (struct sockaddr *)0, cred); if (error) return (error); } inp->in6p_laddr = *addr6; } inp->in6p_faddr = sin6->sin6_addr; inp->inp_fport = sin6->sin6_port; /* update flowinfo - draft-itojun-ipv6-flowlabel-api-00 */ inp->inp_flow &= ~IPV6_FLOWLABEL_MASK; if (inp->inp_flags & IN6P_AUTOFLOWLABEL) inp->inp_flow |= (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); in_pcbrehash(inp); return (0); } void in6_pcbdisconnect(struct inpcb *inp) { INP_INFO_WLOCK_ASSERT(inp->inp_pcbinfo); INP_WLOCK_ASSERT(inp); bzero((caddr_t)&inp->in6p_faddr, sizeof(inp->in6p_faddr)); inp->inp_fport = 0; /* clear flowinfo - draft-itojun-ipv6-flowlabel-api-00 */ inp->inp_flow &= ~IPV6_FLOWLABEL_MASK; in_pcbrehash(inp); } struct sockaddr * in6_sockaddr(in_port_t port, struct in6_addr *addr_p) { struct sockaddr_in6 *sin6; sin6 = malloc(sizeof *sin6, M_SONAME, M_WAITOK); bzero(sin6, sizeof *sin6); sin6->sin6_family = AF_INET6; sin6->sin6_len = sizeof(*sin6); sin6->sin6_port = port; sin6->sin6_addr = *addr_p; (void)sa6_recoverscope(sin6); /* XXX: should catch errors */ return (struct sockaddr *)sin6; } struct sockaddr * in6_v4mapsin6_sockaddr(in_port_t port, struct in_addr *addr_p) { struct sockaddr_in sin; struct sockaddr_in6 *sin6_p; bzero(&sin, sizeof sin); sin.sin_family = AF_INET; sin.sin_len = sizeof(sin); sin.sin_port = port; sin.sin_addr = *addr_p; sin6_p = malloc(sizeof *sin6_p, M_SONAME, M_WAITOK); in6_sin_2_v4mapsin6(&sin, sin6_p); return (struct sockaddr *)sin6_p; } int in6_getsockaddr(struct socket *so, struct sockaddr **nam) { register struct inpcb *inp; struct in6_addr addr; in_port_t port; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in6_getsockaddr: inp == NULL")); INP_RLOCK(inp); port = inp->inp_lport; addr = inp->in6p_laddr; INP_RUNLOCK(inp); *nam = in6_sockaddr(port, &addr); return 0; } int in6_getpeeraddr(struct socket *so, struct sockaddr **nam) { struct inpcb *inp; struct in6_addr addr; in_port_t port; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in6_getpeeraddr: inp == NULL")); INP_RLOCK(inp); port = inp->inp_fport; addr = inp->in6p_faddr; INP_RUNLOCK(inp); *nam = in6_sockaddr(port, &addr); return 0; } int in6_mapped_sockaddr(struct socket *so, struct sockaddr **nam) { struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in6_mapped_sockaddr: inp == NULL")); if ((inp->inp_vflag & (INP_IPV4 | INP_IPV6)) == INP_IPV4) { error = in_getsockaddr(so, nam); if (error == 0) in6_sin_2_v4mapsin6_in_sock(nam); } else { /* scope issues will be handled in in6_getsockaddr(). */ error = in6_getsockaddr(so, nam); } return error; } int in6_mapped_peeraddr(struct socket *so, struct sockaddr **nam) { struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp != NULL, ("in6_mapped_peeraddr: inp == NULL")); if ((inp->inp_vflag & (INP_IPV4 | INP_IPV6)) == INP_IPV4) { error = in_getpeeraddr(so, nam); if (error == 0) in6_sin_2_v4mapsin6_in_sock(nam); } else /* scope issues will be handled in in6_getpeeraddr(). */ error = in6_getpeeraddr(so, nam); return error; } /* * Pass some notification to all connections of a protocol * associated with address dst. The local address and/or port numbers * may be specified to limit the search. The "usual action" will be * taken, depending on the ctlinput cmd. The caller must filter any * cmds that are uninteresting (e.g., no error in the map). * Call the protocol specific routine (if any) to report * any errors for each matching socket. */ void in6_pcbnotify(struct inpcbinfo *pcbinfo, struct sockaddr *dst, u_int fport_arg, const struct sockaddr *src, u_int lport_arg, int cmd, void *cmdarg, struct inpcb *(*notify)(struct inpcb *, int)) { struct inpcb *inp, *inp_temp; struct sockaddr_in6 sa6_src, *sa6_dst; u_short fport = fport_arg, lport = lport_arg; u_int32_t flowinfo; int errno; if ((unsigned)cmd >= PRC_NCMDS || dst->sa_family != AF_INET6) return; sa6_dst = (struct sockaddr_in6 *)dst; if (IN6_IS_ADDR_UNSPECIFIED(&sa6_dst->sin6_addr)) return; /* * note that src can be NULL when we get notify by local fragmentation. */ sa6_src = (src == NULL) ? sa6_any : *(const struct sockaddr_in6 *)src; flowinfo = sa6_src.sin6_flowinfo; /* * Redirects go to all references to the destination, * and use in6_rtchange to invalidate the route cache. * Dead host indications: also use in6_rtchange to invalidate * the cache, and deliver the error to all the sockets. * Otherwise, if we have knowledge of the local port and address, * deliver only to that socket. */ if (PRC_IS_REDIRECT(cmd) || cmd == PRC_HOSTDEAD) { fport = 0; lport = 0; bzero((caddr_t)&sa6_src.sin6_addr, sizeof(sa6_src.sin6_addr)); if (cmd != PRC_HOSTDEAD) notify = in6_rtchange; } errno = inet6ctlerrmap[cmd]; INP_INFO_WLOCK(pcbinfo); LIST_FOREACH_SAFE(inp, pcbinfo->ipi_listhead, inp_list, inp_temp) { INP_WLOCK(inp); if ((inp->inp_vflag & INP_IPV6) == 0) { INP_WUNLOCK(inp); continue; } /* * If the error designates a new path MTU for a destination * and the application (associated with this socket) wanted to * know the value, notify. Note that we notify for all * disconnected sockets if the corresponding application * wanted. This is because some UDP applications keep sending * sockets disconnected. * XXX: should we avoid to notify the value to TCP sockets? */ if (cmd == PRC_MSGSIZE && (inp->inp_flags & IN6P_MTU) != 0 && (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) || IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, &sa6_dst->sin6_addr))) { ip6_notify_pmtu(inp, (struct sockaddr_in6 *)dst, (u_int32_t *)cmdarg); } /* * Detect if we should notify the error. If no source and * destination ports are specifed, but non-zero flowinfo and * local address match, notify the error. This is the case * when the error is delivered with an encrypted buffer * by ESP. Otherwise, just compare addresses and ports * as usual. */ if (lport == 0 && fport == 0 && flowinfo && inp->inp_socket != NULL && flowinfo == (inp->inp_flow & IPV6_FLOWLABEL_MASK) && IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, &sa6_src.sin6_addr)) goto do_notify; else if (!IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, &sa6_dst->sin6_addr) || inp->inp_socket == 0 || (lport && inp->inp_lport != lport) || (!IN6_IS_ADDR_UNSPECIFIED(&sa6_src.sin6_addr) && !IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, &sa6_src.sin6_addr)) || (fport && inp->inp_fport != fport)) { INP_WUNLOCK(inp); continue; } do_notify: if (notify) { if ((*notify)(inp, errno)) INP_WUNLOCK(inp); } else INP_WUNLOCK(inp); } INP_INFO_WUNLOCK(pcbinfo); } /* * Lookup a PCB based on the local address and port. */ struct inpcb * in6_pcblookup_local(struct inpcbinfo *pcbinfo, struct in6_addr *laddr, u_short lport, int wild_okay, struct ucred *cred) { register struct inpcb *inp; int matchwild = 3, wildcard; INP_INFO_WLOCK_ASSERT(pcbinfo); if (!wild_okay) { struct inpcbhead *head; /* * Look for an unconnected (wildcard foreign addr) PCB that * matches the local address and port we're looking for. */ head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport, 0, pcbinfo->ipi_hashmask)]; LIST_FOREACH(inp, head, inp_hash) { /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV6) == 0) continue; if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) && IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) && inp->inp_lport == lport) { /* Found. */ if (cred == NULL || inp->inp_cred->cr_prison == cred->cr_prison) return (inp); } } /* * Not found. */ return (NULL); } else { struct inpcbporthead *porthash; struct inpcbport *phd; struct inpcb *match = NULL; /* * Best fit PCB lookup. * * First see if this local port is in use by looking on the * port hash list. */ porthash = &pcbinfo->ipi_porthashbase[INP_PCBPORTHASH(lport, pcbinfo->ipi_porthashmask)]; LIST_FOREACH(phd, porthash, phd_hash) { if (phd->phd_port == lport) break; } if (phd != NULL) { /* * Port is in use by one or more PCBs. Look for best * fit. */ LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) { wildcard = 0; if (cred != NULL && inp->inp_cred->cr_prison != cred->cr_prison) continue; /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV6) == 0) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) wildcard++; if (!IN6_IS_ADDR_UNSPECIFIED( &inp->in6p_laddr)) { if (IN6_IS_ADDR_UNSPECIFIED(laddr)) wildcard++; else if (!IN6_ARE_ADDR_EQUAL( &inp->in6p_laddr, laddr)) continue; } else { if (!IN6_IS_ADDR_UNSPECIFIED(laddr)) wildcard++; } if (wildcard < matchwild) { match = inp; matchwild = wildcard; if (matchwild == 0) break; } } } return (match); } } void in6_pcbpurgeif0(struct inpcbinfo *pcbinfo, struct ifnet *ifp) { struct inpcb *in6p; struct ip6_moptions *im6o; struct in6_multi_mship *imm, *nimm; INP_INFO_RLOCK(pcbinfo); LIST_FOREACH(in6p, pcbinfo->ipi_listhead, inp_list) { INP_WLOCK(in6p); im6o = in6p->in6p_moptions; if ((in6p->inp_vflag & INP_IPV6) && im6o) { /* * Unselect the outgoing interface if it is being * detached. */ if (im6o->im6o_multicast_ifp == ifp) im6o->im6o_multicast_ifp = NULL; /* * Drop multicast group membership if we joined * through the interface being detached. * XXX controversial - is it really legal for kernel * to force this? */ for (imm = im6o->im6o_memberships.lh_first; imm != NULL; imm = nimm) { nimm = imm->i6mm_chain.le_next; if (imm->i6mm_maddr->in6m_ifp == ifp) { LIST_REMOVE(imm, i6mm_chain); in6_delmulti(imm->i6mm_maddr); free(imm, M_IP6MADDR); } } } INP_WUNLOCK(in6p); } INP_INFO_RUNLOCK(pcbinfo); } /* * Check for alternatives when higher level complains * about service problems. For now, invalidate cached * routing information. If the route was created dynamically * (by a redirect), time to try a default gateway again. */ void in6_losing(struct inpcb *in6p) { /* * We don't store route pointers in the routing table anymore */ return; } /* * After a routing change, flush old routing * and allocate a (hopefully) better one. */ struct inpcb * in6_rtchange(struct inpcb *inp, int errno) { /* * We don't store route pointers in the routing table anymore */ return inp; } /* * Lookup PCB in hash list. */ struct inpcb * in6_pcblookup_hash(struct inpcbinfo *pcbinfo, struct