diff options
Diffstat (limited to 'net')
48 files changed, 3037 insertions, 1335 deletions
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 145f5cde96c..b7486488967 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -120,7 +120,7 @@ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev, unsigned short vid; struct net_device_stats *stats; unsigned short vlan_TCI; - unsigned short proto; + __be16 proto; /* vlan_TCI = ntohs(get_unaligned(&vhdr->h_vlan_TCI)); */ vlan_TCI = ntohs(vhdr->h_vlan_TCI); diff --git a/net/Kconfig b/net/Kconfig index 2bdd5623fdd..60f6f321bd7 100644 --- a/net/Kconfig +++ b/net/Kconfig @@ -140,6 +140,7 @@ config BRIDGE_NETFILTER If unsure, say N. +source "net/netfilter/Kconfig" source "net/ipv4/netfilter/Kconfig" source "net/ipv6/netfilter/Kconfig" source "net/decnet/netfilter/Kconfig" @@ -206,8 +207,6 @@ config NET_PKTGEN To compile this code as a module, choose M here: the module will be called pktgen. -source "net/netfilter/Kconfig" - endmenu endmenu diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c index 2d52fee63a8..d8e36b77512 100644 --- a/net/bridge/br_netfilter.c +++ b/net/bridge/br_netfilter.c @@ -214,9 +214,11 @@ static int br_nf_pre_routing_finish(struct sk_buff *skb) .tos = RT_TOS(iph->tos)} }, .proto = 0}; if (!ip_route_output_key(&rt, &fl)) { - /* Bridged-and-DNAT'ed traffic doesn't - * require ip_forwarding. */ - if (((struct dst_entry *)rt)->dev == dev) { + /* - Bridged-and-DNAT'ed traffic doesn't + * require ip_forwarding. + * - Deal with redirected traffic. */ + if (((struct dst_entry *)rt)->dev == dev || + rt->rt_type == RTN_LOCAL) { skb->dst = (struct dst_entry *)rt; goto bridged_dnat; } diff --git a/net/dccp/Makefile b/net/dccp/Makefile index fb97bb04245..344a8da153f 100644 --- a/net/dccp/Makefile +++ b/net/dccp/Makefile @@ -3,6 +3,8 @@ obj-$(CONFIG_IP_DCCP) += dccp.o dccp-y := ccid.o input.o ipv4.o minisocks.o options.o output.o proto.o \ timer.o +dccp-$(CONFIG_IP_DCCP_ACKVEC) += ackvec.o + obj-$(CONFIG_INET_DCCP_DIAG) += dccp_diag.o dccp_diag-y := diag.o diff --git a/net/dccp/ackvec.c b/net/dccp/ackvec.c new file mode 100644 index 00000000000..6530283eafc --- /dev/null +++ b/net/dccp/ackvec.c @@ -0,0 +1,419 @@ +/* + * net/dccp/ackvec.c + * + * An implementation of the DCCP protocol + * Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@ghostprotocols.net> + * + * 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; version 2 of the License; + */ + +#include "ackvec.h" +#include "dccp.h" + +#include <linux/dccp.h> +#include <linux/skbuff.h> + +#include <net/sock.h> + +int dccp_insert_option_ackvec(struct sock *sk, struct sk_buff *skb) +{ + struct dccp_sock *dp = dccp_sk(sk); + struct dccp_ackvec *av = dp->dccps_hc_rx_ackvec; + int len = av->dccpav_vec_len + 2; + struct timeval now; + u32 elapsed_time; + unsigned char *to, *from; + + dccp_timestamp(sk, &now); + elapsed_time = timeval_delta(&now, &av->dccpav_time) / 10; + + if (elapsed_time != 0) + dccp_insert_option_elapsed_time(sk, skb, elapsed_time); + + if (DCCP_SKB_CB(skb)->dccpd_opt_len + len > DCCP_MAX_OPT_LEN) + return -1; + + /* + * XXX: now we have just one ack vector sent record, so + * we have to wait for it to be cleared. + * + * Of course this is not acceptable, but this is just for + * basic testing now. + */ + if (av->dccpav_ack_seqno != DCCP_MAX_SEQNO + 1) + return -1; + + DCCP_SKB_CB(skb)->dccpd_opt_len += len; + + to = skb_push(skb, len); + *to++ = DCCPO_ACK_VECTOR_0; + *to++ = len; + + len = av->dccpav_vec_len; + from = av->dccpav_buf + av->dccpav_buf_head; + + /* Check if buf_head wraps */ + if (av->dccpav_buf_head + len > av->dccpav_vec_len) { + const u32 tailsize = (av->dccpav_vec_len - av->dccpav_buf_head); + + memcpy(to, from, tailsize); + to += tailsize; + len -= tailsize; + from = av->dccpav_buf; + } + + memcpy(to, from, len); + /* + * From draft-ietf-dccp-spec-11.txt: + * + * For each acknowledgement it sends, the HC-Receiver will add an + * acknowledgement record. ack_seqno will equal the HC-Receiver + * sequence number it used for the ack packet; ack_ptr will equal + * buf_head; ack_ackno will equal buf_ackno; and ack_nonce will + * equal buf_nonce. + * + * This implemention uses just one ack record for now. + */ + av->dccpav_ack_seqno = DCCP_SKB_CB(skb)->dccpd_seq; + av->dccpav_ack_ptr = av->dccpav_buf_head; + av->dccpav_ack_ackno = av->dccpav_buf_ackno; + av->dccpav_ack_nonce = av->dccpav_buf_nonce; + av->dccpav_sent_len = av->dccpav_vec_len; + + dccp_pr_debug("%sACK Vector 0, len=%d, ack_seqno=%llu, " + "ack_ackno=%llu\n", + debug_prefix, av->dccpav_sent_len, + (unsigned long long)av->dccpav_ack_seqno, + (unsigned long long)av->dccpav_ack_ackno); + return -1; +} + +struct dccp_ackvec *dccp_ackvec_alloc(const unsigned int len, + const unsigned int __nocast priority) +{ + struct dccp_ackvec *av = kmalloc(sizeof(*av) + len, priority); + + if (av != NULL) { + av->dccpav_buf_len = len; + av->dccpav_buf_head = + av->dccpav_buf_tail = av->dccpav_buf_len - 1; + av->dccpav_buf_ackno = + av->dccpav_ack_ackno = av->dccpav_ack_seqno = ~0LLU; + av->dccpav_buf_nonce = av->dccpav_buf_nonce = 0; + av->dccpav_ack_ptr = 0; + av->dccpav_time.tv_sec = 0; + av->dccpav_time.tv_usec = 0; + av->dccpav_sent_len = av->dccpav_vec_len = 0; + } + + return av; +} + +void dccp_ackvec_free(struct dccp_ackvec *av) +{ + kfree(av); +} + +static inline u8 dccp_ackvec_state(const struct dccp_ackvec *av, + const unsigned int index) +{ + return av->dccpav_buf[index] & DCCP_ACKVEC_STATE_MASK; +} + +static inline u8 dccp_ackvec_len(const struct dccp_ackvec *av, + const unsigned int index) +{ + return av->dccpav_buf[index] & DCCP_ACKVEC_LEN_MASK; +} + +/* + * If several packets are missing, the HC-Receiver may prefer to enter multiple + * bytes with run length 0, rather than a single byte with a larger run length; + * this simplifies table updates if one of the missing packets arrives. + */ +static inline int dccp_ackvec_set_buf_head_state(struct dccp_ackvec *av, + const unsigned int packets, + const unsigned char state) +{ + unsigned int gap; + signed long new_head; + + if (av->dccpav_vec_len + packets > av->dccpav_buf_len) + return -ENOBUFS; + + gap = packets - 1; + new_head = av->dccpav_buf_head - packets; + + if (new_head < 0) { + if (gap > 0) { + memset(av->dccpav_buf, DCCP_ACKVEC_STATE_NOT_RECEIVED, + gap + new_head + 1); + gap = -new_head; + } + new_head += av->dccpav_buf_len; + } + + av->dccpav_buf_head = new_head; + + if (gap > 0) + memset(av->dccpav_buf + av->dccpav_buf_head + 1, + DCCP_ACKVEC_STATE_NOT_RECEIVED, gap); + + av->dccpav_buf[av->dccpav_buf_head] = state; + av->dccpav_vec_len += packets; + return 0; +} + +/* + * Implements the draft-ietf-dccp-spec-11.txt Appendix A + */ +int dccp_ackvec_add(struct dccp_ackvec *av, const struct sock *sk, + const u64 ackno, const u8 state) +{ + /* + * Check at the right places if the buffer is full, if it is, tell the + * caller to start dropping packets till the HC-Sender acks our ACK + * vectors, when we will free up space in dccpav_buf. + * + * We may well decide to do buffer compression, etc, but for now lets + * just drop. + * + * From Appendix A: + * + * Of course, the circular buffer may overflow, either when the + * HC-Sender is sending data at a very high rate, when the + * HC-Receiver's acknowledgements are not reaching the HC-Sender, + * or when the HC-Sender is forgetting to acknowledge those acks + * (so the HC-Receiver is unable to clean up old state). In this + * case, the HC-Receiver should either compress the buffer (by + * increasing run lengths when possible), transfer its state to + * a larger buffer, or, as a last resort, drop all received + * packets, without processing them whatsoever, until its buffer + * shrinks again. + */ + + /* See if this is the first ackno being inserted */ + if (av->dccpav_vec_len == 0) { + av->dccpav_buf[av->dccpav_buf_head] = state; + av->dccpav_vec_len = 1; + } else if (after48(ackno, av->dccpav_buf_ackno)) { + const u64 delta = dccp_delta_seqno(av->dccpav_buf_ackno, + ackno); + + /* + * Look if the state of this packet is the same as the + * previous ackno and if so if we can bump the head len. + */ + if (delta == 1 && + dccp_ackvec_state(av, av->dccpav_buf_head) == state && + (dccp_ackvec_len(av, av->dccpav_buf_head) < + DCCP_ACKVEC_LEN_MASK)) + av->dccpav_buf[av->dccpav_buf_head]++; + else if (dccp_ackvec_set_buf_head_state(av, delta, state)) + return -ENOBUFS; + } else { + /* + * A.1.2. Old Packets + * + * When a packet with Sequence Number S arrives, and + * S <= buf_ackno, the HC-Receiver will scan the table + * for the byte corresponding to S. (Indexing structures + * could reduce the complexity of this scan.) + */ + u64 delta = dccp_delta_seqno(ackno, av->dccpav_buf_ackno); + unsigned int index = av->dccpav_buf_head; + + while (1) { + const u8 len = dccp_ackvec_len(av, index); + const u8 state = dccp_ackvec_state(av, index); + /* + * valid packets not yet in dccpav_buf have a reserved + * entry, with a len equal to 0. + */ + if (state == DCCP_ACKVEC_STATE_NOT_RECEIVED && + len == 0 && delta == 0) { /* Found our + reserved seat! */ + dccp_pr_debug("Found %llu reserved seat!\n", + (unsigned long long)ackno); + av->dccpav_buf[index] = state; + goto out; + } + /* len == 0 means one packet */ + if (delta < len + 1) + goto out_duplicate; + + delta -= len + 1; + if (++index == av->dccpav_buf_len) + index = 0; + } + } + + av->dccpav_buf_ackno = ackno; + dccp_timestamp(sk, &av->dccpav_time); +out: + dccp_pr_debug(""); + return 0; + +out_duplicate: + /* Duplicate packet */ + dccp_pr_debug("Received a dup or already considered lost " + "packet: %llu\n", (unsigned long long)ackno); + return -EILSEQ; +} + +#ifdef CONFIG_IP_DCCP_DEBUG +void dccp_ackvector_print(const u64 ackno, const unsigned char *vector, int len) +{ + if (!dccp_debug) + return; + + printk("ACK vector len=%d, ackno=%llu |", len, + (unsigned long long)ackno); + + while (len--) { + const u8 state = (*vector & DCCP_ACKVEC_STATE_MASK) >> 6; + const u8 rl = *vector & DCCP_ACKVEC_LEN_MASK; + + printk("%d,%d|", state, rl); + ++vector; + } + + printk("\n"); +} + +void dccp_ackvec_print(const struct dccp_ackvec *av) +{ + dccp_ackvector_print(av->dccpav_buf_ackno, + av->dccpav_buf + av->dccpav_buf_head, + av->dccpav_vec_len); +} +#endif + +static void dccp_ackvec_trow_away_ack_record(struct dccp_ackvec *av) +{ + /* + * As we're keeping track of the ack vector size (dccpav_vec_len) and + * the sent ack vector size (dccpav_sent_len) we don't need + * dccpav_buf_tail at all, but keep this code here as in the future + * we'll implement a vector of ack records, as suggested in + * draft-ietf-dccp-spec-11.txt Appendix A. -acme + */ +#if 0 + av->dccpav_buf_tail = av->dccpav_ack_ptr + 1; + if (av->dccpav_buf_tail >= av->dccpav_vec_len) + av->dccpav_buf_tail -= av->dccpav_vec_len; +#endif + av->dccpav_vec_len -= av->dccpav_sent_len; +} + +void dccp_ackvec_check_rcv_ackno(struct dccp_ackvec *av, struct sock *sk, + const u64 ackno) +{ + /* Check if we actually sent an ACK vector */ + if (av->dccpav_ack_seqno == DCCP_MAX_SEQNO + 1) + return; + + if (ackno == av->dccpav_ack_seqno) { +#ifdef CONFIG_IP_DCCP_DEBUG + struct dccp_sock *dp = dccp_sk(sk); + const char *debug_prefix = dp->dccps_role == DCCP_ROLE_CLIENT ? + "CLIENT rx ack: " : "server rx ack: "; +#endif + dccp_pr_debug("%sACK packet 0, len=%d, ack_seqno=%llu, " + "ack_ackno=%llu, ACKED!\n", + debug_prefix, 1, + (unsigned long long)av->dccpav_ack_seqno, + (unsigned long long)av->dccpav_ack_ackno); + dccp_ackvec_trow_away_ack_record(av); + av->dccpav_ack_seqno = DCCP_MAX_SEQNO + 1; + } +} + +static void dccp_ackvec_check_rcv_ackvector(struct dccp_ackvec *av, + struct sock *sk, u64 ackno, + const unsigned char len, + const unsigned char *vector) +{ + unsigned char i; + + /* Check if we actually sent an ACK vector */ + if (av->dccpav_ack_seqno == DCCP_MAX_SEQNO + 1) + return; + /* + * We're in the receiver half connection, so if the received an ACK + * vector ackno (e.g. 50) before dccpav_ack_seqno (e.g. 52), we're + * not interested. + * + * Extra explanation with example: + * + * if we received an ACK vector with ackno 50, it can only be acking + * 50, 49, 48, etc, not 52 (the seqno for the ACK vector we sent). + */ + /* dccp_pr_debug("is %llu < %llu? ", ackno, av->dccpav_ack_seqno); */ + if (before48(ackno, av->dccpav_ack_seqno)) { + /* dccp_pr_debug_cat("yes\n"); */ + return; + } + /* dccp_pr_debug_cat("no\n"); */ + + i = len; + while (i--) { + const u8 rl = *vector & DCCP_ACKVEC_LEN_MASK; + u64 ackno_end_rl; + + dccp_set_seqno(&ackno_end_rl, ackno - rl); + + /* + * dccp_pr_debug("is %llu <= %llu <= %llu? ", ackno_end_rl, + * av->dccpav_ack_seqno, ackno); + */ + if (between48(av->dccpav_ack_seqno, ackno_end_rl, ackno)) { + const u8 state = (*vector & + DCCP_ACKVEC_STATE_MASK) >> 6; + /* dccp_pr_debug_cat("yes\n"); */ + + if (state != DCCP_ACKVEC_STATE_NOT_RECEIVED) { +#ifdef CONFIG_IP_DCCP_DEBUG + struct dccp_sock *dp = dccp_sk(sk); + const char *debug_prefix = + dp->dccps_role == DCCP_ROLE_CLIENT ? + "CLIENT rx ack: " : "server rx ack: "; +#endif + dccp_pr_debug("%sACK vector 0, len=%d, " + "ack_seqno=%llu, ack_ackno=%llu, " + "ACKED!\n", + debug_prefix, len, + (unsigned long long) + av->dccpav_ack_seqno, + (unsigned long long) + av->dccpav_ack_ackno); + dccp_ackvec_trow_away_ack_record(av); + } + /* + * If dccpav_ack_seqno was not received, no problem + * we'll send another ACK vector. + */ + av->dccpav_ack_seqno = DCCP_MAX_SEQNO + 1; + break; + } + /* dccp_pr_debug_cat("no\n"); */ + + dccp_set_seqno(&ackno, ackno_end_rl - 1); + ++vector; + } +} + +int dccp_ackvec_parse(struct sock *sk, const struct sk_buff *skb, + const u8 opt, const u8 *value, const u8 len) +{ + if (len > DCCP_MAX_ACKVEC_LEN) + return -1; + + /* dccp_ackvector_print(DCCP_SKB_CB(skb)->dccpd_ack_seq, value, len); */ + dccp_ackvec_check_rcv_ackvector(dccp_sk(sk)->dccps_hc_rx_ackvec, sk, + DCCP_SKB_CB(skb)->dccpd_ack_seq, + len, value); + return 0; +} diff --git a/net/dccp/ackvec.h b/net/dccp/ackvec.h new file mode 100644 index 00000000000..8ca51c9191f --- /dev/null +++ b/net/dccp/ackvec.h @@ -0,0 +1,133 @@ +#ifndef _ACKVEC_H +#define _ACKVEC_H +/* + * net/dccp/ackvec.h + * + * An implementation of the DCCP protocol + * Copyright (c) 2005 Arnaldo Carvalho de Melo <acme@mandriva.com> + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#include <linux/config.h> +#include <linux/compiler.h> +#include <linux/time.h> +#include <linux/types.h> + +/* Read about the ECN nonce to see why it is 253 */ +#define DCCP_MAX_ACKVEC_LEN 253 + +#define DCCP_ACKVEC_STATE_RECEIVED 0 +#define DCCP_ACKVEC_STATE_ECN_MARKED (1 << 6) +#define DCCP_ACKVEC_STATE_NOT_RECEIVED (3 << 6) + +#define DCCP_ACKVEC_STATE_MASK 0xC0 /* 11000000 */ +#define DCCP_ACKVEC_LEN_MASK 0x3F /* 00111111 */ + +/** struct dccp_ackvec - ack vector + * + * This data structure is the one defined in the DCCP draft + * Appendix A. + * + * @dccpav_buf_head - circular buffer head + * @dccpav_buf_tail - circular buffer tail + * @dccpav_buf_ackno - ack # of the most recent packet acknowledgeable in the + * buffer (i.e. %dccpav_buf_head) + * @dccpav_buf_nonce - the one-bit sum of the ECN Nonces on all packets acked + * by the buffer with State 0 + * + * Additionally, the HC-Receiver must keep some information about the + * Ack Vectors it has recently sent. For each packet sent carrying an + * Ack Vector, it remembers four variables: + * + * @dccpav_ack_seqno - the Sequence Number used for the packet + * (HC-Receiver seqno) + * @dccpav_ack_ptr - the value of buf_head at the time of acknowledgement. + * @dccpav_ack_ackno - the Acknowledgement Number used for the packet + * (HC-Sender seqno) + * @dccpav_ack_nonce - the one-bit sum of the ECN Nonces for all State 0. + * + * @dccpav_buf_len - circular buffer length + * @dccpav_time - the time in usecs + * @dccpav_buf - circular buffer of acknowledgeable packets + */ +struct dccp_ackvec { + unsigned int dccpav_buf_head; + unsigned int dccpav_buf_tail; + u64 dccpav_buf_ackno; + u64 dccpav_ack_seqno; + u64 dccpav_ack_ackno; + unsigned int dccpav_ack_ptr; + unsigned int dccpav_sent_len; + unsigned int dccpav_vec_len; + unsigned int dccpav_buf_len; + struct timeval dccpav_time; + u8 dccpav_buf_nonce; + u8 dccpav_ack_nonce; + u8 dccpav_buf[0]; +}; + +struct sock; +struct sk_buff; + +#ifdef CONFIG_IP_DCCP_ACKVEC +extern struct dccp_ackvec *dccp_ackvec_alloc(unsigned int len, + const unsigned int __nocast priority); +extern void dccp_ackvec_free(struct dccp_ackvec *av); + +extern int dccp_ackvec_add(struct dccp_ackvec *av, const struct sock *sk, + const u64 ackno, const u8 state); + +extern void dccp_ackvec_check_rcv_ackno(struct dccp_ackvec *av, + struct sock *sk, const u64 ackno); +extern int dccp_ackvec_parse(struct sock *sk, const struct sk_buff *skb, + const u8 opt, const u8 *value, const u8 len); + +extern int dccp_insert_option_ackvec(struct sock *sk, struct sk_buff *skb); + +static inline int dccp_ackvec_pending(const struct dccp_ackvec *av) +{ + return av->dccpav_sent_len != av->dccpav_vec_len; +} +#else /* CONFIG_IP_DCCP_ACKVEC */ +static inline struct dccp_ackvec *dccp_ackvec_alloc(unsigned int len, + const unsigned int __nocast priority) +{ + return NULL; +} + +static inline void dccp_ackvec_free(struct dccp_ackvec *av) +{ +} + +static inline int dccp_ackvec_add(struct dccp_ackvec *av, const struct sock *sk, + const u64 ackno, const u8 state) +{ + return -1; +} + +static inline void dccp_ackvec_check_rcv_ackno(struct dccp_ackvec *av, + struct sock *sk, const u64 ackno) +{ +} + +static inline int dccp_ackvec_parse(struct sock *sk, const struct sk_buff *skb, + const u8 opt, const u8 *value, const u8 len) +{ + return -1; +} + +static inline int dccp_insert_option_ackvec(const struct sock *sk, + const struct sk_buff *skb) +{ + return -1; +} + +static inline int dccp_ackvec_pending(const struct dccp_ackvec *av) +{ + return 0; +} +#endif /* CONFIG_IP_DCCP_ACKVEC */ +#endif /* _ACKVEC_H */ diff --git a/net/dccp/ccid.h b/net/dccp/ccid.h index 962f1e9e2f7..21e55142dcd 100644 --- a/net/dccp/ccid.h +++ b/net/dccp/ccid.h @@ -14,6 +14,7 @@ */ #include <net/sock.h> +#include <linux/compiler.h> #include <linux/dccp.h> #include <linux/list.h> #include <linux/module.h> @@ -54,6 +55,14 @@ struct ccid { struct tcp_info *info); void (*ccid_hc_tx_get_info)(struct sock *sk, struct tcp_info *info); + int (*ccid_hc_rx_getsockopt)(struct sock *sk, + const int optname, int len, + u32 __user *optval, + int __user *optlen); + int (*ccid_hc_tx_getsockopt)(struct sock *sk, + const int optname, int len, + u32 __user *optval, + int __user *optlen); }; extern int ccid_register(struct ccid *ccid); @@ -177,4 +186,26 @@ static inline void ccid_hc_tx_get_info(struct ccid *ccid, struct sock *sk, if (ccid->ccid_hc_tx_get_info != NULL) ccid->ccid_hc_tx_get_info(sk, info); } + +static inline int ccid_hc_rx_getsockopt(struct ccid *ccid, struct sock *sk, + const int optname, int len, + u32 __user *optval, int __user *optlen) +{ + int rc = -ENOPROTOOPT; + if (ccid->ccid_hc_rx_getsockopt != NULL) + rc = ccid->ccid_hc_rx_getsockopt(sk, optname, len, + optval, optlen); + return rc; +} + +static inline int ccid_hc_tx_getsockopt(struct ccid *ccid, struct sock *sk, + const int optname, int len, + u32 __user *optval, int __user *optlen) +{ + int rc = -ENOPROTOOPT; + if (ccid->ccid_hc_tx_getsockopt != NULL) + rc = ccid->ccid_hc_tx_getsockopt(sk, optname, len, + optval, optlen); + return rc; +} #endif /* _CCID_H */ diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index 38aa8498611..aa68e0ab274 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -1120,6 +1120,60 @@ static void ccid3_hc_tx_get_info(struct sock *sk, struct tcp_info *info) info->tcpi_rtt = hctx->ccid3hctx_rtt; } +static int ccid3_hc_rx_getsockopt(struct sock *sk, const int optname, int len, + u32 __user *optval, int __user *optlen) +{ + const struct ccid3_hc_rx_sock *hcrx = ccid3_hc_rx_sk(sk); + const void *val; + + /* Listen socks doesn't have a private CCID block */ + if (sk->sk_state == DCCP_LISTEN) + return -EINVAL; + + switch (optname) { + case DCCP_SOCKOPT_CCID_RX_INFO: + if (len < sizeof(hcrx->ccid3hcrx_tfrc)) + return -EINVAL; + len = sizeof(hcrx->ccid3hcrx_tfrc); + val = &hcrx->ccid3hcrx_tfrc; + break; + default: + return -ENOPROTOOPT; + } + + if (put_user(len, optlen) || copy_to_user(optval, val, len)) + return -EFAULT; + + return 0; +} + +static int ccid3_hc_tx_getsockopt(struct sock *sk, const int optname, int len, + u32 __user *optval, int __user *optlen) +{ + const struct ccid3_hc_tx_sock *hctx = ccid3_hc_tx_sk(sk); + const void *val; + + /* Listen socks doesn't have a private CCID block */ + if (sk->sk_state == DCCP_LISTEN) + return -EINVAL; + + switch (optname) { + case DCCP_SOCKOPT_CCID_TX_INFO: + if (len < sizeof(hctx->ccid3hctx_tfrc)) + return -EINVAL; + len = sizeof(hctx->ccid3hctx_tfrc); + val = &hctx->ccid3hctx_tfrc; + break; + default: + return -ENOPROTOOPT; + } + + if (put_user(len, optlen) || copy_to_user(optval, val, len)) + return -EFAULT; + + return 0; +} + static struct ccid ccid3 = { .ccid_id = 3, .ccid_name = "ccid3", @@ -1139,6 +1193,8 @@ static struct ccid ccid3 = { .ccid_hc_rx_packet_recv = ccid3_hc_rx_packet_recv, .ccid_hc_rx_get_info = ccid3_hc_rx_get_info, .ccid_hc_tx_get_info = ccid3_hc_tx_get_info, + .ccid_hc_rx_getsockopt = ccid3_hc_rx_getsockopt, + .ccid_hc_tx_getsockopt = ccid3_hc_tx_getsockopt, }; module_param(ccid3_debug, int, 0444); diff --git a/net/dccp/ccids/ccid3.h b/net/dccp/ccids/ccid3.h index eb248778eea..0bde4583d09 100644 --- a/net/dccp/ccids/ccid3.h +++ b/net/dccp/ccids/ccid3.h @@ -40,6 +40,7 @@ #include <linux/list.h> #include <linux/time.h> #include <linux/types.h> +#include <linux/tfrc.h> #define TFRC_MIN_PACKET_SIZE 16 #define TFRC_STD_PACKET_SIZE 256 @@ -93,12 +94,15 @@ struct ccid3_options_received { * @ccid3hctx_hist - Packet history */ struct ccid3_hc_tx_sock { - u32 ccid3hctx_x; - u32 ccid3hctx_x_recv; - u32 ccid3hctx_x_calc; + struct tfrc_tx_info ccid3hctx_tfrc; +#define ccid3hctx_x ccid3hctx_tfrc.tfrctx_x +#define ccid3hctx_x_recv ccid3hctx_tfrc.tfrctx_x_recv +#define ccid3hctx_x_calc ccid3hctx_tfrc.tfrctx_x_calc +#define ccid3hctx_rtt ccid3hctx_tfrc.tfrctx_rtt +#define ccid3hctx_p ccid3hctx_tfrc.tfrctx_p +#define ccid3hctx_t_rto ccid3hctx_tfrc.tfrctx_rto +#define ccid3hctx_t_ipi ccid3hctx_tfrc.tfrctx_ipi u16 ccid3hctx_s; - u32 ccid3hctx_rtt; - u32 ccid3hctx_p; u8 ccid3hctx_state; u8 ccid3hctx_last_win_count; u8 ccid3hctx_idle; @@ -106,19 +110,19 @@ struct ccid3_hc_tx_sock { struct timer_list ccid3hctx_no_feedback_timer; struct timeval ccid3hctx_t_ld; struct timeval ccid3hctx_t_nom; - u32 ccid3hctx_t_rto; - u32 ccid3hctx_t_ipi; u32 ccid3hctx_delta; struct list_head ccid3hctx_hist; struct ccid3_options_received ccid3hctx_options_received; }; struct ccid3_hc_rx_sock { + struct tfrc_rx_info ccid3hcrx_tfrc; +#define ccid3hcrx_x_recv ccid3hcrx_tfrc.tfrcrx_x_recv +#define ccid3hcrx_rtt ccid3hcrx_tfrc.tfrcrx_rtt +#define ccid3hcrx_p ccid3hcrx_tfrc.tfrcrx_p u64 ccid3hcrx_seqno_last_counter:48, ccid3hcrx_state:8, ccid3hcrx_last_counter:4; - u32 ccid3hcrx_rtt; - u32 ccid3hcrx_p; u32 ccid3hcrx_bytes_recv; struct timeval ccid3hcrx_tstamp_last_feedback; struct timeval ccid3hcrx_tstamp_last_ack; @@ -127,7 +131,6 @@ struct ccid3_hc_rx_sock { u16 ccid3hcrx_s; u32 ccid3hcrx_pinv; u32 ccid3hcrx_elapsed_time; - u32 ccid3hcrx_x_recv; }; static inline struct ccid3_hc_tx_sock *ccid3_hc_tx_sk(const struct sock *sk) diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 95c4630b3b1..5871c027f9d 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -17,6 +17,7 @@ #include <net/snmp.h> #include <net/sock.h> #include <net/tcp.h> +#include "ackvec.h" #ifdef CONFIG_IP_DCCP_DEBUG extern int dccp_debug; @@ -258,13 +259,12 @@ extern int dccp_v4_send_reset(struct sock *sk, extern void dccp_send_close(struct sock *sk, const int active); struct dccp_skb_cb { - __u8 dccpd_type; - __u8 dccpd_reset_code; - __u8 dccpd_service; - __u8 dccpd_ccval; + __u8 dccpd_type:4; + __u8 dccpd_ccval:4; + __u8 dccpd_reset_code; + __u16 dccpd_opt_len; __u64 dccpd_seq; __u64 dccpd_ack_seq; - int dccpd_opt_len; }; #define DCCP_SKB_CB(__skb) ((struct dccp_skb_cb *)&((__skb)->cb[0])) @@ -359,6 +359,17 @@ static inline void dccp_update_gss(struct sock *sk, u64 seq) (dp->dccps_gss - dp->dccps_options.dccpo_sequence_window + 1)); } + +static inline int dccp_ack_pending(const struct sock *sk) +{ + const struct dccp_sock *dp = dccp_sk(sk); + return dp->dccps_timestamp_echo != 0 || +#ifdef CONFIG_IP_DCCP_ACKVEC + (dp->dccps_options.dccpo_send_ack_vector && + dccp_ackvec_pending(dp->dccps_hc_rx_ackvec)) || +#endif + inet_csk_ack_scheduled(sk); +} extern void dccp_insert_options(struct sock *sk, struct sk_buff *skb); extern void dccp_insert_option_elapsed_time(struct sock *sk, @@ -372,65 +383,6 @@ extern void dccp_insert_option(struct sock *sk, struct sk_buff *skb, extern struct socket *dccp_ctl_socket; -#define DCCP_ACKPKTS_STATE_RECEIVED 0 -#define DCCP_ACKPKTS_STATE_ECN_MARKED (1 << 6) -#define DCCP_ACKPKTS_STATE_NOT_RECEIVED (3 << 6) - -#define DCCP_ACKPKTS_STATE_MASK 0xC0 /* 11000000 */ -#define DCCP_ACKPKTS_LEN_MASK 0x3F /* 00111111 */ - -/** struct dccp_ackpkts - acknowledgeable packets - * - * This data structure is the one defined in the DCCP draft - * Appendix A. - * - * @dccpap_buf_head - circular buffer head - * @dccpap_buf_tail - circular buffer tail - * @dccpap_buf_ackno - ack # of the most recent packet acknowledgeable in the - * buffer (i.e. %dccpap_buf_head) - * @dccpap_buf_nonce - the one-bit sum of the ECN Nonces on all packets acked - * by the buffer with State 0 - * - * Additionally, the HC-Receiver must keep some information about the - * Ack Vectors it has recently sent. For each packet sent carrying an - * Ack Vector, it remembers four variables: - * - * @dccpap_ack_seqno - the Sequence Number used for the packet - * (HC-Receiver seqno) - * @dccpap_ack_ptr - the value of buf_head at the time of acknowledgement. - * @dccpap_ack_ackno - the Acknowledgement Number used for the packet - * (HC-Sender seqno) - * @dccpap_ack_nonce - the one-bit sum of the ECN Nonces for all State 0. - * - * @dccpap_buf_len - circular buffer length - * @dccpap_time - the time in usecs - * @dccpap_buf - circular buffer of acknowledgeable packets - */ -struct dccp_ackpkts { - unsigned int dccpap_buf_head; - unsigned int dccpap_buf_tail; - u64 dccpap_buf_ackno; - u64 dccpap_ack_seqno; - u64 dccpap_ack_ackno; - unsigned int dccpap_ack_ptr; - unsigned int dccpap_buf_vector_len; - unsigned int dccpap_ack_vector_len; - unsigned int dccpap_buf_len; - struct timeval dccpap_time; - u8 dccpap_buf_nonce; - u8 dccpap_ack_nonce; - u8 dccpap_buf[0]; -}; - -extern struct dccp_ackpkts * - dccp_ackpkts_alloc(unsigned int len, - const unsigned int __nocast priority); -extern void dccp_ackpkts_free(struct dccp_ackpkts *ap); -extern int dccp_ackpkts_add(struct dccp_ackpkts *ap, const struct sock *sk, - u64 ackno, u8 state); -extern void dccp_ackpkts_check_rcv_ackno(struct dccp_ackpkts *ap, - struct sock *sk, u64 ackno); - extern void dccp_timestamp(const struct sock *sk, struct timeval *tv); static inline suseconds_t timeval_usecs(const struct timeval *tv) @@ -471,15 +423,4 @@ static inline void timeval_sub_usecs(struct timeval *tv, } } -#ifdef CONFIG_IP_DCCP_DEBUG -extern void dccp_ackvector_print(const u64 ackno, - const unsigned char *vector, int len); -extern void dccp_ackpkts_print(const struct dccp_ackpkts *ap); -#else -static inline void dccp_ackvector_print(const u64 ackno, - const unsigned char *vector, - int len) { } -static inline void dccp_ackpkts_print(const struct dccp_ackpkts *ap) { } -#endif - #endif /* _DCCP_H */ diff --git a/net/dccp/input.c b/net/dccp/input.c index c74034cf7ed..1b6b2cb1237 100644 --- a/net/dccp/input.c +++ b/net/dccp/input.c @@ -16,6 +16,7 @@ #include <net/sock.h> +#include "ackvec.h" #include "ccid.h" #include "dccp.h" @@ -60,8 +61,8 @@ static inline void dccp_event_ack_recv(struct sock *sk, struct sk_buff *skb) struct dccp_sock *dp = dccp_sk(sk); if (dp->dccps_options.dccpo_send_ack_vector) - dccp_ackpkts_check_rcv_ackno(dp->dccps_hc_rx_ackpkts, sk, - DCCP_SKB_CB(skb)->dccpd_ack_seq); + dccp_ackvec_check_rcv_ackno(dp->dccps_hc_rx_ackvec, sk, + DCCP_SKB_CB(skb)->dccpd_ack_seq); } static int dccp_check_seqno(struct sock *sk, struct sk_buff *skb) @@ -164,37 +165,11 @@ int dccp_rcv_established(struct sock *sk, struct sk_buff *skb, if (DCCP_SKB_CB(skb)->dccpd_ack_seq != DCCP_PKT_WITHOUT_ACK_SEQ) dccp_event_ack_recv(sk, skb); - /* - * FIXME: check ECN to see if we should use - * DCCP_ACKPKTS_STATE_ECN_MARKED - */ - if (dp->dccps_options.dccpo_send_ack_vector) { - struct dccp_ackpkts *ap = dp->dccps_hc_rx_ackpkts; - - if (dccp_ackpkts_add(dp->dccps_hc_rx_ackpkts, sk, - DCCP_SKB_CB(skb)->dccpd_seq, - DCCP_ACKPKTS_STATE_RECEIVED)) { - LIMIT_NETDEBUG(KERN_WARNING "DCCP: acknowledgeable " - "packets buffer full!\n"); - ap->dccpap_ack_seqno = DCCP_MAX_SEQNO + 1; - inet_csk_schedule_ack(sk); - inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK, - TCP_DELACK_MIN, - DCCP_RTO_MAX); - goto discard; - } - - /* - * FIXME: this activation is probably wrong, have to study more - * TCP delack machinery and how it fits into DCCP draft, but - * for now it kinda "works" 8) - */ - if (!inet_csk_ack_scheduled(sk)) { - inet_csk_schedule_ack(sk); - inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK, 5 * HZ, - DCCP_RTO_MAX); - } - } + if (dp->dccps_options.dccpo_send_ack_vector && + dccp_ackvec_add(dp->dccps_hc_rx_ackvec, sk, + DCCP_SKB_CB(skb)->dccpd_seq, + DCCP_ACKVEC_STATE_RECEIVED)) + goto discard; ccid_hc_rx_packet_recv(dp->dccps_hc_rx_ccid, sk, skb); ccid_hc_tx_packet_recv(dp->dccps_hc_tx_ccid, sk, skb); @@ -384,9 +359,9 @@ static int dccp_rcv_request_sent_state_process(struct sock *sk, } out_invalid_packet: - return 1; /* dccp_v4_do_rcv will send a reset, but... - FIXME: the reset code should be - DCCP_RESET_CODE_PACKET_ERROR */ + /* dccp_v4_do_rcv will send a reset */ + DCCP_SKB_CB(skb)->dccpd_reset_code = DCCP_RESET_CODE_PACKET_ERROR; + return 1; } static int dccp_rcv_respond_partopen_state_process(struct sock *sk, @@ -433,6 +408,7 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb, struct dccp_hdr *dh, unsigned len) { struct dccp_sock *dp = dccp_sk(sk); + struct dccp_skb_cb *dcb = DCCP_SKB_CB(skb); const int old_state = sk->sk_state; int queued = 0; @@ -473,7 +449,8 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb, if (dh->dccph_type == DCCP_PKT_RESET) goto discard; - /* Caller (dccp_v4_do_rcv) will send Reset(No Connection)*/ + /* Caller (dccp_v4_do_rcv) will send Reset */ + dcb->dccpd_reset_code = DCCP_RESET_CODE_NO_CONNECTION; return 1; } @@ -487,36 +464,17 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb, if (dccp_parse_options(sk, skb)) goto discard; - if (DCCP_SKB_CB(skb)->dccpd_ack_seq != - DCCP_PKT_WITHOUT_ACK_SEQ) + if (dcb->dccpd_ack_seq != DCCP_PKT_WITHOUT_ACK_SEQ) dccp_event_ack_recv(sk, skb); ccid_hc_rx_packet_recv(dp->dccps_hc_rx_ccid, sk, skb); ccid_hc_tx_packet_recv(dp->dccps_hc_tx_ccid, sk, skb); - /* - * FIXME: check ECN to see if we should use - * DCCP_ACKPKTS_STATE_ECN_MARKED - */ - if (dp->dccps_options.dccpo_send_ack_vector) { - if (dccp_ackpkts_add(dp->dccps_hc_rx_ackpkts, sk, - DCCP_SKB_CB(skb)->dccpd_seq, - DCCP_ACKPKTS_STATE_RECEIVED)) - goto discard; - /* - * FIXME: this activation is probably wrong, have to - * study more TCP delack machinery and how it fits into - * DCCP draft, but for now it kinda "works" 8) - */ - if ((dp->dccps_hc_rx_ackpkts->dccpap_ack_seqno == - DCCP_MAX_SEQNO + 1) && - !inet_csk_ack_scheduled(sk)) { - inet_csk_schedule_ack(sk); - inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK, - TCP_DELACK_MIN, - DCCP_RTO_MAX); - } - } + if (dp->dccps_options.dccpo_send_ack_vector && + dccp_ackvec_add(dp->dccps_hc_rx_ackvec, sk, + DCCP_SKB_CB(skb)->dccpd_seq, + DCCP_ACKVEC_STATE_RECEIVED)) + goto discard; } /* @@ -551,8 +509,7 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb, dh->dccph_type == DCCP_PKT_REQUEST) || (sk->sk_state == DCCP_RESPOND && dh->dccph_type == DCCP_PKT_DATA)) { - dccp_send_sync(sk, DCCP_SKB_CB(skb)->dccpd_seq, - DCCP_PKT_SYNC); + dccp_send_sync(sk, dcb->dccpd_seq, DCCP_PKT_SYNC); goto discard; } else if (dh->dccph_type == DCCP_PKT_CLOSEREQ) { dccp_rcv_closereq(sk, skb); @@ -563,13 +520,13 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb, } if (unlikely(dh->dccph_type == DCCP_PKT_SYNC)) { - dccp_send_sync(sk, DCCP_SKB_CB(skb)->dccpd_seq, - DCCP_PKT_SYNCACK); + dccp_send_sync(sk, dcb->dccpd_seq, DCCP_PKT_SYNCACK); goto discard; } switch (sk->sk_state) { case DCCP_CLOSED: + dcb->dccpd_reset_code = DCCP_RESET_CODE_NO_CONNECTION; return 1; case DCCP_REQUESTING: diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 2afaa464e7f..40fe6afacde 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -23,6 +23,7 @@ #include <net/tcp_states.h> #include <net/xfrm.h> +#include "ackvec.h" #include "ccid.h" #include "dccp.h" @@ -246,6 +247,9 @@ static int dccp_v4_connect(struct sock *sk, struct sockaddr *uaddr, dp->dccps_role = DCCP_ROLE_CLIENT; + if (dccp_service_not_initialized(sk)) + return -EPROTO; + if (addr_len < sizeof(struct sockaddr_in)) return -EINVAL; @@ -661,6 +665,16 @@ static inline u64 dccp_v4_init_sequence(const struct sock *sk, dccp_hdr(skb)->dccph_sport); } +static inline int dccp_bad_service_code(const struct sock *sk, + const __u32 service) +{ + const struct dccp_sock *dp = dccp_sk(sk); + + if (dp->dccps_service == service) + return 0; + return !dccp_list_has_service(dp->dccps_service_list, service); +} + int dccp_v4_conn_request(struct sock *sk, struct sk_buff *skb) { struct inet_request_sock *ireq; @@ -669,13 +683,22 @@ int dccp_v4_conn_request(struct sock *sk, struct sk_buff *skb) struct dccp_request_sock *dreq; const __u32 saddr = skb->nh.iph->saddr; const __u32 daddr = skb->nh.iph->daddr; + const __u32 service = dccp_hdr_request(skb)->dccph_req_service; + struct dccp_skb_cb *dcb = DCCP_SKB_CB(skb); + __u8 reset_code = DCCP_RESET_CODE_TOO_BUSY; struct dst_entry *dst = NULL; /* Never answer to DCCP_PKT_REQUESTs send to broadcast or multicast */ if (((struct rtable *)skb->dst)->rt_flags & - (RTCF_BROADCAST | RTCF_MULTICAST)) + (RTCF_BROADCAST | RTCF_MULTICAST)) { + reset_code = DCCP_RESET_CODE_NO_CONNECTION; goto drop; + } + if (dccp_bad_service_code(sk, service)) { + reset_code = DCCP_RESET_CODE_BAD_SERVICE_CODE; + goto drop; + } /* * TW buckets are converted to open requests without * limitations, they conserve resources and peer is @@ -718,9 +741,9 @@ int dccp_v4_conn_request(struct sock *sk, struct sk_buff *skb) * dccp_create_openreq_child. */ dreq = dccp_rsk(req); - dreq->dreq_isr = DCCP_SKB_CB(skb)->dccpd_seq; - dreq->dreq_iss = dccp_v4_init_sequence(sk, skb); - dreq->dreq_service = dccp_hdr_request(skb)->dccph_req_service; + dreq->dreq_isr = dcb->dccpd_seq; + dreq->dreq_iss = dccp_v4_init_sequence(sk, skb); + dreq->dreq_service = service; if (dccp_v4_send_response(sk, req, dst)) goto drop_and_free; @@ -735,6 +758,7 @@ drop_and_free: __reqsk_free(req); drop: DCCP_INC_STATS_BH(DCCP_MIB_ATTEMPTFAILS); + dcb->dccpd_reset_code = reset_code; return -1; } @@ -1005,7 +1029,6 @@ int dccp_v4_do_rcv(struct sock *sk, struct sk_buff *skb) return 0; reset: - DCCP_SKB_CB(skb)->dccpd_reset_code = DCCP_RESET_CODE_NO_CONNECTION; dccp_v4_ctl_send_reset(skb); discard: kfree_skb(skb); @@ -1090,45 +1113,7 @@ int dccp_v4_rcv(struct sk_buff *skb) goto discard_it; dh = dccp_hdr(skb); -#if 0 - /* - * Use something like this to simulate some DATA/DATAACK loss to test - * dccp_ackpkts_add, you'll get something like this on a session that - * sends 10 DATA/DATAACK packets: - * - * ackpkts_print: 281473596467422 |0,0|3,0|0,0|3,0|0,0|3,0|0,0|3,0|0,1| - * - * 0, 0 means: DCCP_ACKPKTS_STATE_RECEIVED, RLE == just this packet - * 0, 1 means: DCCP_ACKPKTS_STATE_RECEIVED, RLE == two adjacent packets - * with the same state - * 3, 0 means: DCCP_ACKPKTS_STATE_NOT_RECEIVED, RLE == just this packet - * - * So... - * - * 281473596467422 was received - * 281473596467421 was not received - * 281473596467420 was received - * 281473596467419 was not received - * 281473596467418 was received - * 281473596467417 was not received - * 281473596467416 was received - * 281473596467415 was not received - * 281473596467414 was received - * 281473596467413 was received (this one was the 3way handshake - * RESPONSE) - * - */ - if (dh->dccph_type == DCCP_PKT_DATA || - dh->dccph_type == DCCP_PKT_DATAACK) { - static int discard = 0; - if (discard) { - discard = 0; - goto discard_it; - } - discard = 1; - } -#endif DCCP_SKB_CB(skb)->dccpd_seq = dccp_hdr_seq(skb); DCCP_SKB_CB(skb)->dccpd_type = dh->dccph_type; @@ -1242,11 +1227,9 @@ static int dccp_v4_init_sock(struct sock *sk) do_gettimeofday(&dp->dccps_epoch); if (dp->dccps_options.dccpo_send_ack_vector) { - dp->dccps_hc_rx_ackpkts = - dccp_ackpkts_alloc(DCCP_MAX_ACK_VECTOR_LEN, - GFP_KERNEL); - - if (dp->dccps_hc_rx_ackpkts == NULL) + dp->dccps_hc_rx_ackvec = dccp_ackvec_alloc(DCCP_MAX_ACKVEC_LEN, + GFP_KERNEL); + if (dp->dccps_hc_rx_ackvec == NULL) return -ENOMEM; } @@ -1258,16 +1241,18 @@ static int dccp_v4_init_sock(struct sock *sk) * setsockopt(CCIDs-I-want/accept). -acme */ if (likely(!dccp_ctl_socket_init)) { - dp->dccps_hc_rx_ccid = ccid_init(dp->dccps_options.dccpo_ccid, + dp->dccps_hc_rx_ccid = ccid_init(dp->dccps_options.dccpo_rx_ccid, sk); - dp->dccps_hc_tx_ccid = ccid_init(dp->dccps_options.dccpo_ccid, + dp->dccps_hc_tx_ccid = ccid_init(dp->dccps_options.dccpo_tx_ccid, sk); if (dp->dccps_hc_rx_ccid == NULL || dp->dccps_hc_tx_ccid == NULL) { ccid_exit(dp->dccps_hc_rx_ccid, sk); ccid_exit(dp->dccps_hc_tx_ccid, sk); - dccp_ackpkts_free(dp->dccps_hc_rx_ackpkts); - dp->dccps_hc_rx_ackpkts = NULL; + if (dp->dccps_options.dccpo_send_ack_vector) { + dccp_ackvec_free(dp->dccps_hc_rx_ackvec); + dp->dccps_hc_rx_ackvec = NULL; + } dp->dccps_hc_rx_ccid = dp->dccps_hc_tx_ccid = NULL; return -ENOMEM; } @@ -1280,6 +1265,7 @@ static int dccp_v4_init_sock(struct sock *sk) sk->sk_write_space = dccp_write_space; dp->dccps_mss_cache = 536; dp->dccps_role = DCCP_ROLE_UNDEFINED; + dp->dccps_service = DCCP_SERVICE_INVALID_VALUE; return 0; } @@ -1301,10 +1287,17 @@ static int dccp_v4_destroy_sock(struct sock *sk) if (inet_csk(sk)->icsk_bind_hash != NULL) inet_put_port(&dccp_hashinfo, sk); + if (dp->dccps_service_list != NULL) { + kfree(dp->dccps_service_list); + dp->dccps_service_list = NULL; + } + ccid_hc_rx_exit(dp->dccps_hc_rx_ccid, sk); ccid_hc_tx_exit(dp->dccps_hc_tx_ccid, sk); - dccp_ackpkts_free(dp->dccps_hc_rx_ackpkts); - dp->dccps_hc_rx_ackpkts = NULL; + if (dp->dccps_options.dccpo_send_ack_vector) { + dccp_ackvec_free(dp->dccps_hc_rx_ackvec); + dp->dccps_hc_rx_ackvec = NULL; + } ccid_exit(dp->dccps_hc_rx_ccid, sk); ccid_exit(dp->dccps_hc_tx_ccid, sk); dp->dccps_hc_rx_ccid = dp->dccps_hc_tx_ccid = NULL; diff --git a/net/dccp/minisocks.c b/net/dccp/minisocks.c index 18461bc04cb..1393461898b 100644 --- a/net/dccp/minisocks.c +++ b/net/dccp/minisocks.c @@ -19,6 +19,7 @@ #include <net/xfrm.h> #include <net/inet_timewait_sock.h> +#include "ackvec.h" #include "ccid.h" #include "dccp.h" @@ -93,22 +94,24 @@ struct sock *dccp_create_openreq_child(struct sock *sk, struct inet_connection_sock *newicsk = inet_csk(sk); struct dccp_sock *newdp = dccp_sk(newsk); - newdp->dccps_hc_rx_ackpkts = NULL; - newdp->dccps_role = DCCP_ROLE_SERVER; - newicsk->icsk_rto = DCCP_TIMEOUT_INIT; + newdp->dccps_role = DCCP_ROLE_SERVER; + newdp->dccps_hc_rx_ackvec = NULL; + newdp->dccps_service_list = NULL; + newdp->dccps_service = dreq->dreq_service; + newicsk->icsk_rto = DCCP_TIMEOUT_INIT; do_gettimeofday(&newdp->dccps_epoch); if (newdp->dccps_options.dccpo_send_ack_vector) { - newdp->dccps_hc_rx_ackpkts = - dccp_ackpkts_alloc(DCCP_MAX_ACK_VECTOR_LEN, - GFP_ATOMIC); + newdp->dccps_hc_rx_ackvec = + dccp_ackvec_alloc(DCCP_MAX_ACKVEC_LEN, + GFP_ATOMIC); /* * XXX: We're using the same CCIDs set on the parent, * i.e. sk_clone copied the master sock and left the * CCID pointers for this child, that is why we do the * __ccid_get calls. */ - if (unlikely(newdp->dccps_hc_rx_ackpkts == NULL)) + if (unlikely(newdp->dccps_hc_rx_ackvec == NULL)) goto out_free; } @@ -116,7 +119,7 @@ struct sock *dccp_create_openreq_child(struct sock *sk, newsk) != 0 || ccid_hc_tx_init(newdp->dccps_hc_tx_ccid, newsk) != 0)) { - dccp_ackpkts_free(newdp->dccps_hc_rx_ackpkts); + dccp_ackvec_free(newdp->dccps_hc_rx_ackvec); ccid_hc_rx_exit(newdp->dccps_hc_rx_ccid, newsk); ccid_hc_tx_exit(newdp->dccps_hc_tx_ccid, newsk); out_free: diff --git a/net/dccp/options.c b/net/dccp/options.c index d4c4242d8dd..0a76426c9ae 100644 --- a/net/dccp/options.c +++ b/net/dccp/options.c @@ -18,19 +18,15 @@ #include <linux/kernel.h> #include <linux/skbuff.h> +#include "ackvec.h" #include "ccid.h" #include "dccp.h" -static void dccp_ackpkts_check_rcv_ackvector(struct dccp_ackpkts *ap, - struct sock *sk, - const u64 ackno, - const unsigned char len, - const unsigned char *vector); - /* stores the default values for new connection. may be changed with sysctl */ static const struct dccp_options dccpo_default_values = { .dccpo_sequence_window = DCCPF_INITIAL_SEQUENCE_WINDOW, - .dccpo_ccid = DCCPF_INITIAL_CCID, + .dccpo_rx_ccid = DCCPF_INITIAL_CCID, + .dccpo_tx_ccid = DCCPF_INITIAL_CCID, .dccpo_send_ack_vector = DCCPF_INITIAL_SEND_ACK_VECTOR, .dccpo_send_ndp_count = DCCPF_INITIAL_SEND_NDP_COUNT, }; @@ -113,25 +109,13 @@ int dccp_parse_options(struct sock *sk, struct sk_buff *skb) opt_recv->dccpor_ndp); break; case DCCPO_ACK_VECTOR_0: - if (len > DCCP_MAX_ACK_VECTOR_LEN) - goto out_invalid_option; - + case DCCPO_ACK_VECTOR_1: if (pkt_type == DCCP_PKT_DATA) continue; - opt_recv->dccpor_ack_vector_len = len; - opt_recv->dccpor_ack_vector_idx = value - options; - - dccp_pr_debug("%sACK vector 0, len=%d, ack_ackno=%llu\n", - debug_prefix, len, - (unsigned long long) - DCCP_SKB_CB(skb)->dccpd_ack_seq); - dccp_ackvector_print(DCCP_SKB_CB(skb)->dccpd_ack_seq, - value, len); - dccp_ackpkts_check_rcv_ackvector(dp->dccps_hc_rx_ackpkts, - sk, - DCCP_SKB_CB(skb)->dccpd_ack_seq, - len, value); + if (dp->dccps_options.dccpo_send_ack_vector && + dccp_ackvec_parse(sk, skb, opt, value, len)) + goto out_invalid_option; break; case DCCPO_TIMESTAMP: if (len != 4) @@ -352,86 +336,6 @@ void dccp_insert_option_elapsed_time(struct sock *sk, EXPORT_SYMBOL_GPL(dccp_insert_option_elapsed_time); -static void dccp_insert_option_ack_vector(struct sock *sk, struct sk_buff *skb) -{ - struct dccp_sock *dp = dccp_sk(sk); -#ifdef CONFIG_IP_DCCP_DEBUG - const char *debug_prefix = dp->dccps_role == DCCP_ROLE_CLIENT ? - "CLIENT TX opt: " : "server TX opt: "; -#endif - struct dccp_ackpkts *ap = dp->dccps_hc_rx_ackpkts; - int len = ap->dccpap_buf_vector_len + 2; - struct timeval now; - u32 elapsed_time; - unsigned char *to, *from; - - dccp_timestamp(sk, &now); - elapsed_time = timeval_delta(&now, &ap->dccpap_time) / 10; - - if (elapsed_time != 0) - dccp_insert_option_elapsed_time(sk, skb, elapsed_time); - - if (DCCP_SKB_CB(skb)->dccpd_opt_len + len > DCCP_MAX_OPT_LEN) { - LIMIT_NETDEBUG(KERN_INFO "DCCP: packet too small to " - "insert ACK Vector!