freebsd-dev/sys/netinet/udp_usrreq.c

1155 lines
29 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
* The Regents of the University of California.
* All rights reserved.
1994-05-24 10:09:53 +00:00
*
* 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
1999-08-28 01:08:13 +00:00
* $FreeBSD$
1994-05-24 10:09:53 +00:00
*/
#include "opt_ipfw.h"
#include "opt_ipsec.h"
#include "opt_inet6.h"
#include "opt_mac.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/domain.h>
#include <sys/eventhandler.h>
#include <sys/jail.h>
1997-02-24 20:31:25 +00:00
#include <sys/kernel.h>
#include <sys/lock.h>
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/priv.h>
#include <sys/proc.h>
1994-05-24 10:09:53 +00:00
#include <sys/protosw.h>
#include <sys/signalvar.h>
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/sx.h>
#include <sys/sysctl.h>
#include <sys/syslog.h>
#include <vm/uma.h>
1994-05-24 10:09:53 +00:00
#include <net/if.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/in_pcb.h>
#include <netinet/in_var.h>
1994-05-24 10:09:53 +00:00
#include <netinet/ip.h>
#ifdef INET6
#include <netinet/ip6.h>
#endif
#include <netinet/ip_icmp.h>
#include <netinet/icmp_var.h>
1994-05-24 10:09:53 +00:00
#include <netinet/ip_var.h>
#include <netinet/ip_options.h>
#ifdef INET6
#include <netinet6/ip6_var.h>
#endif
1994-05-24 10:09:53 +00:00
#include <netinet/udp.h>
#include <netinet/udp_var.h>
#ifdef FAST_IPSEC
#include <netipsec/ipsec.h>
#endif
#ifdef IPSEC
#include <netinet6/ipsec.h>
#endif
#include <machine/in_cksum.h>
#include <security/mac/mac_framework.h>
1994-05-24 10:09:53 +00:00
/*
* UDP protocol implementation.
* Per RFC 768, August, 1980.
*/
/*
* 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 by bad checksums may cause
* problems (especially for NFS data blocks).
*/
static int udpcksum = 1;
SYSCTL_INT(_net_inet_udp, UDPCTL_CHECKSUM, checksum, CTLFLAG_RW, &udpcksum,
0, "");
1994-05-24 10:09:53 +00:00
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");
static int blackhole = 0;
SYSCTL_INT(_net_inet_udp, OID_AUTO, blackhole, CTLFLAG_RW, &blackhole, 0,
"Do not send port unreachables for refused connects");
static int strict_mcast_mship = 0;
SYSCTL_INT(_net_inet_udp, OID_AUTO, strict_mcast_mship, CTLFLAG_RW,
&strict_mcast_mship, 0, "Only send multicast to member sockets");
2007-05-07 13:47:39 +00:00
struct inpcbhead udb; /* from udp_var.h */
struct inpcbinfo udbinfo;
#ifndef UDBHASHSIZE
2007-05-07 13:47:39 +00:00
#define UDBHASHSIZE 16
#endif
2007-05-07 13:47:39 +00:00
struct udpstat udpstat; /* from udp_var.h */
SYSCTL_STRUCT(_net_inet_udp, UDPCTL_STATS, stats, CTLFLAG_RW, &udpstat,
udpstat, "UDP statistics (struct udpstat, netinet/udp_var.h)");
static void udp_append(struct inpcb *last, struct ip *ip, struct mbuf *n,
int off, struct sockaddr_in *udp_in);
1994-05-24 10:09:53 +00:00
static void udp_detach(struct socket *so);
static int udp_output(struct inpcb *, struct mbuf *, struct sockaddr *,
struct mbuf *, struct thread *);
1994-05-24 10:09:53 +00:00
static void
udp_zone_change(void *tag)
{
uma_zone_set_max(udbinfo.ipi_zone, maxsockets);
}
static int
udp_inpcb_init(void *mem, int size, int flags)
{
2007-05-07 13:47:39 +00:00
struct inpcb *inp;
2007-05-07 13:47:39 +00:00
inp = mem;
INP_LOCK_INIT(inp, "inp", "udpinp");
return (0);
}
1994-05-24 10:09:53 +00:00
void
2007-05-07 13:47:39 +00:00
udp_init(void)
1994-05-24 10:09:53 +00:00
{
2007-05-07 13:47:39 +00:00
INP_INFO_LOCK_INIT(&udbinfo, "udp");
LIST_INIT(&udb);
udbinfo.ipi_listhead = &udb;
udbinfo.ipi_hashbase = hashinit(UDBHASHSIZE, M_PCB,
&udbinfo.ipi_hashmask);
udbinfo.ipi_porthashbase = hashinit(UDBHASHSIZE, M_PCB,
&udbinfo.ipi_porthashmask);
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(udbinfo.ipi_zone, maxsockets);
EVENTHANDLER_REGISTER(maxsockets_change, udp_zone_change, NULL,
2007-05-07 13:47:39 +00:00
EVENTHANDLER_PRI_ANY);
1994-05-24 10:09:53 +00:00
}
void
udp_input(struct mbuf *m, int off)
1994-05-24 10:09:53 +00:00
{
int iphlen = off;
struct ip *ip;
struct udphdr *uh;
struct inpcb *inp;
1994-05-24 10:09:53 +00:00
int len;
struct ip save_ip;
struct sockaddr_in udp_in;
#ifdef IPFIREWALL_FORWARD
struct m_tag *fwd_tag;
#endif
1994-05-24 10:09:53 +00:00
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.
