/* * INET An implementation of the TCP/IP protocol suite for the LINUX * operating system. INET is implemented using the BSD Socket * interface as the means of communication with the user level. * * RAW - implementation of IP "raw" sockets. * * Version: $Id: raw.c,v 1.43 1999/08/20 11:05:57 davem Exp $ * * Authors: Ross Biro, * Fred N. van Kempen, * * Fixes: * Alan Cox : verify_area() fixed up * Alan Cox : ICMP error handling * Alan Cox : EMSGSIZE if you send too big a packet * Alan Cox : Now uses generic datagrams and shared skbuff * library. No more peek crashes, no more backlogs * Alan Cox : Checks sk->broadcast. * Alan Cox : Uses skb_free_datagram/skb_copy_datagram * Alan Cox : Raw passes ip options too * Alan Cox : Setsocketopt added * Alan Cox : Fixed error return for broadcasts * Alan Cox : Removed wake_up calls * Alan Cox : Use ttl/tos * Alan Cox : Cleaned up old debugging * Alan Cox : Use new kernel side addresses * Arnt Gulbrandsen : Fixed MSG_DONTROUTE in raw sockets. * Alan Cox : BSD style RAW socket demultiplexing. * Alan Cox : Beginnings of mrouted support. * Alan Cox : Added IP_HDRINCL option. * Alan Cox : Skip broadcast check if BSDism set. * David S. Miller : New socket lookup architecture. * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. */ #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 sock *raw_v4_htable[RAWV4_HTABLE_SIZE]; rwlock_t raw_v4_lock = RW_LOCK_UNLOCKED; static void raw_v4_hash(struct sock *sk) { struct sock **skp = &raw_v4_htable[sk->num & (RAWV4_HTABLE_SIZE - 1)]; write_lock_bh(&raw_v4_lock); if ((sk->next = *skp) != NULL) (*skp)->pprev = &sk->next; *skp = sk; sk->pprev = skp; sk->prot->inuse++; if(sk->prot->highestinuse < sk->prot->inuse) sk->prot->highestinuse = sk->prot->inuse; sock_hold(sk); write_unlock_bh(&raw_v4_lock); } static void raw_v4_unhash(struct sock *sk) { write_lock_bh(&raw_v4_lock); if (sk->pprev) { if (sk->next) sk->next->pprev = sk->pprev; *sk->pprev = sk->next; sk->pprev = NULL; sk->prot->inuse--; __sock_put(sk); } write_unlock_bh(&raw_v4_lock); } struct sock *__raw_v4_lookup(struct sock *sk, unsigned short num, unsigned long raddr, unsigned long laddr, int dif) { struct sock *s = sk; for(s = sk; s; s = s->next) { if((s->num == num) && !(s->daddr && s->daddr != raddr) && !(s->rcv_saddr && s->rcv_saddr != laddr) && !(s->bound_dev_if && s->bound_dev_if != dif)) break; /* gotcha */ } return s; } /* * 0 - deliver * 1 - block */ static __inline__ int icmp_filter(struct sock *sk, struct sk_buff *skb) { int type; type = skb->h.icmph->type; if (type < 32) return test_bit(type, &sk->tp_pinfo.tp_raw4.filter); /* Do not block unknown ICMP types */ return 0; } /* IP input processing comes here for RAW socket delivery. * This is fun as to avoid copies we want to make no surplus * copies. * * RFC 1122: SHOULD pass TOS value up to the transport layer. * -> It does. And not only TOS, but all IP header. */ struct sock *raw_v4_input(struct sk_buff *skb, struct iphdr *iph, int hash) { struct sock *sk; read_lock(&raw_v4_lock); if ((sk = raw_v4_htable[hash]) == NULL) goto out; sk = __raw_v4_lookup(sk, iph->protocol, iph->saddr, iph->daddr, skb->dev->ifindex); while(sk != NULL) { struct sock *sknext = __raw_v4_lookup(sk->next, iph->protocol, iph->saddr, iph->daddr, skb->dev->ifindex); if (iph->protocol != IPPROTO_ICMP || ! icmp_filter(sk, skb)) { struct sk_buff *clone; if(sknext == NULL) break; clone = skb_clone(skb, GFP_ATOMIC); /* Not releasing hash table! */ if(clone) raw_rcv(sk, clone); } sk = sknext; } out: if (sk) sock_hold(sk); read_unlock(&raw_v4_lock); return sk; } void raw_err (struct sock *sk, struct sk_buff *skb) { int type = skb->h.icmph->type; int code = skb->h.icmph->code; u32 info = 0; int err = 0; int harderr = 0; /* Report error on raw socket, if: 1. User requested ip_recverr. 