in6_addr *faddr, u_int fport_arg, struct in6_addr *laddr, u_int lport_arg, int wildcard, struct ifnet *ifp) { struct inpcbhead *head; struct inpcb *inp, *tmpinp; u_short fport = fport_arg, lport = lport_arg; int faith; INP_INFO_LOCK_ASSERT(pcbinfo); if (faithprefix_p != NULL) faith = (*faithprefix_p)(laddr); else faith = 0; /* * First look for an exact match. */ tmpinp = NULL; head = &pcbinfo->ipi_hashbase[ INP_PCBHASH(faddr->s6_addr32[3] /* XXX */, lport, fport, pcbinfo->ipi_hashmask)]; LIST_FOREACH(inp, head, inp_hash) { /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV6) == 0) continue; if (IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, faddr) && IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) && inp->inp_fport == fport && inp->inp_lport == lport) { /* * XXX We should be able to directly return * the inp here, without any checks. * Well unless both bound with SO_REUSEPORT? */ if (jailed(inp->inp_cred)) return (inp); if (tmpinp == NULL) tmpinp = inp; } } if (tmpinp != NULL) return (tmpinp); /* * Then look for a wildcard match, if requested. */ if (wildcard == INPLOOKUP_WILDCARD) { struct inpcb *local_wild = NULL, *local_exact = NULL; struct inpcb *jail_wild = NULL; int injail; /* * Order of socket selection - we always prefer jails. * 1. jailed, non-wild. * 2. jailed, wild. * 3. non-jailed, non-wild. * 4. non-jailed, wild. */ head = &pcbinfo->ipi_hashbase[INP_PCBHASH(INADDR_ANY, lport, 0, pcbinfo->ipi_hashmask)]; LIST_FOREACH(inp, head, inp_hash) { /* XXX inp locking */ if ((inp->inp_vflag & INP_IPV6) == 0) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) || inp->inp_lport != lport) { continue; } /* XXX inp locking */ if (faith && (inp->inp_flags & INP_FAITH) == 0) continue; injail = jailed(inp->inp_cred); if (injail) { - if (!prison_check_ip6(inp->inp_cred, laddr)) + if (prison_check_ip6(inp->inp_cred, + laddr) != 0) continue; } else { if (local_exact != NULL) continue; } if (IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr)) { if (injail) return (inp); else local_exact = inp; } else if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) { if (injail) jail_wild = inp; else local_wild = inp; } } /* LIST_FOREACH */ if (jail_wild != NULL) return (jail_wild); if (local_exact != NULL) return (local_exact); if (local_wild != NULL) return (local_wild); } /* if (wildcard == INPLOOKUP_WILDCARD) */ /* * Not found. */ return (NULL); } void init_sin6(struct sockaddr_in6 *sin6, struct mbuf *m) { struct ip6_hdr *ip; ip = mtod(m, struct ip6_hdr *); bzero(sin6, sizeof(*sin6)); sin6->sin6_len = sizeof(*sin6); sin6->sin6_family = AF_INET6; sin6->sin6_addr = ip->ip6_src; (void)sa6_recoverscope(sin6); /* XXX: should catch errors... */ return; } diff --git a/sys/netinet6/in6_src.c b/sys/netinet6/in6_src.c index ca65bc9a8568..80304169490e 100644 --- a/sys/netinet6/in6_src.c +++ b/sys/netinet6/in6_src.c @@ -1,1179 +1,1178 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 3. Neither the name of the project 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 PROJECT 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 PROJECT 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: in6_src.c,v 1.132 2003/08/26 04:42:27 keiichi Exp $ */ /*- * Copyright (c) 1982, 1986, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_mpath.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef RADIX_MPATH #include #endif #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static struct mtx addrsel_lock; #define ADDRSEL_LOCK_INIT() mtx_init(&addrsel_lock, "addrsel_lock", NULL, MTX_DEF) #define ADDRSEL_LOCK() mtx_lock(&addrsel_lock) #define ADDRSEL_UNLOCK() mtx_unlock(&addrsel_lock) #define ADDRSEL_LOCK_ASSERT() mtx_assert(&addrsel_lock, MA_OWNED) static struct sx addrsel_sxlock; #define ADDRSEL_SXLOCK_INIT() sx_init(&addrsel_sxlock, "addrsel_sxlock") #define ADDRSEL_SLOCK() sx_slock(&addrsel_sxlock) #define ADDRSEL_SUNLOCK() sx_sunlock(&addrsel_sxlock) #define ADDRSEL_XLOCK() sx_xlock(&addrsel_sxlock) #define ADDRSEL_XUNLOCK() sx_xunlock(&addrsel_sxlock) #define ADDR_LABEL_NOTAPP (-1) #ifdef VIMAGE_GLOBALS struct in6_addrpolicy defaultaddrpolicy; int ip6_prefer_tempaddr; #endif static int selectroute __P((struct sockaddr_in6 *, struct ip6_pktopts *, struct ip6_moptions *, struct route_in6 *, struct ifnet **, struct rtentry **, int)); static int in6_selectif __P((struct sockaddr_in6 *, struct ip6_pktopts *, struct ip6_moptions *, struct route_in6 *ro, struct ifnet **)); static struct in6_addrpolicy *lookup_addrsel_policy(struct sockaddr_in6 *); static void init_policy_queue(void); static int add_addrsel_policyent(struct in6_addrpolicy *); static int delete_addrsel_policyent(struct in6_addrpolicy *); static int walk_addrsel_policy __P((int (*)(struct in6_addrpolicy *, void *), void *)); static int dump_addrsel_policyent(struct in6_addrpolicy *, void *); static struct in6_addrpolicy *match_addrsel_policy(struct sockaddr_in6 *); /* * Return an IPv6 address, which is the most appropriate for a given * destination and user specified options. * If necessary, this function lookups the routing table and returns * an entry to the caller for later use. */ #define REPLACE(r) do {\ if ((r) < sizeof(V_ip6stat.ip6s_sources_rule) / \ sizeof(V_ip6stat.ip6s_sources_rule[0])) /* check for safety */ \ V_ip6stat.ip6s_sources_rule[(r)]++; \ /* { \ char ip6buf[INET6_ADDRSTRLEN], ip6b[INET6_ADDRSTRLEN]; \ printf("in6_selectsrc: replace %s with %s by %d\n", ia_best ? ip6_sprintf(ip6buf, &ia_best->ia_addr.sin6_addr) : "none", ip6_sprintf(ip6b, &ia->ia_addr.sin6_addr), (r)); \ } */ \ goto replace; \ } while(0) #define NEXT(r) do {\ if ((r) < sizeof(V_ip6stat.ip6s_sources_rule) / \ sizeof(V_ip6stat.ip6s_sources_rule[0])) /* check for safety */ \ V_ip6stat.ip6s_sources_rule[(r)]++; \ /* { \ char ip6buf[INET6_ADDRSTRLEN], ip6b[INET6_ADDRSTRLEN]; \ printf("in6_selectsrc: keep %s against %s by %d\n", ia_best ? ip6_sprintf(ip6buf, &ia_best->ia_addr.sin6_addr) : "none", ip6_sprintf(ip6b, &ia->ia_addr.sin6_addr), (r)); \ } */ \ goto next; /* XXX: we can't use 'continue' here */ \ } while(0) #define BREAK(r) do { \ if ((r) < sizeof(V_ip6stat.ip6s_sources_rule) / \ sizeof(V_ip6stat.ip6s_sources_rule[0])) /* check for safety */ \ V_ip6stat.ip6s_sources_rule[(r)]++; \ goto out; /* XXX: we can't use 'break' here */ \ } while(0) struct in6_addr * in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts, struct inpcb *inp, struct route_in6 *ro, struct ucred *cred, struct ifnet **ifpp, int *errorp) { INIT_VNET_INET6(curvnet); struct in6_addr dst; struct ifnet *ifp = NULL; struct in6_ifaddr *ia = NULL, *ia_best = NULL; struct in6_pktinfo *pi = NULL; int dst_scope = -1, best_scope = -1, best_matchlen = -1; struct in6_addrpolicy *dst_policy = NULL, *best_policy = NULL; u_int32_t odstzone; int prefer_tempaddr; struct ip6_moptions *mopts; dst = dstsock->sin6_addr; /* make a copy for local operation */ *errorp = 0; if (ifpp) *ifpp = NULL; if (inp != NULL) { INP_LOCK_ASSERT(inp); mopts = inp->in6p_moptions; } else { mopts = NULL; } /* * If the source address is explicitly specified by the caller, * check if the requested source address is indeed a unicast address * assigned to the node, and can be used as the packet's source * address. If everything is okay, use the address as source. */ if (opts && (pi = opts->ip6po_pktinfo) && !IN6_IS_ADDR_UNSPECIFIED(&pi->ipi6_addr)) { struct sockaddr_in6 srcsock; struct in6_ifaddr *ia6; /* get the outgoing interface */ if ((*errorp = in6_selectif(dstsock, opts, mopts, ro, &ifp)) != 0) { return (NULL); } /* * determine the appropriate zone id of the source based on * the zone of the destination and the outgoing interface. * If the specified address is ambiguous wrt the scope zone, * the interface must be specified; otherwise, ifa_ifwithaddr() * will fail matching the address. */ bzero(&srcsock, sizeof(srcsock)); srcsock.sin6_family = AF_INET6; srcsock.sin6_len = sizeof(srcsock); srcsock.sin6_addr = pi->ipi6_addr; if (ifp) { *errorp = in6_setscope(&srcsock.sin6_addr, ifp, NULL); if (*errorp != 0) return (NULL); } - if (cred != NULL && prison_local_ip6(cred, &srcsock.sin6_addr, - (inp != NULL && (inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) { - *errorp = EADDRNOTAVAIL; + if (cred != NULL && (*errorp = prison_local_ip6(cred, + &srcsock.sin6_addr, (inp != NULL && + (inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0) return (NULL); - } ia6 = (struct in6_ifaddr *)ifa_ifwithaddr((struct sockaddr *)(&srcsock)); if (ia6 == NULL || (ia6->ia6_flags & (IN6_IFF_ANYCAST | IN6_IFF_NOTREADY))) { *errorp = EADDRNOTAVAIL; return (NULL); } pi->ipi6_addr = srcsock.sin6_addr; /* XXX: this overrides pi */ if (ifpp) *ifpp = ifp; return (&ia6->ia_addr.sin6_addr); } /* * Otherwise, if the socket has already bound the source, just use it. */ if (inp != NULL && !IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) { - if (cred != NULL && prison_local_ip6(cred, &inp->in6p_laddr, - ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) { - *errorp = EADDRNOTAVAIL; + if (cred != NULL && + (*errorp = prison_local_ip6(cred, &inp->in6p_laddr, + ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0))) != 0) return (NULL); - } return (&inp->in6p_laddr); } /* * If the address is not specified, choose the best one based on * the outgoing interface and the destination address. */ /* get the outgoing interface */ if ((*errorp = in6_selectif(dstsock, opts, mopts, ro, &ifp)) != 0) return (NULL); #ifdef DIAGNOSTIC if (ifp == NULL) /* this should not happen */ panic("in6_selectsrc: NULL ifp"); #endif *errorp = in6_setscope(&dst, ifp, &odstzone); if (*errorp != 0) return (NULL); for (ia = V_in6_ifaddr; ia; ia = ia->ia_next) { int new_scope = -1, new_matchlen = -1; struct in6_addrpolicy *new_policy = NULL; u_int32_t srczone, osrczone, dstzone; struct in6_addr src; struct ifnet *ifp1 = ia->ia_ifp; /* * We'll never take an address that breaks the scope zone * of the destination. We also skip an address if its zone * does not contain the outgoing interface. * XXX: we should probably use sin6_scope_id