1994-05-24 10:09:53 +00:00
*/
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) {
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;
1994-05-24 10:09:53 +00:00
/*
2007-05-07 13:47:39 +00:00
* Make mbuf data length reflect UDP length. If not enough data to
* reflect UDP length, drop.
1994-05-24 10:09:53 +00:00
*/
len = ntohs((u_short)uh->uh_ulen);
if (ip->ip_len != len) {
if (len > ip->ip_len || len < sizeof(struct udphdr)) {
1994-05-24 10:09:53 +00:00
udpstat.udps_badlen++;
goto badunlocked;
1994-05-24 10:09:53 +00:00
}
m_adj(m, len - ip->ip_len);
/* ip->ip_len = len; */
}
1994-05-24 10:09:53 +00:00
/*
* Save a copy of the IP header in case we want restore it for
* sending an ICMP error message in response.
1994-05-24 10:09:53 +00:00
*/
if (!blackhole)
save_ip = *ip;
1994-05-24 10:09:53 +00:00
/*
* 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];
2007-05-07 13:47:39 +00:00
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) {
1994-05-24 10:09:53 +00:00
udpstat.udps_badsum++;
m_freem(m);
return;
}
} else
udpstat.udps_nosum++;
1994-05-24 10:09:53 +00:00
#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(&udbinfo);
1994-05-24 10:09:53 +00:00
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
in_broadcast(ip->ip_dst, m->m_pkthdr.rcvif)) {
struct inpcb *last;
1994-05-24 10:09:53 +00:00
/*
* Deliver a multicast or broadcast datagram to *all* sockets
* for which the local and remote addresses and ports match
* those of the incoming datagram. This allows more than one
* process to receive multi/broadcasts on the same port.
1994-05-24 10:09:53 +00:00
* (This really ought to be done for unicast datagrams as
* well, but that would cause problems with existing
* applications that open both address-specific sockets and a
* wildcard socket listening to the same port -- they would
1994-05-24 10:09:53 +00:00
* end up receiving duplicates of every unicast datagram.
* Those applications open the multiple sockets to overcome
* an inadequacy of the UDP socket interface, but for
* backwards compatibility we avoid the problem here rather
* than fixing the interface. Maybe 4.5BSD will remedy
* this?)