\n"); - return; - } - - /* - * XXX: now we have just one ack vector sent record, so - * we have to wait for it to be cleared. - * - * Of course this is not acceptable, but this is just for - * basic testing now. - */ - if (ap->dccpap_ack_seqno != DCCP_MAX_SEQNO + 1) - return; - - DCCP_SKB_CB(skb)->dccpd_opt_len += len; - - to = skb_push(skb, len); - *to++ = DCCPO_ACK_VECTOR_0; - *to++ = len; - - len = ap->dccpap_buf_vector_len; - from = ap->dccpap_buf + ap->dccpap_buf_head; - - /* Check if buf_head wraps */ - if (ap->dccpap_buf_head + len > ap->dccpap_buf_len) { - const unsigned int tailsize = (ap->dccpap_buf_len - - ap->dccpap_buf_head); - - memcpy(to, from, tailsize); - to += tailsize; - len -= tailsize; - from = ap->dccpap_buf; - } - - memcpy(to, from, len); - /* - * From draft-ietf-dccp-spec-11.txt: - * - * For each acknowledgement it sends, the HC-Receiver will add an - * acknowledgement record. ack_seqno will equal the HC-Receiver - * sequence number it used for the ack packet; ack_ptr will equal - * buf_head; ack_ackno will equal buf_ackno; and ack_nonce will - * equal buf_nonce. - * - * This implemention uses just one ack record for now. - */ - ap->dccpap_ack_seqno = DCCP_SKB_CB(skb)->dccpd_seq; - ap->dccpap_ack_ptr = ap->dccpap_buf_head; - ap->dccpap_ack_ackno = ap->dccpap_buf_ackno; - ap->dccpap_ack_nonce = ap->dccpap_buf_nonce; - ap->dccpap_ack_vector_len = ap->dccpap_buf_vector_len; - - dccp_pr_debug("%sACK Vector 0, len=%d, ack_seqno=%llu, " - "ack_ackno=%llu\n", - debug_prefix, ap->dccpap_ack_vector_len, - (unsigned long long) ap->dccpap_ack_seqno, - (unsigned long long) ap->dccpap_ack_ackno); -} - void dccp_timestamp(const struct sock *sk, struct timeval *tv) { const struct dccp_sock *dp = dccp_sk(sk); @@ -528,9 +432,8 @@ void dccp_insert_options(struct sock *sk, struct sk_buff *skb) if (!dccp_packet_without_ack(skb)) { if (dp->dccps_options.dccpo_send_ack_vector && - (dp->dccps_hc_rx_ackpkts->dccpap_buf_ackno != - DCCP_MAX_SEQNO + 1)) - dccp_insert_option_ack_vector(sk, skb); + dccp_ackvec_pending(dp->dccps_hc_rx_ackvec)) + dccp_insert_option_ackvec(sk, skb); if (dp->dccps_timestamp_echo != 0) dccp_insert_option_timestamp_echo(sk, skb); } @@ -557,331 +460,3 @@ void dccp_insert_options(struct sock *sk, struct sk_buff *skb) } } } - -struct dccp_ackpkts *dccp_ackpkts_alloc(const unsigned int len, - const unsigned int __nocast priority) -{ - struct dccp_ackpkts *ap = kmalloc(sizeof(*ap) + len, priority); - - if (ap != NULL) { -#ifdef CONFIG_IP_DCCP_DEBUG - memset(ap->dccpap_buf, 0xFF, len); -#endif - ap->dccpap_buf_len = len; - ap->dccpap_buf_head = - ap->dccpap_buf_tail = - ap->dccpap_buf_len - 1; - ap->dccpap_buf_ackno = - ap->dccpap_ack_ackno = - ap->dccpap_ack_seqno = DCCP_MAX_SEQNO + 1; - ap->dccpap_buf_nonce = ap->dccpap_buf_nonce = 0; - ap->dccpap_ack_ptr = 0; - ap->dccpap_time.tv_sec = 0; - ap->dccpap_time.tv_usec = 0; - ap->dccpap_buf_vector_len = ap->dccpap_ack_vector_len = 0; - } - - return ap; -} - -void dccp_ackpkts_free(struct dccp_ackpkts *ap) -{ - if (ap != NULL) { -#ifdef CONFIG_IP_DCCP_DEBUG - memset(ap, 0xFF, sizeof(*ap) + ap->dccpap_buf_len); -#endif - kfree(ap); - } -} - -static inline u8 dccp_ackpkts_state(const struct dccp_ackpkts *ap, - const unsigned int index) -{ - return ap->dccpap_buf[index] & DCCP_ACKPKTS_STATE_MASK; -} - -static inline u8 dccp_ackpkts_len(const struct dccp_ackpkts *ap, - const unsigned int index) -{ - return ap->dccpap_buf[index] & DCCP_ACKPKTS_LEN_MASK; -} - -/* - * If several packets are missing, the HC-Receiver may prefer to enter multiple - * bytes with run length 0, rather than a single byte with a larger run length; - * this simplifies table updates if one of the missing packets arrives. - */ -static inline int dccp_ackpkts_set_buf_head_state(struct dccp_ackpkts *ap, - const unsigned int packets, - const unsigned char state) -{ - unsigned int gap; - signed long new_head; - - if (ap->dccpap_buf_vector_len + packets > ap->dccpap_buf_len) - return -ENOBUFS; - - gap = packets - 1; - new_head = ap->dccpap_buf_head - packets; - - if (new_head < 0) { - if (gap > 0) { - memset(ap->dccpap_buf, DCCP_ACKPKTS_STATE_NOT_RECEIVED, - gap + new_head + 1); - gap = -new_head; - } - new_head += ap->dccpap_buf_len; - } - - ap->dccpap_buf_head = new_head; - - if (gap > 0) - memset(ap->dccpap_buf + ap->dccpap_buf_head + 1, - DCCP_ACKPKTS_STATE_NOT_RECEIVED, gap); - - ap->dccpap_buf[ap->dccpap_buf_head] = state; - ap->dccpap_buf_vector_len += packets; - return 0; -} - -/* - * Implements the draft-ietf-dccp-spec-11.txt Appendix A - */ -int dccp_ackpkts_add(struct dccp_ackpkts *ap, const struct sock *sk, - u64 ackno, u8 state) -{ - /* - * Check at the right places if the buffer is full, if it is, tell the - * caller to start dropping packets till the HC-Sender acks our ACK - * vectors, when we will free up space in dccpap_buf. - * - * We may well decide to do buffer compression, etc, but for now lets - * just drop. - * - * From Appendix A: - * - * Of course, the circular buffer may overflow, either when the - * HC-Sender is sending data at a very high rate, when the - * HC-Receiver's acknowledgements are not reaching the HC-Sender, - * or when the HC-Sender is forgetting to acknowledge those acks - * (so the HC-Receiver is unable to clean up old state). In this - * case, the HC-Receiver should either compress the buffer (by - * increasing run lengths when possible), transfer its state to - * a larger buffer, or, as a last resort, drop all received - * packets, without processing them whatsoever, until its buffer - * shrinks again. - */ - - /* See if this is the first ackno being inserted */ - if (ap->dccpap_buf_vector_len == 0) { - ap->dccpap_buf[ap->dccpap_buf_head] = state; - ap->dccpap_buf_vector_len = 1; - } else if (after48(ackno, ap->dccpap_buf_ackno)) { - const u64 delta = dccp_delta_seqno(ap->dccpap_buf_ackno, - ackno); - - /* - * Look if the state of this packet is the same as the - * previous ackno and if so if we can bump the head len. - */ - if (delta == 1 && - dccp_ackpkts_state(ap, ap->dccpap_buf_head) == state && - (dccp_ackpkts_len(ap, ap->dccpap_buf_head) < - DCCP_ACKPKTS_LEN_MASK)) - ap->dccpap_buf[ap->dccpap_buf_head]++; - else if (dccp_ackpkts_set_buf_head_state(ap, delta, state)) - return -ENOBUFS; - } else { - /* - * A.1.2. Old Packets - * - * When a packet with Sequence Number S arrives, and - * S <= buf_ackno, the HC-Receiver will scan the table - * for the byte corresponding to S. (Indexing structures - * could reduce the complexity of this scan.) - */ - u64 delta = dccp_delta_seqno(ackno, ap->dccpap_buf_ackno); - unsigned int index = ap->dccpap_buf_head; - - while (1) { - const u8 len = dccp_ackpkts_len(ap, index); - const u8 state = dccp_ackpkts_state(ap, index); - /* - * valid packets not yet in dccpap_buf have a reserved - * entry, with a len equal to 0. - */ - if (state == DCCP_ACKPKTS_STATE_NOT_RECEIVED && - len == 0 && delta == 0) { /* Found our - reserved seat! */ - dccp_pr_debug("Found %llu reserved seat!\n", - (unsigned long long) ackno); - ap->dccpap_buf[index] = state; - goto out; - } - /* len == 0 means one packet */ - if (delta < len + 1) - goto out_duplicate; - - delta -= len + 1; - if (++index == ap->dccpap_buf_len) - index = 0; - } - } - - ap->dccpap_buf_ackno = ackno; - dccp_timestamp(sk, &ap->dccpap_time); -out: - dccp_pr_debug(""); - dccp_ackpkts_print(ap); - return 0; - -out_duplicate: - /* Duplicate packet */ - dccp_pr_debug("Received a dup or already considered lost " - "packet: %llu\n", (unsigned long long) ackno); - return -EILSEQ; -} - -#ifdef CONFIG_IP_DCCP_DEBUG -void dccp_ackvector_print(const u64 ackno, const unsigned char *vector, - int len) -{ - if (!dccp_debug) - return; - - printk("ACK vector len=%d, ackno=%llu |", len, - (unsigned long long) ackno); - - while (len--) { - const u8 state = (*vector & DCCP_ACKPKTS_STATE_MASK) >> 6; - const u8 rl = (*vector & DCCP_ACKPKTS_LEN_MASK); - - printk("%d,%d|", state, rl); - ++vector; - } - - printk("\n"); -} - -void dccp_ackpkts_print(const struct dccp_ackpkts *ap) -{ - dccp_ackvector_print(ap->dccpap_buf_ackno, - ap->dccpap_buf + ap->dccpap_buf_head, - ap->dccpap_buf_vector_len); -} -#endif - -static void dccp_ackpkts_trow_away_ack_record(struct dccp_ackpkts *ap) -{ - /* - * As we're keeping track of the ack vector size - * (dccpap_buf_vector_len) and the sent ack vector size - * (dccpap_ack_vector_len) we don't need dccpap_buf_tail at all, but - * keep this code here as in the future we'll implement a vector of - * ack records, as suggested in draft-ietf-dccp-spec-11.txt - * Appendix A. -acme - */ -#if 0 - ap->dccpap_buf_tail = ap->dccpap_ack_ptr + 1; - if (ap->dccpap_buf_tail >= ap->dccpap_buf_len) - ap->dccpap_buf_tail -= ap->dccpap_buf_len; -#endif - ap->dccpap_buf_vector_len -= ap->dccpap_ack_vector_len; -} - -void dccp_ackpkts_check_rcv_ackno(struct dccp_ackpkts *ap, struct sock *sk, - u64 ackno) -{ - /* Check if we actually sent an ACK vector */ - if (ap->dccpap_ack_seqno == DCCP_MAX_SEQNO + 1) - return; - - if (ackno == ap->dccpap_ack_seqno) { -#ifdef CONFIG_IP_DCCP_DEBUG - struct dccp_sock *dp = dccp_sk(sk); - const char *debug_prefix = dp->dccps_role == DCCP_ROLE_CLIENT ? - "CLIENT rx ack: " : "server rx ack: "; -#endif - dccp_pr_debug("%sACK packet 0, len=%d, ack_seqno=%llu, " - "ack_ackno=%llu, ACKED!\n", - debug_prefix, 1, - (unsigned long long) ap->dccpap_ack_seqno, - (unsigned long long) ap->dccpap_ack_ackno); - dccp_ackpkts_trow_away_ack_record(ap); - ap->dccpap_ack_seqno = DCCP_MAX_SEQNO + 1; - } -} - -static void dccp_ackpkts_check_rcv_ackvector(struct dccp_ackpkts *ap, - struct sock *sk, u64 ackno, - const unsigned char len, - const unsigned char *vector) -{ - unsigned char i; - - /* Check if we actually sent an ACK vector */ - if (ap->dccpap_ack_seqno == DCCP_MAX_SEQNO + 1) - return; - /* - * We're in the receiver half connection, so if the received an ACK - * vector ackno (e.g. 50) before dccpap_ack_seqno (e.g. 52), we're - * not interested. - * - * Extra explanation with example: - * - * if we received an ACK vector with ackno 50, it can only be acking - * 50, 49, 48, etc, not 52 (the seqno for the ACK vector we sent). - */ - /* dccp_pr_debug("is %llu < %llu? ", ackno, ap->dccpap_ack_seqno); */ - if (before48(ackno, ap->dccpap_ack_seqno)) { - /* dccp_pr_debug_cat("yes\n"); */ - return; - } - /* dccp_pr_debug_cat("no\n"); */ - - i = len; - while (i--) { - const u8 rl = (*vector & DCCP_ACKPKTS_LEN_MASK); - u64 ackno_end_rl; - - dccp_set_seqno(&ackno_end_rl, ackno - rl); - - /* - * dccp_pr_debug("is %llu <= %llu <= %llu? ", ackno_end_rl, - * ap->dccpap_ack_seqno, ackno); - */ - if (between48(ap->dccpap_ack_seqno, ackno_end_rl, ackno)) { - const u8 state = (*vector & - DCCP_ACKPKTS_STATE_MASK) >> 6; - /* dccp_pr_debug_cat("yes\n"); */ - - if (state != DCCP_ACKPKTS_STATE_NOT_RECEIVED) { -#ifdef CONFIG_IP_DCCP_DEBUG - struct dccp_sock *dp = dccp_sk(sk); - const char *debug_prefix = - dp->dccps_role == DCCP_ROLE_CLIENT ? - "CLIENT rx ack: " : "server rx ack: "; -#endif - dccp_pr_debug("%sACK vector 0, len=%d, " - "ack_seqno=%llu, ack_ackno=%llu, " - "ACKED!\n", - debug_prefix, len, - (unsigned long long) - ap->dccpap_ack_seqno, - (unsigned long long) - ap->dccpap_ack_ackno); - dccp_ackpkts_trow_away_ack_record(ap); - } - /* - * If dccpap_ack_seqno was not received, no problem - * we'll send another ACK vector. - */ - ap->dccpap_ack_seqno = DCCP_MAX_SEQNO + 1; - break; - } - /* dccp_pr_debug_cat("no\n"); */ - - dccp_set_seqno(&ackno, ackno_end_rl - 1); - ++vector; - } -} diff --git a/net/dccp/output.c b/net/dccp/output.c index ea6d0e91e51..4786bdcddcc 100644 --- a/net/dccp/output.c +++ b/net/dccp/output.c @@ -16,6 +16,7 @@ #include <net/sock.h> +#include "ackvec.h" #include "ccid.h" #include "dccp.h" @@ -85,7 +86,7 @@ int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) switch (dcb->dccpd_type) { case DCCP_PKT_REQUEST: dccp_hdr_request(skb)->dccph_req_service = - dcb->dccpd_service; + dp->dccps_service; break; case DCCP_PKT_RESET: dccp_hdr_reset(skb)->dccph_reset_code = @@ -225,7 +226,6 @@ int dccp_write_xmit(struct sock *sk, struct sk_buff *skb, long *timeo) err = dccp_wait_for_ccid(sk, skb, timeo); if (err == 0) { - const struct dccp_ackpkts *ap = dp->dccps_hc_rx_ackpkts; struct dccp_skb_cb *dcb = DCCP_SKB_CB(skb); const int len = skb->len; @@ -236,15 +236,7 @@ int dccp_write_xmit(struct sock *sk, struct sk_buff *skb, long *timeo) inet_csk(sk)->icsk_rto, DCCP_RTO_MAX); dcb->dccpd_type = DCCP_PKT_DATAACK; - /* - * FIXME: we really should have a - * dccps_ack_pending or use icsk. - */ - } else if (inet_csk_ack_scheduled(sk) || - dp->dccps_timestamp_echo != 0 || - (dp->dccps_options.dccpo_send_ack_vector && - ap->dccpap_buf_ackno != DCCP_MAX_SEQNO + 1 && - ap->dccpap_ack_seqno == DCCP_MAX_SEQNO + 1)) + } else if (dccp_ack_pending(sk)) dcb->dccpd_type = DCCP_PKT_DATAACK; else dcb->dccpd_type = DCCP_PKT_DATA; @@ -270,6 +262,7 @@ struct sk_buff *dccp_make_response(struct sock *sk, struct dst_entry *dst, struct request_sock *req) { struct dccp_hdr *dh; + struct dccp_request_sock *dreq; const int dccp_header_size = sizeof(struct dccp_hdr) + sizeof(struct dccp_hdr_ext) + sizeof(struct dccp_hdr_response); @@ -285,8 +278,9 @@ struct sk_buff *dccp_make_response(struct sock *sk, struct dst_entry *dst, skb->dst = dst_clone(dst); skb->csum = 0; + dreq = dccp_rsk(req); DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_RESPONSE; - DCCP_SKB_CB(skb)->dccpd_seq = dccp_rsk(req)->dreq_iss; + DCCP_SKB_CB(skb)->dccpd_seq = dreq->dreq_iss; dccp_insert_options(sk, skb); skb->h.raw = skb_push(skb, dccp_header_size); @@ -300,8 +294,9 @@ struct sk_buff *dccp_make_response(struct sock *sk, struct dst_entry *dst, DCCP_SKB_CB(skb)->dccpd_opt_len) / 4; dh->dccph_type = DCCP_PKT_RESPONSE; dh->dccph_x = 1; - dccp_hdr_set_seq(dh, dccp_rsk(req)->dreq_iss); - dccp_hdr_set_ack(dccp_hdr_ack_bits(skb), dccp_rsk(req)->dreq_isr); + dccp_hdr_set_seq(dh, dreq->dreq_iss); + dccp_hdr_set_ack(dccp_hdr_ack_bits(skb), dreq->dreq_isr); + dccp_hdr_response(skb)->dccph_resp_service = dreq->dreq_service; dh->dccph_checksum = dccp_v4_checksum(skb, inet_rsk(req)->loc_addr, inet_rsk(req)->rmt_addr); @@ -397,9 +392,6 @@ int dccp_connect(struct sock *sk) skb_reserve(skb, MAX_DCCP_HEADER); DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_REQUEST; - /* FIXME: set service to something meaningful, coming - * from userspace*/ - DCCP_SKB_CB(skb)->dccpd_service = 0; skb->csum = 0; skb_set_owner_w(skb, sk); diff --git a/net/dccp/proto.c b/net/dccp/proto.c index 18a0e69c9dc..a1cfd0e9e3b 100644 --- a/net/dccp/proto.c +++ b/net/dccp/proto.c @@ -94,7 +94,15 @@ EXPORT_SYMBOL_GPL(dccp_state_name); static inline int dccp_listen_start(struct sock *sk) { - dccp_sk(sk)->dccps_role = DCCP_ROLE_LISTEN; + struct dccp_sock *dp = dccp_sk(sk); + + dp->dccps_role = DCCP_ROLE_LISTEN; + /* + * Apps need to use setsockopt(DCCP_SOCKOPT_SERVICE) + * before calling listen() + */ + if (dccp_service_not_initialized(sk)) + return -EPROTO; return inet_csk_listen_start(sk, TCP_SYNQ_HSIZE); } @@ -202,6 +210,42 @@ int dccp_ioctl(struct sock *sk, int cmd, unsigned long arg) return -ENOIOCTLCMD; } +static int dccp_setsockopt_service(struct sock *sk, const u32 service, + char __user *optval, int optlen) +{ + struct dccp_sock *dp = dccp_sk(sk); + struct dccp_service_list *sl = NULL; + + if (service == DCCP_SERVICE_INVALID_VALUE || + optlen > DCCP_SERVICE_LIST_MAX_LEN * sizeof(u32)) + return -EINVAL; + + if (optlen > sizeof(service)) { + sl = kmalloc(optlen, GFP_KERNEL); + if (sl == NULL) + return -ENOMEM; + + sl->dccpsl_nr = optlen / sizeof(u32) - 1; + if (copy_from_user(sl->dccpsl_list, + optval + sizeof(service), + optlen - sizeof(service)) || + dccp_list_has_service(sl, DCCP_SERVICE_INVALID_VALUE)) { + kfree(sl); + return -EFAULT; + } + } + + lock_sock(sk); + dp->dccps_service = service; + + if (dp->dccps_service_list != NULL) + kfree(dp->dccps_service_list); + + dp->dccps_service_list = sl; + release_sock(sk); + return 0; +} + int dccp_setsockopt(struct sock *sk, int level, int optname, char __user *optval, int optlen) { @@ -218,8 +262,10 @@ int dccp_setsockopt(struct sock *sk, int level, int optname, if (get_user(val, (int __user *)optval)) return -EFAULT; - lock_sock(sk); + if (optname == DCCP_SOCKOPT_SERVICE) + return dccp_setsockopt_service(sk, val, optval, optlen); + lock_sock(sk); dp = dccp_sk(sk); err = 0; @@ -236,6 +282,37 @@ int dccp_setsockopt(struct sock *sk, int level, int optname, return err; } +static int dccp_getsockopt_service(struct sock *sk, int len, + u32 __user *optval, + int __user *optlen) +{ + const struct dccp_sock *dp = dccp_sk(sk); + const struct dccp_service_list *sl; + int err = -ENOENT, slen = 0, total_len = sizeof(u32); + + lock_sock(sk); + if (dccp_service_not_initialized(sk)) + goto out; + + if ((sl = dp->dccps_service_list) != NULL) { + slen = sl->dccpsl_nr * sizeof(u32); + total_len += slen; + } + + err = -EINVAL; + if (total_len > len) + goto out; + + err = 0; + if (put_user(total_len, optlen) || + put_user(dp->dccps_service, optval) || + (sl != NULL && copy_to_user(optval + 1, sl->dccpsl_list, slen))) + err = -EFAULT; +out: + release_sock(sk); + return err; +} + int dccp_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen) { @@ -248,8 +325,7 @@ int dccp_getsockopt(struct sock *sk, int level, int optname, if (get_user(len, optlen)) return -EFAULT; - len = min_t(unsigned int, len, sizeof(int)); - if (len < 0) + if (len < sizeof(int)) return -EINVAL; dp = dccp_sk(sk); @@ -257,7 +333,17 @@ int dccp_getsockopt(struct sock *sk, int level, int optname, switch (optname) { case DCCP_SOCKOPT_PACKET_SIZE: val = dp->dccps_packet_size; + len = sizeof(dp->dccps_packet_size); break; + case DCCP_SOCKOPT_SERVICE: + return dccp_getsockopt_service(sk, len, + (u32 __user *)optval, optlen); + case 128 ... 191: + return ccid_hc_rx_getsockopt(dp->dccps_hc_rx_ccid, sk, optname, + len, (u32 __user *)optval, optlen); + case 192 ... 