here. */ if (in6_setscope(&dst, ifp1, &dstzone) || odstzone != dstzone) { continue; } src = ia->ia_addr.sin6_addr; if (in6_setscope(&src, ifp, &osrczone) || in6_setscope(&src, ifp1, &srczone) || osrczone != srczone) { continue; } /* avoid unusable addresses */ if ((ia->ia6_flags & (IN6_IFF_NOTREADY | IN6_IFF_ANYCAST | IN6_IFF_DETACHED))) { continue; } if (!V_ip6_use_deprecated && IFA6_IS_DEPRECATED(ia)) continue; if (cred != NULL && prison_local_ip6(cred, &ia->ia_addr.sin6_addr, (inp != NULL && (inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) continue; /* Rule 1: Prefer same address */ if (IN6_ARE_ADDR_EQUAL(&dst, &ia->ia_addr.sin6_addr)) { ia_best = ia; BREAK(1); /* there should be no better candidate */ } if (ia_best == NULL) REPLACE(0); /* Rule 2: Prefer appropriate scope */ if (dst_scope < 0) dst_scope = in6_addrscope(&dst); new_scope = in6_addrscope(&ia->ia_addr.sin6_addr); if (IN6_ARE_SCOPE_CMP(best_scope, new_scope) < 0) { if (IN6_ARE_SCOPE_CMP(best_scope, dst_scope) < 0) REPLACE(2); NEXT(2); } else if (IN6_ARE_SCOPE_CMP(new_scope, best_scope) < 0) { if (IN6_ARE_SCOPE_CMP(new_scope, dst_scope) < 0) NEXT(2); REPLACE(2); } /* * Rule 3: Avoid deprecated addresses. Note that the case of * !ip6_use_deprecated is already rejected above. */ if (!IFA6_IS_DEPRECATED(ia_best) && IFA6_IS_DEPRECATED(ia)) NEXT(3); if (IFA6_IS_DEPRECATED(ia_best) && !IFA6_IS_DEPRECATED(ia)) REPLACE(3); /* Rule 4: Prefer home addresses */ /* * XXX: This is a TODO. We should probably merge the MIP6 * case above. */ /* Rule 5: Prefer outgoing interface */ if (ia_best->ia_ifp == ifp && ia->ia_ifp != ifp) NEXT(5); if (ia_best->ia_ifp != ifp && ia->ia_ifp == ifp) REPLACE(5); /* * Rule 6: Prefer matching label * Note that best_policy should be non-NULL here. */ if (dst_policy == NULL) dst_policy = lookup_addrsel_policy(dstsock); if (dst_policy->label != ADDR_LABEL_NOTAPP) { new_policy = lookup_addrsel_policy(&ia->ia_addr); if (dst_policy->label == best_policy->label && dst_policy->label != new_policy->label) NEXT(6); if (dst_policy->label != best_policy->label && dst_policy->label == new_policy->label) REPLACE(6); } /* * Rule 7: Prefer public addresses. * We allow users to reverse the logic by configuring * a sysctl variable, so that privacy conscious users can * always prefer temporary addresses. */ if (opts == NULL || opts->ip6po_prefer_tempaddr == IP6PO_TEMPADDR_SYSTEM) { prefer_tempaddr = V_ip6_prefer_tempaddr; } else if (opts->ip6po_prefer_tempaddr == IP6PO_TEMPADDR_NOTPREFER) { prefer_tempaddr = 0; } else prefer_tempaddr = 1; if (!(ia_best->ia6_flags & IN6_IFF_TEMPORARY) && (ia->ia6_flags & IN6_IFF_TEMPORARY)) { if (prefer_tempaddr) REPLACE(7); else NEXT(7); } if ((ia_best->ia6_flags & IN6_IFF_TEMPORARY) && !(ia->ia6_flags & IN6_IFF_TEMPORARY)) { if (prefer_tempaddr) NEXT(7); else REPLACE(7); } /* * Rule 8: prefer addresses on alive interfaces. * This is a KAME specific rule. */ if ((ia_best->ia_ifp->if_flags & IFF_UP) && !(ia->ia_ifp->if_flags & IFF_UP)) NEXT(8); if (!(ia_best->ia_ifp->if_flags & IFF_UP) && (ia->ia_ifp->if_flags & IFF_UP)) REPLACE(8); /* * Rule 14: Use longest matching prefix. * Note: in the address selection draft, this rule is * documented as "Rule 8". However, since it is also * documented that this rule can be overridden, we assign * a large number so that it is easy to assign smaller numbers * to more preferred rules. */ new_matchlen = in6_matchlen(&ia->ia_addr.sin6_addr, &dst); if (best_matchlen < new_matchlen) REPLACE(14); if (new_matchlen < best_matchlen) NEXT(14); /* Rule 15 is reserved. */ /* * Last resort: just keep the current candidate. * Or, do we need more rules? */ continue; replace: ia_best = ia; best_scope = (new_scope >= 0 ? new_scope : in6_addrscope(&ia_best->ia_addr.sin6_addr)); best_policy = (new_policy ? new_policy : lookup_addrsel_policy(&ia_best->ia_addr)); best_matchlen = (new_matchlen >= 0 ? new_matchlen : in6_matchlen(&ia_best->ia_addr.sin6_addr, &dst)); next: continue; out: break; } if ((ia = ia_best) == NULL) { *errorp = EADDRNOTAVAIL; return (NULL); } if (ifpp) *ifpp = ifp; return (&ia->ia_addr.sin6_addr); } /* * clone - meaningful only for bsdi and freebsd */ static int selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts, struct ip6_moptions *mopts, struct route_in6 *ro, struct ifnet **retifp, struct rtentry **retrt, int norouteok) { INIT_VNET_INET6(curvnet); int error = 0; struct ifnet *ifp = NULL; struct rtentry *rt = NULL; struct sockaddr_in6 *sin6_next; struct in6_pktinfo *pi = NULL; struct in6_addr *dst = &dstsock->sin6_addr; #if 0 char ip6buf[INET6_ADDRSTRLEN]; if (dstsock->sin6_addr.s6_addr32[0] == 0 && dstsock->sin6_addr.s6_addr32[1] == 0 && !IN6_IS_ADDR_LOOPBACK(&dstsock->sin6_addr)) { printf("in6_selectroute: strange destination %s\n", ip6_sprintf(ip6buf, &dstsock->sin6_addr)); } else { printf("in6_selectroute: destination = %s%%%d\n", ip6_sprintf(ip6buf, &dstsock->sin6_addr), dstsock->sin6_scope_id); /* for debug */ } #endif /* If the caller specify the outgoing interface explicitly, use it. */ if (opts && (pi = opts->ip6po_pktinfo) != NULL && pi->ipi6_ifindex) { /* XXX boundary check is assumed to be already done. */ ifp = ifnet_byindex(pi->ipi6_ifindex); if (ifp != NULL && (norouteok || retrt == NULL || IN6_IS_ADDR_MULTICAST(dst))) { /* * we do not have to check or get the route for * multicast. */ goto done; } else goto getroute; } /* * If the destination address is a multicast address and the outgoing * interface for the address is specified by the caller, use it. */ if (IN6_IS_ADDR_MULTICAST(dst) && mopts != NULL && (ifp = mopts->im6o_multicast_ifp) != NULL) { goto done; /* we do not need a route for multicast. */ } getroute: /* * If the next hop address for the packet is specified by the caller, * use it as the gateway. */ if (opts && opts->ip6po_nexthop) { struct route_in6 *ron; struct llentry *la; sin6_next = satosin6(opts->ip6po_nexthop); /* at this moment, we only support AF_INET6 next hops */ if (sin6_next->sin6_family != AF_INET6) { error = EAFNOSUPPORT; /* or should we proceed? */ goto done; } /* * If the next hop is an IPv6 address, then the node identified * by that address must be a neighbor of the sending host. */ ron = &opts->ip6po_nextroute; /* * XXX what do we do here? * PLZ to be fixing */ if (ron->ro_rt == NULL) { rtalloc((struct route *)ron); /* multi path case? */ if (ron->ro_rt == NULL) { if (ron->ro_rt) { RTFREE(ron->ro_rt); ron->ro_rt = NULL; } error = EHOSTUNREACH; goto done; } } rt = ron->ro_rt; ifp = rt->rt_ifp; IF_AFDATA_LOCK(ifp); la = lla_lookup(LLTABLE6(ifp), 0, (struct sockaddr *)&sin6_next->sin6_addr); IF_AFDATA_UNLOCK(ifp); if (la != NULL) LLE_RUNLOCK(la); else { error = EHOSTUNREACH; goto done; } #if 0 if ((ron->ro_rt && (ron->ro_rt->rt_flags & (RTF_UP | RTF_LLINFO)) != (RTF_UP | RTF_LLINFO)) || !IN6_ARE_ADDR_EQUAL(&satosin6(&ron->ro_dst)->sin6_addr, &sin6_next->sin6_addr)) { if (ron->ro_rt) { RTFREE(ron->ro_rt); ron->ro_rt = NULL; } *satosin6(&ron->ro_dst) = *sin6_next; } if (ron->ro_rt == NULL) { rtalloc((struct route *)ron); /* multi path case? */ if (ron->ro_rt == NULL || !(ron->ro_rt->rt_flags & RTF_LLINFO)) { if (ron->ro_rt) { RTFREE(ron->ro_rt); ron->ro_rt = NULL; } error = EHOSTUNREACH; goto done; } } #endif /* * When cloning is required, try to allocate a route to the * destination so that the caller can store path MTU * information. */ goto done; } /* * Use a cached route if it exists and is valid, else try to allocate * a new one. Note that we should check the address family of the * cached destination, in case of sharing the cache with IPv4. */ if (ro) { if (ro->ro_rt && (!(ro->ro_rt->rt_flags & RTF_UP) || ((struct sockaddr *)(&ro->ro_dst))->sa_family != AF_INET6 || !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, dst))) { RTFREE(ro->ro_rt); ro->ro_rt = (struct rtentry *)NULL; } if (ro->ro_rt == (struct rtentry *)NULL) { struct sockaddr_in6 *sa6; /* No route yet, so try to acquire one */ bzero(&ro->ro_dst, sizeof(struct sockaddr_in6)); sa6 = (struct sockaddr_in6 *)&ro->ro_dst; *sa6 = *dstsock; sa6->sin6_scope_id = 0; #ifdef RADIX_MPATH rtalloc_mpath((struct route *)ro, ntohl(sa6->sin6_addr.s6_addr32[3])); #else ro->ro_rt = rtalloc1(&((struct route *)ro) ->ro_dst, 0, 0UL); if (ro->ro_rt) RT_UNLOCK(ro->ro_rt); #endif } /* * do not care about the result if we have the nexthop * explicitly specified. */ if (opts && opts->ip6po_nexthop) goto done; if (ro->ro_rt) { ifp = ro->ro_rt->rt_ifp; if (ifp == NULL) { /* can this really happen? */ RTFREE(ro->ro_rt); ro->ro_rt = NULL; } } if (ro->ro_rt == NULL) error = EHOSTUNREACH; rt = ro->ro_rt; /* * Check if the outgoing interface conflicts with * the interface specified by ipi6_ifindex (if specified). * Note that loopback interface is always okay. * (this may happen when we are sending a packet to one of * our own addresses.) */ if (ifp && opts && opts->ip6po_pktinfo && opts->ip6po_pktinfo->ipi6_ifindex) { if (!