1994-05-24 10:09:53 +00:00
*/
last = NULL;
LIST_FOREACH(inp, &udb, inp_list) {
if (inp->inp_lport != uh->uh_dport)
continue;
#ifdef INET6
1999-12-21 11:14:12 +00:00
if ((inp->inp_vflag & INP_IPV4) == 0)
continue;
#endif
1994-05-24 10:09:53 +00:00
if (inp->inp_laddr.s_addr != INADDR_ANY) {
if (inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
continue;
1994-05-24 10:09:53 +00:00
}
if (inp->inp_faddr.s_addr != INADDR_ANY) {
if (inp->inp_faddr.s_addr !=
ip->ip_src.s_addr ||
inp->inp_fport != uh->uh_sport)
continue;
1994-05-24 10:09:53 +00:00
}
/*
* Check multicast packets to make sure they are only
* sent to sockets with multicast memberships for the
* packet's destination address and arrival interface
*/
2007-05-07 13:47:39 +00:00
#define MSHIP(_inp, n) ((_inp)->inp_moptions->imo_membership[(n)])
#define NMSHIPS(_inp) ((_inp)->inp_moptions->imo_num_memberships)
INP_LOCK(inp);
if (strict_mcast_mship && inp->inp_moptions != NULL) {
int mship, foundmship = 0;
for (mship = 0; mship < NMSHIPS(inp);
mship++) {
if (MSHIP(inp, mship)->inm_addr.s_addr
== ip->ip_dst.s_addr &&
MSHIP(inp, mship)->inm_ifp
== m->m_pkthdr.rcvif) {
foundmship = 1;
break;
}
}
if (foundmship == 0) {
INP_UNLOCK(inp);
continue;
}
}
#undef NMSHIPS
#undef MSHIP
1994-05-24 10:09:53 +00:00
if (last != NULL) {
struct mbuf *n;
n = m_copy(m, 0, M_COPYALL);
if (n != NULL)
udp_append(last, ip, n, iphlen +
sizeof(struct udphdr), &udp_in);
INP_UNLOCK(last);
1994-05-24 10:09:53 +00:00
}
last = inp;
1994-05-24 10:09:53 +00:00
/*
* 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.
1994-05-24 10:09:53 +00:00
*/
if ((last->inp_socket->so_options &
(SO_REUSEPORT|SO_REUSEADDR)) == 0)
1994-05-24 10:09:53 +00:00
break;
}
if (last == NULL) {
/*
* No matching pcb found; discard datagram. (No need
* to send an ICMP Port Unreachable for a broadcast
* or multicast datgram.)
1994-05-24 10:09:53 +00:00
*/
udpstat.udps_noportbcast++;
goto badheadlocked;
1994-05-24 10:09:53 +00:00
}
udp_append(last, ip, m, iphlen + sizeof(struct udphdr),
&udp_in);
INP_UNLOCK(last);
INP_INFO_RUNLOCK(&udbinfo);
1994-05-24 10:09:53 +00:00
return;
}
1994-05-24 10:09:53 +00:00
/*
* Locate pcb for datagram.
1994-05-24 10:09:53 +00:00
*/
Improved connection establishment performance by doing local port lookups via a hashed port list. In the new scheme, in_pcblookup() goes away and is replaced by a new routine, in_pcblookup_local() for doing the local port check. Note that this implementation is space inefficient in that the PCB struct is now too large to fit into 128 bytes. I might deal with this in the future by using the new zone allocator, but I wanted these changes to be extensively tested in their current form first. Also: 1) Fixed off-by-one errors in the port lookup loops in in_pcbbind(). 2) Got rid of some unneeded rehashing. Adding a new routine, in_pcbinshash() to do the initialial hash insertion. 3) Renamed in_pcblookuphash() to in_pcblookup_hash() for easier readability. 4) Added a new routine, in_pcbremlists() to remove the PCB from the various hash lists. 5) Added/deleted comments where appropriate. 6) Removed unnecessary splnet() locking. In general, the PCB functions should be called at splnet()...there are unfortunately a few exceptions, however. 7) Reorganized a few structs for better cache line behavior. 8) Killed my TCP_ACK_HACK kludge. It may come back in a different form in the future, however. These changes have been tested on wcarchive for more than a month. In tests done here, connection establishment overhead is reduced by more than 50 times, thus getting rid of one of the major networking scalability problems. Still to do: make tcp_fastimo/tcp_slowtimo scale well for systems with a large number of connections. tcp_fastimo is easy; tcp_slowtimo is difficult. WARNING: Anything that knows about inpcb and tcpcb structs will have to be recompiled; at the very least, this includes netstat(1).