255: + return ccid_hc_tx_getsockopt(dp->dccps_hc_tx_ccid, sk, optname, + len, (u32 __user *)optval, optlen); default: return -ENOPROTOOPT; } diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index 1b63b482416..50c0519cd70 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -43,7 +43,7 @@ * 2 of the License, or (at your option) any later version. */ -#define VERSION "0.403" +#define VERSION "0.404" #include <linux/config.h> #include <asm/uaccess.h> @@ -224,7 +224,7 @@ static inline int tkey_mismatch(t_key a, int offset, t_key b) Consider a node 'n' and its parent 'tp'. If n is a leaf, every bit in its key is significant. Its presence is - necessitaded by path compression, since during a tree traversal (when + necessitated by path compression, since during a tree traversal (when searching for a leaf - unless we are doing an insertion) we will completely ignore all skipped bits we encounter. Thus we need to verify, at the end of a potentially successful search, that we have indeed been walking the @@ -836,11 +836,12 @@ static void trie_init(struct trie *t) #endif } -/* readside most use rcu_read_lock currently dump routines +/* readside must use rcu_read_lock currently dump routines via get_fa_head and dump */ -static struct leaf_info *find_leaf_info(struct hlist_head *head, int plen) +static struct leaf_info *find_leaf_info(struct leaf *l, int plen) { + struct hlist_head *head = &l->list; struct hlist_node *node; struct leaf_info *li; @@ -853,7 +854,7 @@ static struct leaf_info *find_leaf_info(struct hlist_head *head, int plen) static inline struct list_head * get_fa_head(struct leaf *l, int plen) { - struct leaf_info *li = find_leaf_info(&l->list, plen); + struct leaf_info *li = find_leaf_info(l, plen); if (!li) return NULL; @@ -1085,7 +1086,7 @@ fib_insert_node(struct trie *t, int *err, u32 key, int plen) } if (tp && tp->pos + tp->bits > 32) - printk("ERROR tp=%p pos=%d, bits=%d, key=%0x plen=%d\n", + printk(KERN_WARNING "fib_trie tp=%p pos=%d, bits=%d, key=%0x plen=%d\n", tp, tp->pos, tp->bits, key, plen); /* Rebalance the trie */ @@ -1248,7 +1249,7 @@ err: } -/* should be clalled with rcu_read_lock */ +/* should be called with rcu_read_lock */ static inline int check_leaf(struct trie *t, struct leaf *l, t_key key, int *plen, const struct flowi *flp, struct fib_result *res) @@ -1590,7 +1591,7 @@ fn_trie_delete(struct fib_table *tb, struct rtmsg *r, struct kern_rta *rta, rtmsg_fib(RTM_DELROUTE, htonl(key), fa, plen, tb->tb_id, nlhdr, req); l = fib_find_node(t, key); - li = find_leaf_info(&l->list, plen); + li = find_leaf_info(l, plen); list_del_rcu(&fa->fa_list); @@ -1714,7 +1715,6 @@ static int fn_trie_flush(struct fib_table *tb) t->revision++; - rcu_read_lock(); for (h = 0; (l = nextleaf(t, l)) != NULL; h++) { found += trie_flush_leaf(t, l); @@ -1722,7 +1722,6 @@ static int fn_trie_flush(struct fib_table *tb) trie_leaf_remove(t, ll->key); ll = l; } - rcu_read_unlock(); if (ll && hlist_empty(&ll->list)) trie_leaf_remove(t, ll->key); @@ -1833,16 +1832,7 @@ static int fn_trie_dump_fa(t_key key, int plen, struct list_head *fah, struct fi i++; continue; } - if (fa->fa_info->fib_nh == NULL) { - printk("Trie error _fib_nh=NULL in fa[%d] k=%08x plen=%d\n", i, key, plen); - i++; - continue; - } - if (fa->fa_info == NULL) { - printk("Trie error fa_info=NULL in fa[%d] k=%08x plen=%d\n", i, key, plen); - i++; - continue; - } + BUG_ON(!fa->fa_info); if (fib_dump_info(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq, @@ -1965,7 +1955,7 @@ struct fib_table * __init fib_hash_init(int id) trie_main = t; if (id == RT_TABLE_LOCAL) - printk("IPv4 FIB: Using LC-trie version %s\n", VERSION); + printk(KERN_INFO "IPv4 FIB: Using LC-trie version %s\n", VERSION); return tb; } @@ -2029,7 +2019,7 @@ static struct node *fib_trie_get_first(struct fib_trie_iter *iter, iter->tnode = (struct tnode *) n; iter->trie = t; iter->index = 0; - iter->depth = 0; + iter->depth = 1; return n; } return NULL; @@ -2274,11 +2264,12 @@ static int fib_trie_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "<local>:\n"); else seq_puts(seq, "<main>:\n"); - } else { - seq_indent(seq, iter->depth-1); - seq_printf(seq, " +-- %d.%d.%d.%d/%d\n", - NIPQUAD(prf), tn->pos); - } + } + seq_indent(seq, iter->depth-1); + seq_printf(seq, " +-- %d.%d.%d.%d/%d %d %d %d\n", + NIPQUAD(prf), tn->pos, tn->bits, tn->full_children, + tn->empty_children); + } else { struct leaf *l = (struct leaf *) n; int i; @@ -2287,7 +2278,7 @@ static int fib_trie_seq_show(struct seq_file *seq, void *v) seq_indent(seq, iter->depth); seq_printf(seq, " |-- %d.%d.%d.%d\n", NIPQUAD(val)); for (i = 32; i >= 0; i--) { - struct leaf_info *li = find_leaf_info(&l->list, i); + struct leaf_info *li = find_leaf_info(l, i); if (li) { struct fib_alias *fa; list_for_each_entry_rcu(fa, &li->falh, fa_list) { @@ -2383,7 +2374,7 @@ static int fib_route_seq_show(struct seq_file *seq, void *v) return 0; for (i=32; i>=0; i--) { - struct leaf_info *li = find_leaf_info(&l->list, i); + struct leaf_info *li = find_leaf_info(l, i); struct fib_alias *fa; u32 mask, prefix; diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 44607f4767b..70c44e4c3ce 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -1603,7 +1603,7 @@ static void ip_mc_clear_src(struct ip_mc_list *pmc) } pmc->sources = NULL; pmc->sfmode = MCAST_EXCLUDE; - pmc->sfcount[MCAST_EXCLUDE] = 0; + pmc->sfcount[MCAST_INCLUDE] = 0; pmc->sfcount[MCAST_EXCLUDE] = 1; } diff --git a/net/ipv4/ipvs/ip_vs_conn.c b/net/ipv4/ipvs/ip_vs_conn.c index e11952ea17a..f828fa2eb7d 100644 --- a/net/ipv4/ipvs/ip_vs_conn.c +++ b/net/ipv4/ipvs/ip_vs_conn.c @@ -196,6 +196,7 @@ static inline struct ip_vs_conn *__ip_vs_conn_in_get list_for_each_entry(cp, &ip_vs_conn_tab[hash], c_list) { if (s_addr==cp->caddr && s_port==cp->cport && d_port==cp->vport && d_addr==cp->vaddr && + ((!s_port) ^ (!(cp->flags & IP_VS_CONN_F_NO_CPORT))) && protocol==cp->protocol) { /* HIT */ atomic_inc(&cp->refcnt); @@ -227,6 +228,40 @@ struct ip_vs_conn *ip_vs_conn_in_get return cp; } +/* Get reference to connection template */ +struct ip_vs_conn *ip_vs_ct_in_get +(int protocol, __u32 s_addr, __u16 s_port, __u32 d_addr, __u16 d_port) +{ + unsigned hash; + struct ip_vs_conn *cp; + + hash = ip_vs_conn_hashkey(protocol, s_addr, s_port); + + ct_read_lock(hash); + + list_for_each_entry(cp, &ip_vs_conn_tab[hash], c_list) { + if (s_addr==cp->caddr && s_port==cp->cport && + d_port==cp->vport && d_addr==cp->vaddr && + cp->flags & IP_VS_CONN_F_TEMPLATE && + protocol==cp->protocol) { + /* HIT */ + atomic_inc(&cp->refcnt); + goto out; + } + } + cp = NULL; + + out: + ct_read_unlock(hash); + + IP_VS_DBG(7, "template lookup/in %s %u.%u.%u.%u:%d->%u.%u.%u.%u:%d %s\n", + ip_vs_proto_name(protocol), + NIPQUAD(s_addr), ntohs(s_port), + NIPQUAD(d_addr), ntohs(d_port), + cp?"hit":"not hit"); + + return cp; +} /* * Gets ip_vs_conn associated with supplied parameters in the ip_vs_conn_tab. @@ -367,7 +402,7 @@ ip_vs_bind_dest(struct ip_vs_conn *cp, struct ip_vs_dest *dest) atomic_read(&dest->refcnt)); /* Update the connection counters */ - if (cp->cport || (cp->flags & IP_VS_CONN_F_NO_CPORT)) { + if (!(cp->flags & IP_VS_CONN_F_TEMPLATE)) { /* It is a normal connection, so increase the inactive connection counter because it is in TCP SYNRECV state (inactive) or other protocol inacive state */ @@ -406,7 +441,7 @@ static inline void ip_vs_unbind_dest(struct ip_vs_conn *cp) atomic_read(&dest->refcnt)); /* Update the connection counters */ - if (cp->cport || (cp->flags & IP_VS_CONN_F_NO_CPORT)) { + if (!(cp->flags & IP_VS_CONN_F_TEMPLATE)) { /* It is a normal connection, so decrease the inactconns or activeconns counter */ if (cp->flags & IP_VS_CONN_F_INACTIVE) { @@ -467,7 +502,7 @@ int ip_vs_check_template(struct ip_vs_conn *ct) /* * Invalidate the connection template */ - if (ct->cport) { + if (ct->vport != 65535) { if (ip_vs_conn_unhash(ct)) { ct->dport = 65535; ct->vport = 65535; @@ -776,7 +811,7 @@ void ip_vs_random_dropentry(void) ct_write_lock_bh(hash); list_for_each_entry(cp, &ip_vs_conn_tab[hash], c_list) { - if (!cp->cport && !(cp->flags & IP_VS_CONN_F_NO_CPORT)) + if (cp->flags & IP_VS_CONN_F_TEMPLATE) /* connection template */ continue; diff --git a/net/ipv4/ipvs/ip_vs_core.c b/net/ipv4/ipvs/ip_vs_core.c index 3ac7eeca04a..981cc3244ef 100644 --- a/net/ipv4/ipvs/ip_vs_core.c +++ b/net/ipv4/ipvs/ip_vs_core.c @@ -243,10 +243,10 @@ ip_vs_sched_persist(struct ip_vs_service *svc, if (ports[1] == svc->port) { /* Check if a template already exists */ if (svc->port != FTPPORT) - ct = ip_vs_conn_in_get(iph->protocol, snet, 0, + ct = ip_vs_ct_in_get(iph->protocol, snet, 0, iph->daddr, ports[1]); else - ct = ip_vs_conn_in_get(iph->protocol, snet, 0, + ct = ip_vs_ct_in_get(iph->protocol, snet, 0, iph->daddr, 0); if (!ct || !ip_vs_check_template(ct)) { @@ -272,14 +272,14 @@ ip_vs_sched_persist(struct ip_vs_service *svc, iph->daddr, ports[1], dest->addr, dest->port, - 0, + IP_VS_CONN_F_TEMPLATE, dest); else ct = ip_vs_conn_new(iph->protocol, snet, 0, iph->daddr, 0, dest->addr, 0, - 0, + IP_VS_CONN_F_TEMPLATE, dest); if (ct == NULL) return NULL; @@ -298,10 +298,10 @@ ip_vs_sched_persist(struct ip_vs_service *svc, * port zero template: <protocol,caddr,0,vaddr,0,daddr,0> */ if (svc->fwmark) - ct = ip_vs_conn_in_get(IPPROTO_IP, snet, 0, + ct = ip_vs_ct_in_get(IPPROTO_IP, snet, 0, htonl(svc->fwmark), 0); else - ct = ip_vs_conn_in_get(iph->protocol, snet, 0, + ct = ip_vs_ct_in_get(iph->protocol, snet, 0, iph->daddr, 0); if (!ct || !ip_vs_check_template(ct)) { @@ -326,14 +326,14 @@ ip_vs_sched_persist(struct ip_vs_service *svc, snet, 0, htonl(svc->fwmark), 0, dest->addr, 0, - 0, + IP_VS_CONN_F_TEMPLATE, dest); else ct = ip_vs_conn_new(iph->protocol, snet, 0, iph->daddr, 0, dest->addr, 0, - 0, + IP_VS_CONN_F_TEMPLATE, dest); if (ct == NULL) return NULL; diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index 574d1f509b4..2e5ced3d806 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -297,16 +297,24 @@ static void ip_vs_process_message(const char *buffer, const size_t buflen) p = (char *)buffer + sizeof(struct ip_vs_sync_mesg); for (i=0; i<m->nr_conns; i++) { + unsigned flags; + s = (struct ip_vs_sync_conn *)p; - cp = ip_vs_conn_in_get(s->protocol, - s->caddr, s->cport, - s->vaddr, s->vport); + flags = ntohs(s->flags); + if (!(flags & IP_VS_CONN_F_TEMPLATE)) + cp = ip_vs_conn_in_get(s->protocol, + s->caddr, s->cport, + s->vaddr, s->vport); + else + cp = ip_vs_ct_in_get(s->protocol, + s->caddr, s->cport, + s->vaddr, s->vport); if (!cp) { cp = ip_vs_conn_new(s->protocol, s->caddr, s->cport, s->vaddr, s->vport, s->daddr, s->dport, - ntohs(s->flags), NULL); + flags, NULL); if (!cp) { IP_VS_ERR("ip_vs_conn_new failed\n"); return; @@ -315,11 +323,11 @@ static void ip_vs_process_message(const char *buffer, const size_t buflen) } else if (!cp->dest) { /* it is an entry created by the synchronization */ cp->state = ntohs(s->state); - cp->flags = ntohs(s->flags) | IP_VS_CONN_F_HASHED; + cp->flags = flags | IP_VS_CONN_F_HASHED; } /* Note that we don't touch its state and flags if it is a normal entry. */ - if (ntohs(s->flags) & IP_VS_CONN_F_SEQ_MASK) { + if (flags & IP_VS_CONN_F_SEQ_MASK) { opt = (struct ip_vs_sync_conn_options *)&s[1]; memcpy(&cp->in_seq, opt, sizeof(*opt)); p += FULL_CONN_SIZE; diff --git a/net/ipv4/netfilter/Kconfig b/net/ipv4/netfilter/Kconfig index 30aa8e2ee21..3cf9b451675 100644 --- a/net/ipv4/netfilter/Kconfig +++ b/net/ipv4/netfilter/Kconfig @@ -51,6 +51,14 @@ config IP_NF_CONNTRACK_EVENTS IF unsure, say `N'. +config IP_NF_CONNTRACK_NETLINK + tristate 'Connection tracking netlink interface' + depends on IP_NF_CONNTRACK && NETFILTER_NETLINK + depends on IP_NF_CONNTRACK!=y || NETFILTER_NETLINK!=m + help + This option enables support for a netlink-based userspace interface + + config IP_NF_CT_PROTO_SCTP tristate 'SCTP protocol connection tracking support (EXPERIMENTAL)' depends on IP_NF_CONNTRACK && EXPERIMENTAL @@ -129,6 +137,22 @@ config IP_NF_AMANDA To compile it as a module, choose M here. If unsure, say Y. +config IP_NF_PPTP + tristate 'PPTP protocol support' + help + This module adds support for PPTP (Point to Point Tunnelling + Protocol, RFC2637) conncection tracking and NAT. + + If you are running PPTP sessions over a stateful firewall or NAT + box, you may want to enable this feature. + + Please note that not all PPTP modes of operation are supported yet. + For more info, read top of the file + net/ipv4/netfilter/ip_conntrack_pptp.c + + If you want to compile it as a module, say M here and read + Documentation/modules.txt. If unsure, say `N'. + config IP_NF_QUEUE tristate "IP Userspace queueing via NETLINK (OBSOLETE)" help @@ -613,6 +637,12 @@ config IP_NF_NAT_AMANDA default IP_NF_NAT if IP_NF_AMANDA=y default m if IP_NF_AMANDA=m +config IP_NF_NAT_PPTP + tristate + depends on IP_NF_NAT!=n && IP_NF_PPTP!=n + default IP_NF_NAT if IP_NF_PPTP=y + default m if IP_NF_PPTP=m + # mangle + specific targets config IP_NF_MANGLE tristate "Packet mangling" @@ -774,11 +804,5 @@ config IP_NF_ARP_MANGLE Allows altering the ARP packet payload: source and destination hardware and network addresses. -config IP_NF_CONNTRACK_NETLINK - tristate 'Connection tracking netlink interface' - depends on IP_NF_CONNTRACK && NETFILTER_NETLINK - help - This option enables support for a netlink-based userspace interface - endmenu diff --git a/net/ipv4/netfilter/Makefile b/net/ipv4/netfilter/Makefile index 1ba0db74681..3d45d3c0283 100644 --- a/net/ipv4/netfilter/Makefile +++ b/net/ipv4/netfilter/Makefile @@ -6,6 +6,9 @@ ip_conntrack-objs := ip_conntrack_standalone.o ip_conntrack_core.o ip_conntrack_proto_generic.o ip_conntrack_proto_tcp.o ip_conntrack_proto_udp.o ip_conntrack_proto_icmp.o iptable_nat-objs := ip_nat_standalone.o ip_nat_rule.o ip_nat_core.o ip_nat_helper.o ip_nat_proto_unknown.o ip_nat_proto_tcp.o ip_nat_proto_udp.o ip_nat_proto_icmp.o +ip_conntrack_pptp-objs := ip_conntrack_helper_pptp.o ip_conntrack_proto_gre.o +ip_nat_pptp-objs := ip_nat_helper_pptp.o ip_nat_proto_gre.o + # connection tracking obj-$(CONFIG_IP_NF_CONNTRACK) += ip_conntrack.o @@ -17,6 +20,7 @@ obj-$(CONFIG_IP_NF_CONNTRACK_NETLINK) += ip_conntrack_netlink.o obj-$(CONFIG_IP_NF_CT_PROTO_SCTP) += ip_conntrack_proto_sctp.o # connection tracking helpers +obj-$(CONFIG_IP_NF_PPTP) += ip_conntrack_pptp.o obj-$(CONFIG_IP_NF_AMANDA) += ip_conntrack_amanda.o obj-$(CONFIG_IP_NF_TFTP) += ip_conntrack_tftp.o obj-$(CONFIG_IP_NF_FTP) += ip_conntrack_ftp.o @@ -24,6 +28,7 @@ obj-$(CONFIG_IP_NF_IRC) += ip_conntrack_irc.o obj-$(CONFIG_IP_NF_NETBIOS_NS) += ip_conntrack_netbios_ns.o # NAT helpers +obj-$(CONFIG_IP_NF_NAT_PPTP) += ip_nat_pptp.o obj-$(CONFIG_IP_NF_NAT_AMANDA) += ip_nat_amanda.o obj-$(CONFIG_IP_NF_NAT_TFTP) += ip_nat_tftp.o obj-$(CONFIG_IP_NF_NAT_FTP) += ip_nat_ftp.o diff --git a/net/ipv4/netfilter/ip_conntrack_core.c b/net/ipv4/netfilter/ip_conntrack_core.c index 19cba16e6e1..c1f82e0c81c 100644 --- a/net/ipv4/netfilter/ip_conntrack_core.c +++ b/net/ipv4/netfilter/ip_conntrack_core.c @@ -233,7 +233,7 @@ __ip_conntrack_expect_find(const struct ip_conntrack_tuple *tuple) /* Just find a expectation corresponding to a tuple. */ struct ip_conntrack_expect * -ip_conntrack_expect_find_get(const struct ip_conntrack_tuple *tuple) +ip_conntrack_expect_find(const struct ip_conntrack_tuple *tuple) { struct ip_conntrack_expect *i; @@ -1143,7 +1143,10 @@ void ip_ct_refresh_acct(struct ip_conntrack *ct, if (del_timer(&ct->timeout)) { ct->timeout.expires = jiffies + extra_jiffies; add_timer(&ct->timeout); - ip_conntrack_event_cache(IPCT_REFRESH, skb); + /* FIXME: We loose some REFRESH events if this function + * is called without an skb. I'll fix this later -HW */ + if (skb) + ip_conntrack_event_cache(IPCT_REFRESH, skb); } ct_add_counters(ct, ctinfo, skb); write_unlock_bh(&ip_conntrack_lock); diff --git a/net/ipv4/netfilter/ip_conntrack_helper_pptp.c b/net/ipv4/netfilter/ip_conntrack_helper_pptp.c new file mode 100644 index 00000000000..79db5b70d5f --- /dev/null +++ b/net/ipv4/netfilter/ip_conntrack_helper_pptp.c @@ -0,0 +1,805 @@ +/* + * ip_conntrack_pptp.c - Version 3.0 + * + * Connection tracking support for PPTP (Point to Point Tunneling Protocol). + * PPTP is a a protocol for creating virtual private networks. + * It is a specification defined by Microsoft and some vendors + * working with Microsoft. PPTP is built on top of a modified + * version of the Internet Generic Routing Encapsulation Protocol. + * GRE is defined in RFC 1701 and RFC 1702. Documentation of + * PPTP can be found in RFC 2637 + * + * (C) 2000-2005 by Harald Welte <laforge@gnumonks.org> + * + * Development of this code funded by Astaro AG (http://www.astaro.com/) + * + * Limitations: + * - We blindly assume that control connections are always + * established in PNS->PAC direction. This is a violation + * of RFFC2673 + * - We can only support one single call within each session + * + * TODO: + * - testing of incoming PPTP calls + * + * Changes: + * 2002-02-05 - Version 1.3 + * - Call ip_conntrack_unexpect_related() from + * pptp_destroy_siblings() to destroy expectations in case + * CALL_DISCONNECT_NOTIFY or tcp fin packet was seen + * (Philip Craig <philipc@snapgear.com>) + * - Add Version information at module loadtime + * 2002-02-10 - Version 1.6 + * - move to C99 style initializers + * - remove second expectation if first arrives + * 2004-10-22 - Version 2.0 + * - merge Mandrake's 2.6.x port with recent 2.6.x API changes + * - fix lots of linear skb assumptions from Mandrake's port + * 2005-06-10 - Version 2.1 + * - use ip_conntrack_expect_free() instead of kfree() on the + * expect's (which are from the slab for quite some time) + * 2005-06-10 - Version 3.0 + * - port helper to post-2.6.11 API changes, + * funded by Oxcoda NetBox Blue (http://www.netboxblue.com/) + * 2005-07-30 - Version 3.1 + * - port helper to 2.6.13 API changes + * + */ + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/netfilter.h> +#include <linux/ip.h> +#include <net/checksum.h> +#include <net/tcp.h> + +#include <linux/netfilter_ipv4/ip_conntrack.h> +#include <linux/netfilter_ipv4/ip_conntrack_core.h> +#include <linux/netfilter_ipv4/ip_conntrack_helper.h> +#include <linux/netfilter_ipv4/ip_conntrack_proto_gre.h> +#include <linux/netfilter_ipv4/ip_conntrack_pptp.h> + +#define IP_CT_PPTP_VERSION "3.1" + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Harald Welte <laforge@gnumonks.org>"); +MODULE_DESCRIPTION("Netfilter connection tracking helper module for PPTP"); + +static DEFINE_SPINLOCK(ip_pptp_lock); + +int +(*ip_nat_pptp_hook_outbound)(struct sk_buff **pskb, + struct ip_conntrack *ct, + enum ip_conntrack_info ctinfo, + struct PptpControlHeader *ctlh, + union pptp_ctrl_union *pptpReq); + +int +(*ip_nat_pptp_hook_inbound)(struct sk_buff **pskb, + struct ip_conntrack *ct, + enum ip_conntrack_info ctinfo, + struct PptpControlHeader *ctlh, + union pptp_ctrl_union *pptpReq); + +int +(*ip_nat_pptp_hook_exp_gre)(struct ip_conntrack_expect *expect_orig, + struct ip_conntrack_expect *expect_reply); + +void +(*ip_nat_pptp_hook_expectfn)(struct ip_conntrack *ct, + struct ip_conntrack_expect *exp); + +#if 0 +/* PptpControlMessageType names */ +const char *pptp_msg_name[] = { + "UNKNOWN_MESSAGE", + "START_SESSION_REQUEST", + "START_SESSION_REPLY", + "STOP_SESSION_REQUEST", + "STOP_SESSION_REPLY", + "ECHO_REQUEST", + "ECHO_REPLY", + "OUT_CALL_REQUEST", + "OUT_CALL_REPLY", + "IN_CALL_REQUEST", + "IN_CALL_REPLY", + "IN_CALL_CONNECT", + "CALL_CLEAR_REQUEST", + "CALL_DISCONNECT_NOTIFY", + "WAN_ERROR_NOTIFY", + "SET_LINK_INFO" +}; +EXPORT_SYMBOL(pptp_msg_name); +#define DEBUGP(format, args...) printk(KERN_DEBUG "%s:%s: " format, __FILE__, __FUNCTION__, ## args) +#else +#define DEBUGP(format, args...) +#endif + +#define SECS *HZ +#define MINS * 60 SECS +#define HOURS * 60 MINS + +#define PPTP_GRE_TIMEOUT (10 MINS) +#define PPTP_GRE_STREAM_TIMEOUT (5 HOURS) + +static void pptp_expectfn(struct ip_conntrack *ct, + struct ip_conntrack_expect *exp) +{ + DEBUGP("increasing timeouts\n"); + + /* increase timeout of GRE data channel conntrack entry */ + ct->proto.gre.timeout = PPTP_GRE_TIMEOUT; + ct->proto.gre.stream_timeout = PPTP_GRE_STREAM_TIMEOUT; + + /* Can you see how rusty this code is, compared with the pre-2.6.11 + * one? That's what happened to my shiny newnat of 2002 ;( -HW */ + + if (!ip_nat_pptp_hook_expectfn) { + struct ip_conntrack_tuple inv_t; + struct ip_conntrack_expect *exp_other; + + /* obviously this tuple inversion only works until you do NAT */ + invert_tuplepr(&inv_t, &exp->tuple); + DEBUGP("trying to unexpect other dir: "); + DUMP_TUPLE(&inv_t); + + exp_other = ip_conntrack_expect_find(&inv_t); + if (exp_other) { + /* delete other expectation. */ + DEBUGP("found\n"); + ip_conntrack_unexpect_related(exp_other); + ip_conntrack_expect_put(exp_other); + } else { + DEBUGP("not found\n"); + } + } else { + /* we need more than simple inversion */ + ip_nat_pptp_hook_expectfn(ct, exp); + } +} + +static int destroy_sibling_or_exp(const struct ip_conntrack_tuple *t) +{ + struct ip_conntrack_tuple_hash *h; + struct ip_conntrack_expect *exp; + + DEBUGP("trying to timeout ct or exp for tuple "); + DUMP_TUPLE(t); + + h = ip_conntrack_find_get(t, NULL); + if (h) { + struct ip_conntrack *sibling = tuplehash_to_ctrack(h); + DEBUGP("setting timeout of conntrack %p to 0\n", sibling); + sibling->proto.gre.timeout = 0; + sibling->proto.gre.stream_timeout = 0; + /* refresh_acct will not modify counters if skb == NULL */ + if (del_timer(&sibling->timeout)) + sibling->timeout.function((unsigned long)sibling); + ip_conntrack_put(sibling); + return 1; + } else { + exp = ip_conntrack_expect_find(t); + if (exp) { + DEBUGP("unexpect_related of expect %p\n", exp); + ip_conntrack_unexpect_related(exp); + ip_conntrack_expect_put(exp); + return 1; + } + } + + return 0; +} + + +/* timeout GRE data connections */ +static void pptp_destroy_siblings(struct ip_conntrack *ct) +{ + struct ip_conntrack_tuple t; + + /* Since ct->sibling_list has literally rusted away in 2.6.11, + * we now need another way to find out about our sibling + * contrack and expects... -HW */ + + /* try original (pns->pac) tuple */ + memcpy(&t, &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple, sizeof(t)); + t.dst.protonum = IPPROTO_GRE; + t.src.u.gre.key = htons(ct->help.ct_pptp_info.pns_call_id); + t.dst.u.gre.key = htons(ct->help.ct_pptp_info.pac_call_id); + + if (!destroy_sibling_or_exp(&t)) + DEBUGP("failed to timeout original pns->pac ct/exp\n"); + + /* try reply (pac->pns) tuple */ + memcpy(&t, &ct->tuplehash[IP_CT_DIR_REPLY].tuple, sizeof(t)); + t.dst.protonum = IPPROTO_GRE; + t.src.u.gre.key = htons(ct->help.ct_pptp_info.pac_call_id); + t.dst.u.gre.key = htons(ct->help.ct_pptp_info.pns_call_id); + + if (!destroy_sibling_or_exp(&t)) + DEBUGP("failed to timeout reply pac->pns ct/exp\n"); +} + +/* expect GRE connections (PNS->PAC and PAC->PNS direction) */ +static inline int +exp_gre(struct ip_conntrack *master, + u_int32_t seq, + u_int16_t callid, + u_int16_t peer_callid) +{ + struct ip_conntrack_tuple inv_tuple; + struct ip_conntrack_tuple exp_tuples[] = { + /* tuple in original direction, PNS->PAC */ + { .src = { .ip = master->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.ip, + .u = { .gre = { .key = peer_callid } } + }, + .dst = { .ip = master->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.ip, + .u = { .gre = { .key = callid } }, + .protonum = IPPROTO_GRE + }, + }, + /* tuple in reply direction, PAC->PNS */ + { .src = { .ip = master->tuplehash[IP_CT_DIR_REPLY].tuple.src.ip, + .u = { .gre = { .key = callid } } + }, + .dst = { .ip = master->tuplehash[IP_CT_DIR_REPLY].tuple.dst.ip, + .u = { .gre = { .key = peer_callid } }, + .protonum = IPPROTO_GRE + }, + } + }; + struct ip_conntrack_expect *exp_orig, *exp_reply; + int ret = 1; + + exp_orig = ip_conntrack_expect_alloc(master); + if (exp_orig == NULL) + goto out; + + exp_reply = ip_conntrack_expect_alloc(master); + if (exp_reply == NULL) + goto out_put_orig; + + memcpy(&exp_orig->tuple, &exp_tuples[0], sizeof(exp_orig->tuple)); + + exp_orig->mask.src.ip = 0xffffffff; + exp_orig->mask.src.u.all = 0; + exp_orig->mask.dst.u.all = 0; + exp_orig->mask.dst.u.gre.key = 0xffff; + exp_orig->mask.dst.ip = 0xffffffff; + exp_orig->mask.dst.protonum = 0xff; + + exp_orig->master = master; + exp_orig->expectfn = pptp_expectfn; + exp_orig->flags = 0; + + exp_orig->dir = IP_CT_DIR_ORIGINAL; + + /* both expectations are identical apart from tuple */ + memcpy(exp_reply, exp_orig, sizeof(*exp_reply)); + memcpy(&exp_reply->tuple, &exp_tuples[1], sizeof(exp_reply->tuple)); + + exp_reply->dir = !exp_orig->dir; + + if (ip_nat_pptp_hook_exp_gre) + ret = ip_nat_pptp_hook_exp_gre(exp_orig, exp_reply); + else { + + DEBUGP("calling expect_related PNS->PAC"); + DUMP_TUPLE(&exp_orig->tuple); + + if (ip_conntrack_expect_related(exp_orig) != 0) { + DEBUGP("cannot expect_related()\n"); + goto out_put_both; + } + + DEBUGP("calling expect_related PAC->PNS"); + DUMP_TUPLE(&exp_reply->tuple); + + if (ip_conntrack_expect_related(exp_reply) != 0) { + DEBUGP("cannot expect_related()\n"); + goto out_unexpect_orig; + } + + /* Add GRE keymap entries */ + if (ip_ct_gre_keymap_add(master, &exp_reply->tuple, 0) != 0) { + DEBUGP("cannot keymap_add() exp\n"); + goto out_unexpect_both; + } + + invert_tuplepr(&inv_tuple, &exp_reply->tuple); + if (ip_ct_gre_keymap_add(master, &inv_tuple, 1) != 0) { + ip_ct_gre_keymap_destroy(master); + DEBUGP("cannot keymap_add() exp_inv\n"); + goto out_unexpect_both; + } + ret = 0; + } + +out_put_both: + ip_conntrack_expect_put(exp_reply); +out_put_orig: + ip_conntrack_expect_put(exp_orig); +out: + return ret; + +out_unexpect_both: + ip_conntrack_unexpect_related(exp_reply); +out_unexpect_orig: + ip_conntrack_unexpect_related(exp_orig); + goto out_put_both; +} + +static inline int +pptp_inbound_pkt(struct sk_buff **pskb, + struct tcphdr *tcph, + unsigned int nexthdr_off, + unsigned int datalen, + struct ip_conntrack *ct, + enum ip_conntrack_info ctinfo) +{ + struct PptpControlHeader _ctlh, *ctlh; + unsigned int reqlen; + union pptp_ctrl_union _pptpReq, *pptpReq; + struct ip_ct_pptp_master *info = &ct->help.ct_pptp_info; + u_int16_t msg, *cid, *pcid; + u_int32_t seq; + + ctlh = skb_header_pointer(*pskb, nexthdr_off, sizeof(_ctlh), &_ctlh); + if (!ctlh) { + DEBUGP("error during skb_header_pointer\n"); + return NF_ACCEPT; + } + nexthdr_off += sizeof(_ctlh); + datalen -= sizeof(_ctlh); + + reqlen = datalen; + if (reqlen > sizeof(*pptpReq)) + reqlen = sizeof(*pptpReq); + pptpReq = skb_header_pointer(*pskb, nexthdr_off, reqlen, &_pptpReq); + if (!pptpReq) { + DEBUGP("error during skb_header_pointer\n"); + return NF_ACCEPT; + } + + msg = ntohs(ctlh->messageType); + DEBUGP("inbound control message %s\n", pptp_msg_name[msg]); + + switch (msg) { + case PPTP_START_SESSION_REPLY: + if (reqlen < sizeof(_pptpReq.srep)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* server confirms new control session */ + if (info->sstate < PPTP_SESSION_REQUESTED) { + DEBUGP("%s without START_SESS_REQUEST\n", + pptp_msg_name[msg]); + break; + } + if (pptpReq->srep.resultCode == PPTP_START_OK) + info->sstate = PPTP_SESSION_CONFIRMED; + else + info->sstate = PPTP_SESSION_ERROR; + break; + + case PPTP_STOP_SESSION_REPLY: + if (reqlen < sizeof(_pptpReq.strep)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* server confirms end of control session */ + if (info->sstate > PPTP_SESSION_STOPREQ) { + DEBUGP("%s without STOP_SESS_REQUEST\n", + pptp_msg_name[msg]); + break; + } + if (pptpReq->strep.resultCode == PPTP_STOP_OK) + info->sstate = PPTP_SESSION_NONE; + else + info->sstate = PPTP_SESSION_ERROR; + break; + + case PPTP_OUT_CALL_REPLY: + if (reqlen < sizeof(_pptpReq.ocack)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* server accepted call, we now expect GRE frames */ + if (info->sstate != PPTP_SESSION_CONFIRMED) { + DEBUGP("%s but no session\n", pptp_msg_name[msg]); + break; + } + if (info->cstate != PPTP_CALL_OUT_REQ && + info->cstate != PPTP_CALL_OUT_CONF) { + DEBUGP("%s without OUTCALL_REQ\n", pptp_msg_name[msg]); + break; + } + if (pptpReq->ocack.resultCode != PPTP_OUTCALL_CONNECT) { + info->cstate = PPTP_CALL_NONE; + break; + } + + cid = &pptpReq->ocack.callID; + pcid = &pptpReq->ocack.peersCallID; + + info->pac_call_id = ntohs(*cid); + + if (htons(info->pns_call_id) != *pcid) { + DEBUGP("%s for unknown callid %u\n", + pptp_msg_name[msg], ntohs(*pcid)); + break; + } + + DEBUGP("%s, CID=%X, PCID=%X\n", pptp_msg_name[msg], + ntohs(*cid), ntohs(*pcid)); + + info->cstate = PPTP_CALL_OUT_CONF; + + seq = ntohl(tcph->seq) + sizeof(struct pptp_pkt_hdr) + + sizeof(struct PptpControlHeader) + + ((void *)pcid - (void *)pptpReq); + + if (exp_gre(ct, seq, *cid, *pcid) != 0) + printk("ip_conntrack_pptp: error during exp_gre\n"); + break; + + case PPTP_IN_CALL_REQUEST: + if (reqlen < sizeof(_pptpReq.icack)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* server tells us about incoming call request */ + if (info->sstate != PPTP_SESSION_CONFIRMED) { + DEBUGP("%s but no session\n", pptp_msg_name[msg]); + break; + } + pcid = &pptpReq->icack.peersCallID; + DEBUGP("%s, PCID=%X\n", pptp_msg_name[msg], ntohs(*pcid)); + info->cstate = PPTP_CALL_IN_REQ; + info->pac_call_id = ntohs(*pcid); + break; + + case PPTP_IN_CALL_CONNECT: + if (reqlen < sizeof(_pptpReq.iccon)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* server tells us about incoming call established */ + if (info->sstate != PPTP_SESSION_CONFIRMED) { + DEBUGP("%s but no session\n", pptp_msg_name[msg]); + break; + } + if (info->sstate != PPTP_CALL_IN_REP + && info->sstate != PPTP_CALL_IN_CONF) { + DEBUGP("%s but never sent IN_CALL_REPLY\n", + pptp_msg_name[msg]); + break; + } + + pcid = &pptpReq->iccon.peersCallID; + cid = &info->pac_call_id; + + if (info->pns_call_id != ntohs(*pcid)) { + DEBUGP("%s for unknown CallID %u\n", + pptp_msg_name[msg], ntohs(*cid)); + break; + } + + DEBUGP("%s, PCID=%X\n", pptp_msg_name[msg], ntohs(*pcid)); + info->cstate = PPTP_CALL_IN_CONF; + + /* we expect a GRE connection from PAC to PNS */ + seq = ntohl(tcph->seq) + sizeof(struct pptp_pkt_hdr) + + sizeof(struct PptpControlHeader) + + ((void *)pcid - (void *)pptpReq); + + if (exp_gre(ct, seq, *cid, *pcid) != 0) + printk("ip_conntrack_pptp: error during exp_gre\n"); + + break; + + case PPTP_CALL_DISCONNECT_NOTIFY: + if (reqlen < sizeof(_pptpReq.disc)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* server confirms disconnect */ + cid = &pptpReq->disc.callID; + DEBUGP("%s, CID=%X\n", pptp_msg_name[msg], ntohs(*cid)); + info->cstate = PPTP_CALL_NONE; + + /* untrack this call id, unexpect GRE packets */ + pptp_destroy_siblings(ct); + break; + + case PPTP_WAN_ERROR_NOTIFY: + break; + + case PPTP_ECHO_REQUEST: + case PPTP_ECHO_REPLY: + /* I don't have to explain these ;) */ + break; + default: + DEBUGP("invalid %s (TY=%d)\n", (msg <= PPTP_MSG_MAX) + ? pptp_msg_name[msg]:pptp_msg_name[0], msg); + break; + } + + + if (ip_nat_pptp_hook_inbound) + return ip_nat_pptp_hook_inbound(pskb, ct, ctinfo, ctlh, + pptpReq); + + return NF_ACCEPT; + +} + +static inline int +pptp_outbound_pkt(struct sk_buff **pskb, + struct tcphdr *tcph, + unsigned int nexthdr_off, + unsigned int datalen, + struct ip_conntrack *ct, + enum ip_conntrack_info ctinfo) +{ + struct PptpControlHeader _ctlh, *ctlh; + unsigned int reqlen; + union pptp_ctrl_union _pptpReq, *pptpReq; + struct ip_ct_pptp_master *info = &ct->help.ct_pptp_info; + u_int16_t msg, *cid, *pcid; + + ctlh = skb_header_pointer(*pskb, nexthdr_off, sizeof(_ctlh), &_ctlh); + if (!ctlh) + return NF_ACCEPT; + nexthdr_off += sizeof(_ctlh); + datalen -= sizeof(_ctlh); + + reqlen = datalen; + if (reqlen > sizeof(*pptpReq)) + reqlen = sizeof(*pptpReq); + pptpReq = skb_header_pointer(*pskb, nexthdr_off, reqlen, &_pptpReq); + if (!pptpReq) + return NF_ACCEPT; + + msg = ntohs(ctlh->messageType); + DEBUGP("outbound control message %s\n", pptp_msg_name[msg]); + + switch (msg) { + case PPTP_START_SESSION_REQUEST: + /* client requests for new control session */ + if (info->sstate != PPTP_SESSION_NONE) { + DEBUGP("%s but we already have one", + pptp_msg_name[msg]); + } + info->sstate = PPTP_SESSION_REQUESTED; + break; + case PPTP_STOP_SESSION_REQUEST: + /* client requests end of control session */ + info->sstate = PPTP_SESSION_STOPREQ; + break; + + case PPTP_OUT_CALL_REQUEST: + if (reqlen < sizeof(_pptpReq.ocreq)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + /* FIXME: break; */ + } + + /* client initiating connection to server */ + if (info->sstate != PPTP_SESSION_CONFIRMED) { + DEBUGP("%s but no session\n", + pptp_msg_name[msg]); + break; + } + info->cstate = PPTP_CALL_OUT_REQ; + /* track PNS call id */ + cid = &pptpReq->ocreq.callID; + DEBUGP("%s, CID=%X\n", pptp_msg_name[msg], ntohs(*cid)); + info->pns_call_id = ntohs(*cid); + break; + case PPTP_IN_CALL_REPLY: + if (reqlen < sizeof(_pptpReq.icack)) { + DEBUGP("%s: short packet\n", pptp_msg_name[msg]); + break; + } + + /* client answers incoming call */ + if (info->cstate != PPTP_CALL_IN_REQ + && info->cstate != PPTP_CALL_IN_REP) { + DEBUGP("%s without incall_req\n", + pptp_msg_name[msg]); + break; + } + if (pptpReq->icack.resultCode != PPTP_INCALL_ACCEPT) { + info->cstate = PPTP_CALL_NONE; + break; + } + pcid = &pptpReq->icack.peersCallID; + if (info->pac_call_id != ntohs(*pcid)) { + DEBUGP("%s for unknown call %u\n", + pptp_msg_name[msg], ntohs(*pcid)); + break; + } + DEBUGP("%s, CID=%X\n", pptp_msg_name[msg], ntohs(*pcid)); + /* part two of the three-way handshake */ + info->cstate = PPTP_CALL_IN_REP; + info->pns_call_id = ntohs(pptpReq->icack.callID); + break; + + case PPTP_CALL_CLEAR_REQUEST: + /* client requests hangup of call */ + if (info->sstate != PPTP_SESSION_CONFIRMED) { + DEBUGP("CLEAR_CALL but no session\n"); + break; + } + /* FUTURE: iterate over all calls and check if + * call ID is valid. We don't do this without newnat, + * because we only know about last call */ + info->cstate = PPTP_CALL_CLEAR_REQ; + break; + case PPTP_SET_LINK_INFO: + break; + case PPTP_ECHO_REQUEST: + case PPTP_ECHO_REPLY: + /* I don't have to explain these ;) */ + break; + default: + DEBUGP("invalid %s (TY=%d)\n", (msg <= PPTP_MSG_MAX)? + pptp_msg_name[msg]:pptp_msg_name[0], msg); + /* unknown: no need to create GRE masq table entry */ + break; + } + + if (ip_nat_pptp_hook_outbound) + return ip_nat_pptp_hook_outbound(pskb, ct, ctinfo, ctlh, + pptpReq); + + return NF_ACCEPT; +} + + +/* track caller id inside control connection, call expect_related */ +static int +conntrack_pptp_help(struct sk_buff **pskb, + struct ip_conntrack *ct, enum ip_conntrack_info ctinfo) + +{ + struct pptp_pkt_hdr _pptph, *pptph; + struct tcphdr _tcph, *tcph; + u_int32_t tcplen = (*pskb)->len - (*pskb)->nh.iph->ihl * 4; + u_int32_t datalen; + int dir = CTINFO2DIR(ctinfo); + struct ip_ct_pptp_master *info = &ct->help.ct_pptp_info; + unsigned int nexthdr_off; + + int oldsstate, oldcstate; + int ret; + + /* don't do any tracking before tcp handshake complete */ + if (ctinfo != IP_CT_ESTABLISHED + && ctinfo != IP_CT_ESTABLISHED+IP_CT_IS_REPLY) { + DEBUGP("ctinfo = %u, skipping\n", ctinfo); + return NF_ACCEPT; + } + + nexthdr_off = (*pskb)->nh.iph->ihl*4; + tcph = skb_header_pointer(*pskb, nexthdr_off, sizeof(_tcph), &_tcph); + BUG_ON(!tcph); + nexthdr_off += tcph->doff * 4; + datalen = tcplen - tcph->doff * 4; + + if (tcph->fin || tcph->rst) { + DEBUGP("RST/FIN received, timeouting GRE\n"); + /* can't do this after real newnat */ + info->cstate = PPTP_CALL_NONE; + + /* untrack this call id, unexpect GRE packets */ + pptp_destroy_siblings(ct); + } + + pptph = skb_header_pointer(*pskb, nexthdr_off, sizeof(_pptph), &_pptph); + if (!pptph) { + DEBUGP("no full PPTP header, can't track\n"); + return NF_ACCEPT; + } + nexthdr_off += sizeof(_pptph); + datalen -= sizeof(_pptph); + + /* if it's not a control message we can't do anything with it */ + if (ntohs(pptph->packetType) != PPTP_PACKET_CONTROL || + ntohl(pptph->magicCookie) != PPTP_MAGIC_COOKIE) { + DEBUGP("not a control packet\n"); + return NF_ACCEPT; + } + + oldsstate = info->sstate; + oldcstate = info->cstate; + + spin_lock_bh(&ip_pptp_lock); + + /* FIXME: We just blindly assume that the control connection is always + * established from PNS->PAC. However, RFC makes no guarantee */ + if (dir == IP_CT_DIR_ORIGINAL) + /* client -> server (PNS -> PAC) */ + ret = pptp_outbound_pkt(pskb, tcph, nexthdr_off, datalen, ct, + ctinfo); + else + /* server -> client (PAC -> PNS) */ + ret = pptp_inbound_pkt(pskb, tcph, nexthdr_off, datalen, ct, + ctinfo); + DEBUGP("sstate: %d->%d, cstate: %d->%d\n", + oldsstate, info->sstate, oldcstate, info->cstate); + spin_unlock_bh(&ip_pptp_lock); + + return ret; +} + +/* control protocol helper */ +static struct ip_conntrack_helper pptp = { + .list = { NULL, NULL }, + .name = "pptp", + .me = THIS_MODULE, + .max_expected = 2, + .timeout = 5 * 60, + .tuple = { .src = { .ip = 0, + .u = { .tcp = { .port = + __constant_htons(PPTP_CONTROL_PORT) } } + }, + .dst = { .ip = 0, + .u = { .all = 0 }, + .protonum = IPPROTO_TCP + } + }, + .mask = { .src = { .ip = 0, + .u = { .tcp = { .port = 0xffff } } + }, + .dst = { .ip = 0, + .u = { .all = 0 }, + .protonum = 0xff + } + }, + .help = conntrack_pptp_help +}; + +extern void __exit ip_ct_proto_gre_fini(void); +extern int __init ip_ct_proto_gre_init(void); + +/* ip_conntrack_pptp initialization */ +static int __init init(void) +{ + int retcode; + + retcode = ip_ct_proto_gre_init(); + if (retcode < 0) + return retcode; + + DEBUGP(" registering helper\n"); + if ((retcode = ip_conntrack_helper_register(&pptp))) { + printk(KERN_ERR "Unable to register conntrack application " + "helper for pptp: %d\n", retcode); + ip_ct_proto_gre_fini(); + return retcode; + } + + printk("ip_conntrack_pptp version %s loaded\n", IP_CT_PPTP_VERSION); + return 0; +} + +static void __exit fini(void) +{ + ip_conntrack_helper_unregister(&pptp); + ip_ct_proto_gre_fini(); + printk("ip_conntrack_pptp version %s unloaded\n", IP_CT_PPTP_VERSION); +} + +module_init(init); +module_exit(fini); + +EXPORT_SYMBOL(ip_nat_pptp_hook_outbound); +EXPORT_SYMBOL(ip_nat_pptp_hook_inbound); +EXPORT_SYMBOL(ip_nat_pptp_hook_exp_gre); +EXPORT_SYMBOL(ip_nat_pptp_hook_expectfn); diff --git a/net/ipv4/netfilter/ip_conntrack_netlink.c b/net/ipv4/netfilter/ip_conntrack_netlink.c index 15aef356474..b08a432efcf 100644 --- a/net/ipv4/netfilter/ip_conntrack_netlink.c +++ b/net/ipv4/netfilter/ip_conntrack_netlink.c @@ -1270,7 +1270,7 @@ ctnetlink_get_expect(struct sock *ctnl, struct sk_buff *skb, if (err < 0) return err; - exp = ip_conntrack_expect_find_get(&tuple); + exp = ip_conntrack_expect_find(&tuple); if (!exp) return -ENOENT; @@ -1318,7 +1318,7 @@ ctnetlink_del_expect(struct sock *ctnl, struct sk_buff *skb, return err; /* bump usage count to 2 */ - exp = ip_conntrack_expect_find_get(&tuple); + exp = ip_conntrack_expect_find(&tuple); if (!exp) return -ENOENT; diff --git a/net/ipv4/netfilter/ip_conntrack_proto_gre.c b/net/ipv4/netfilter/ip_conntrack_proto_gre.c new file mode 100644 index 00000000000..de3cb9db6f8 --- /dev/null +++ b/net/ipv4/netfilter/ip_conntrack_proto_gre.c @@ -0,0 +1,327 @@ +/* + * ip_conntrack_proto_gre.c - Version 3.0 + * + * Connection tracking protocol helper module for GRE. + * + * GRE is a generic encapsulation protocol, which is generally not very + * suited for NAT, as it has no protocol-specific part as port numbers. + * + * It has an optional key field, which may help us distinguishing two + * connections between the same two hosts. + * + * GRE is defined in RFC 1701 and RFC 1702, as well as RFC 2784 + * + * PPTP is built on top of a modified version of GRE, and has a mandatory + * field called "CallID", which serves us for the same purpose as the key + * field in plain GRE. + * + * Documentation about PPTP can be found in RFC 2637 + * + * (C) 2000-2005 by Harald Welte <laforge@gnumonks.org> + * + * Development of this code funded by Astaro AG (http://www.astaro.com/) + * + */ + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/types.h> +#include <linux/timer.h> +#include <linux/netfilter.h> +#include <linux/ip.h> +#include <linux/in.h> +#include <linux/list.h> + +static DEFINE_RWLOCK(ip_ct_gre_lock); +#define ASSERT_READ_LOCK(x) +#define ASSERT_WRITE_LOCK(x) + +#include <linux/netfilter_ipv4/listhelp.h> +#include <linux/netfilter_ipv4/ip_conntrack_protocol.h> +#include <linux/netfilter_ipv4/ip_conntrack_helper.h> +#include <linux/netfilter_ipv4/ip_conntrack_core.h> + +#include <linux/netfilter_ipv4/ip_conntrack_proto_gre.h> +#include <linux/netfilter_ipv4/ip_conntrack_pptp.h> + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Harald Welte <laforge@gnumonks.org>"); +MODULE_DESCRIPTION("netfilter connection tracking protocol helper for GRE"); + +/* shamelessly stolen from ip_conntrack_proto_udp.c */ +#define GRE_TIMEOUT (30*HZ) +#define GRE_STREAM_TIMEOUT (180*HZ) + +#if 0 +#define DEBUGP(format, args...) printk(KERN_DEBUG "%s:%s: " format, __FILE__, __FUNCTION__, ## args) +#define DUMP_TUPLE_GRE(x) printk("%u.%u.%u.%u:0x%x -> %u.%u.%u.