(ifp->if_flags & IFF_LOOPBACK) && ifp->if_index != opts->ip6po_pktinfo->ipi6_ifindex) { error = EHOSTUNREACH; goto done; } } } done: if (ifp == NULL && rt == NULL) { /* * This can happen if the caller did not pass a cached route * nor any other hints. We treat this case an error. */ error = EHOSTUNREACH; } if (error == EHOSTUNREACH) V_ip6stat.ip6s_noroute++; if (retifp != NULL) *retifp = ifp; if (retrt != NULL) *retrt = rt; /* rt may be NULL */ return (error); } static int in6_selectif(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts, struct ip6_moptions *mopts, struct route_in6 *ro, struct ifnet **retifp) { int error; struct route_in6 sro; struct rtentry *rt = NULL; if (ro == NULL) { bzero(&sro, sizeof(sro)); ro = &sro; } if ((error = selectroute(dstsock, opts, mopts, ro, retifp, &rt, 1)) != 0) { if (ro == &sro && rt && rt == sro.ro_rt) RTFREE(rt); return (error); } /* * do not use a rejected or black hole route. * XXX: this check should be done in the L2 output routine. * However, if we skipped this check here, we'd see the following * scenario: * - install a rejected route for a scoped address prefix * (like fe80::/10) * - send a packet to a destination that matches the scoped prefix, * with ambiguity about the scope zone. * - pick the outgoing interface from the route, and disambiguate the * scope zone with the interface. * - ip6_output() would try to get another route with the "new" * destination, which may be valid. * - we'd see no error on output. * Although this may not be very harmful, it should still be confusing. * We thus reject the case here. */ if (rt && (rt->rt_flags & (RTF_REJECT | RTF_BLACKHOLE))) { int flags = (rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH); if (ro == &sro && rt && rt == sro.ro_rt) RTFREE(rt); return (flags); } /* * Adjust the "outgoing" interface. If we're going to loop the packet * back to ourselves, the ifp would be the loopback interface. * However, we'd rather know the interface associated to the * destination address (which should probably be one of our own * addresses.) */ if (rt && rt->rt_ifa && rt->rt_ifa->ifa_ifp) *retifp = rt->rt_ifa->ifa_ifp; if (ro == &sro && rt && rt == sro.ro_rt) RTFREE(rt); return (0); } /* * clone - meaningful only for bsdi and freebsd */ int in6_selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts, struct ip6_moptions *mopts, struct route_in6 *ro, struct ifnet **retifp, struct rtentry **retrt) { return (selectroute(dstsock, opts, mopts, ro, retifp, retrt, 0)); } /* * Default hop limit selection. The precedence is as follows: * 1. Hoplimit value specified via ioctl. * 2. (If the outgoing interface is detected) the current * hop limit of the interface specified by router advertisement. * 3. The system default hoplimit. */ int in6_selecthlim(struct inpcb *in6p, struct ifnet *ifp) { INIT_VNET_INET6(curvnet); if (in6p && in6p->in6p_hops >= 0) return (in6p->in6p_hops); else if (ifp) return (ND_IFINFO(ifp)->chlim); else if (in6p && !IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr)) { struct route_in6 ro6; struct ifnet *lifp; bzero(&ro6, sizeof(ro6)); ro6.ro_dst.sin6_family = AF_INET6; ro6.ro_dst.sin6_len = sizeof(struct sockaddr_in6); ro6.ro_dst.sin6_addr = in6p->in6p_faddr; rtalloc((struct route *)&ro6); if (ro6.ro_rt) { lifp = ro6.ro_rt->rt_ifp; RTFREE(ro6.ro_rt); if (lifp) return (ND_IFINFO(lifp)->chlim); } else return (V_ip6_defhlim); } return (V_ip6_defhlim); } /* * XXX: this is borrowed from in6_pcbbind(). If possible, we should * share this function by all *bsd*... */ int in6_pcbsetport(struct in6_addr *laddr, struct inpcb *inp, struct ucred *cred) { INIT_VNET_INET(curvnet); struct socket *so = inp->inp_socket; u_int16_t lport = 0, first, last, *lastport; - int count, error = 0, wild = 0, dorandom; + int count, error, wild = 0, dorandom; struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; INP_INFO_WLOCK_ASSERT(pcbinfo); INP_WLOCK_ASSERT(inp); - if (prison_local_ip6(cred, laddr, - ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) - return(EINVAL); + error = prison_local_ip6(cred, laddr, + ((inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)); + if (error) + return(error); /* XXX: this is redundant when called from in6_pcbbind */ if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) wild = INPLOOKUP_WILDCARD; inp->inp_flags |= INP_ANONPORT; if (inp->inp_flags & INP_HIGHPORT) { first = V_ipport_hifirstauto; /* sysctl */ last = V_ipport_hilastauto; lastport = &pcbinfo->ipi_lasthi; } else if (inp->inp_flags & INP_LOWPORT) { error = priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT, 0); if (error) return error; first = V_ipport_lowfirstauto; /* 1023 */ last = V_ipport_lowlastauto; /* 600 */ lastport = &pcbinfo->ipi_lastlow; } else { first = V_ipport_firstauto; /* sysctl */ last = V_ipport_lastauto; lastport = &pcbinfo->ipi_lastport; } /* * For UDP, use random port allocation as long as the user * allows it. For TCP (and as of yet unknown) connections, * use random port allocation only if the user allows it AND * ipport_tick() allows it. */ if (V_ipport_randomized && (!V_ipport_stoprandom || pcbinfo == &V_udbinfo)) dorandom = 1; else dorandom = 0; /* * It makes no sense to do random port allocation if * we have the only port available. */ if (first == last) dorandom = 0; /* Make sure to not include UDP packets in the count. */ if (pcbinfo != &V_udbinfo) V_ipport_tcpallocs++; /* * Instead of having two loops further down counting up or down * make sure that first is always <= last and go with only one * code path implementing all logic. */ if (first > last) { u_int16_t aux; aux = first; first = last; last = aux; } if (dorandom) *lastport = first + (arc4random() % (last - first)); count = last - first; do { if (count-- < 0) { /* completely used? */ /* Undo an address bind that may have occurred. */ inp->in6p_laddr = in6addr_any; return (EADDRNOTAVAIL); } ++*lastport; if (*lastport < first || *lastport > last) *lastport = first; lport = htons(*lastport); } while (in6_pcblookup_local(pcbinfo, &inp->in6p_laddr, lport, wild, cred)); inp->inp_lport = lport; if (in_pcbinshash(inp) != 0) { inp->in6p_laddr = in6addr_any; inp->inp_lport = 0; return (EAGAIN); } return (0); } void addrsel_policy_init(void) { ADDRSEL_LOCK_INIT(); ADDRSEL_SXLOCK_INIT(); INIT_VNET_INET6(curvnet); V_ip6_prefer_tempaddr = 0; init_policy_queue(); /* initialize the "last resort" policy */ bzero(&V_defaultaddrpolicy, sizeof(V_defaultaddrpolicy)); V_defaultaddrpolicy.label = ADDR_LABEL_NOTAPP; } static struct in6_addrpolicy * lookup_addrsel_policy(struct sockaddr_in6 *key) { INIT_VNET_INET6(curvnet); struct in6_addrpolicy *match = NULL; ADDRSEL_LOCK(); match = match_addrsel_policy(key); if (match == NULL) match = &V_defaultaddrpolicy; else match->use++; ADDRSEL_UNLOCK(); return (match); } /* * Subroutines to manage the address selection policy table via sysctl. */ struct walkarg { struct sysctl_req *w_req; }; static int in6_src_sysctl(SYSCTL_HANDLER_ARGS); SYSCTL_DECL(_net_inet6_ip6); SYSCTL_NODE(_net_inet6_ip6, IPV6CTL_ADDRCTLPOLICY, addrctlpolicy, CTLFLAG_RD, in6_src_sysctl, ""); static int in6_src_sysctl(SYSCTL_HANDLER_ARGS) { struct walkarg w; if (req->newptr) return EPERM; bzero(&w, sizeof(w)); w.w_req = req; return (walk_addrsel_policy(dump_addrsel_policyent, &w)); } int in6_src_ioctl(u_long cmd, caddr_t data) { int i; struct in6_addrpolicy ent0; if (cmd != SIOCAADDRCTL_POLICY && cmd != SIOCDADDRCTL_POLICY) return (EOPNOTSUPP); /* check for safety */ ent0 = *(struct in6_addrpolicy *)data; if (ent0.label == ADDR_LABEL_NOTAPP) return (EINVAL); /* check if the prefix mask is consecutive. */ if (in6_mask2len(&ent0.addrmask.sin6_addr, NULL) < 0) return (EINVAL); /* clear trailing garbages (if any) of the prefix address. */ for (i = 0; i < 4; i++) { ent0.addr.sin6_addr.s6_addr32[i] &= ent0.addrmask.sin6_addr.s6_addr32[i]; } ent0.use = 0; switch (cmd) { case SIOCAADDRCTL_POLICY: return (add_addrsel_policyent(&ent0)); case SIOCDADDRCTL_POLICY: return (delete_addrsel_policyent(&ent0)); } return (0); /* XXX: compromise compilers */ } /* * The followings are implementation of the policy table using a * simple tail queue. * XXX such details should be hidden. * XXX implementation using binary tree should be more efficient. */ struct addrsel_policyent { TAILQ_ENTRY(addrsel_policyent) ape_entry; struct in6_addrpolicy ape_policy; }; TAILQ_HEAD(addrsel_policyhead, addrsel_policyent); #ifdef VIMAGE_GLOBALS struct addrsel_policyhead addrsel_policytab; #endif static void init_policy_queue(void) { INIT_VNET_INET6(curvnet); TAILQ_INIT(&V_addrsel_policytab); } static int add_addrsel_policyent(struct in6_addrpolicy *newpolicy) { INIT_VNET_INET6(curvnet); struct addrsel_policyent *new, *pol; new = malloc(sizeof(*new), M_IFADDR, M_WAITOK); ADDRSEL_XLOCK(); ADDRSEL_LOCK(); /* duplication check */ TAILQ_FOREACH(pol, &V_addrsel_policytab, ape_entry) { if (IN6_ARE_ADDR_EQUAL(&newpolicy->addr.sin6_addr, &pol->ape_policy.addr.sin6_addr) && IN6_ARE_ADDR_EQUAL(&newpolicy->addrmask.sin6_addr, &pol->ape_policy.addrmask.sin6_addr)) { ADDRSEL_UNLOCK(); ADDRSEL_XUNLOCK(); free(new, M_IFADDR); return (EEXIST); /* or override it? */ } } bzero(new, sizeof(*new)); /* XXX: should validate entry */ new->ape_policy = *newpolicy; TAILQ_INSERT_TAIL(&V_addrsel_policytab, new, ape_entry); ADDRSEL_UNLOCK(); ADDRSEL_XUNLOCK(); return (0); } static int delete_addrsel_policyent(struct in6_addrpolicy *key) { INIT_VNET_INET6(curvnet); struct addrsel_policyent *pol; ADDRSEL_XLOCK(); ADDRSEL_LOCK(); /* search for the entry in the table */ TAILQ_FOREACH(pol, &V_addrsel_policytab, ape_entry) { if (IN6_ARE_ADDR_EQUAL(&key->addr.sin6_addr, &pol->ape_policy.addr.sin6_addr) && IN6_ARE_ADDR_EQUAL(&key->addrmask.sin6_addr, &pol->ape_policy.addrmask.sin6_addr)) { break; } } if (pol == NULL) { ADDRSEL_UNLOCK(); ADDRSEL_XUNLOCK(); return (ESRCH); } TAILQ_REMOVE(&V_addrsel_policytab, pol, ape_entry); ADDRSEL_UNLOCK(); ADDRSEL_XUNLOCK(); return (0); } static int walk_addrsel_policy(int (*callback)(struct in6_addrpolicy *, void *), void *w) { INIT_VNET_INET6(curvnet); struct addrsel_policyent *pol; int error = 0; ADDRSEL_SLOCK(); TAILQ_FOREACH(pol, &V_addrsel_policytab, ape_entry) { if ((error = (*callback)(&pol->ape_policy, w)) != 0) { ADDRSEL_SUNLOCK(); return (error); } } ADDRSEL_SUNLOCK(); return (error); } static int dump_addrsel_policyent(struct in6_addrpolicy *pol, void *arg) { int error = 0; struct walkarg *w = arg; error = SYSCTL_OUT(w->w_req, pol, sizeof(*pol)); return (error); } static struct in6_addrpolicy * match_addrsel_policy(struct sockaddr_in6 *key) { INIT_VNET_INET6(curvnet); struct addrsel_policyent *pent; struct in6_addrpolicy *bestpol = NULL, *pol; int matchlen, bestmatchlen = -1; u_char *mp, *ep, *k, *p, m; TAILQ_FOREACH(pent, &V_addrsel_policytab, ape_entry) { matchlen = 0; pol = &pent->ape_policy; mp = (u_char *)&pol->addrmask.sin6_addr; ep = mp + 16; /* XXX: scope field? */ k = (u_char *)&key->sin6_addr; p = (u_char *)&pol->addr.sin6_addr; for (; mp < ep && *mp; mp++, k++, p++) { m = *mp; if ((*k & m) != *p) goto next; /* not match */ if (m == 0xff) /* short cut for a typical case */ matchlen += 8; else { while (m >= 0x80) { matchlen++; m <<= 1; } } } /* matched. check if this is better than the current best. */ if (bestpol == NULL || matchlen > bestmatchlen) { bestpol = pol; bestmatchlen = matchlen; } next: continue; } return (bestpol); } diff --git a/sys/netinet6/raw_ip6.c b/sys/netinet6/raw_ip6.c index 18b8131fc6eb..da781a06a027 100644 --- a/sys/netinet6/raw_ip6.c +++ b/sys/netinet6/raw_ip6.c @@ -1,852 +1,848 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 3. Neither the name of the project 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 PROJECT 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 PROJECT 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. */ /*- * Copyright (c) 1982, 1986, 1988, 1993 * The Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)raw_ip.c 8.2 (Berkeley) 1/4/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_ipsec.