1998-01-27 09:15:13 +00:00
inp = in_pcblookup_hash(&udbinfo, ip->ip_src, uh->uh_sport,
ip->ip_dst, uh->uh_dport, 1, m->m_pkthdr.rcvif);
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));
}
1994-05-24 10:09:53 +00:00
udpstat.udps_noport++;
if (m->m_flags & (M_BCAST | M_MCAST)) {
udpstat.udps_noportbcast++;
goto badheadlocked;
1994-05-24 10:09:53 +00:00
}
if (blackhole)
goto badheadlocked;
if (badport_bandlim(BANDLIM_ICMP_UNREACH) < 0)
goto badheadlocked;
Fixed broken ICMP error generation, unified conversion of IP header fields between host and network byte order. The details: o icmp_error() now does not add IP header length. This fixes the problem when icmp_error() is called from ip_forward(). In this case the ip_len of the original IP datagram returned with ICMP error was wrong. o icmp_error() expects all three fields, ip_len, ip_id and ip_off in host byte order, so DTRT and convert these fields back to network byte order before sending a message. This fixes the problem described in PR 16240 and PR 20877 (ip_id field was returned in host byte order). o ip_ttl decrement operation in ip_forward() was moved down to make sure that it does not corrupt the copy of original IP datagram passed later to icmp_error(). o A copy of original IP datagram in ip_forward() was made a read-write, independent copy. This fixes the problem I first reported to Garrett Wollman and Bill Fenner and later put in audit trail of PR 16240: ip_output() (not always) converts fields of original datagram to network byte order, but because copy (mcopy) and its original (m) most likely share the same mbuf cluster, ip_output()'s manipulations on original also corrupted the copy. o ip_output() now expects all three fields, ip_len, ip_off and (what is significant) ip_id in host byte order. It was a headache for years that ip_id was handled differently. The only compatibility issue here is the raw IP socket interface with IP_HDRINCL socket option set and a non-zero ip_id field, but ip.4 manual page was unclear on whether in this case ip_id field should be in host or network byte order.
2000-09-01 12:33:03 +00:00
*ip = save_ip;
ip->ip_len += iphlen;
icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PORT, 0, 0);
INP_INFO_RUNLOCK(&udbinfo);
1994-05-24 10:09:53 +00:00
return;
}
/*
* Check the minimum TTL for socket.
*/
INP_LOCK(inp);
if (inp->inp_ip_minttl && inp->inp_ip_minttl > ip->ip_ttl)
goto badheadlocked;
udp_append(inp, ip, m, iphlen + sizeof(struct udphdr), &udp_in);
INP_UNLOCK(inp);
INP_INFO_RUNLOCK(&udbinfo);
1994-05-24 10:09:53 +00:00
return;
badheadlocked:
if (inp)
INP_UNLOCK(inp);
INP_INFO_RUNLOCK(&udbinfo);
badunlocked:
1994-05-24 10:09:53 +00:00
m_freem(m);
}
/*
* 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;
2004-11-04 07:14:03 +00:00
#ifdef INET6
struct sockaddr_in6 udp_in6;
2004-11-04 07:14:03 +00:00
#endif
INP_LOCK_ASSERT(inp);
#if defined(IPSEC) || defined(FAST_IPSEC)
/* check AH/ESP integrity. */
if (ipsec4_in_reject(n, inp)) {
#ifdef IPSEC
ipsecstat.in_polvio++;
#endif
m_freem(n);
return;
}
#endif /*IPSEC || FAST_IPSEC*/
#ifdef MAC
if (mac_check_inpcb_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) {
int savedflags;
savedflags = inp->inp_flags;
inp->inp_flags &= ~INP_UNMAPPABLEOPTS;
ip6_savecontrol(inp, n, &opts);
inp->inp_flags = savedflags;
} 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) {
m_freem(n);
if (opts)
m_freem(opts);
udpstat.udps_fullsock++;
SOCKBUF_UNLOCK(&so->so_rcv);
} else
sorwakeup_locked(so);
1994-05-24 10:09:53 +00:00
}
/*
* Notify a udp user of an asynchronous error; just wake up so that they can
* collect error status.