%u:0x%x\n", \ + NIPQUAD((x)->src.ip), ntohs((x)->src.u.gre.key), \ + NIPQUAD((x)->dst.ip), ntohs((x)->dst.u.gre.key)) +#else +#define DEBUGP(x, args...) +#define DUMP_TUPLE_GRE(x) +#endif + +/* GRE KEYMAP HANDLING FUNCTIONS */ +static LIST_HEAD(gre_keymap_list); + +static inline int gre_key_cmpfn(const struct ip_ct_gre_keymap *km, + const struct ip_conntrack_tuple *t) +{ + return ((km->tuple.src.ip == t->src.ip) && + (km->tuple.dst.ip == t->dst.ip) && + (km->tuple.dst.protonum == t->dst.protonum) && + (km->tuple.dst.u.all == t->dst.u.all)); +} + +/* look up the source key for a given tuple */ +static u_int32_t gre_keymap_lookup(struct ip_conntrack_tuple *t) +{ + struct ip_ct_gre_keymap *km; + u_int32_t key = 0; + + read_lock_bh(&ip_ct_gre_lock); + km = LIST_FIND(&gre_keymap_list, gre_key_cmpfn, + struct ip_ct_gre_keymap *, t); + if (km) + key = km->tuple.src.u.gre.key; + read_unlock_bh(&ip_ct_gre_lock); + + DEBUGP("lookup src key 0x%x up key for ", key); + DUMP_TUPLE_GRE(t); + + return key; +} + +/* add a single keymap entry, associate with specified master ct */ +int +ip_ct_gre_keymap_add(struct ip_conntrack *ct, + struct ip_conntrack_tuple *t, int reply) +{ + struct ip_ct_gre_keymap **exist_km, *km, *old; + + if (!ct->helper || strcmp(ct->helper->name, "pptp")) { + DEBUGP("refusing to add GRE keymap to non-pptp session\n"); + return -1; + } + + if (!reply) + exist_km = &ct->help.ct_pptp_info.keymap_orig; + else + exist_km = &ct->help.ct_pptp_info.keymap_reply; + + if (*exist_km) { + /* check whether it's a retransmission */ + old = LIST_FIND(&gre_keymap_list, gre_key_cmpfn, + struct ip_ct_gre_keymap *, t); + if (old == *exist_km) { + DEBUGP("retransmission\n"); + return 0; + } + + DEBUGP("trying to override keymap_%s for ct %p\n", + reply? "reply":"orig", ct); + return -EEXIST; + } + + km = kmalloc(sizeof(*km), GFP_ATOMIC); + if (!km) + return -ENOMEM; + + memcpy(&km->tuple, t, sizeof(*t)); + *exist_km = km; + + DEBUGP("adding new entry %p: ", km); + DUMP_TUPLE_GRE(&km->tuple); + + write_lock_bh(&ip_ct_gre_lock); + list_append(&gre_keymap_list, km); + write_unlock_bh(&ip_ct_gre_lock); + + return 0; +} + +/* destroy the keymap entries associated with specified master ct */ +void ip_ct_gre_keymap_destroy(struct ip_conntrack *ct) +{ + DEBUGP("entering for ct %p\n", ct); + + if (!ct->helper || strcmp(ct->helper->name, "pptp")) { + DEBUGP("refusing to destroy GRE keymap to non-pptp session\n"); + return; + } + + write_lock_bh(&ip_ct_gre_lock); + if (ct->help.ct_pptp_info.keymap_orig) { + DEBUGP("removing %p from list\n", + ct->help.ct_pptp_info.keymap_orig); + list_del(&ct->help.ct_pptp_info.keymap_orig->list); + kfree(ct->help.ct_pptp_info.keymap_orig); + ct->help.ct_pptp_info.keymap_orig = NULL; + } + if (ct->help.ct_pptp_info.keymap_reply) { + DEBUGP("removing %p from list\n", + ct->help.ct_pptp_info.keymap_reply); + list_del(&ct->help.ct_pptp_info.keymap_reply->list); + kfree(ct->help.ct_pptp_info.keymap_reply); + ct->help.ct_pptp_info.keymap_reply = NULL; + } + write_unlock_bh(&ip_ct_gre_lock); +} + + +/* PUBLIC CONNTRACK PROTO HELPER FUNCTIONS */ + +/* invert gre part of tuple */ +static int gre_invert_tuple(struct ip_conntrack_tuple *tuple, + const struct ip_conntrack_tuple *orig) +{ + tuple->dst.u.gre.key = orig->src.u.gre.key; + tuple->src.u.gre.key = orig->dst.u.gre.key; + + return 1; +} + +/* gre hdr info to tuple */ +static int gre_pkt_to_tuple(const struct sk_buff *skb, + unsigned int dataoff, + struct ip_conntrack_tuple *tuple) +{ + struct gre_hdr_pptp _pgrehdr, *pgrehdr; + u_int32_t srckey; + struct gre_hdr _grehdr, *grehdr; + + /* first only delinearize old RFC1701 GRE header */ + grehdr = skb_header_pointer(skb, dataoff, sizeof(_grehdr), &_grehdr); + if (!grehdr || grehdr->version != GRE_VERSION_PPTP) { + /* try to behave like "ip_conntrack_proto_generic" */ + tuple->src.u.all = 0; + tuple->dst.u.all = 0; + return 1; + } + + /* PPTP header is variable length, only need up to the call_id field */ + pgrehdr = skb_header_pointer(skb, dataoff, 8, &_pgrehdr); + if (!pgrehdr) + return 1; + + if (ntohs(grehdr->protocol) != GRE_PROTOCOL_PPTP) { + DEBUGP("GRE_VERSION_PPTP but unknown proto\n"); + return 0; + } + + tuple->dst.u.gre.key = pgrehdr->call_id; + srckey = gre_keymap_lookup(tuple); + tuple->src.u.gre.key = srckey; + + return 1; +} + +/* print gre part of tuple */ +static int gre_print_tuple(struct seq_file *s, + const struct ip_conntrack_tuple *tuple) +{ + return seq_printf(s, "srckey=0x%x dstkey=0x%x ", + ntohs(tuple->src.u.gre.key), + ntohs(tuple->dst.u.gre.key)); +} + +/* print private data for conntrack */ +static int gre_print_conntrack(struct seq_file *s, + const struct ip_conntrack *ct) +{ + return seq_printf(s, "timeout=%u, stream_timeout=%u ", + (ct->proto.gre.timeout / HZ), + (ct->proto.gre.stream_timeout / HZ)); +} + +/* Returns verdict for packet, and may modify conntrack */ +static int gre_packet(struct ip_conntrack *ct, + const struct sk_buff *skb, + enum ip_conntrack_info conntrackinfo) +{ + /* If we've seen traffic both ways, this is a GRE connection. + * Extend timeout. */ + if (ct->status & IPS_SEEN_REPLY) { + ip_ct_refresh_acct(ct, conntrackinfo, skb, + ct->proto.gre.stream_timeout); + /* Also, more likely to be important, and not a probe. */ + set_bit(IPS_ASSURED_BIT, &ct->status); + } else + ip_ct_refresh_acct(ct, conntrackinfo, skb, + ct->proto.gre.timeout); + + return NF_ACCEPT; +} + +/* Called when a new connection for this protocol found. */ +static int gre_new(struct ip_conntrack *ct, + const struct sk_buff *skb) +{ + DEBUGP(": "); + DUMP_TUPLE_GRE(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple); + + /* initialize to sane value. Ideally a conntrack helper + * (e.g. in case of pptp) is increasing them */ + ct->proto.gre.stream_timeout = GRE_STREAM_TIMEOUT; + ct->proto.gre.timeout = GRE_TIMEOUT; + + return 1; +} + +/* Called when a conntrack entry has already been removed from the hashes + * and is about to be deleted from memory */ +static void gre_destroy(struct ip_conntrack *ct) +{ + struct ip_conntrack *master = ct->master; + DEBUGP(" entering\n"); + + if (!master) + DEBUGP("no master !?!\n"); + else + ip_ct_gre_keymap_destroy(master); +} + +/* protocol helper struct */ +static struct ip_conntrack_protocol gre = { + .proto = IPPROTO_GRE, + .name = "gre", + .pkt_to_tuple = gre_pkt_to_tuple, + .invert_tuple = gre_invert_tuple, + .print_tuple = gre_print_tuple, + .print_conntrack = gre_print_conntrack, + .packet = gre_packet, + .new = gre_new, + .destroy = gre_destroy, + .me = THIS_MODULE, +#if defined(CONFIG_IP_NF_CONNTRACK_NETLINK) || \ + defined(CONFIG_IP_NF_CONNTRACK_NETLINK_MODULE) + .tuple_to_nfattr = ip_ct_port_tuple_to_nfattr, + .nfattr_to_tuple = ip_ct_port_nfattr_to_tuple, +#endif +}; + +/* ip_conntrack_proto_gre initialization */ +int __init ip_ct_proto_gre_init(void) +{ + return ip_conntrack_protocol_register(&gre); +} + +void __exit ip_ct_proto_gre_fini(void) +{ + struct list_head *pos, *n; + + /* delete all keymap entries */ + write_lock_bh(&ip_ct_gre_lock); + list_for_each_safe(pos, n, &gre_keymap_list) { + DEBUGP("deleting keymap %p at module unload time\n", pos); + list_del(pos); + kfree(pos); + } + write_unlock_bh(&ip_ct_gre_lock); + + ip_conntrack_protocol_unregister(&gre); +} + +EXPORT_SYMBOL(ip_ct_gre_keymap_add); +EXPORT_SYMBOL(ip_ct_gre_keymap_destroy); diff --git a/net/ipv4/netfilter/ip_conntrack_standalone.c b/net/ipv4/netfilter/ip_conntrack_standalone.c index ae3e3e655db..d3c7808010e 100644 --- a/net/ipv4/netfilter/ip_conntrack_standalone.c +++ b/net/ipv4/netfilter/ip_conntrack_standalone.c @@ -993,11 +993,11 @@ EXPORT_SYMBOL(ip_ct_refresh_acct); EXPORT_SYMBOL(ip_conntrack_expect_alloc); EXPORT_SYMBOL(ip_conntrack_expect_put); -EXPORT_SYMBOL_GPL(ip_conntrack_expect_find_get); +EXPORT_SYMBOL_GPL(__ip_conntrack_expect_find); +EXPORT_SYMBOL_GPL(ip_conntrack_expect_find); EXPORT_SYMBOL(ip_conntrack_expect_related); EXPORT_SYMBOL(ip_conntrack_unexpect_related); EXPORT_SYMBOL_GPL(ip_conntrack_expect_list); -EXPORT_SYMBOL_GPL(__ip_conntrack_expect_find); EXPORT_SYMBOL_GPL(ip_ct_unlink_expect); EXPORT_SYMBOL(ip_conntrack_tuple_taken); diff --git a/net/ipv4/netfilter/ip_nat_core.c b/net/ipv4/netfilter/ip_nat_core.c index 1adedb743f6..c3ea891d38e 100644 --- a/net/ipv4/netfilter/ip_nat_core.c +++ b/net/ipv4/netfilter/ip_nat_core.c @@ -578,6 +578,8 @@ ip_nat_port_nfattr_to_range(struct nfattr *tb[], struct ip_nat_range *range) return ret; } +EXPORT_SYMBOL_GPL(ip_nat_port_nfattr_to_range); +EXPORT_SYMBOL_GPL(ip_nat_port_range_to_nfattr); #endif int __init ip_nat_init(void) diff --git a/net/ipv4/netfilter/ip_nat_helper_pptp.c b/net/ipv4/netfilter/ip_nat_helper_pptp.c new file mode 100644 index 00000000000..3cdd0684d30 --- /dev/null +++ b/net/ipv4/netfilter/ip_nat_helper_pptp.c @@ -0,0 +1,401 @@ +/* + * ip_nat_pptp.c - Version 3.0 + * + * NAT support for PPTP (Point to Point Tunneling Protocol). + * PPTP is a a protocol for creating virtual private networks. + * It is a specification defined by Microsoft and some vendors + * working with Microsoft. PPTP is built on top of a modified + * version of the Internet Generic Routing Encapsulation Protocol. + * GRE is defined in RFC 1701 and RFC 1702. Documentation of + * PPTP can be found in RFC 2637 + * + * (C) 2000-2005 by Harald Welte <laforge@gnumonks.org> + * + * Development of this code funded by Astaro AG (http://www.astaro.com/) + * + * TODO: - NAT to a unique tuple, not to TCP source port + * (needs netfilter tuple reservation) + * + * Changes: + * 2002-02-10 - Version 1.3 + * - Use ip_nat_mangle_tcp_packet() because of cloned skb's + * in local connections (Philip Craig <philipc@snapgear.com>) + * - add checks for magicCookie and pptp version + * - make argument list of pptp_{out,in}bound_packet() shorter + * - move to C99 style initializers + * - print version number at module loadtime + * 2003-09-22 - Version 1.5 + * - use SNATed tcp sourceport as callid, since we get called before + * TCP header is mangled (Philip Craig <philipc@snapgear.com>) + * 2004-10-22 - Version 2.0 + * - kernel 2.6.x version + * 2005-06-10 - Version 3.0 + * - kernel >= 2.6.11 version, + * funded by Oxcoda NetBox Blue (http://www.netboxblue.com/) + * + */ + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/ip.h> +#include <linux/tcp.h> +#include <net/tcp.h> + +#include <linux/netfilter_ipv4/ip_nat.h> +#include <linux/netfilter_ipv4/ip_nat_rule.h> +#include <linux/netfilter_ipv4/ip_nat_helper.h> +#include <linux/netfilter_ipv4/ip_nat_pptp.h> +#include <linux/netfilter_ipv4/ip_conntrack_core.h> +#include <linux/netfilter_ipv4/ip_conntrack_helper.h> +#include <linux/netfilter_ipv4/ip_conntrack_proto_gre.h> +#include <linux/netfilter_ipv4/ip_conntrack_pptp.h> + +#define IP_NAT_PPTP_VERSION "3.0" + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Harald Welte <laforge@gnumonks.org>"); +MODULE_DESCRIPTION("Netfilter NAT helper module for PPTP"); + + +#if 0 +extern const char *pptp_msg_name[]; +#define DEBUGP(format, args...) printk(KERN_DEBUG "%s:%s: " format, __FILE__, \ + __FUNCTION__, ## args) +#else +#define DEBUGP(format, args...) +#endif + +static void pptp_nat_expected(struct ip_conntrack *ct, + struct ip_conntrack_expect *exp) +{ + struct ip_conntrack *master = ct->master; + struct ip_conntrack_expect *other_exp; + struct ip_conntrack_tuple t; + struct ip_ct_pptp_master *ct_pptp_info; + struct ip_nat_pptp *nat_pptp_info; + + ct_pptp_info = &master->help.ct_pptp_info; + nat_pptp_info = &master->nat.help.nat_pptp_info; + + /* And here goes the grand finale of corrosion... */ + + if (exp->dir == IP_CT_DIR_ORIGINAL) { + DEBUGP("we are PNS->PAC\n"); + /* therefore, build tuple for PAC->PNS */ + t.src.ip = master->tuplehash[IP_CT_DIR_REPLY].tuple.src.ip; + t.src.u.gre.key = htons(master->help.ct_pptp_info.pac_call_id); + t.dst.ip = master->tuplehash[IP_CT_DIR_REPLY].tuple.dst.ip; + t.dst.u.gre.key = htons(master->help.ct_pptp_info.pns_call_id); + t.dst.protonum = IPPROTO_GRE; + } else { + DEBUGP("we are PAC->PNS\n"); + /* build tuple for PNS->PAC */ + t.src.ip = master->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.ip; + t.src.u.gre.key = + htons(master->nat.help.nat_pptp_info.pns_call_id); + t.dst.ip = master->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.ip; + t.dst.u.gre.key = + htons(master->nat.help.nat_pptp_info.pac_call_id); + t.dst.protonum = IPPROTO_GRE; + } + + DEBUGP("trying to unexpect other dir: "); + DUMP_TUPLE(&t); + other_exp = ip_conntrack_expect_find(&t); + if (other_exp) { + ip_conntrack_unexpect_related(other_exp); + ip_conntrack_expect_put(other_exp); + DEBUGP("success\n"); + } else { + DEBUGP("not found!\n"); + } + + ip_nat_follow_master(ct, exp); +} + +/* outbound packets == from PNS to PAC */ +static int +pptp_outbound_pkt(struct sk_buff **pskb, + struct ip_conntrack *ct, + enum ip_conntrack_info ctinfo, + struct PptpControlHeader *ctlh, + union pptp_ctrl_union *pptpReq) + +{ + struct ip_ct_pptp_master *ct_pptp_info = &ct->help.ct_pptp_info; + struct ip_nat_pptp *nat_pptp_info = &ct->nat.help.nat_pptp_info; + + u_int16_t msg, *cid = NULL, new_callid; + + new_callid = htons(ct_pptp_info->pns_call_id); + + switch (msg = ntohs(ctlh->messageType)) { + case PPTP_OUT_CALL_REQUEST: + cid = &pptpReq->ocreq.callID; + /* FIXME: ideally we would want to reserve a call ID + * here. current netfilter NAT core is not able to do + * this :( For now we use TCP source port. This breaks + * multiple calls within one control session */ + + /* save original call ID in nat_info */ + nat_pptp_info->pns_call_id = ct_pptp_info->pns_call_id; + + /* don't use tcph->source since we are at a DSTmanip + * hook (e.g. PREROUTING) and pkt is not mangled yet */ + new_callid = ct->tuplehash[IP_CT_DIR_REPLY].tuple.dst.u.tcp.port; + + /* save new call ID in ct info */ + ct_pptp_info->pns_call_id = ntohs(new_callid); + break; + case PPTP_IN_CALL_REPLY: + cid = &pptpReq->icreq.callID; + break; + case PPTP_CALL_CLEAR_REQUEST: + cid = &pptpReq->clrreq.callID; + break; + default: + DEBUGP("unknown outbound packet 0x%04x:%s\n", msg, + (msg <= PPTP_MSG_MAX)? + pptp_msg_name[msg]:pptp_msg_name[0]); + /* fall through */ + + case PPTP_SET_LINK_INFO: + /* only need to NAT in case PAC is behind NAT box */ + case PPTP_START_SESSION_REQUEST: + case PPTP_START_SESSION_REPLY: + case PPTP_STOP_SESSION_REQUEST: + case PPTP_STOP_SESSION_REPLY: + case PPTP_ECHO_REQUEST: + case PPTP_ECHO_REPLY: + /* no need to alter packet */ + return NF_ACCEPT; + } + + /* only OUT_CALL_REQUEST, IN_CALL_REPLY, CALL_CLEAR_REQUEST pass + * down to here */ + + IP_NF_ASSERT(cid); + + DEBUGP("altering call id from 0x%04x to 0x%04x\n", + ntohs(*cid), ntohs(new_callid)); + + /* mangle packet */ + if (ip_nat_mangle_tcp_packet(pskb, ct, ctinfo, + (void *)cid - ((void *)ctlh - sizeof(struct pptp_pkt_hdr)), + sizeof(new_callid), + (char *)&new_callid, + sizeof(new_callid)) == 0) + return NF_DROP; + + return NF_ACCEPT; +} + +static int +pptp_exp_gre(struct ip_conntrack_expect *expect_orig, + struct ip_conntrack_expect *expect_reply) +{ + struct ip_ct_pptp_master *ct_pptp_info = + &expect_orig->master->help.ct_pptp_info; + struct ip_nat_pptp *nat_pptp_info = + &expect_orig->master->nat.help.nat_pptp_info; + + struct ip_conntrack *ct = expect_orig->master; + + struct ip_conntrack_tuple inv_t; + struct ip_conntrack_tuple *orig_t, *reply_t; + + /* save original PAC call ID in nat_info */ + nat_pptp_info->pac_call_id = ct_pptp_info->pac_call_id; + + /* alter expectation */ + orig_t = &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple; + reply_t = &ct->tuplehash[IP_CT_DIR_REPLY].tuple; + + /* alter expectation for PNS->PAC direction */ + invert_tuplepr(&inv_t, &expect_orig->tuple); + expect_orig->saved_proto.gre.key = htons(nat_pptp_info->pac_call_id); + expect_orig->tuple.src.u.gre.key = htons(nat_pptp_info->pns_call_id); + expect_orig->tuple.dst.u.gre.key = htons(ct_pptp_info->pac_call_id); + inv_t.src.ip = reply_t->src.ip; + inv_t.dst.ip = reply_t->dst.ip; + inv_t.src.u.gre.key = htons(nat_pptp_info->pac_call_id); + inv_t.dst.u.gre.key = htons(ct_pptp_info->pns_call_id); + + if (!ip_conntrack_expect_related(expect_orig)) { + DEBUGP("successfully registered expect\n"); + } else { + DEBUGP("can't expect_related(expect_orig)\n"); + return 1; + } + + /* alter expectation for PAC->PNS direction */ + invert_tuplepr(&inv_t, &expect_reply->tuple); + expect_reply->saved_proto.gre.key = htons(nat_pptp_info->pns_call_id); + expect_reply->tuple.src.u.gre.key = htons(nat_pptp_info->pac_call_id); + expect_reply->tuple.dst.u.gre.key = htons(ct_pptp_info->pns_call_id); + inv_t.src.ip = orig_t->src.ip; + inv_t.dst.ip = orig_t->dst.ip; + inv_t.src.u.gre.key = htons(nat_pptp_info->pns_call_id); + inv_t.dst.u.gre.key = htons(ct_pptp_info->pac_call_id); + + if (!ip_conntrack_expect_related(expect_reply)) { + DEBUGP("successfully registered expect\n"); + } else { + DEBUGP("can't expect_related(expect_reply)\n"); + ip_conntrack_unexpect_related(expect_orig); + return 1; + } + + if (ip_ct_gre_keymap_add(ct, &expect_reply->tuple, 0) < 0) { + DEBUGP("can't register original keymap\n"); + ip_conntrack_unexpect_related(expect_orig); + ip_conntrack_unexpect_related(expect_reply); + return 1; + } + + if (ip_ct_gre_keymap_add(ct, &inv_t, 1) < 0) { + DEBUGP("can't register reply keymap\n"); + ip_conntrack_unexpect_related(expect_orig); + ip_conntrack_unexpect_related(expect_reply); + ip_ct_gre_keymap_destroy(ct); + return 1; + } + + return 0; +} + +/* inbound packets == from PAC to PNS */ +static int +pptp_inbound_pkt(struct sk_buff **pskb, + struct ip_conntrack *ct, + enum ip_conntrack_info ctinfo, + struct PptpControlHeader *ctlh, + union pptp_ctrl_union *pptpReq) +{ + struct ip_nat_pptp *nat_pptp_info = &ct->nat.help.nat_pptp_info; + u_int16_t msg, new_cid = 0, new_pcid, *pcid = NULL, *cid = NULL; + + int ret = NF_ACCEPT, rv; + + new_pcid = htons(nat_pptp_info->pns_call_id); + + switch (msg = ntohs(ctlh->messageType)) { + case PPTP_OUT_CALL_REPLY: + pcid = &pptpReq->ocack.peersCallID; + cid = &pptpReq->ocack.callID; + break; + case PPTP_IN_CALL_CONNECT: + pcid = &pptpReq->iccon.peersCallID; + break; + case PPTP_IN_CALL_REQUEST: + /* only need to nat in case PAC is behind NAT box */ + break; + case PPTP_WAN_ERROR_NOTIFY: + pcid = &pptpReq->wanerr.peersCallID; + break; + case PPTP_CALL_DISCONNECT_NOTIFY: + pcid = &pptpReq->disc.callID; + break; + case PPTP_SET_LINK_INFO: + pcid = &pptpReq->setlink.peersCallID; + break; + + default: + DEBUGP("unknown inbound packet %s\n", (msg <= PPTP_MSG_MAX)? + pptp_msg_name[msg]:pptp_msg_name[0]); + /* fall through */ + + case PPTP_START_SESSION_REQUEST: + case PPTP_START_SESSION_REPLY: + case PPTP_STOP_SESSION_REQUEST: + case PPTP_STOP_SESSION_REPLY: + case PPTP_ECHO_REQUEST: + case PPTP_ECHO_REPLY: + /* no need to alter packet */ + return NF_ACCEPT; + } + + /* only OUT_CALL_REPLY, IN_CALL_CONNECT, IN_CALL_REQUEST, + * WAN_ERROR_NOTIFY, CALL_DISCONNECT_NOTIFY pass down here */ + + /* mangle packet */ + IP_NF_ASSERT(pcid); + DEBUGP("altering peer call id from 0x%04x to 0x%04x\n", + ntohs(*pcid), ntohs(new_pcid)); + + rv = ip_nat_mangle_tcp_packet(pskb, ct, ctinfo, + (void *)pcid - ((void *)ctlh - sizeof(struct pptp_pkt_hdr)), + sizeof(new_pcid), (char *)&new_pcid, + sizeof(new_pcid)); + if (rv != NF_ACCEPT) + return rv; + + if (new_cid) { + IP_NF_ASSERT(cid); + DEBUGP("altering call id from 0x%04x to 0x%04x\n", + ntohs(*cid), ntohs(new_cid)); + rv = ip_nat_mangle_tcp_packet(pskb, ct, ctinfo, + (void *)cid - ((void *)ctlh - sizeof(struct pptp_pkt_hdr)), + sizeof(new_cid), + (char *)&new_cid, + sizeof(new_cid)); + if (rv != NF_ACCEPT) + return rv; + } + + /* check for earlier return value of 'switch' above */ + if (ret != NF_ACCEPT) + return ret; + + /* great, at least we don't need to resize packets */ + return NF_ACCEPT; +} + + +extern int __init ip_nat_proto_gre_init(void); +extern void __exit ip_nat_proto_gre_fini(void); + +static int __init init(void) +{ + int ret; + + DEBUGP("%s: registering NAT helper\n", __FILE__); + + ret = ip_nat_proto_gre_init(); + if (ret < 0) + return ret; + + BUG_ON(ip_nat_pptp_hook_outbound); + ip_nat_pptp_hook_outbound = &pptp_outbound_pkt; + + BUG_ON(ip_nat_pptp_hook_inbound); + ip_nat_pptp_hook_inbound = &pptp_inbound_pkt; + + BUG_ON(ip_nat_pptp_hook_exp_gre); + ip_nat_pptp_hook_exp_gre = &pptp_exp_gre; + + BUG_ON(ip_nat_pptp_hook_expectfn); + ip_nat_pptp_hook_expectfn = &pptp_nat_expected; + + printk("ip_nat_pptp version %s loaded\n", IP_NAT_PPTP_VERSION); + return 0; +} + +static void __exit fini(void) +{ + DEBUGP("cleanup_module\n" ); + + ip_nat_pptp_hook_expectfn = NULL; + ip_nat_pptp_hook_exp_gre = NULL; + ip_nat_pptp_hook_inbound = NULL; + ip_nat_pptp_hook_outbound = NULL; + + ip_nat_proto_gre_fini(); + /* Make sure noone calls it, meanwhile */ + synchronize_net(); + + printk("ip_nat_pptp version %s unloaded\n", IP_NAT_PPTP_VERSION); +} + +module_init(init); +module_exit(fini); diff --git a/net/ipv4/netfilter/ip_nat_proto_gre.c b/net/ipv4/netfilter/ip_nat_proto_gre.c new file mode 100644 index 00000000000..7c128540167 --- /dev/null +++ b/net/ipv4/netfilter/ip_nat_proto_gre.c @@ -0,0 +1,214 @@ +/* + * ip_nat_proto_gre.c - Version 2.0 + * + * NAT protocol helper module for GRE. + * + * GRE is a generic encapsulation protocol, which is generally not very + * suited for NAT, as it has no protocol-specific part as port numbers. + * + * It has an optional key field, which may help us distinguishing two + * connections between the same two hosts. + * + * GRE is defined in RFC 1701 and RFC 1702, as well as RFC 2784 + * + * PPTP is built on top of a modified version of GRE, and has a mandatory + * field called "CallID", which serves us for the same purpose as the key + * field in plain GRE. + * + * Documentation about PPTP can be found in RFC 2637 + * + * (C) 2000-2005 by Harald Welte <laforge@gnumonks.org> + * + * Development of this code funded by Astaro AG (http://www.astaro.com/) + * + */ + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/ip.h> +#include <linux/netfilter_ipv4/ip_nat.h> +#include <linux/netfilter_ipv4/ip_nat_rule.h> +#include <linux/netfilter_ipv4/ip_nat_protocol.h> +#include <linux/netfilter_ipv4/ip_conntrack_proto_gre.h> + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Harald Welte <laforge@gnumonks.org>"); +MODULE_DESCRIPTION("Netfilter NAT protocol helper module for GRE"); + +#if 0 +#define DEBUGP(format, args...) printk(KERN_DEBUG "%s:%s: " format, __FILE__, \ + __FUNCTION__, ## args) +#else +#define DEBUGP(x, args...) +#endif + +/* is key in given range between min and max */ +static int +gre_in_range(const struct ip_conntrack_tuple *tuple, + enum ip_nat_manip_type maniptype, + const union ip_conntrack_manip_proto *min, + const union ip_conntrack_manip_proto *max) +{ + u_int32_t key; + + if (maniptype == IP_NAT_MANIP_SRC) + key = tuple->src.u.gre.key; + else + key = tuple->dst.u.gre.key; + + return ntohl(key) >= ntohl(min->gre.key) + && ntohl(key) <= ntohl(max->gre.key); +} + +/* generate unique tuple ... */ +static int +gre_unique_tuple(struct ip_conntrack_tuple *tuple, + const struct ip_nat_range *range, + enum ip_nat_manip_type maniptype, + const struct ip_conntrack *conntrack) +{ + static u_int16_t key; + u_int16_t *keyptr; + unsigned int min, i, range_size; + + if (maniptype == IP_NAT_MANIP_SRC) + keyptr = &tuple->src.u.gre.key; + else + keyptr = &tuple->dst.u.gre.key; + + if (!