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef IPSEC #include #include #endif /* IPSEC */ #include #define satosin6(sa) ((struct sockaddr_in6 *)(sa)) #define ifatoia6(ifa) ((struct in6_ifaddr *)(ifa)) /* * Raw interface to IP6 protocol. */ #ifdef VIMAGE_GLOBALS extern struct inpcbhead ripcb; extern struct inpcbinfo ripcbinfo; struct rip6stat rip6stat; #endif extern u_long rip_sendspace; extern u_long rip_recvspace; /* * Hooks for multicast forwarding. */ struct socket *ip6_mrouter = NULL; int (*ip6_mrouter_set)(struct socket *, struct sockopt *); int (*ip6_mrouter_get)(struct socket *, struct sockopt *); int (*ip6_mrouter_done)(void); int (*ip6_mforward)(struct ip6_hdr *, struct ifnet *, struct mbuf *); int (*mrt6_ioctl)(int, caddr_t); /* * Setup generic address and protocol structures for raw_input routine, then * pass them along with mbuf chain. */ int rip6_input(struct mbuf **mp, int *offp, int proto) { INIT_VNET_INET(curvnet); INIT_VNET_INET6(curvnet); #ifdef IPSEC INIT_VNET_IPSEC(curvnet); #endif struct mbuf *m = *mp; register struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); register struct inpcb *in6p; struct inpcb *last = 0; struct mbuf *opts = NULL; struct sockaddr_in6 fromsa; V_rip6stat.rip6s_ipackets++; if (faithprefix_p != NULL && (*faithprefix_p)(&ip6->ip6_dst)) { /* XXX Send icmp6 host/port unreach? */ m_freem(m); return (IPPROTO_DONE); } init_sin6(&fromsa, m); /* general init */ INP_INFO_RLOCK(&V_ripcbinfo); LIST_FOREACH(in6p, &V_ripcb, inp_list) { /* XXX inp locking */ if ((in6p->inp_vflag & INP_IPV6) == 0) continue; if (in6p->inp_ip_p && in6p->inp_ip_p != proto) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) && !IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &ip6->ip6_dst)) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) && !IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &ip6->ip6_src)) continue; - if (jailed(in6p->inp_cred)) { - if (!prison_check_ip6(in6p->inp_cred, &ip6->ip6_dst)) - continue; - } + if (prison_check_ip6(in6p->inp_cred, &ip6->ip6_dst) != 0) + continue; INP_RLOCK(in6p); if (in6p->in6p_cksum != -1) { V_rip6stat.rip6s_isum++; if (in6_cksum(m, proto, *offp, m->m_pkthdr.len - *offp)) { INP_RUNLOCK(in6p); V_rip6stat.rip6s_badsum++; continue; } } if (last != NULL) { struct mbuf *n = m_copy(m, 0, (int)M_COPYALL); #ifdef IPSEC /* * Check AH/ESP integrity. */ if (n && ipsec6_in_reject(n, last)) { m_freem(n); V_ipsec6stat.in_polvio++; /* Do not inject data into pcb. */ } else #endif /* IPSEC */ if (n) { if (last->inp_flags & INP_CONTROLOPTS || last->inp_socket->so_options & SO_TIMESTAMP) ip6_savecontrol(last, n, &opts); /* strip intermediate headers */ m_adj(n, *offp); if (sbappendaddr(&last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, n, opts) == 0) { m_freem(n); if (opts) m_freem(opts); V_rip6stat.rip6s_fullsock++; } else sorwakeup(last->inp_socket); opts = NULL; } INP_RUNLOCK(last); } last = in6p; } INP_INFO_RUNLOCK(&V_ripcbinfo); #ifdef IPSEC /* * Check AH/ESP integrity. */ if ((last != NULL) && ipsec6_in_reject(m, last)) { m_freem(m); V_ipsec6stat.in_polvio++; V_ip6stat.ip6s_delivered--; /* Do not inject data into pcb. */ INP_RUNLOCK(last); } else #endif /* IPSEC */ if (last != NULL) { if (last->inp_flags & INP_CONTROLOPTS || last->inp_socket->so_options & SO_TIMESTAMP) ip6_savecontrol(last, m, &opts); /* Strip intermediate headers. */ m_adj(m, *offp); if (sbappendaddr(&last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, m, opts) == 0) { m_freem(m); if (opts) m_freem(opts); V_rip6stat.rip6s_fullsock++; } else sorwakeup(last->inp_socket); INP_RUNLOCK(last); } else { V_rip6stat.rip6s_nosock++; if (m->m_flags & M_MCAST) V_rip6stat.rip6s_nosockmcast++; if (proto == IPPROTO_NONE) m_freem(m); else { char *prvnxtp = ip6_get_prevhdr(m, *offp); /* XXX */ icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_NEXTHEADER, prvnxtp - mtod(m, char *)); } V_ip6stat.ip6s_delivered--; } return (IPPROTO_DONE); } void rip6_ctlinput(int cmd, struct sockaddr *sa, void *d) { INIT_VNET_INET(curvnet); struct ip6_hdr *ip6; struct mbuf *m; int off = 0; struct ip6ctlparam *ip6cp = NULL; const struct sockaddr_in6 *sa6_src = NULL; void *cmdarg; struct inpcb *(*notify)(struct inpcb *, int) = in6_rtchange; if (sa->sa_family != AF_INET6 || sa->sa_len != sizeof(struct sockaddr_in6)) return; if ((unsigned)cmd >= PRC_NCMDS) return; if (PRC_IS_REDIRECT(cmd)) notify = in6_rtchange, d = NULL; else if (cmd == PRC_HOSTDEAD) d = NULL; else if (inet6ctlerrmap[cmd] == 0) return; /* * If the parameter is from icmp6, decode it. */ if (d != NULL) { ip6cp = (struct ip6ctlparam *)d; m = ip6cp->ip6c_m; ip6 = ip6cp->ip6c_ip6; off = ip6cp->ip6c_off; cmdarg = ip6cp->ip6c_cmdarg; sa6_src = ip6cp->ip6c_src; } else { m = NULL; ip6 = NULL; cmdarg = NULL; sa6_src = &sa6_any; } (void) in6_pcbnotify(&V_ripcbinfo, sa, 0, (const struct sockaddr *)sa6_src, 0, cmd, cmdarg, notify); } /* * Generate IPv6 header and pass packet to ip6_output. Tack on options user * may have setup with control call. */ int #if __STDC__ rip6_output(struct mbuf *m, ...) #else rip6_output(m, va_alist) struct mbuf *m; va_dcl #endif { INIT_VNET_INET6(curvnet); struct mbuf *control; struct socket *so; struct sockaddr_in6 *dstsock; struct in6_addr *dst; struct ip6_hdr *ip6; struct inpcb *in6p; u_int plen = m->m_pkthdr.len; int error = 0; struct ip6_pktopts opt, *optp; struct ifnet *oifp = NULL; int type = 0, code = 0; /* for ICMPv6 output statistics only */ int scope_ambiguous = 0; struct in6_addr *in6a; va_list ap; va_start(ap, m); so = va_arg(ap, struct socket *); dstsock = va_arg(ap, struct sockaddr_in6 *); control = va_arg(ap, struct mbuf *); va_end(ap); in6p = sotoinpcb(so); INP_WLOCK(in6p); dst = &dstsock->sin6_addr; if (control != NULL) { if ((error = ip6_setpktopts(control, &opt, in6p->in6p_outputopts, so->so_cred, so->so_proto->pr_protocol)) != 0) { goto bad; } optp = &opt; } else optp = in6p->in6p_outputopts; /* * Check and convert scope zone ID into internal form. * * XXX: we may still need to determine the zone later. */ if (!(so->so_state & SS_ISCONNECTED)) { if (dstsock->sin6_scope_id == 0 && !V_ip6_use_defzone) scope_ambiguous = 1; if ((error = sa6_embedscope(dstsock, V_ip6_use_defzone)) != 0) goto bad; } /* * For an ICMPv6 packet, we should know its type and code to update * statistics. */ if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) { struct icmp6_hdr *icmp6; if (m->m_len < sizeof(struct icmp6_hdr) && (m = m_pullup(m, sizeof(struct icmp6_hdr))) == NULL) { error = ENOBUFS; goto bad; } icmp6 = mtod(m, struct icmp6_hdr *); type = icmp6->icmp6_type; code = icmp6->icmp6_code; } M_PREPEND(m, sizeof(*ip6), M_DONTWAIT); if (m == NULL) { error = ENOBUFS; goto bad; } ip6 = mtod(m, struct ip6_hdr *); /* * Source address selection. */ if ((in6a = in6_selectsrc(dstsock, optp, in6p, NULL, so->so_cred, &oifp, &error)) == NULL) { if (error == 0) error = EADDRNOTAVAIL; goto bad; } - if (jailed(in6p->inp_cred)) - if (prison_get_ip6(in6p->inp_cred, in6a) != 0) { - error = EPERM; - goto bad; - } + error = prison_get_ip6(in6p->inp_cred, in6a); + if (error != 0) + goto bad; ip6->ip6_src = *in6a; if (oifp && scope_ambiguous) { /* * Application should provide a proper zone ID or the use of * default zone IDs should be enabled. Unfortunately, some * applications do not behave as it should, so we need a * workaround. Even if an appropriate ID is not determined * (when it's required), if we can determine the outgoing * interface. determine the zone ID based on the interface. */ error = in6_setscope(&dstsock->sin6_addr, oifp, NULL); if (error != 0) goto bad; } ip6->ip6_dst = dstsock->sin6_addr; /* * Fill in the rest of the IPv6 header fields. */ ip6->ip6_flow = (ip6->ip6_flow & ~IPV6_FLOWINFO_MASK) | (in6p->inp_flow & IPV6_FLOWINFO_MASK); ip6->ip6_vfc = (ip6->ip6_vfc & ~IPV6_VERSION_MASK) | (IPV6_VERSION & IPV6_VERSION_MASK); /* * ip6_plen will be filled in ip6_output, so not fill it here. */ ip6->ip6_nxt = in6p->inp_ip_p; ip6->ip6_hlim = in6_selecthlim(in6p, oifp); if (so->so_proto->pr_protocol == IPPROTO_ICMPV6 || in6p->in6p_cksum != -1) { struct mbuf *n; int off; u_int16_t *p; /* Compute checksum. */ if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) off = offsetof(struct icmp6_hdr, icmp6_cksum); else off = in6p->in6p_cksum; if (plen < off + 1) { error = EINVAL; goto bad; } off += sizeof(struct ip6_hdr); n = m; while (n && n->m_len <= off) { off -= n->m_len; n = n->m_next; } if (!n) goto bad; p = (u_int16_t *)(mtod(n, caddr_t) + off); *p = 0; *p = in6_cksum(m, ip6->ip6_nxt, sizeof(*ip6), plen); } error = ip6_output(m, optp, NULL, 0, in6p->in6p_moptions, &oifp, in6p); if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) { if (oifp) icmp6_ifoutstat_inc(oifp, type, code); V_icmp6stat.icp6s_outhist[type]++; } else V_rip6stat.rip6s_opackets++; goto freectl; bad: if (m) m_freem(m); freectl: if (control != NULL) { ip6_clearpktopts(&opt, -1); m_freem(control); } INP_WUNLOCK(in6p); return (error); } /* * Raw IPv6 socket option processing. */ int rip6_ctloutput(struct socket *so, struct sockopt *sopt) { int error; if (sopt->sopt_level == IPPROTO_ICMPV6) /* * XXX: is it better to call icmp6_ctloutput() directly * from protosw? */ return (icmp6_ctloutput(so, sopt)); else if (sopt->sopt_level != IPPROTO_IPV6) return (EINVAL); error = 0; switch (sopt->sopt_dir) { case SOPT_GET: switch (sopt->sopt_name) { case MRT6_INIT: case MRT6_DONE: case MRT6_ADD_MIF: case MRT6_DEL_MIF: case MRT6_ADD_MFC: case MRT6_DEL_MFC: case MRT6_PIM: error = ip6_mrouter_get ? ip6_mrouter_get(so, sopt) : EOPNOTSUPP; break; case IPV6_CHECKSUM: error = ip6_raw_ctloutput(so, sopt); break; default: error = ip6_ctloutput(so, sopt); break; } break; case SOPT_SET: switch (sopt->sopt_name) { case MRT6_INIT: case MRT6_DONE: case MRT6_ADD_MIF: case MRT6_DEL_MIF: case MRT6_ADD_MFC: case MRT6_DEL_MFC: case MRT6_PIM: error = ip6_mrouter_set ? ip6_mrouter_set(so, sopt) : EOPNOTSUPP; break; case IPV6_CHECKSUM: error = ip6_raw_ctloutput(so, sopt); break; default: error = ip6_ctloutput(so, sopt); break; } break; } return (error); } static int rip6_attach(struct socket *so, int proto, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; struct icmp6_filter *filter; int error; inp = sotoinpcb(so); KASSERT(inp == NULL, ("rip6_attach: inp != NULL")); error = priv_check(td, PRIV_NETINET_RAW); if (error) return (error); error = soreserve(so, rip_sendspace, rip_recvspace); if (error) return (error); filter = malloc(sizeof(struct icmp6_filter), M_PCB, M_NOWAIT); if (filter == NULL) return (ENOMEM); INP_INFO_WLOCK(&V_ripcbinfo); error = in_pcballoc(so, &V_ripcbinfo); if (error) { INP_INFO_WUNLOCK(&V_ripcbinfo); free(filter, M_PCB); return (error); } inp = (struct inpcb *)so->so_pcb; INP_INFO_WUNLOCK(&V_ripcbinfo); inp->inp_vflag |= INP_IPV6; inp->inp_ip_p = (long)proto; inp->in6p_hops = -1; /* use kernel default */ inp->in6p_cksum = -1; inp->in6p_icmp6filt = filter; ICMP6_FILTER_SETPASSALL(inp->in6p_icmp6filt); INP_WUNLOCK(inp); return (0); } static void rip6_detach(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_detach: inp == NULL")); if (so == ip6_mrouter && ip6_mrouter_done) ip6_mrouter_done(); /* xxx: RSVP */ INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); free(inp->in6p_icmp6filt, M_PCB); in_pcbdetach(inp); in_pcbfree(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); } /* XXXRW: This can't ever be called. */ static void rip6_abort(struct socket *so) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_abort: inp == NULL")); soisdisconnected(so); } static void rip6_close(struct socket *so) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_close: inp == NULL")); soisdisconnected(so); } static int rip6_disconnect(struct socket *so) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_disconnect: inp == NULL")); if ((so->so_state & SS_ISCONNECTED) == 0) return (ENOTCONN); inp->in6p_faddr = in6addr_any; rip6_abort(so); return (0); } static int rip6_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_NET(so->so_vnet); INIT_VNET_INET(so->so_vnet); INIT_VNET_INET6(so->so_vnet); struct inpcb *inp; struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam; struct ifaddr *ia = NULL; int error = 0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_bind: inp == NULL")); if (nam->sa_len != sizeof(*addr)) return (EINVAL); - if (!prison_check_ip6(td->td_ucred, &addr->sin6_addr)) - return (EADDRNOTAVAIL); + if ((error = prison_check_ip6(td->td_ucred, &addr->sin6_addr)) != 0) + return (error); if (TAILQ_EMPTY(&V_ifnet) || addr->sin6_family != AF_INET6) return (EADDRNOTAVAIL); if ((error = sa6_embedscope(addr, V_ip6_use_defzone)) != 0) return (error); if (!IN6_IS_ADDR_UNSPECIFIED(&addr->sin6_addr) && (ia = ifa_ifwithaddr((struct sockaddr *)addr)) == 0) return (EADDRNOTAVAIL); if (ia && ((struct in6_ifaddr *)ia)->ia6_flags & (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY| IN6_IFF_DETACHED|IN6_IFF_DEPRECATED)) { return (EADDRNOTAVAIL); } INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); inp->in6p_laddr = addr->sin6_addr; INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (0); } static int rip6_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_NET(so->so_vnet); INIT_VNET_INET(so->so_vnet); INIT_VNET_INET6(so->so_vnet); struct inpcb *inp; struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam; struct in6_addr *in6a = NULL; struct ifnet *ifp = NULL; int error = 0, scope_ambiguous = 0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_connect: inp == NULL")); if (nam->sa_len != sizeof(*addr)) return (EINVAL); if (TAILQ_EMPTY(&V_ifnet)) return (EADDRNOTAVAIL); if (addr->sin6_family != AF_INET6) return (EAFNOSUPPORT); /* * Application should provide a proper zone ID or the use of default * zone IDs should be enabled. Unfortunately, some applications do * not behave as it should, so we need a workaround. Even if an * appropriate ID is not determined, we'll see if we can determine * the outgoing interface. If we can, determine the zone ID based on * the interface below. */ if (addr->sin6_scope_id == 0 && !V_ip6_use_defzone) scope_ambiguous = 1; if ((error = sa6_embedscope(addr, V_ip6_use_defzone)) != 0) return (error); INP_INFO_WLOCK(&V_ripcbinfo); INP_WLOCK(inp); /* Source address selection. XXX: need pcblookup? */ in6a = in6_selectsrc(addr, inp->in6p_outputopts, inp, NULL, so->so_cred, &ifp, &error); if (in6a == NULL) { INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (error ? error : EADDRNOTAVAIL); } /* XXX: see above */ if (ifp && scope_ambiguous && (error = in6_setscope(&addr->sin6_addr, ifp, NULL)) != 0) { INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (error); } inp->in6p_faddr = addr->sin6_addr; inp->in6p_laddr = *in6a; soisconnected(so); INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_ripcbinfo); return (0); } static int rip6_shutdown(struct socket *so) { struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_shutdown: inp == NULL")); INP_WLOCK(inp); socantsendmore(so); INP_WUNLOCK(inp); return (0); } static int rip6_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, struct mbuf *control, struct thread *td) { struct inpcb *inp; struct sockaddr_in6 tmp; struct sockaddr_in6 *dst; int ret; inp = sotoinpcb(so); KASSERT(inp != NULL, ("rip6_send: inp == NULL")); /* Always copy sockaddr to avoid overwrites. */ /* Unlocked read. */ if (so->so_state & SS_ISCONNECTED) { if (nam) { m_freem(m); return (EISCONN); } /* XXX */ bzero(&tmp, sizeof(tmp)); tmp.sin6_family = AF_INET6; tmp.sin6_len = sizeof(struct sockaddr_in6); INP_RLOCK(inp); bcopy(&inp->in6p_faddr, &tmp.sin6_addr, sizeof(struct in6_addr)); INP_RUNLOCK(inp); dst = &tmp; } else { if (nam == NULL) { m_freem(m); return (ENOTCONN); } if (nam->sa_len != sizeof(struct sockaddr_in6)) { m_freem(m); return (EINVAL); } tmp = *(struct sockaddr_in6 *)nam; dst = &tmp; if (dst->sin6_family == AF_UNSPEC) { /* * XXX: we allow this case for backward * compatibility to buggy applications that * rely on old (and wrong) kernel behavior. */ log(LOG_INFO, "rip6 SEND: address family is " "unspec. Assume AF_INET6\n"); dst->sin6_family = AF_INET6; } else if (dst->sin6_family != AF_INET6) { m_freem(m); return(EAFNOSUPPORT); } } ret = rip6_output(m, so, dst, control); return (ret); } struct pr_usrreqs rip6_usrreqs = { .pru_abort = rip6_abort, .pru_attach = rip6_attach, .pru_bind = rip6_bind, .pru_connect = rip6_connect, .pru_control = in6_control, .pru_detach = rip6_detach, .pru_disconnect = rip6_disconnect, .pru_peeraddr = in6_getpeeraddr, .pru_send = rip6_send, .pru_shutdown = rip6_shutdown, .pru_sockaddr = in6_getsockaddr, .pru_close = rip6_close, }; diff --git a/sys/netinet6/udp6_usrreq.c b/sys/netinet6/udp6_usrreq.c index dea8ed220134..d0bcfc02132d 100644 --- a/sys/netinet6/udp6_usrreq.c +++ b/sys/netinet6/udp6_usrreq.c @@ -1,1104 +1,1100 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 3. Neither the name of the project 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 PROJECT 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 PROJECT 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: udp6_usrreq.c,v 1.27 2001/05/21 05:45:10 jinmei Exp $ * $KAME: udp6_output.c,v 1.31 2001/05/21 16:39:15 jinmei Exp $ */ /*- * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995 * The Regents of the University of California. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. 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. * 4. Neither the name of the University 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 REGENTS 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 REGENTS 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. * * @(#)udp_usrreq.c 8.6 (Berkeley) 5/23/95 */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_mac.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef IPSEC #include #include #endif /* IPSEC */ #include /* * UDP protocol implementation. * Per RFC 768, August, 1980. */ extern struct protosw inetsw[]; static void udp6_detach(struct socket *so); static void udp6_append(struct inpcb *inp, struct mbuf *n, int off, struct sockaddr_in6 *fromsa) { INIT_VNET_INET(inp->inp_vnet); struct socket *so; struct mbuf *opts; INP_LOCK_ASSERT(inp); #ifdef IPSEC /* Check AH/ESP integrity. */ if (ipsec6_in_reject(n, inp)) { INIT_VNET_IPSEC(inp->inp_vnet); m_freem(n); V_ipsec6stat.in_polvio++; return; } #endif /* IPSEC */ #ifdef MAC if (mac_inpcb_check_deliver(inp, n) != 0) { m_freem(n); return; } #endif opts = NULL; if (inp->inp_flags & INP_CONTROLOPTS || inp->inp_socket->so_options & SO_TIMESTAMP) ip6_savecontrol(inp, n, &opts); m_adj(n, off + sizeof(struct udphdr)); so = inp->inp_socket; SOCKBUF_LOCK(&so->so_rcv); if (sbappendaddr_locked(&so->so_rcv, (struct sockaddr *)fromsa, n, opts) == 0) { SOCKBUF_UNLOCK(&so->so_rcv); m_freem(n); if (opts) m_freem(opts); V_udpstat.udps_fullsock++; } else sorwakeup_locked(so); } int udp6_input(struct mbuf **mp, int *offp, int proto) { INIT_VNET_INET(curvnet); INIT_VNET_INET6(curvnet); struct mbuf *m = *mp; struct ip6_hdr *ip6; struct udphdr *uh; struct inpcb *inp; int off = *offp; int plen, ulen; struct sockaddr_in6 fromsa; ip6 = mtod(m, struct ip6_hdr *); if (faithprefix_p != NULL && (*faithprefix_p)(&ip6->ip6_dst)) { /* XXX send icmp6 host/port unreach? */ m_freem(m); return (IPPROTO_DONE); } #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, sizeof(struct udphdr), IPPROTO_DONE); ip6 = mtod(m, struct ip6_hdr *); uh = (struct udphdr *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(uh, struct udphdr *, m, off, sizeof(*uh)); if (!uh) return (IPPROTO_DONE); #endif V_udpstat.udps_ipackets++; /* * Destination port of 0 is illegal, based on RFC768. */ if (uh->uh_dport == 0) goto badunlocked; plen = ntohs(ip6->ip6_plen) - off + sizeof(*ip6); ulen = ntohs((u_short)uh->uh_ulen); if (plen != ulen) { V_udpstat.udps_badlen++; goto badunlocked; } /* * Checksum extended UDP header and data. */ if (uh->uh_sum == 0) { V_udpstat.udps_nosum++; goto badunlocked; } if (in6_cksum(m, IPPROTO_UDP, off, ulen) != 0) { V_udpstat.udps_badsum++; goto badunlocked; } /* * Construct sockaddr format source address. */ init_sin6(&fromsa, m); fromsa.sin6_port = uh->uh_sport; INP_INFO_RLOCK(&V_udbinfo); if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { struct inpcb *last; /* * In the event that laddr should be set to the link-local * address (this happens in RIPng), the multicast address * specified in the received packet will not match laddr. To * handle this situation, matching is relaxed if the * receiving interface is the same as one specified in the * socket and if the destination multicast address matches * one of the multicast groups specified in the socket. */ /* * KAME note: traditionally we dropped udpiphdr from mbuf * here. We need udphdr for IPsec processing so we do that * later. */ last = NULL; LIST_FOREACH(inp, &V_udb, inp_list) { if ((inp->inp_vflag & INP_IPV6) == 0) continue; if (inp->inp_lport != uh->uh_dport) continue; /* * XXX: Do not check source port of incoming datagram * unless inp_connect() has been called to bind the * fport part of the 4-tuple; the source could be * trying to talk to us with an ephemeral port. */ if (inp->inp_fport != 0 && inp->inp_fport != uh->uh_sport) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) { if (!IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, &ip6->ip6_dst)) continue; } if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { if (!IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, &ip6->ip6_src) || inp->inp_fport != uh->uh_sport) continue; } if (last != NULL) { struct mbuf *n; if ((n = m_copy(m, 0, M_COPYALL)) != NULL) { INP_RLOCK(last); if (last->inp_ppcb != NULL) { /* * Engage the tunneling * protocol we will have to * leave the info_lock up, * since we are hunting * through multiple UDP's. * */ udp_tun_func_t tunnel_func; tunnel_func = (udp_tun_func_t)last->inp_ppcb; tunnel_func(n, off, last); INP_RUNLOCK(last); } else { udp6_append(last, n, off, &fromsa); INP_RUNLOCK(last); } } } last = inp; /* * Don't look for additional matches if this one does * not have either the SO_REUSEPORT or SO_REUSEADDR * socket options set. This heuristic avoids * searching through all pcbs in the common case of a * non-shared port. It assumes that an application * will never clear these options after setting them. */ if ((last->inp_socket->so_options & (SO_REUSEPORT|SO_REUSEADDR)) == 0) break; } if (last == NULL) { /* * No matching pcb found; discard datagram. (No need * to send an ICMP Port Unreachable for a broadcast * or multicast datgram.) */ V_udpstat.udps_noport++; V_udpstat.udps_noportmcast++; goto badheadlocked; } INP_RLOCK(last); INP_INFO_RUNLOCK(&V_udbinfo); if (last->inp_ppcb != NULL) { /* * Engage the tunneling protocol. */ udp_tun_func_t tunnel_func; tunnel_func = (udp_tun_func_t)inp->inp_ppcb; tunnel_func(m, off, last); INP_RUNLOCK(last); return (IPPROTO_DONE); } udp6_append(last, m, off, &fromsa); INP_RUNLOCK(last); return (IPPROTO_DONE); } /* * Locate pcb for datagram. */ inp = in6_pcblookup_hash(&V_udbinfo, &ip6->ip6_src, uh->uh_sport, &ip6->ip6_dst, uh->uh_dport, 1, m->m_pkthdr.rcvif); if (inp == NULL) { if (udp_log_in_vain) { char ip6bufs[INET6_ADDRSTRLEN]; char ip6bufd[INET6_ADDRSTRLEN]; log(LOG_INFO, "Connection attempt to UDP [%s]:%d from [%s]:%d\n", ip6_sprintf(ip6bufd, &ip6->ip6_dst), ntohs(uh->uh_dport), ip6_sprintf(ip6bufs, &ip6->ip6_src), ntohs(uh->uh_sport)); } V_udpstat.udps_noport++; if (m->m_flags & M_MCAST) { printf("UDP6: M_MCAST is set in a unicast packet.\n"); V_udpstat.udps_noportmcast++; goto badheadlocked; } INP_INFO_RUNLOCK(&V_udbinfo); if (V_udp_blackhole) goto badunlocked; if (badport_bandlim(BANDLIM_ICMP6_UNREACH) < 0) goto badunlocked; icmp6_error(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_NOPORT, 0); return (IPPROTO_DONE); } INP_RLOCK(inp); INP_INFO_RUNLOCK(&V_udbinfo); if (inp->inp_ppcb != NULL) { /* * Engage the tunneling protocol. */ udp_tun_func_t tunnel_func; tunnel_func = (udp_tun_func_t)inp->inp_ppcb; tunnel_func(m, off, inp); INP_RUNLOCK(inp); return (IPPROTO_DONE); } udp6_append(inp, m, off, &fromsa); INP_RUNLOCK(inp); return (IPPROTO_DONE); badheadlocked: INP_INFO_RUNLOCK(&V_udbinfo); badunlocked: if (m) m_freem(m); return (IPPROTO_DONE); } void udp6_ctlinput(int cmd, struct sockaddr *sa, void *d) { INIT_VNET_INET(curvnet); struct udphdr uh; struct ip6_hdr *ip6; struct mbuf *m; int off = 0; struct ip6ctlparam *ip6cp = NULL; const struct sockaddr_in6 *sa6_src = NULL; void *cmdarg; struct inpcb *(*notify)(struct inpcb *, int) = udp_notify; struct udp_portonly { u_int16_t uh_sport; u_int16_t uh_dport; } *uhp; if (sa->sa_family != AF_INET6 || sa->sa_len != sizeof(struct sockaddr_in6)) return; if ((unsigned)cmd >= PRC_NCMDS) return; if (PRC_IS_REDIRECT(cmd)) notify = in6_rtchange, d = NULL; else if (cmd == PRC_HOSTDEAD) d = NULL; else if (inet6ctlerrmap[cmd] == 0) return; /* if the parameter is from icmp6, decode it. */ if (d != NULL) { ip6cp = (struct ip6ctlparam *)d; m = ip6cp->ip6c_m; ip6 = ip6cp->ip6c_ip6; off = ip6cp->ip6c_off; cmdarg = ip6cp->ip6c_cmdarg; sa6_src = ip6cp->ip6c_src; } else { m = NULL; ip6 = NULL; cmdarg = NULL; sa6_src = &sa6_any; } if (ip6) { /* * XXX: We assume that when IPV6 is non NULL, * M and OFF are valid. */ /* Check if we can safely examine src and dst ports. */ if (m->m_pkthdr.len < off + sizeof(*uhp)) return; bzero(&uh, sizeof(uh)); m_copydata(m, off, sizeof(*uhp), (caddr_t)&uh); (void) in6_pcbnotify(&V_udbinfo, sa, uh.uh_dport, (struct sockaddr *)ip6cp->ip6c_src, uh.uh_sport, cmd, cmdarg, notify); } else (void) in6_pcbnotify(&V_udbinfo, sa, 0, (const struct sockaddr *)sa6_src, 0, cmd, cmdarg, notify); } static int udp6_getcred(SYSCTL_HANDLER_ARGS) { INIT_VNET_INET(curvnet); INIT_VNET_INET6(curvnet); struct xucred xuc; struct sockaddr_in6 addrs[2]; struct inpcb *inp; int error; error = priv_check(req->td, PRIV_NETINET_GETCRED); if (error) return (error); if (req->newlen != sizeof(addrs)) return (EINVAL); if (req->oldlen != sizeof(struct xucred)) return (EINVAL); error = SYSCTL_IN(req, addrs, sizeof(addrs)); if (error) return (error); if ((error = sa6_embedscope(&addrs[0], V_ip6_use_defzone)) != 0 || (error = sa6_embedscope(&addrs[1], V_ip6_use_defzone)) != 0) { return (error); } INP_INFO_RLOCK(&V_udbinfo); inp = in6_pcblookup_hash(&V_udbinfo, &addrs[1].sin6_addr, addrs[1].sin6_port, &addrs[0].sin6_addr, addrs[0].sin6_port, 1, NULL); if (inp != NULL) { INP_RLOCK(inp); INP_INFO_RUNLOCK(&V_udbinfo); if (inp->inp_socket == NULL) error = ENOENT; if (error == 0) error = cr_canseesocket(req->td->td_ucred, inp->inp_socket); if (error == 0) cru2x(inp->inp_cred, &xuc); INP_RUNLOCK(inp); } else { INP_INFO_RUNLOCK(&V_udbinfo); error = ENOENT; } if (error == 0) error = SYSCTL_OUT(req, &xuc, sizeof(struct xucred)); return (error); } SYSCTL_PROC(_net_inet6_udp6, OID_AUTO, getcred, CTLTYPE_OPAQUE|CTLFLAG_RW, 0, 0, udp6_getcred, "S,xucred", "Get the xucred of a UDP6 connection"); static int udp6_output(struct inpcb *inp, struct mbuf *m, struct sockaddr *addr6, struct mbuf *control, struct thread *td) { INIT_VNET_INET(curvnet); INIT_VNET_INET6(curvnet); u_int32_t ulen = m->m_pkthdr.len; u_int32_t plen = sizeof(struct udphdr) + ulen; struct ip6_hdr *ip6; struct udphdr *udp6; struct in6_addr *laddr, *faddr; struct sockaddr_in6 *sin6 = NULL; struct ifnet *oifp = NULL; int scope_ambiguous = 0; u_short fport; int error = 0; struct ip6_pktopts *optp, opt; int af = AF_INET6, hlen = sizeof(struct ip6_hdr); int flags; struct sockaddr_in6 tmp; INP_WLOCK_ASSERT(inp); if (addr6) { /* addr6 has been validated in udp6_send(). */ sin6 = (struct sockaddr_in6 *)addr6; /* protect *sin6 from overwrites */ tmp = *sin6; sin6 = &tmp; /* * Application should provide a proper zone ID or the use of * default zone IDs should be enabled. Unfortunately, some * applications do not behave as it should, so we need a * workaround. Even if an appropriate ID is not determined, * we'll see if we can determine the outgoing interface. If we * can, determine the zone ID based on the interface below. */ if (sin6->sin6_scope_id == 0 && !V_ip6_use_defzone) scope_ambiguous = 1; if ((error = sa6_embedscope(sin6, V_ip6_use_defzone)) != 0) return (error); } if (control) { if ((error = ip6_setpktopts(control, &opt, inp->in6p_outputopts, td->td_ucred, IPPROTO_UDP)) != 0) goto release; optp = &opt; } else optp = inp->in6p_outputopts; if (sin6) { faddr = &sin6->sin6_addr; /* * IPv4 version of udp_output calls in_pcbconnect in this case, * which needs splnet and affects performance. * Since we saw no essential reason for calling in_pcbconnect, * we get rid of such kind of logic, and call in6_selectsrc * and in6_pcbsetport in order to fill in the local address * and the local port. */ if (sin6->sin6_port == 0) { error = EADDRNOTAVAIL; goto release; } if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { /* how about ::ffff:0.0.0.0 case? */ error = EISCONN; goto release; } fport = sin6->sin6_port; /* allow 0 port */ if (IN6_IS_ADDR_V4MAPPED(faddr)) { if ((inp->inp_flags & IN6P_IPV6_V6ONLY)) { /* * I believe we should explicitly discard the * packet when mapped addresses are disabled, * rather than send the packet as an IPv6 one. * If we chose the latter approach, the packet * might be sent out on the wire based on the * default route, the situation which we'd * probably want to