1994-05-24 10:09:53 +00:00
*/
struct inpcb *
udp_notify(struct inpcb *inp, int errno)
1994-05-24 10:09:53 +00:00
{
1994-05-24 10:09:53 +00:00
inp->inp_socket->so_error = errno;
sorwakeup(inp->inp_socket);
sowwakeup(inp->inp_socket);
return (inp);
1994-05-24 10:09:53 +00:00
}
void
udp_ctlinput(int cmd, struct sockaddr *sa, void *vip)
1994-05-24 10:09:53 +00:00
{
struct ip *ip = vip;
struct udphdr *uh;
struct inpcb *(*notify)(struct inpcb *, int) = udp_notify;
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;
1994-05-24 10:09:53 +00:00
/*
* 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)
2007-05-07 13:47:39 +00:00
ip = NULL;
else if ((unsigned)cmd >= PRC_NCMDS || inetctlerrmap[cmd] == 0)
1994-05-24 10:09:53 +00:00
return;
2007-05-07 13:47:39 +00:00
if (ip != NULL) {
1994-05-24 10:09:53 +00:00
uh = (struct udphdr *)((caddr_t)ip + (ip->ip_hl << 2));
INP_INFO_RLOCK(&udbinfo);
inp = in_pcblookup_hash(&udbinfo, faddr, uh->uh_dport,
ip->ip_src, uh->uh_sport, 0, NULL);
if (inp != NULL) {
INP_LOCK(inp);
if (inp->inp_socket != NULL) {
(*notify)(inp, inetctlerrmap[cmd]);
}
INP_UNLOCK(inp);
}
INP_INFO_RUNLOCK(&udbinfo);
1994-05-24 10:09:53 +00:00
} else
in_pcbnotifyall(&udbinfo, faddr, inetctlerrmap[cmd], notify);
1994-05-24 10:09:53 +00:00
}
static int
udp_pcblist(SYSCTL_HANDLER_ARGS)
{
2005-06-01 11:24:00 +00:00
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 = 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(&udbinfo);
gencnt = udbinfo.ipi_gencnt;
n = udbinfo.ipi_count;
INP_INFO_RUNLOCK(&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(&udbinfo);
for (inp = LIST_FIRST(udbinfo.ipi_listhead), i = 0; inp && i < n;
inp = LIST_NEXT(inp, inp_list)) {
INP_LOCK(inp);
if (inp->inp_gencnt <= gencnt &&
cr_canseesocket(req->td->td_ucred, inp->inp_socket) == 0)
inp_list[i++] = inp;
INP_UNLOCK(inp);
}
INP_INFO_RUNLOCK(&udbinfo);
n = i;
error = 0;
for (i = 0; i < n; i++) {
inp = inp_list[i];
INP_LOCK(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_UNLOCK(inp);
error = SYSCTL_OUT(req, &xi, sizeof xi);
} else
INP_UNLOCK(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(&udbinfo);
xig.xig_gen = udbinfo.ipi_gencnt;
xig.xig_sogen = so_gencnt;
xig.xig_count = udbinfo.ipi_count;
INP_INFO_RUNLOCK(&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)
{
struct xucred xuc;
struct sockaddr_in addrs[2];
struct inpcb *inp;
2005-06-01 11:24:00 +00:00
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(&udbinfo);
inp = in_pcblookup_hash(&udbinfo, addrs[1].sin_addr, addrs[1].sin_port,
addrs[0].sin_addr, addrs[0].sin_port, 1, NULL);
if (inp == NULL || inp->inp_socket == NULL) {
error = ENOENT;
goto out;
}
error = cr_canseesocket(req->td->td_ucred, inp->inp_socket);
if (error)
goto out;
cru2x(inp->inp_socket->so_cred, &xuc);
out:
INP_INFO_RUNLOCK(&udbinfo);
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)
1994-05-24 10:09:53 +00:00
{
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;
1994-05-24 10:09:53 +00:00
/*
* udp_output() may need to temporarily bind or connect the current
* inpcb. As such, we don't know up front what inpcb locks we will
* need. Do any work to decide what is needed up front before
* acquiring 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 *);
2007-05-07 13:47:39 +00:00
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;
2007-05-07 13:47:39 +00:00
src.sin_addr =
*(struct in_addr *)CMSG_DATA(cm);
break;
2007-05-07 13:47:39 +00:00
default:
error = ENOPROTOOPT;
break;
}
if (error)
break;
}
m_freem(control);
}
if (error) {
m_freem(m);
return (error);
}
if (src.sin_family == AF_INET || addr != NULL) {
INP_INFO_WLOCK(&udbinfo);
unlock_udbinfo = 1;
} else
unlock_udbinfo = 0;
INP_LOCK(inp);
#ifdef MAC
mac_create_mbuf_from_inpcb(inp, m);
#endif
/*
* 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) {
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;
}
1994-05-24 10:09:53 +00:00
if (addr) {
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
sin = (struct sockaddr_in *)addr;
if (jailed(td->td_ucred))
prison_remote_ip(td->td_ucred, 0,
&sin->sin_addr.s_addr);
1994-05-24 10:09:53 +00:00
if (inp->inp_faddr.s_addr != INADDR_ANY) {
error = EISCONN;
goto release;
}
error = in_pcbconnect_setup(inp, addr, &laddr.s_addr, &lport,
&faddr.s_addr, &fport, NULL, td->td_ucred);
if (error)
1994-05-24 10:09:53 +00:00
goto release;
/* Commit the local port if newly assigned. */
if (inp->inp_laddr.s_addr == INADDR_ANY &&
inp->inp_lport == 0) {
/*
* 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;
1994-05-24 10:09:53 +00:00
}
} else {
faddr = inp->inp_faddr;
fport = inp->inp_fport;
if (faddr.s_addr == INADDR_ANY) {
1994-05-24 10:09:53 +00:00
error = ENOTCONN;
goto release;
}
}
1994-05-24 10:09:53 +00:00
/*
* 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.