(range->flags & IP_NAT_RANGE_PROTO_SPECIFIED)) { + DEBUGP("%p: NATing GRE PPTP\n", conntrack); + min = 1; + range_size = 0xffff; + } else { + min = ntohl(range->min.gre.key); + range_size = ntohl(range->max.gre.key) - min + 1; + } + + DEBUGP("min = %u, range_size = %u\n", min, range_size); + + for (i = 0; i < range_size; i++, key++) { + *keyptr = htonl(min + key % range_size); + if (!ip_nat_used_tuple(tuple, conntrack)) + return 1; + } + + DEBUGP("%p: no NAT mapping\n", conntrack); + + return 0; +} + +/* manipulate a GRE packet according to maniptype */ +static int +gre_manip_pkt(struct sk_buff **pskb, + unsigned int iphdroff, + const struct ip_conntrack_tuple *tuple, + enum ip_nat_manip_type maniptype) +{ + struct gre_hdr *greh; + struct gre_hdr_pptp *pgreh; + struct iphdr *iph = (struct iphdr *)((*pskb)->data + iphdroff); + unsigned int hdroff = iphdroff + iph->ihl*4; + + /* pgreh includes two optional 32bit fields which are not required + * to be there. That's where the magic '8' comes from */ + if (!skb_make_writable(pskb, hdroff + sizeof(*pgreh)-8)) + return 0; + + greh = (void *)(*pskb)->data + hdroff; + pgreh = (struct gre_hdr_pptp *) greh; + + /* we only have destination manip of a packet, since 'source key' + * is not present in the packet itself */ + if (maniptype == IP_NAT_MANIP_DST) { + /* key manipulation is always dest */ + switch (greh->version) { + case 0: + if (!greh->key) { + DEBUGP("can't nat GRE w/o key\n"); + break; + } + if (greh->csum) { + /* FIXME: Never tested this code... */ + *(gre_csum(greh)) = + ip_nat_cheat_check(~*(gre_key(greh)), + tuple->dst.u.gre.key, + *(gre_csum(greh))); + } + *(gre_key(greh)) = tuple->dst.u.gre.key; + break; + case GRE_VERSION_PPTP: + DEBUGP("call_id -> 0x%04x\n", + ntohl(tuple->dst.u.gre.key)); + pgreh->call_id = htons(ntohl(tuple->dst.u.gre.key)); + break; + default: + DEBUGP("can't nat unknown GRE version\n"); + return 0; + break; + } + } + return 1; +} + +/* print out a nat tuple */ +static unsigned int +gre_print(char *buffer, + const struct ip_conntrack_tuple *match, + const struct ip_conntrack_tuple *mask) +{ + unsigned int len = 0; + + if (mask->src.u.gre.key) + len += sprintf(buffer + len, "srckey=0x%x ", + ntohl(match->src.u.gre.key)); + + if (mask->dst.u.gre.key) + len += sprintf(buffer + len, "dstkey=0x%x ", + ntohl(match->src.u.gre.key)); + + return len; +} + +/* print a range of keys */ +static unsigned int +gre_print_range(char *buffer, const struct ip_nat_range *range) +{ + if (range->min.gre.key != 0 + || range->max.gre.key != 0xFFFF) { + if (range->min.gre.key == range->max.gre.key) + return sprintf(buffer, "key 0x%x ", + ntohl(range->min.gre.key)); + else + return sprintf(buffer, "keys 0x%u-0x%u ", + ntohl(range->min.gre.key), + ntohl(range->max.gre.key)); + } else + return 0; +} + +/* nat helper struct */ +static struct ip_nat_protocol gre = { + .name = "GRE", + .protonum = IPPROTO_GRE, + .manip_pkt = gre_manip_pkt, + .in_range = gre_in_range, + .unique_tuple = gre_unique_tuple, + .print = gre_print, + .print_range = gre_print_range, +#if defined(CONFIG_IP_NF_CONNTRACK_NETLINK) || \ + defined(CONFIG_IP_NF_CONNTRACK_NETLINK_MODULE) + .range_to_nfattr = ip_nat_port_range_to_nfattr, + .nfattr_to_range = ip_nat_port_nfattr_to_range, +#endif +}; + +int __init ip_nat_proto_gre_init(void) +{ + return ip_nat_protocol_register(&gre); +} + +void __exit ip_nat_proto_gre_fini(void) +{ + ip_nat_protocol_unregister(&gre); +} diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c index 7d38913754b..9bcb398fbc1 100644 --- a/net/ipv4/netfilter/ipt_CLUSTERIP.c +++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c @@ -13,6 +13,7 @@ #include <linux/config.h> #include <linux/proc_fs.h> #include <linux/jhash.h> +#include <linux/bitops.h> #include <linux/skbuff.h> #include <linux/ip.h> #include <linux/tcp.h> @@ -30,7 +31,7 @@ #include <linux/netfilter_ipv4/ipt_CLUSTERIP.h> #include <linux/netfilter_ipv4/ip_conntrack.h> -#define CLUSTERIP_VERSION "0.7" +#define CLUSTERIP_VERSION "0.8" #define DEBUG_CLUSTERIP @@ -49,13 +50,14 @@ MODULE_DESCRIPTION("iptables target for CLUSTERIP"); struct clusterip_config { struct list_head list; /* list of all configs */ atomic_t refcount; /* reference count */ + atomic_t entries; /* number of entries/rules + * referencing us */ u_int32_t clusterip; /* the IP address */ u_int8_t clustermac[ETH_ALEN]; /* the MAC address */ struct net_device *dev; /* device */ u_int16_t num_total_nodes; /* total number of nodes */ - u_int16_t num_local_nodes; /* number of local nodes */ - u_int16_t local_nodes[CLUSTERIP_MAX_NODES]; /* node number array */ + unsigned long local_nodes; /* node number array */ #ifdef CONFIG_PROC_FS struct proc_dir_entry *pde; /* proc dir entry */ @@ -66,8 +68,7 @@ struct clusterip_config { static LIST_HEAD(clusterip_configs); -/* clusterip_lock protects the clusterip_configs list _AND_ the configurable - * data within all structurses (num_local_nodes, local_nodes[]) */ +/* clusterip_lock protects the clusterip_configs list */ static DEFINE_RWLOCK(clusterip_lock); #ifdef CONFIG_PROC_FS @@ -76,23 +77,48 @@ static struct proc_dir_entry *clusterip_procdir; #endif static inline void -clusterip_config_get(struct clusterip_config *c) { +clusterip_config_get(struct clusterip_config *c) +{ atomic_inc(&c->refcount); } static inline void -clusterip_config_put(struct clusterip_config *c) { - if (atomic_dec_and_test(&c->refcount)) { +clusterip_config_put(struct clusterip_config *c) +{ + if (atomic_dec_and_test(&c->refcount)) + kfree(c); +} + +/* increase the count of entries(rules) using/referencing this config */ +static inline void +clusterip_config_entry_get(struct clusterip_config *c) +{ + atomic_inc(&c->entries); +} + +/* decrease the count of entries using/referencing this config. If last + * entry(rule) is removed, remove the config from lists, but don't free it + * yet, since proc-files could still be holding references */ +static inline void +clusterip_config_entry_put(struct clusterip_config *c) +{ + if (atomic_dec_and_test(&c->entries)) { write_lock_bh(&clusterip_lock); list_del(&c->list); write_unlock_bh(&clusterip_lock); + dev_mc_delete(c->dev, c->clustermac, ETH_ALEN, 0); dev_put(c->dev); - kfree(c); + + /* In case anyone still accesses the file, the open/close + * functions are also incrementing the refcount on their own, + * so it's safe to remove the entry even if it's in use. */ +#ifdef CONFIG_PROC_FS + remove_proc_entry(c->pde->name, c->pde->parent); +#endif } } - static struct clusterip_config * __clusterip_config_find(u_int32_t clusterip) { @@ -111,7 +137,7 @@ __clusterip_config_find(u_int32_t clusterip) } static inline struct clusterip_config * -clusterip_config_find_get(u_int32_t clusterip) +clusterip_config_find_get(u_int32_t clusterip, int entry) { struct clusterip_config *c; @@ -122,11 +148,24 @@ clusterip_config_find_get(u_int32_t clusterip) return NULL; } atomic_inc(&c->refcount); + if (entry) + atomic_inc(&c->entries); read_unlock_bh(&clusterip_lock); return c; } +static void +clusterip_config_init_nodelist(struct clusterip_config *c, + const struct ipt_clusterip_tgt_info *i) +{ + int n; + + for (n = 0; n < i->num_local_nodes; n++) { + set_bit(i->local_nodes[n] - 1, &c->local_nodes); + } +} + static struct clusterip_config * clusterip_config_init(struct ipt_clusterip_tgt_info *i, u_int32_t ip, struct net_device *dev) @@ -143,11 +182,11 @@ clusterip_config_init(struct ipt_clusterip_tgt_info *i, u_int32_t ip, c->clusterip = ip; memcpy(&c->clustermac, &i->clustermac, ETH_ALEN); c->num_total_nodes = i->num_total_nodes; - c->num_local_nodes = i->num_local_nodes; - memcpy(&c->local_nodes, &i->local_nodes, sizeof(c->local_nodes)); + clusterip_config_init_nodelist(c, i); c->hash_mode = i->hash_mode; c->hash_initval = i->hash_initval; atomic_set(&c->refcount, 1); + atomic_set(&c->entries, 1); #ifdef CONFIG_PROC_FS /* create proc dir entry */ @@ -171,53 +210,28 @@ clusterip_config_init(struct ipt_clusterip_tgt_info *i, u_int32_t ip, static int clusterip_add_node(struct clusterip_config *c, u_int16_t nodenum) { - int i; - - write_lock_bh(&clusterip_lock); - if (c->num_local_nodes >= CLUSTERIP_MAX_NODES - || nodenum > CLUSTERIP_MAX_NODES) { - write_unlock_bh(&clusterip_lock); + if (nodenum == 0 || + nodenum > c->num_total_nodes) return 1; - } - - /* check if we alrady have this number in our array */ - for (i = 0; i < c->num_local_nodes; i++) { - if (c->local_nodes[i] == nodenum) { - write_unlock_bh(&clusterip_lock); - return 1; - } - } - c->local_nodes[c->num_local_nodes++] = nodenum; + /* check if we already have this number in our bitfield */ + if (test_and_set_bit(nodenum - 1, &c->local_nodes)) + return 1; - write_unlock_bh(&clusterip_lock); return 0; } static int clusterip_del_node(struct clusterip_config *c, u_int16_t nodenum) { - int i; - - write_lock_bh(&clusterip_lock); - - if (c->num_local_nodes <= 1 || nodenum > CLUSTERIP_MAX_NODES) { - write_unlock_bh(&clusterip_lock); + if (nodenum == 0 || + nodenum > c->num_total_nodes) return 1; - } - for (i = 0; i < c->num_local_nodes; i++) { - if (c->local_nodes[i] == nodenum) { - int size = sizeof(u_int16_t)*(c->num_local_nodes-(i+1)); - memmove(&c->local_nodes[i], &c->local_nodes[i+1], size); - c->num_local_nodes--; - write_unlock_bh(&clusterip_lock); - return 0; - } - } + if (test_and_clear_bit(nodenum - 1, &c->local_nodes)) + return 0; - write_unlock_bh(&clusterip_lock); return 1; } @@ -285,25 +299,7 @@ clusterip_hashfn(struct sk_buff *skb, struct clusterip_config *config) static inline int clusterip_responsible(struct clusterip_config *config, u_int32_t hash) { - int i; - - read_lock_bh(&clusterip_lock); - - if (config->num_local_nodes == 0) { - read_unlock_bh(&clusterip_lock); - return 0; - } - - for (i = 0; i < config->num_local_nodes; i++) { - if (config->local_nodes[i] == hash) { - read_unlock_bh(&clusterip_lock); - return 1; - } - } - - read_unlock_bh(&clusterip_lock); - - return 0; + return test_bit(hash - 1, &config->local_nodes); } /*********************************************************************** @@ -415,8 +411,26 @@ checkentry(const char *tablename, /* FIXME: further sanity checks */ - config = clusterip_config_find_get(e->ip.dst.s_addr); - if (!config) { + config = clusterip_config_find_get(e->ip.dst.s_addr, 1); + if (config) { + if (cipinfo->config != NULL) { + /* Case A: This is an entry that gets reloaded, since + * it still has a cipinfo->config pointer. Simply + * increase the entry refcount and return */ + if (cipinfo->config != config) { + printk(KERN_ERR "CLUSTERIP: Reloaded entry " + "has invalid config pointer!\n"); + return 0; + } + clusterip_config_entry_get(cipinfo->config); + } else { + /* Case B: This is a new rule referring to an existing + * clusterip config. */ + cipinfo->config = config; + clusterip_config_entry_get(cipinfo->config); + } + } else { + /* Case C: This is a completely new clusterip config */ if (!(cipinfo->flags & CLUSTERIP_FLAG_NEW)) { printk(KERN_WARNING "CLUSTERIP: no config found for %u.%u.%u.%u, need 'new'\n", NIPQUAD(e->ip.dst.s_addr)); return 0; @@ -443,10 +457,9 @@ checkentry(const char *tablename, } dev_mc_add(config->dev,config->clustermac, ETH_ALEN, 0); } + cipinfo->config = config; } - cipinfo->config = config; - return 1; } @@ -455,13 +468,10 @@ static void destroy(void *matchinfo, unsigned int matchinfosize) { struct ipt_clusterip_tgt_info *cipinfo = matchinfo; - /* we first remove the proc entry and then drop the reference - * count. In case anyone still accesses the file, the open/close - * functions are also incrementing the refcount on their own */ -#ifdef CONFIG_PROC_FS - remove_proc_entry(cipinfo->config->pde->name, - cipinfo->config->pde->parent); -#endif + /* if no more entries are referencing the config, remove it + * from the list and destroy the proc entry */ + clusterip_config_entry_put(cipinfo->config); + clusterip_config_put(cipinfo->config); } @@ -533,7 +543,7 @@ arp_mangle(unsigned int hook, /* if there is no clusterip configuration for the arp reply's * source ip, we don't want to mangle it */ - c = clusterip_config_find_get(payload->src_ip); + c = clusterip_config_find_get(payload->src_ip, 0); if (!c) return NF_ACCEPT; @@ -574,56 +584,69 @@ static struct nf_hook_ops cip_arp_ops = { #ifdef CONFIG_PROC_FS +struct clusterip_seq_position { + unsigned int pos; /* position */ + unsigned int weight; /* number of bits set == size */ + unsigned int bit; /* current bit */ + unsigned long val; /* current value */ +}; + static void *clusterip_seq_start(struct seq_file *s, loff_t *pos) { struct proc_dir_entry *pde = s->private; struct clusterip_config *c = pde->data; - unsigned int *nodeidx; - - read_lock_bh(&clusterip_lock); - if (*pos >= c->num_local_nodes) + unsigned int weight; + u_int32_t local_nodes; + struct clusterip_seq_position *idx; + + /* FIXME: possible race */ + local_nodes = c->local_nodes; + weight = hweight32(local_nodes); + if (*pos >= weight) return NULL; - nodeidx = kmalloc(sizeof(unsigned int), GFP_KERNEL); - if (!nodeidx) + idx = kmalloc(sizeof(struct clusterip_seq_position), GFP_KERNEL); + if (!idx) return ERR_PTR(-ENOMEM); - *nodeidx = *pos; - return nodeidx; + idx->pos = *pos; + idx->weight = weight; + idx->bit = ffs(local_nodes); + idx->val = local_nodes; + clear_bit(idx->bit - 1, &idx->val); + + return idx; } static void *clusterip_seq_next(struct seq_file *s, void *v, loff_t *pos) { - struct proc_dir_entry *pde = s->private; - struct clusterip_config *c = pde->data; - unsigned int *nodeidx = (unsigned int *)v; + struct clusterip_seq_position *idx = (struct clusterip_seq_position *)v; - *pos = ++(*nodeidx); - if (*pos >= c->num_local_nodes) { + *pos = ++idx->pos; + if (*pos >= idx->weight) { kfree(v); return NULL; } - return nodeidx; + idx->bit = ffs(idx->val); + clear_bit(idx->bit - 1, &idx->val); + return idx; } static void clusterip_seq_stop(struct seq_file *s, void *v) { kfree(v); - - read_unlock_bh(&clusterip_lock); } static int clusterip_seq_show(struct seq_file *s, void *v) { - struct proc_dir_entry *pde = s->private; - struct clusterip_config *c = pde->data; - unsigned int *nodeidx = (unsigned int *)v; + struct clusterip_seq_position *idx = (struct clusterip_seq_position *)v; - if (*nodeidx != 0) + if (idx->pos != 0) seq_putc(s, ','); - seq_printf(s, "%u", c->local_nodes[*nodeidx]); - if (*nodeidx == c->num_local_nodes-1) + seq_printf(s, "%u", idx->bit); + + if (idx->pos == idx->weight - 1) seq_putc(s, '\n'); return 0; diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 304bb0a1d4f..4b0d7e4d626 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -361,7 +361,7 @@ static void raw_probe_proto_opt(struct flowi *fl, struct msghdr *msg) if (type && code) { get_user(fl->fl_icmp_type, type); - __get_user(fl->fl_icmp_code, code); + get_user(fl->fl_icmp_code, code); probed = 1; } break; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 29222b96495..a7537c7bbd0 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -979,14 +979,19 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_ if (!before(TCP_SKB_CB(skb)->seq, end_seq)) break; + in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq) && + !before(end_seq, TCP_SKB_CB(skb)->end_seq); + pcount = tcp_skb_pcount(skb); - if (pcount > 1 && - (after(start_seq, TCP_SKB_CB(skb)->seq) || - before(end_seq, TCP_SKB_CB(skb)->end_seq))) { + if (pcount > 1 && !in_sack && + after(TCP_SKB_CB(skb)->end_seq, start_seq)) { unsigned int pkt_len; - if (after(start_seq, TCP_SKB_CB(skb)->seq)) + in_sack = !after(start_seq, + TCP_SKB_CB(skb)->seq); + + if (!in_sack) pkt_len = (start_seq - TCP_SKB_CB(skb)->seq); else @@ -999,9 +1004,6 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_ fack_count += pcount; - in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq) && - !before(end_seq, TCP_SKB_CB(skb)->end_seq); - sacked = TCP_SKB_CB(skb)->sacked; /* Account D-SACK for retransmitted packet. */ diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index a88db28b0af..b1a63b2c6b4 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -384,7 +384,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req, newtp->frto_counter = 0; newtp->frto_highmark = 0; - newicsk->icsk_ca_ops = &tcp_reno; + newicsk->icsk_ca_ops = &tcp_init_congestion_ops; tcp_set_ca_state(newsk, TCP_CA_Open); tcp_init_xmit_timers(newsk); diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index c10e4435e3b..5dd6dd7d091 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -435,6 +435,8 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss int nsize, old_factor; u16 flags; + BUG_ON(len >= skb->len); + nsize = skb_headlen(skb) - len; if (nsize < 0) nsize = 0; @@ -459,9 +461,7 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss flags = TCP_SKB_CB(skb)->flags; TCP_SKB_CB(skb)->flags = flags & ~(TCPCB_FLAG_FIN|TCPCB_FLAG_PSH); TCP_SKB_CB(buff)->flags = flags; - TCP_SKB_CB(buff)->sacked = - (TCP_SKB_CB(skb)->sacked & - (TCPCB_LOST | TCPCB_EVER_RETRANS | TCPCB_AT_TAIL)); + TCP_SKB_CB(buff)->sacked = TCP_SKB_CB(skb)->sacked; TCP_SKB_CB(skb)->sacked &= ~TCPCB_AT_TAIL; if (!skb_shinfo(skb)->nr_frags && skb->ip_summed != CHECKSUM_HW) { @@ -499,6 +499,12 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss tcp_skb_pcount(buff); tp->packets_out -= diff; + + if (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED) + tp->sacked_out -= diff; + if (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_RETRANS) + tp->retrans_out -= diff; + if (TCP_SKB_CB(skb)->sacked & TCPCB_LOST) { tp->lost_out -= diff; tp->left_out -= diff; diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index 29fed6e58d0..519899fb11d 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -1968,7 +1968,7 @@ static void ip6_mc_clear_src(struct ifmcaddr6 *pmc) } pmc->mca_sources = NULL; pmc->mca_sfmode = MCAST_EXCLUDE; - pmc->mca_sfcount[MCAST_EXCLUDE] = 0; + pmc->mca_sfcount[MCAST_INCLUDE] = 0; pmc->mca_sfcount[MCAST_EXCLUDE] = 1; } diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c index 1cb8adb2787..2da514b16d9 100644 --- a/net/ipv6/netfilter/ip6_tables.c +++ b/net/ipv6/netfilter/ip6_tables.c @@ -1955,6 +1955,57 @@ static void __exit fini(void) #endif } +/* + * find specified header up to transport protocol header. + * If found target header, the offset to the header is set to *offset + * and return 0. otherwise, return -1. + * + * Notes: - non-1st Fragment Header isn't skipped. + * - ESP header isn't skipped. + * - The target header may be trancated. + */ +int ipv6_find_hdr(const struct sk_buff *skb, unsigned int *offset, u8 target) +{ + unsigned int start = (u8*)(skb->nh.ipv6h + 1) - skb->data; + u8 nexthdr = skb->nh.ipv6h->nexthdr; + unsigned int len = skb->len - start; + + while (nexthdr != target) { + struct ipv6_opt_hdr _hdr, *hp; + unsigned int hdrlen; + + if ((!ipv6_ext_hdr(nexthdr)) || nexthdr == NEXTHDR_NONE) + return -1; + hp = skb_header_pointer(skb, start, sizeof(_hdr), &_hdr); + if (hp == NULL) + return -1; + if (nexthdr == NEXTHDR_FRAGMENT) { + unsigned short _frag_off, *fp; + fp = skb_header_pointer(skb, + start+offsetof(struct frag_hdr, + frag_off), + sizeof(_frag_off), + &_frag_off); + if (fp == NULL) + return -1; + + if (ntohs(*fp) & ~0x7) + return -1; + hdrlen = 8; + } else if (nexthdr == NEXTHDR_AUTH) + hdrlen = (hp->hdrlen + 2) << 2; + else + hdrlen = ipv6_optlen(hp); + + nexthdr = hp->nexthdr; + len -= hdrlen; + start += hdrlen; + } + + *offset = start; + return 0; +} + EXPORT_SYMBOL(ip6t_register_table); EXPORT_SYMBOL(ip6t_unregister_table); EXPORT_SYMBOL(ip6t_do_table); @@ -1963,6 +2014,7 @@ EXPORT_SYMBOL(ip6t_unregister_match); EXPORT_SYMBOL(ip6t_register_target); EXPORT_SYMBOL(ip6t_unregister_target); EXPORT_SYMBOL(ip6t_ext_hdr); +EXPORT_SYMBOL(ipv6_find_hdr); module_init(init); module_exit(fini); diff --git a/net/ipv6/netfilter/ip6t_ah.c b/net/ipv6/netfilter/ip6t_ah.c index d5b94f142bb..dde37793d20 100644 --- a/net/ipv6/netfilter/ip6t_ah.c +++ b/net/ipv6/netfilter/ip6t_ah.c @@ -48,92 +48,21 @@ match(const struct sk_buff *skb, unsigned int protoff, int *hotdrop) { - struct ip_auth_hdr *ah = NULL, _ah; + struct ip_auth_hdr *ah, _ah; const struct ip6t_ah *ahinfo = matchinfo; - unsigned int temp; - int len; - u8 nexthdr; unsigned int ptr; unsigned int hdrlen = 0; - /*DEBUGP("IPv6 AH entered\n");*/ - /* if (opt->auth == 0) return 0; - * It does not filled on output */ - - /* type of the 1st exthdr */ - nexthdr = skb->nh.ipv6h->nexthdr; - /* pointer to the 1st exthdr */ - ptr = sizeof(struct ipv6hdr); - /* available length */ - len = skb->len - ptr; - temp = 0; - - while (ip6t_ext_hdr(nexthdr)) { - struct