avoid. * (20010421 jinmei@kame.net) */ error = EINVAL; goto release; } if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr) && !IN6_IS_ADDR_V4MAPPED(&inp->in6p_laddr)) { /* * when remote addr is an IPv4-mapped address, * local addr should not be an IPv6 address, * since you cannot determine how to map IPv6 * source address to IPv4. */ error = EINVAL; goto release; } af = AF_INET; } if (!IN6_IS_ADDR_V4MAPPED(faddr)) { laddr = in6_selectsrc(sin6, optp, inp, NULL, td->td_ucred, &oifp, &error); if (oifp && scope_ambiguous && (error = in6_setscope(&sin6->sin6_addr, oifp, NULL))) { goto release; } } else laddr = &inp->in6p_laddr; /* XXX */ if (laddr == NULL) { if (error == 0) error = EADDRNOTAVAIL; goto release; } if (inp->inp_lport == 0 && (error = in6_pcbsetport(laddr, inp, td->td_ucred)) != 0) goto release; } else { if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { error = ENOTCONN; goto release; } if (IN6_IS_ADDR_V4MAPPED(&inp->in6p_faddr)) { if ((inp->inp_flags & IN6P_IPV6_V6ONLY)) { /* * XXX: this case would happen when the * application sets the V6ONLY flag after * connecting the foreign address. * Such applications should be fixed, * so we bark here. */ log(LOG_INFO, "udp6_output: IPV6_V6ONLY " "option was set for a connected socket\n"); error = EINVAL; goto release; } else af = AF_INET; } laddr = &inp->in6p_laddr; faddr = &inp->in6p_faddr; fport = inp->inp_fport; } if (af == AF_INET) hlen = sizeof(struct ip); /* * Calculate data length and get a mbuf * for UDP and IP6 headers. */ M_PREPEND(m, hlen + sizeof(struct udphdr), M_DONTWAIT); if (m == 0) { error = ENOBUFS; goto release; } /* * Stuff checksum and output datagram. */ udp6 = (struct udphdr *)(mtod(m, caddr_t) + hlen); udp6->uh_sport = inp->inp_lport; /* lport is always set in the PCB */ udp6->uh_dport = fport; if (plen <= 0xffff) udp6->uh_ulen = htons((u_short)plen); else udp6->uh_ulen = 0; udp6->uh_sum = 0; switch (af) { case AF_INET6: ip6 = mtod(m, struct ip6_hdr *); ip6->ip6_flow = inp->inp_flow & IPV6_FLOWINFO_MASK; ip6->ip6_vfc &= ~IPV6_VERSION_MASK; ip6->ip6_vfc |= IPV6_VERSION; #if 0 /* ip6_plen will be filled in ip6_output. */ ip6->ip6_plen = htons((u_short)plen); #endif ip6->ip6_nxt = IPPROTO_UDP; ip6->ip6_hlim = in6_selecthlim(inp, NULL); ip6->ip6_src = *laddr; ip6->ip6_dst = *faddr; if ((udp6->uh_sum = in6_cksum(m, IPPROTO_UDP, sizeof(struct ip6_hdr), plen)) == 0) { udp6->uh_sum = 0xffff; } flags = 0; V_udpstat.udps_opackets++; error = ip6_output(m, optp, NULL, flags, inp->in6p_moptions, NULL, inp); break; case AF_INET: error = EAFNOSUPPORT; goto release; } goto releaseopt; release: m_freem(m); releaseopt: if (control) { ip6_clearpktopts(&opt, -1); m_freem(control); } return (error); } static void udp6_abort(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_abort: inp == NULL")); #ifdef INET if (inp->inp_vflag & INP_IPV4) { struct pr_usrreqs *pru; pru = inetsw[ip_protox[IPPROTO_UDP]].pr_usrreqs; (*pru->pru_abort)(so); return; } #endif INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { in6_pcbdisconnect(inp); inp->in6p_laddr = in6addr_any; soisdisconnected(so); } INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); } static int udp6_attach(struct socket *so, int proto, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp == NULL, ("udp6_attach: inp != NULL")); if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { error = soreserve(so, udp_sendspace, udp_recvspace); if (error) return (error); } INP_INFO_WLOCK(&V_udbinfo); error = in_pcballoc(so, &V_udbinfo); if (error) { INP_INFO_WUNLOCK(&V_udbinfo); return (error); } inp = (struct inpcb *)so->so_pcb; INP_INFO_WUNLOCK(&V_udbinfo); inp->inp_vflag |= INP_IPV6; if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) inp->inp_vflag |= INP_IPV4; inp->in6p_hops = -1; /* use kernel default */ inp->in6p_cksum = -1; /* just to be sure */ /* * XXX: ugly!! * IPv4 TTL initialization is necessary for an IPv6 socket as well, * because the socket may be bound to an IPv6 wildcard address, * which may match an IPv4-mapped IPv6 address. */ inp->inp_ip_ttl = V_ip_defttl; INP_WUNLOCK(inp); return (0); } static int udp6_bind(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_bind: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); inp->inp_vflag &= ~INP_IPV4; inp->inp_vflag |= INP_IPV6; if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) { struct sockaddr_in6 *sin6_p; sin6_p = (struct sockaddr_in6 *)nam; if (IN6_IS_ADDR_UNSPECIFIED(&sin6_p->sin6_addr)) inp->inp_vflag |= INP_IPV4; else if (IN6_IS_ADDR_V4MAPPED(&sin6_p->sin6_addr)) { struct sockaddr_in sin; in6_sin6_2_sin(&sin, sin6_p); inp->inp_vflag |= INP_IPV4; inp->inp_vflag &= ~INP_IPV6; error = in_pcbbind(inp, (struct sockaddr *)&sin, td->td_ucred); goto out; } } error = in6_pcbbind(inp, nam, td->td_ucred); out: INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (error); } static void udp6_close(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_close: inp == NULL")); #ifdef INET if (inp->inp_vflag & INP_IPV4) { struct pr_usrreqs *pru; pru = inetsw[ip_protox[IPPROTO_UDP]].pr_usrreqs; (*pru->pru_disconnect)(so); return; } #endif INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { in6_pcbdisconnect(inp); inp->in6p_laddr = in6addr_any; soisdisconnected(so); } INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); } static int udp6_connect(struct socket *so, struct sockaddr *nam, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_connect: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) { struct sockaddr_in6 *sin6_p; sin6_p = (struct sockaddr_in6 *)nam; if (IN6_IS_ADDR_V4MAPPED(&sin6_p->sin6_addr)) { struct sockaddr_in sin; if (inp->inp_faddr.s_addr != INADDR_ANY) { error = EISCONN; goto out; } in6_sin6_2_sin(&sin, sin6_p); - if (td && jailed(td->td_ucred)) - if (prison_remote_ip4(td->td_ucred, - &sin.sin_addr) != 0) { - error = EAFNOSUPPORT; - goto out; - } + if (td && (error = prison_remote_ip4(td->td_ucred, + &sin.sin_addr)) != 0) + goto out; error = in_pcbconnect(inp, (struct sockaddr *)&sin, td->td_ucred); if (error == 0) { inp->inp_vflag |= INP_IPV4; inp->inp_vflag &= ~INP_IPV6; soisconnected(so); } goto out; } } if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { error = EISCONN; goto out; } - if (td && jailed(td->td_ucred)) { + if (td) { struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam; - if (prison_remote_ip6(td->td_ucred, &sin6->sin6_addr) != 0) { - error = EAFNOSUPPORT; + if ((error = prison_remote_ip6(td->td_ucred, + &sin6->sin6_addr)) != 0) goto out; - } } error = in6_pcbconnect(inp, nam, td->td_ucred); if (error == 0) { if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) { /* should be non mapped addr */ inp->inp_vflag &= ~INP_IPV4; inp->inp_vflag |= INP_IPV6; } soisconnected(so); } out: INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (error); } static void udp6_detach(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_detach: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); in_pcbdetach(inp); in_pcbfree(inp); INP_INFO_WUNLOCK(&V_udbinfo); } static int udp6_disconnect(struct socket *so) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_disconnect: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); #ifdef INET if (inp->inp_vflag & INP_IPV4) { struct pr_usrreqs *pru; pru = inetsw[ip_protox[IPPROTO_UDP]].pr_usrreqs; error = (*pru->pru_disconnect)(so); goto out; } #endif if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) { error = ENOTCONN; goto out; } in6_pcbdisconnect(inp); inp->in6p_laddr = in6addr_any; SOCK_LOCK(so); so->so_state &= ~SS_ISCONNECTED; /* XXX */ SOCK_UNLOCK(so); out: INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (0); } static int udp6_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *addr, struct mbuf *control, struct thread *td) { INIT_VNET_INET(so->so_vnet); struct inpcb *inp; int error = 0; inp = sotoinpcb(so); KASSERT(inp != NULL, ("udp6_send: inp == NULL")); INP_INFO_WLOCK(&V_udbinfo); INP_WLOCK(inp); if (addr) { if (addr->sa_len != sizeof(struct sockaddr_in6)) { error = EINVAL; goto bad; } if (addr->sa_family != AF_INET6) { error = EAFNOSUPPORT; goto bad; } } #ifdef INET if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0) { int hasv4addr; struct sockaddr_in6 *sin6 = 0; if (addr == 0) hasv4addr = (inp->inp_vflag & INP_IPV4); else { sin6 = (struct sockaddr_in6 *)addr; hasv4addr = IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr) ? 1 : 0; } if (hasv4addr) { struct pr_usrreqs *pru; if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr) && !IN6_IS_ADDR_V4MAPPED(&inp->in6p_laddr)) { /* * When remote addr is IPv4-mapped address, * local addr should not be an IPv6 address; * since you cannot determine how to map IPv6 * source address to IPv4. */ error = EINVAL; goto out; } /* * XXXRW: We release UDP-layer locks before calling * udp_send() in order to avoid recursion. However, * this does mean there is a short window where inp's * fields are unstable. Could this lead to a * potential race in which the factors causing us to * select the UDPv4 output routine are invalidated? */ INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); if (sin6) in6_sin6_2_sin_in_sock(addr); pru = inetsw[ip_protox[IPPROTO_UDP]].pr_usrreqs; /* addr will just be freed in sendit(). */ return ((*pru->pru_send)(so, flags, m, addr, control, td)); } } #endif #ifdef MAC mac_inpcb_create_mbuf(inp, m); #endif error = udp6_output(inp, m, addr, control, td); out: INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); return (error); bad: INP_WUNLOCK(inp); INP_INFO_WUNLOCK(&V_udbinfo); m_freem(m); return (error); } struct pr_usrreqs udp6_usrreqs = { .pru_abort = udp6_abort, .pru_attach = udp6_attach, .pru_bind = udp6_bind, .pru_connect = udp6_connect, .pru_control = in6_control, .pru_detach = udp6_detach, .pru_disconnect = udp6_disconnect, .pru_peeraddr = in6_mapped_peeraddr, .pru_send = udp6_send, .pru_shutdown = udp_shutdown, .pru_sockaddr = in6_mapped_sockaddr, .pru_soreceive = soreceive_dgram, .pru_sosend = sosend_dgram, .pru_sosetlabel = in_pcbsosetlabel, .pru_close = udp6_close };