1994-05-24 10:09:53 +00:00
*/
M_PREPEND(m, sizeof(struct udpiphdr) + max_linkhdr, M_DONTWAIT);
if (m == NULL) {
1994-05-24 10:09:53 +00:00
error = ENOBUFS;
goto release;
1994-05-24 10:09:53 +00:00
}
m->m_data += max_linkhdr;
m->m_len -= max_linkhdr;
m->m_pkthdr.len -= max_linkhdr;
1994-05-24 10:09:53 +00:00
/*
* Fill in mbuf with extended UDP header and addresses and length put
* into network format.
1994-05-24 10:09:53 +00:00
*/
ui = mtod(m, struct udpiphdr *);
bzero(ui->ui_x1, sizeof(ui->ui_x1)); /* XXX still needed? */
1994-05-24 10:09:53 +00:00
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));
1994-05-24 10:09:53 +00:00
/*
* 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;
1994-05-24 10:09:53 +00:00
/*
* Set up checksum and output datagram.
1994-05-24 10:09:53 +00:00
*/
if (udpcksum) {
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;
1994-05-24 10:09:53 +00:00
((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 */
1994-05-24 10:09:53 +00:00
udpstat.udps_opackets++;
if (unlock_udbinfo)
INP_INFO_WUNLOCK(&udbinfo);
error = ip_output(m, inp->inp_options, NULL, ipflags,
inp->inp_moptions, inp);
INP_UNLOCK(inp);
1994-05-24 10:09:53 +00:00
return (error);
release:
INP_UNLOCK(inp);
if (unlock_udbinfo)
INP_INFO_WUNLOCK(&udbinfo);
1994-05-24 10:09:53 +00:00
m_freem(m);
return (error);
}
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
);
2007-05-07 13:47:39 +00:00
SYSCTL_ULONG(_net_inet_udp, UDPCTL_RECVSPACE, recvspace, CTLFLAG_RW,
&udp_recvspace, 0, "Maximum space for incoming UDP datagrams");
1994-05-24 10:09:53 +00:00
static void
udp_abort(struct socket *so)
1994-05-24 10:09:53 +00:00
{
struct inpcb *inp;
1994-05-24 10:09:53 +00:00
inp = sotoinpcb(so);
KASSERT(inp != NULL, ("udp_abort: inp == NULL"));
INP_INFO_WLOCK(&udbinfo);
INP_LOCK(inp);
if (inp->inp_faddr.s_addr != INADDR_ANY) {
in_pcbdisconnect(inp);
inp->inp_laddr.s_addr = INADDR_ANY;
soisdisconnected(so);
}
INP_UNLOCK(inp);
INP_INFO_WUNLOCK(&udbinfo);
}
1994-05-24 10:09:53 +00:00
static int
udp_attach(struct socket *so, int proto, struct thread *td)
{
struct inpcb *inp;
2005-06-01 11:24:00 +00:00
int error;
1994-05-24 10:09:53 +00:00
inp = sotoinpcb(so);
KASSERT(inp == NULL, ("udp_attach: inp != NULL"));
error = soreserve(so, udp_sendspace, udp_recvspace);
if (error)
return (error);
INP_INFO_WLOCK(&udbinfo);
error = in_pcballoc(so, &udbinfo);
if (error) {
INP_INFO_WUNLOCK(&udbinfo);
return (error);
}
inp = (struct inpcb *)so->so_pcb;
INP_INFO_WUNLOCK(&udbinfo);
inp->inp_vflag |= INP_IPV4;
inp->inp_ip_ttl = ip_defttl;
INP_UNLOCK(inp);
return (0);
}
1994-05-24 10:09:53 +00:00
static int
udp_bind(struct socket *so, struct sockaddr *nam, struct thread *td)