2. Socket is connected (otherwise the error indication is useless without ip_recverr and error is hard. */ if (!sk->protinfo.af_inet.recverr && sk->state != TCP_ESTABLISHED) return; switch (type) { default: case ICMP_TIME_EXCEEDED: err = EHOSTUNREACH; break; case ICMP_SOURCE_QUENCH: return; case ICMP_PARAMETERPROB: err = EPROTO; info = ntohl(skb->h.icmph->un.gateway)>>24; harderr = 1; break; case ICMP_DEST_UNREACH: err = EHOSTUNREACH; if (code > NR_ICMP_UNREACH) break; err = icmp_err_convert[code].errno; harderr = icmp_err_convert[code].fatal; if (code == ICMP_FRAG_NEEDED) { harderr = (sk->protinfo.af_inet.pmtudisc != IP_PMTUDISC_DONT); err = EMSGSIZE; info = ntohs(skb->h.icmph->un.frag.mtu); } } if (sk->protinfo.af_inet.recverr) ip_icmp_error(sk, skb, err, 0, info, (u8 *)(skb->h.icmph + 1)); if (sk->protinfo.af_inet.recverr || harderr) { sk->err = err; sk->error_report(sk); } } static int raw_rcv_skb(struct sock * sk, struct sk_buff * skb) { /* Charge it to the socket. */ if (sock_queue_rcv_skb(sk,skb)<0) { ip_statistics.IpInDiscards++; kfree_skb(skb); return -1; } ip_statistics.IpInDelivers++; return 0; } /* * This should be the easiest of all, all we do is * copy it into a buffer. All demultiplexing is done * in ip.c */ int raw_rcv(struct sock *sk, struct sk_buff *skb) { /* Now we need to copy this into memory. */ skb_trim(skb, ntohs(skb->nh.iph->tot_len)); skb->h.raw = skb->nh.raw; raw_rcv_skb(sk, skb); return 0; } struct rawfakehdr { struct iovec *iov; u32 saddr; }; /* * Send a RAW IP packet. */ /* * Callback support is trivial for SOCK_RAW */ static int raw_getfrag(const void *p, char *to, unsigned int offset, unsigned int fraglen) { struct rawfakehdr *rfh = (struct rawfakehdr *) p; return memcpy_fromiovecend(to, rfh->iov, offset, fraglen); } /* * IPPROTO_RAW needs extra work. */ static int raw_getrawfrag(const void *p, char *to, unsigned int offset, unsigned int fraglen) { struct rawfakehdr *rfh = (struct rawfakehdr *) p; if (memcpy_fromiovecend(to, rfh->iov, offset, fraglen)) return -EFAULT; if (offset==0) { struct iphdr *iph = (struct iphdr *)to; if (!iph->saddr) iph->saddr = rfh->saddr; iph->check=0; iph->tot_len=htons(fraglen); /* This is right as you can't frag RAW packets */ /* * Deliberate breach of modularity to keep * ip_build_xmit clean (well less messy). */ if (!iph->id) iph->id = htons(ip_id_count++); iph->check=ip_fast_csum((unsigned char *)iph, iph->ihl); } return 0; } static int raw_sendmsg(struct sock *sk, struct msghdr *msg, int len) { struct ipcm_cookie ipc; struct rawfakehdr rfh; struct rtable *rt = NULL; int free = 0; u32 daddr; u8 tos; int err; /* This check is ONLY to check for arithmetic overflow on integer(!) len. Not more! Real check will be made in ip_build_xmit --ANK BTW socket.c -> af_*.c -> ... make multiple invalid conversions size_t -> int. We MUST repair it f.e. by replacing all of them with size_t and revise all the places sort of len += sizeof(struct iphdr) If len was ULONG_MAX-10 it would be cathastrophe --ANK */ if (len < 0 || len > 0xFFFF) return -EMSGSIZE; /* * Check the flags. */ if (msg->msg_flags & MSG_OOB) /* Mirror BSD error message compatibility */ return -EOPNOTSUPP; /* * Get and verify the address. */ if (msg->msg_namelen) { struct sockaddr_in *usin = (struct sockaddr_in*)msg->msg_name; if (msg->msg_namelen < sizeof(*usin)) return(-EINVAL); if (usin->sin_family != AF_INET) { static int complained; if (!complained++) printk(KERN_INFO "%s forgot to set AF_INET in raw sendmsg. Fix it!