ipv6_opt_hdr _hdr, *hp; - - DEBUGP("ipv6_ah header iteration \n"); - - /* Is there enough space for the next ext header? */ - if (len < sizeof(struct ipv6_opt_hdr)) - return 0; - /* No more exthdr -> evaluate */ - if (nexthdr == NEXTHDR_NONE) - break; - /* ESP -> evaluate */ - if (nexthdr == NEXTHDR_ESP) - break; - - hp = skb_header_pointer(skb, ptr, sizeof(_hdr), &_hdr); - BUG_ON(hp == NULL); - - /* Calculate the header length */ - if (nexthdr == NEXTHDR_FRAGMENT) - hdrlen = 8; - else if (nexthdr == NEXTHDR_AUTH) - hdrlen = (hp->hdrlen+2)<<2; - else - hdrlen = ipv6_optlen(hp); - - /* AH -> evaluate */ - if (nexthdr == NEXTHDR_AUTH) { - temp |= MASK_AH; - break; - } - - - /* set the flag */ - switch (nexthdr) { - case NEXTHDR_HOP: - case NEXTHDR_ROUTING: - case NEXTHDR_FRAGMENT: - case NEXTHDR_AUTH: - case NEXTHDR_DEST: - break; - default: - DEBUGP("ipv6_ah match: unknown nextheader %u\n",nexthdr); - return 0; - } - - nexthdr = hp->nexthdr; - len -= hdrlen; - ptr += hdrlen; - if (ptr > skb->len) { - DEBUGP("ipv6_ah: new pointer too large! \n"); - break; - } - } - - /* AH header not found */ - if (temp != MASK_AH) + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_AUTH) < 0) return 0; - if (len < sizeof(struct ip_auth_hdr)){ + ah = skb_header_pointer(skb, ptr, sizeof(_ah), &_ah); + if (ah == NULL) { *hotdrop = 1; return 0; } - ah = skb_header_pointer(skb, ptr, sizeof(_ah), &_ah); - BUG_ON(ah == NULL); + hdrlen = (ah->hdrlen + 2) << 2; DEBUGP("IPv6 AH LEN %u %u ", hdrlen, ah->hdrlen); DEBUGP("RES %04X ", ah->reserved); diff --git a/net/ipv6/netfilter/ip6t_dst.c b/net/ipv6/netfilter/ip6t_dst.c index 540925e4a7a..c450a635e54 100644 --- a/net/ipv6/netfilter/ip6t_dst.c +++ b/net/ipv6/netfilter/ip6t_dst.c @@ -63,8 +63,6 @@ match(const struct sk_buff *skb, struct ipv6_opt_hdr _optsh, *oh; const struct ip6t_opts *optinfo = matchinfo; unsigned int temp; - unsigned int len; - u8 nexthdr; unsigned int ptr; unsigned int hdrlen = 0; unsigned int ret = 0; @@ -72,97 +70,25 @@ match(const struct sk_buff *skb, u8 _optlen, *lp = NULL; unsigned int optlen; - /* type of the 1st exthdr */ - nexthdr = skb->nh.ipv6h->nexthdr; - /* pointer to the 1st exthdr */ - ptr = sizeof(struct ipv6hdr); - /* available length */ - len = skb->len - ptr; - temp = 0; - - while (ip6t_ext_hdr(nexthdr)) { - struct ipv6_opt_hdr _hdr, *hp; - - DEBUGP("ipv6_opts header iteration \n"); - - /* Is there enough space for the next ext header? */ - if (len < (int)sizeof(struct ipv6_opt_hdr)) - return 0; - /* No more exthdr -> evaluate */ - if (nexthdr == NEXTHDR_NONE) { - break; - } - /* ESP -> evaluate */ - if (nexthdr == NEXTHDR_ESP) { - break; - } - - hp = skb_header_pointer(skb, ptr, sizeof(_hdr), &_hdr); - BUG_ON(hp == NULL); - - /* Calculate the header length */ - if (nexthdr == NEXTHDR_FRAGMENT) { - hdrlen = 8; - } else if (nexthdr == NEXTHDR_AUTH) - hdrlen = (hp->hdrlen+2)<<2; - else - hdrlen = ipv6_optlen(hp); - - /* OPTS -> evaluate */ #if HOPBYHOP - if (nexthdr == NEXTHDR_HOP) { - temp |= MASK_HOPOPTS; + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_HOP) < 0) #else - if (nexthdr == NEXTHDR_DEST) { - temp |= MASK_DSTOPTS; + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_DEST) < 0) #endif - break; - } - + return 0; - /* set the flag */ - switch (nexthdr){ - case NEXTHDR_HOP: - case NEXTHDR_ROUTING: - case NEXTHDR_FRAGMENT: - case NEXTHDR_AUTH: - case NEXTHDR_DEST: - break; - default: - DEBUGP("ipv6_opts match: unknown nextheader %u\n",nexthdr); - return 0; - break; - } - - nexthdr = hp->nexthdr; - len -= hdrlen; - ptr += hdrlen; - if ( ptr > skb->len ) { - DEBUGP("ipv6_opts: new pointer is too large! \n"); - break; - } - } - - /* OPTIONS header not found */ -#if HOPBYHOP - if ( temp != MASK_HOPOPTS ) return 0; -#else - if ( temp != MASK_DSTOPTS ) return 0; -#endif - - if (len < (int)sizeof(struct ipv6_opt_hdr)){ + oh = skb_header_pointer(skb, ptr, sizeof(_optsh), &_optsh); + if (oh == NULL){ *hotdrop = 1; return 0; } - if (len < hdrlen){ + hdrlen = ipv6_optlen(oh); + if (skb->len - ptr < hdrlen){ /* Packet smaller than it's length field */ return 0; } - oh = skb_header_pointer(skb, ptr, sizeof(_optsh), &_optsh); - BUG_ON(oh == NULL); - DEBUGP("IPv6 OPTS LEN %u %u ", hdrlen, oh->hdrlen); DEBUGP("len %02X %04X %02X ", diff --git a/net/ipv6/netfilter/ip6t_esp.c b/net/ipv6/netfilter/ip6t_esp.c index e39dd236fd8..24bc0cde43a 100644 --- a/net/ipv6/netfilter/ip6t_esp.c +++ b/net/ipv6/netfilter/ip6t_esp.c @@ -48,87 +48,22 @@ match(const struct sk_buff *skb, unsigned int protoff, int *hotdrop) { - struct ip_esp_hdr _esp, *eh = NULL; + struct ip_esp_hdr _esp, *eh; const struct ip6t_esp *espinfo = matchinfo; - unsigned int temp; - int len; - u8 nexthdr; unsigned int ptr; /* Make sure this isn't an evil packet */ /*DEBUGP("ipv6_esp entered \n");*/ - /* type of the 1st exthdr */ - nexthdr = skb->nh.ipv6h->nexthdr; - /* pointer to the 1st exthdr */ - ptr = sizeof(struct ipv6hdr); - /* available length */ - len = skb->len - ptr; - temp = 0; - - while (ip6t_ext_hdr(nexthdr)) { - struct ipv6_opt_hdr _hdr, *hp; - int hdrlen; - - DEBUGP("ipv6_esp header iteration \n"); - - /* Is there enough space for the next ext header? */ - if (len < sizeof(struct ipv6_opt_hdr)) - return 0; - /* No more exthdr -> evaluate */ - if (nexthdr == NEXTHDR_NONE) - break; - /* ESP -> evaluate */ - if (nexthdr == NEXTHDR_ESP) { - temp |= MASK_ESP; - break; - } - - hp = skb_header_pointer(skb, ptr, sizeof(_hdr), &_hdr); - BUG_ON(hp == NULL); - - /* Calculate the header length */ - if (nexthdr == NEXTHDR_FRAGMENT) - hdrlen = 8; - else if (nexthdr == NEXTHDR_AUTH) - hdrlen = (hp->hdrlen+2)<<2; - else - hdrlen = ipv6_optlen(hp); - - /* set the flag */ - switch (nexthdr) { - case NEXTHDR_HOP: - case NEXTHDR_ROUTING: - case NEXTHDR_FRAGMENT: - case NEXTHDR_AUTH: - case NEXTHDR_DEST: - break; - default: - DEBUGP("ipv6_esp match: unknown nextheader %u\n",nexthdr); - return 0; - } - - nexthdr = hp->nexthdr; - len -= hdrlen; - ptr += hdrlen; - if (ptr > skb->len) { - DEBUGP("ipv6_esp: new pointer too large! \n"); - break; - } - } - - /* ESP header not found */ - if (temp != MASK_ESP) + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_ESP) < 0) return 0; - if (len < sizeof(struct ip_esp_hdr)) { + eh = skb_header_pointer(skb, ptr, sizeof(_esp), &_esp); + if (eh == NULL) { *hotdrop = 1; return 0; } - eh = skb_header_pointer(skb, ptr, sizeof(_esp), &_esp); - BUG_ON(eh == NULL); - DEBUGP("IPv6 ESP SPI %u %08X\n", ntohl(eh->spi), ntohl(eh->spi)); return (eh != NULL) diff --git a/net/ipv6/netfilter/ip6t_frag.c b/net/ipv6/netfilter/ip6t_frag.c index 4bfa30a9bc8..085d5f8eea2 100644 --- a/net/ipv6/netfilter/ip6t_frag.c +++ b/net/ipv6/netfilter/ip6t_frag.c @@ -48,90 +48,18 @@ match(const struct sk_buff *skb, unsigned int protoff, int *hotdrop) { - struct frag_hdr _frag, *fh = NULL; + struct frag_hdr _frag, *fh; const struct ip6t_frag *fraginfo = matchinfo; - unsigned int temp; - int len; - u8 nexthdr; unsigned int ptr; - unsigned int hdrlen = 0; - - /* type of the 1st exthdr */ - nexthdr = skb->nh.ipv6h->nexthdr; - /* pointer to the 1st exthdr */ - ptr = sizeof(struct ipv6hdr); - /* available length */ - len = skb->len - ptr; - temp = 0; - - while (ip6t_ext_hdr(nexthdr)) { - struct ipv6_opt_hdr _hdr, *hp; - - DEBUGP("ipv6_frag header iteration \n"); - - /* Is there enough space for the next ext header? */ - if (len < (int)sizeof(struct ipv6_opt_hdr)) - return 0; - /* No more exthdr -> evaluate */ - if (nexthdr == NEXTHDR_NONE) { - break; - } - /* ESP -> evaluate */ - if (nexthdr == NEXTHDR_ESP) { - break; - } - - hp = skb_header_pointer(skb, ptr, sizeof(_hdr), &_hdr); - BUG_ON(hp == NULL); - - /* Calculate the header length */ - if (nexthdr == NEXTHDR_FRAGMENT) { - hdrlen = 8; - } else if (nexthdr == NEXTHDR_AUTH) - hdrlen = (hp->hdrlen+2)<<2; - else - hdrlen = ipv6_optlen(hp); - - /* FRAG -> evaluate */ - if (nexthdr == NEXTHDR_FRAGMENT) { - temp |= MASK_FRAGMENT; - break; - } - - - /* set the flag */ - switch (nexthdr){ - case NEXTHDR_HOP: - case NEXTHDR_ROUTING: - case NEXTHDR_FRAGMENT: - case NEXTHDR_AUTH: - case NEXTHDR_DEST: - break; - default: - DEBUGP("ipv6_frag match: unknown nextheader %u\n",nexthdr); - return 0; - break; - } - - nexthdr = hp->nexthdr; - len -= hdrlen; - ptr += hdrlen; - if ( ptr > skb->len ) { - DEBUGP("ipv6_frag: new pointer too large! \n"); - break; - } - } - - /* FRAG header not found */ - if ( temp != MASK_FRAGMENT ) return 0; - - if (len < sizeof(struct frag_hdr)){ - *hotdrop = 1; - return 0; - } - fh = skb_header_pointer(skb, ptr, sizeof(_frag), &_frag); - BUG_ON(fh == NULL); + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_FRAGMENT) < 0) + return 0; + + fh = skb_header_pointer(skb, ptr, sizeof(_frag), &_frag); + if (fh == NULL){ + *hotdrop = 1; + return 0; + } DEBUGP("INFO %04X ", fh->frag_off); DEBUGP("OFFSET %04X ", ntohs(fh->frag_off) & ~0x7); diff --git a/net/ipv6/netfilter/ip6t_hbh.c b/net/ipv6/netfilter/ip6t_hbh.c index 27f3650d127..1d09485111d 100644 --- a/net/ipv6/netfilter/ip6t_hbh.c +++ b/net/ipv6/netfilter/ip6t_hbh.c @@ -63,8 +63,6 @@ match(const struct sk_buff *skb, struct ipv6_opt_hdr _optsh, *oh; const struct ip6t_opts *optinfo = matchinfo; unsigned int temp; - unsigned int len; - u8 nexthdr; unsigned int ptr; unsigned int hdrlen = 0; unsigned int ret = 0; @@ -72,97 +70,25 @@ match(const struct sk_buff *skb, u8 _optlen, *lp = NULL; unsigned int optlen; - /* type of the 1st exthdr */ - nexthdr = skb->nh.ipv6h->nexthdr; - /* pointer to the 1st exthdr */ - ptr = sizeof(struct ipv6hdr); - /* available length */ - len = skb->len - ptr; - temp = 0; - - while (ip6t_ext_hdr(nexthdr)) { - struct ipv6_opt_hdr _hdr, *hp; - - DEBUGP("ipv6_opts header iteration \n"); - - /* Is there enough space for the next ext header? */ - if (len < (int)sizeof(struct ipv6_opt_hdr)) - return 0; - /* No more exthdr -> evaluate */ - if (nexthdr == NEXTHDR_NONE) { - break; - } - /* ESP -> evaluate */ - if (nexthdr == NEXTHDR_ESP) { - break; - } - - hp = skb_header_pointer(skb, ptr, sizeof(_hdr), &_hdr); - BUG_ON(hp == NULL); - - /* Calculate the header length */ - if (nexthdr == NEXTHDR_FRAGMENT) { - hdrlen = 8; - } else if (nexthdr == NEXTHDR_AUTH) - hdrlen = (hp->hdrlen+2)<<2; - else - hdrlen = ipv6_optlen(hp); - - /* OPTS -> evaluate */ #if HOPBYHOP - if (nexthdr == NEXTHDR_HOP) { - temp |= MASK_HOPOPTS; + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_HOP) < 0) #else - if (nexthdr == NEXTHDR_DEST) { - temp |= MASK_DSTOPTS; + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_DEST) < 0) #endif - break; - } - + return 0; - /* set the flag */ - switch (nexthdr){ - case NEXTHDR_HOP: - case NEXTHDR_ROUTING: - case NEXTHDR_FRAGMENT: - case NEXTHDR_AUTH: - case NEXTHDR_DEST: - break; - default: - DEBUGP("ipv6_opts match: unknown nextheader %u\n",nexthdr); - return 0; - break; - } - - nexthdr = hp->nexthdr; - len -= hdrlen; - ptr += hdrlen; - if ( ptr > skb->len ) { - DEBUGP("ipv6_opts: new pointer is too large! \n"); - break; - } - } - - /* OPTIONS header not found */ -#if HOPBYHOP - if ( temp != MASK_HOPOPTS ) return 0; -#else - if ( temp != MASK_DSTOPTS ) return 0; -#endif - - if (len < (int)sizeof(struct ipv6_opt_hdr)){ + oh = skb_header_pointer(skb, ptr, sizeof(_optsh), &_optsh); + if (oh == NULL){ *hotdrop = 1; return 0; } - if (len < hdrlen){ + hdrlen = ipv6_optlen(oh); + if (skb->len - ptr < hdrlen){ /* Packet smaller than it's length field */ return 0; } - oh = skb_header_pointer(skb, ptr, sizeof(_optsh), &_optsh); - BUG_ON(oh == NULL); - DEBUGP("IPv6 OPTS LEN %u %u ", hdrlen, oh->hdrlen); DEBUGP("len %02X %04X %02X ", diff --git a/net/ipv6/netfilter/ip6t_rt.c b/net/ipv6/netfilter/ip6t_rt.c index 2bb670037df..beb2fd5cebb 100644 --- a/net/ipv6/netfilter/ip6t_rt.c +++ b/net/ipv6/netfilter/ip6t_rt.c @@ -50,98 +50,29 @@ match(const struct sk_buff *skb, unsigned int protoff, int *hotdrop) { - struct ipv6_rt_hdr _route, *rh = NULL; + struct ipv6_rt_hdr _route, *rh; const struct ip6t_rt *rtinfo = matchinfo; unsigned int temp; - unsigned int len; - u8 nexthdr; unsigned int ptr; unsigned int hdrlen = 0; unsigned int ret = 0; struct in6_addr *ap, _addr; - /* type of the 1st exthdr */ - nexthdr = skb->nh.ipv6h->nexthdr; - /* pointer to the 1st exthdr */ - ptr = sizeof(struct ipv6hdr); - /* available length */ - len = skb->len - ptr; - temp = 0; + if (ipv6_find_hdr(skb, &ptr, NEXTHDR_ROUTING) < 0) + return 0; - while (ip6t_ext_hdr(nexthdr)) { - struct ipv6_opt_hdr _hdr, *hp; - - DEBUGP("ipv6_rt header iteration \n"); - - /* Is there enough space for the next ext header? */ - if (len < (int)sizeof(struct ipv6_opt_hdr)) - return 0; - /* No more exthdr -> evaluate */ - if (nexthdr == NEXTHDR_NONE) { - break; - } - /* ESP -> evaluate */ - if (nexthdr == NEXTHDR_ESP) { - break; - } - - hp = skb_header_pointer(skb, ptr, sizeof(_hdr), &_hdr); - BUG_ON(hp == NULL); - - /* Calculate the header length */ - if (nexthdr == NEXTHDR_FRAGMENT) { - hdrlen = 8; - } else if (nexthdr == NEXTHDR_AUTH) - hdrlen = (hp->hdrlen+2)<<2; - else - hdrlen = ipv6_optlen(hp); - - /* ROUTING -> evaluate */ - if (nexthdr == NEXTHDR_ROUTING) { - temp |= MASK_ROUTING; - break; - } - - - /* set the flag */ - switch (nexthdr){ - case NEXTHDR_HOP: - case NEXTHDR_ROUTING: - case NEXTHDR_FRAGMENT: - case NEXTHDR_AUTH: - case NEXTHDR_DEST: - break; - default: - DEBUGP("ipv6_rt match: unknown nextheader %u\n",nexthdr); - return 0; - break; - } - - nexthdr = hp->nexthdr; - len -= hdrlen; - ptr += hdrlen; - if ( ptr > skb->len ) { - DEBUGP("ipv6_rt: new pointer is too large! \n"); - break; - } - } - - /* ROUTING header not found */ - if ( temp != MASK_ROUTING ) return 0; - - if (len < (int)sizeof(struct ipv6_rt_hdr)){ + rh = skb_header_pointer(skb, ptr, sizeof(_route), &_route); + if (rh == NULL){ *hotdrop = 1; return 0; } - if (len < hdrlen){ + hdrlen = ipv6_optlen(rh); + if (skb->len - ptr < hdrlen){ /* Pcket smaller than its length field */ return 0; } - rh = skb_header_pointer(skb, ptr, sizeof(_route), &_route); - BUG_ON(rh == NULL); - DEBUGP("IPv6 RT LEN %u %u ", hdrlen, rh->hdrlen); DEBUGP("TYPE %04X ", rh->type); DEBUGP("SGS_LEFT %u %02X\n", rh->segments_left, rh->segments_left); diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 5aa3691c578..a1265a320b1 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -627,7 +627,7 @@ static void rawv6_probe_proto_opt(struct flowi *fl, struct msghdr *msg) if (type && code) { get_user(fl->fl_icmp_type, type); - __get_user(fl->fl_icmp_code, code); + get_user(fl->fl_icmp_code, code); probed = 1; } break; diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 69b146843a2..6001948600f 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -405,9 +405,8 @@ static struct sock *udp_v6_mcast_next(struct sock *sk, continue; if (!ipv6_addr_any(&np->rcv_saddr)) { - if (ipv6_addr_equal(&np->rcv_saddr, loc_addr)) - return s; - continue; + if (!ipv6_addr_equal(&np->rcv_saddr, loc_addr)) + continue; } if(!inet6_mc_check(s, loc_addr, rmt_addr)) continue; @@ -640,6 +639,7 @@ static int udpv6_sendmsg(struct kiocb *iocb, struct sock *sk, int tclass = -1; int corkreq = up->corkflag || msg->msg_flags&MSG_MORE; int err; + int connected = 0; /* destination address check */ if (sin6) { @@ -749,6 +749,7 @@ do_udp_sendmsg: fl->fl_ip_dport = inet->dport; daddr = &np->daddr; fl->fl6_flowlabel = np->flow_label; + connected = 1; } if (!fl->oif) @@ -771,6 +772,7 @@ do_udp_sendmsg: } if (!(opt->opt_nflen|opt->opt_flen)) opt = NULL; + connected = 0; } if (opt == NULL) opt = np->opt; @@ -788,10 +790,13 @@ do_udp_sendmsg: ipv6_addr_copy(&final, &fl->fl6_dst); ipv6_addr_copy(&fl->fl6_dst, rt0->addr); final_p = &final; + connected = 0; } - if (!fl->oif && ipv6_addr_is_multicast(&fl->fl6_dst)) + if (!fl->oif && ipv6_addr_is_multicast(&fl->fl6_dst)) { fl->oif = np->mcast_oif; + connected = 0; + } err = ip6_dst_lookup(sk, &dst, fl); if (err) @@ -847,7 +852,7 @@ do_append_data: else if (!corkreq) err = udp_v6_push_pending_frames(sk, up); - if (dst) + if (dst && connected) ip6_dst_store(sk, dst, ipv6_addr_equal(&fl->fl6_dst, &np->daddr) ? &np->daddr : NULL); diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 8690f171c1e..ee865d88183 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -36,6 +36,11 @@ * Michal Ostrowski : Module initialization cleanup. * Ulises Alonso : Frame number limit removal and * packet_set_ring memory leak. + * Eric Biederman : Allow for > 8 byte hardware addresses. + * The convention is that longer addresses + * will simply extend the hardware address + * byte arrays at the end of sockaddr_ll + * and packet_mreq. * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License @@ -161,7 +166,17 @@ struct packet_mclist int count; unsigned short type; unsigned short alen; - unsigned char addr[8]; + unsigned char addr[MAX_ADDR_LEN]; +}; +/* identical to struct packet_mreq except it has + * a longer address field. + */ +struct packet_mreq_max +{ + int mr_ifindex; + unsigned short mr_type; + unsigned short mr_alen; + unsigned char mr_address[MAX_ADDR_LEN]; }; #endif #ifdef CONFIG_PACKET_MMAP @@ -716,6 +731,8 @@ static int packet_sendmsg(struct kiocb *iocb, struct socket *sock, err = -EINVAL; if (msg->msg_namelen < sizeof(struct sockaddr_ll)) goto out; + if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr))) + goto out; ifindex = saddr->sll_ifindex; proto = saddr->sll_protocol; addr = saddr->sll_addr; @@ -744,6 +761,12 @@ static int packet_sendmsg(struct kiocb *iocb, struct socket *sock, if (dev->hard_header) { int res; err = -EINVAL; + if (saddr) { + if (saddr->sll_halen != dev->addr_len) + goto out_free; + if (saddr->sll_hatype != dev->type) + goto out_free; + } res = dev->hard_header(skb, dev, ntohs(proto), addr, NULL, len); if (sock->type != SOCK_DGRAM) { skb->tail = skb->data; @@ -1045,6 +1068,7 @@ static int packet_recvmsg(struct kiocb *iocb, struct socket *sock, struct sock *sk = sock->sk; struct sk_buff *skb; int copied, err; + struct sockaddr_ll *sll; err = -EINVAL; if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT)) @@ -1057,16 +1081,6 @@ static int packet_recvmsg(struct kiocb *iocb, struct socket *sock, #endif /* - * If the address length field is there to be filled in, we fill - * it in now. - */ - - if (sock->type == SOCK_PACKET) - msg->msg_namelen = sizeof(struct sockaddr_pkt); - else - msg->msg_namelen = sizeof(struct sockaddr_ll); - - /* * Call the generic datagram receiver. This handles all sorts * of horrible races and re-entrancy so we can forget about it * in the protocol layers. @@ -1087,6 +1101,17 @@ static int packet_recvmsg(struct kiocb *iocb, struct socket *sock, goto out; /* + * If the address length field is there to be filled in, we fill + * it in now. + */ + + sll = (struct sockaddr_ll*)skb->cb; + if (sock->type == SOCK_PACKET) + msg->msg_namelen = sizeof(struct sockaddr_pkt); + else + msg->msg_namelen = sll->sll_halen + offsetof(struct sockaddr_ll, sll_addr); + + /* * You lose any data beyond the buffer you gave. If it worries a * user program they can ask the device for its MTU anyway. */ @@ -1166,7 +1191,7 @@ static int packet_getname(struct socket *sock, struct sockaddr *uaddr, sll->sll_hatype = 0; /* Bad: we have no ARPHRD_UNSPEC */ sll->sll_halen = 0; } - *uaddr_len = sizeof(*sll); + *uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen; return 0; } @@ -1199,7 +1224,7 @@ static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, i } } -static int packet_mc_add(struct sock *sk, struct packet_mreq *mreq) +static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq) { struct packet_sock *po = pkt_sk(sk); struct packet_mclist *ml, *i; @@ -1249,7 +1274,7 @@ done: return err; } -static int packet_mc_drop(struct sock *sk, struct packet_mreq *mreq) +static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq) { struct packet_mclist *ml, **mlp; @@ -1315,11 +1340,17 @@ packet_setsockopt(struct socket *sock, int level, int optname, char __user *optv case PACKET_ADD_MEMBERSHIP: case PACKET_DROP_MEMBERSHIP: { - struct packet_mreq mreq; - if (optlen<sizeof(mreq)) + struct packet_mreq_max mreq; + int len = optlen; + memset(&mreq, 0, sizeof(mreq)); + if (len < sizeof(struct packet_mreq)) return -EINVAL; - if (copy_from_user(&mreq,optval,sizeof(mreq))) + if (len > sizeof(mreq)) + len = sizeof(mreq); + if (copy_from_user(&mreq,optval,len)) return -EFAULT; + if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address))) + return -EINVAL; if (optname == PACKET_ADD_MEMBERSHIP) ret = packet_mc_add(sk, &mreq); else diff --git a/net/socket.c b/net/socket.c index c699e93c33d..f9264472377 100644 --- a/net/socket.c +++ b/net/socket.c @@ -1862,7 +1862,8 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, unsigned int flag if (err < 0) goto out_freeiov; } - err = __put_user(msg_sys.msg_flags, COMPAT_FLAGS(msg)); + err = __put_user((msg_sys.msg_flags & ~MSG_CMSG_COMPAT), + COMPAT_FLAGS(msg)); if (err) goto out_freeiov; if (MSG_CMSG_COMPAT & flags) |