{
struct inpcb *inp;
2005-06-01 11:24:00 +00:00
int error;
1994-05-24 10:09:53 +00:00
inp = sotoinpcb(so);
KASSERT(inp != NULL, ("udp_bind: inp == NULL"));
INP_INFO_WLOCK(&udbinfo);
INP_LOCK(inp);
error = in_pcbbind(inp, nam, td->td_ucred);
INP_UNLOCK(inp);
INP_INFO_WUNLOCK(&udbinfo);
return (error);
}
1994-05-24 10:09:53 +00:00
static void
udp_close(struct socket *so)
{
struct inpcb *inp;
inp = sotoinpcb(so);
KASSERT(inp != NULL, ("udp_close: inp == NULL"));
INP_INFO_WLOCK(&udbinfo);
INP_LOCK(inp);
if (inp->inp_faddr.s_addr != INADDR_ANY) {
in_pcbdisconnect(inp);
inp->inp_laddr.s_addr = INADDR_ANY;
soisdisconnected(so);
}
INP_UNLOCK(inp);
INP_INFO_WUNLOCK(&udbinfo);
}
static int
udp_connect(struct socket *so, struct sockaddr *nam, struct thread *td)
{
struct inpcb *inp;
2005-06-01 11:24:00 +00:00
int error;
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
struct sockaddr_in *sin;
inp = sotoinpcb(so);
KASSERT(inp != NULL, ("udp_connect: inp == NULL"));
INP_INFO_WLOCK(&udbinfo);
INP_LOCK(inp);
if (inp->inp_faddr.s_addr != INADDR_ANY) {
INP_UNLOCK(inp);
INP_INFO_WUNLOCK(&udbinfo);
return (EISCONN);
}
sin = (struct sockaddr_in *)nam;
if (jailed(td->td_ucred))
prison_remote_ip(td->td_ucred, 0, &sin->sin_addr.s_addr);
error = in_pcbconnect(inp, nam, td->td_ucred);
if (error == 0)
soisconnected(so);
INP_UNLOCK(inp);
INP_INFO_WUNLOCK(&udbinfo);
return (error);
}
1994-05-24 10:09:53 +00:00
static void
udp_detach(struct socket *so)
{
struct inpcb *inp;
1994-05-24 10:09:53 +00:00
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(&udbinfo);
INP_LOCK(inp);
in_pcbdetach(inp);
in_pcbfree(inp);
INP_INFO_WUNLOCK(&udbinfo);
}
1994-05-24 10:09:53 +00:00
static int
udp_disconnect(struct socket *so)
{
struct inpcb *inp;
1994-05-24 10:09:53 +00:00
inp = sotoinpcb(so);
KASSERT(inp != NULL, ("udp_disconnect: inp == NULL"));
INP_INFO_WLOCK(&udbinfo);
INP_LOCK(inp);
if (inp->inp_faddr.s_addr == INADDR_ANY) {
INP_INFO_WUNLOCK(&udbinfo);
INP_UNLOCK(inp);
return (ENOTCONN);
}
1994-05-24 10:09:53 +00:00
in_pcbdisconnect(inp);
inp->inp_laddr.s_addr = INADDR_ANY;
SOCK_LOCK(so);
so->so_state &= ~SS_ISCONNECTED; /* XXX */
SOCK_UNLOCK(so);
INP_UNLOCK(inp);
INP_INFO_WUNLOCK(&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));
1994-05-24 10:09:53 +00:00
}
int
udp_shutdown(struct socket *so)
{
1994-05-24 10:09:53 +00:00
struct inpcb *inp;
inp = sotoinpcb(so);
KASSERT(inp != NULL, ("udp_shutdown: inp == NULL"));
INP_LOCK(inp);
socantsendmore(so);
INP_UNLOCK(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_sosend = sosend_dgram,
.pru_shutdown = udp_shutdown,
.pru_sockaddr = in_getsockaddr,
.pru_sosetlabel = in_pcbsosetlabel,
.pru_close = udp_close,
};