\n", current->comm); if (usin->sin_family) return -EINVAL; } daddr = usin->sin_addr.s_addr; /* ANK: I did not forget to get protocol from port field. * I just do not know, who uses this weirdness. * IP_HDRINCL is much more convenient. */ } else { if (sk->state != TCP_ESTABLISHED) return(-EINVAL); daddr = sk->daddr; } ipc.addr = sk->saddr; ipc.opt = NULL; ipc.oif = sk->bound_dev_if; if (msg->msg_controllen) { int tmp = ip_cmsg_send(msg, &ipc); if (tmp) return tmp; if (ipc.opt) free=1; } rfh.saddr = ipc.addr; ipc.addr = daddr; if (!ipc.opt) ipc.opt = sk->protinfo.af_inet.opt; if (ipc.opt) { err = -EINVAL; /* Linux does not mangle headers on raw sockets, * so that IP options + IP_HDRINCL is non-sense. */ if (sk->protinfo.af_inet.hdrincl) goto done; if (ipc.opt->srr) { if (!daddr) goto done; daddr = ipc.opt->faddr; } } tos = RT_TOS(sk->protinfo.af_inet.tos) | sk->localroute; if (msg->msg_flags&MSG_DONTROUTE) tos |= RTO_ONLINK; if (MULTICAST(daddr)) { if (!ipc.oif) ipc.oif = sk->protinfo.af_inet.mc_index; if (!rfh.saddr) rfh.saddr = sk->protinfo.af_inet.mc_addr; } err = ip_route_output(&rt, daddr, rfh.saddr, tos, ipc.oif); if (err) goto done; err = -EACCES; if (rt->rt_flags&RTCF_BROADCAST && !sk->broadcast) goto done; if (msg->msg_flags&MSG_CONFIRM) goto do_confirm; back_from_confirm: rfh.iov = msg->msg_iov; rfh.saddr = rt->rt_src; if (!ipc.addr) ipc.addr = rt->rt_dst; err=ip_build_xmit(sk, sk->protinfo.af_inet.hdrincl ? raw_getrawfrag : raw_getfrag, &rfh, len, &ipc, rt, msg->msg_flags); done: if (free) kfree(ipc.opt); ip_rt_put(rt); return err<0 ? err : len; do_confirm: dst_confirm(&rt->u.dst); if (!(msg->msg_flags&MSG_PROBE) || len) goto back_from_confirm; err = 0; goto done; } static void raw_close(struct sock *sk, long timeout) { /* * Raw sockets may have direct kernel refereneces. Kill them. */ ip_ra_control(sk, 0, NULL); inet_sock_release(sk); } /* This gets rid of all the nasties in af_inet. -DaveM */ static int raw_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len) { struct sockaddr_in *addr = (struct sockaddr_in *) uaddr; int chk_addr_ret; if((sk->state != TCP_CLOSE) || (addr_len < sizeof(struct sockaddr_in))) return -EINVAL; chk_addr_ret = inet_addr_type(addr->sin_addr.s_addr); if(addr->sin_addr.s_addr != 0 && chk_addr_ret != RTN_LOCAL && chk_addr_ret != RTN_MULTICAST && chk_addr_ret != RTN_BROADCAST) return -EADDRNOTAVAIL; sk->rcv_saddr = sk->saddr = addr->sin_addr.s_addr; if(chk_addr_ret == RTN_MULTICAST || chk_addr_ret == RTN_BROADCAST) sk->saddr = 0; /* Use device */ sk_dst_reset(sk); return 0; } /* * This should be easy, if there is something there * we return it, otherwise we block. */ int raw_recvmsg(struct sock *sk, struct msghdr *msg, int len, int noblock, int flags,int *addr_len) { int copied=0; struct sk_buff *skb; int err; struct sockaddr_in *sin=(struct sockaddr_in *)msg->msg_name; if (flags & MSG_OOB) return -EOPNOTSUPP; if (addr_len) *addr_len=sizeof(*sin); if (flags & MSG_ERRQUEUE) return ip_recv_error(sk, msg, len); skb=skb_recv_datagram(sk,flags,noblock,&err); if(skb==NULL) return err; copied = skb->len; if (len < copied) { msg->msg_flags |= MSG_TRUNC; copied = len; } err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied); if (err) goto done; sk->stamp=skb->stamp; /* Copy the address. */ if (sin) { sin->sin_family = AF_INET; sin->sin_addr.s_addr = skb->nh.iph->saddr; } if (sk->protinfo.af_inet.cmsg_flags) ip_cmsg_recv(msg, skb); done: skb_free_datagram(sk, skb); return (err ? : copied); } static int raw_init(struct sock *sk) { struct raw_opt *tp = &(sk->tp_pinfo.tp_raw4); if (sk->num == IPPROTO_ICMP) memset(&tp->filter, 0, sizeof(tp->filter)); return 0; } static int raw_seticmpfilter(struct sock *sk, char *optval, int optlen) { if (optlen > sizeof(struct icmp_filter)) optlen = sizeof(struct icmp_filter); if (copy_from_user(&sk->tp_pinfo.tp_raw4.filter, optval, optlen)) return -EFAULT; return 0; } static int raw_geticmpfilter(struct sock *sk, char *optval, int *optlen) { int len; if (get_user(len,optlen)) return -EFAULT; if (len > sizeof(struct icmp_filter)) len = sizeof(struct icmp_filter); if (put_user(len, optlen)) return -EFAULT; if (copy_to_user(optval, &sk->tp_pinfo.tp_raw4.filter, len)) return -EFAULT; return 0; } static int raw_setsockopt(struct sock *sk, int level, int optname, char *optval, int optlen) { if (level != SOL_RAW) return ip_setsockopt(sk, level, optname, optval, optlen); switch (optname) { case ICMP_FILTER: if (sk->num != IPPROTO_ICMP) return -EOPNOTSUPP; return raw_seticmpfilter(sk, optval, optlen); }; return -ENOPROTOOPT; } static int raw_getsockopt(struct sock *sk, int level, int optname, char *optval, int *optlen) { if (level != SOL_RAW) return ip_getsockopt(sk, level, optname, optval, optlen); switch (optname) { case ICMP_FILTER: if (sk->num != IPPROTO_ICMP) return -EOPNOTSUPP; return raw_geticmpfilter(sk, optval, optlen); }; return -ENOPROTOOPT; } static void get_raw_sock(struct sock *sp, char *tmpbuf, int i) { unsigned int dest, src; __u16 destp, srcp; int timer_active; unsigned long timer_expires; dest = sp->daddr; src = sp->rcv_saddr; destp = 0; srcp = sp->num; timer_active = (sp->timer.prev != NULL) ? 2 : 0; timer_expires = (timer_active == 2 ? sp->timer.expires : jiffies); sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X" " %02X %08X:%08X %02X:%08lX %08X %5d %8d %ld %d %p", i, src, srcp, dest, destp, sp->state, atomic_read(&sp->wmem_alloc), atomic_read(&sp->rmem_alloc), timer_active, timer_expires-jiffies, 0, sp->socket->inode->i_uid, 0, sp->socket ? sp->socket->inode->i_ino : 0, atomic_read(&sp->refcnt), sp); } int raw_get_info(char *buffer, char **start, off_t offset, int length, int dummy) { int len = 0, num = 0, i; off_t pos = 0; off_t begin; char tmpbuf[129]; if (offset < 128) len += sprintf(buffer, "%-127s\n", " sl local_address rem_address st tx_queue " "rx_queue tr tm->when retrnsmt uid timeout inode"); pos = 128; read_lock(&raw_v4_lock); for (i = 0; i < RAWV4_HTABLE_SIZE; i++) { struct sock *sk; for (sk = raw_v4_htable[i]; sk; sk = sk->next, num++) { if (sk->family != PF_INET) continue; pos += 128; if (pos < offset) continue; get_raw_sock(sk, tmpbuf, i); len += sprintf(buffer+len, "%-127s\n", tmpbuf); if(len >= length) goto out; } } out: read_unlock(&raw_v4_lock); begin = len - (pos - offset); *start = buffer + begin; len -= begin; if(len > length) len = length; if (len < 0) len = 0; return len; } struct proto raw_prot = { raw_close, /* close */ udp_connect, /* connect */ udp_disconnect, /* disconnect */ NULL, /* accept */ NULL, /* retransmit */ NULL, /* write_wakeup */ NULL, /* read_wakeup */ datagram_poll, /* poll */ #ifdef CONFIG_IP_MROUTE ipmr_ioctl, /* ioctl */ #else NULL, /* ioctl */ #endif raw_init, /* init */ NULL, /* destroy */ NULL, /* shutdown */ raw_setsockopt, /* setsockopt */ raw_getsockopt, /* getsockopt */ raw_sendmsg, /* sendmsg */ raw_recvmsg, /* recvmsg */ raw_bind, /* bind */ raw_rcv_skb, /* backlog_rcv */ raw_v4_hash, /* hash */ raw_v4_unhash, /* unhash */ NULL, /* get_port */ 128, /* max_header */ 0, /* retransmits */ "RAW", /* name */ 0, /* inuse */ 0 /* highestinuse */ };