diff -ruN linux-2.4.0.orig/Documentation/Configure.help linux-2.4.0/Documentation/Configure.help --- linux-2.4.0.orig/Documentation/Configure.help Thu Jan 4 16:00:55 2001 +++ linux-2.4.0/Documentation/Configure.help Wed Jan 24 12:07:54 2001 @@ -2022,6 +2022,34 @@ If you want to compile it as a module, say M here and read Documentation/modules.txt. If unsure, say `N'. +TCPMSS target support +CONFIG_IP_NF_TARGET_TCPMSS + This option adds a `TCPMSS' target, which allows you to alter the + MSS value of TCP SYN packets, to control the maximum size for that + connection (usually limiting it to your outgoing interface's MTU + minus 40). + + THIS IS A HACK, used to overcome criminally braindead ISPs or + servers which block ICMP Fragmentation Needed packets. The symptoms + of this problem are that everything works fine from your Linux + firewall/router, but machines behind it can never exchange large + packets: + 1) Web browsers connect, then hang with no data received. + 2) Small mail works fine, but large EMails hang. + 3) ssh works fine, but scp hangs after initial handshaking. + + If you want to compile it as a module, say M here and read + Documentation/modules.txt. If unsure, say `N'. + +tcpmss match support +CONFIG_IP_NF_MATCH_TCPMSS + This option adds a `tcpmss' match, which allows you to examine the + MSS value of TCP SYN packets, which control the maximum packet size + for that connection. + + If you want to compile it as a module, say M here and read + Documentation/modules.txt. If unsure, say `N'. + LOG target support CONFIG_IP_NF_TARGET_LOG This option adds a `LOG' target, which allows you to create rules in diff -ruN linux-2.4.0.orig/include/linux/netfilter_ipv4/ipt_TCPMSS.h linux-2.4.0/include/linux/netfilter_ipv4/ipt_TCPMSS.h --- linux-2.4.0.orig/include/linux/netfilter_ipv4/ipt_TCPMSS.h Wed Dec 31 19:00:00 1969 +++ linux-2.4.0/include/linux/netfilter_ipv4/ipt_TCPMSS.h Wed Jan 24 12:07:54 2001 @@ -0,0 +1,10 @@ +#ifndef _IPT_TCPMSS_H +#define _IPT_TCPMSS_H + +struct ipt_tcpmss_info { + u_int16_t mss; +}; + +#define IPT_TCPMSS_CLAMP_PMTU 0xffff + +#endif /*_IPT_TCPMSS_H*/ diff -ruN linux-2.4.0.orig/include/linux/netfilter_ipv4/ipt_tcpmss.h linux-2.4.0/include/linux/netfilter_ipv4/ipt_tcpmss.h --- linux-2.4.0.orig/include/linux/netfilter_ipv4/ipt_tcpmss.h Wed Dec 31 19:00:00 1969 +++ linux-2.4.0/include/linux/netfilter_ipv4/ipt_tcpmss.h Wed Jan 24 12:07:54 2001 @@ -0,0 +1,9 @@ +#ifndef _IPT_TCPMSS_MATCH_H +#define _IPT_TCPMSS_MATCH_H + +struct ipt_tcpmss_match_info { + u_int16_t mss_min, mss_max; + u_int8_t invert; +}; + +#endif /*_IPT_TCPMSS_MATCH_H*/ diff -ruN linux-2.4.0.orig/net/ipv4/netfilter/Config.in linux-2.4.0/net/ipv4/netfilter/Config.in --- linux-2.4.0.orig/net/ipv4/netfilter/Config.in Mon Mar 27 13:35:56 2000 +++ linux-2.4.0/net/ipv4/netfilter/Config.in Wed Jan 24 12:07:54 2001 @@ -20,6 +20,7 @@ dep_tristate ' netfilter MARK match support' CONFIG_IP_NF_MATCH_MARK $CONFIG_IP_NF_IPTABLES dep_tristate ' Multiple port match support' CONFIG_IP_NF_MATCH_MULTIPORT $CONFIG_IP_NF_IPTABLES dep_tristate ' TOS match support' CONFIG_IP_NF_MATCH_TOS $CONFIG_IP_NF_IPTABLES + dep_tristate ' tcpmss match support' CONFIG_IP_NF_MATCH_TCPMSS $CONFIG_IP_NF_IPTABLES if [ "$CONFIG_IP_NF_CONNTRACK" != "n" ]; then dep_tristate ' Connection state match support' CONFIG_IP_NF_MATCH_STATE $CONFIG_IP_NF_CONNTRACK $CONFIG_IP_NF_IPTABLES fi @@ -51,6 +52,7 @@ dep_tristate ' MARK target support' CONFIG_IP_NF_TARGET_MARK $CONFIG_IP_NF_MANGLE fi dep_tristate ' LOG target support' CONFIG_IP_NF_TARGET_LOG $CONFIG_IP_NF_IPTABLES + dep_tristate ' TCPMSS target support' CONFIG_IP_NF_TARGET_TCPMSS $CONFIG_IP_NF_IPTABLES fi # Backwards compatibility modules: only if you don't build in the others. diff -ruN linux-2.4.0.orig/net/ipv4/netfilter/Makefile linux-2.4.0/net/ipv4/netfilter/Makefile --- linux-2.4.0.orig/net/ipv4/netfilter/Makefile Fri Dec 29 17:07:24 2000 +++ linux-2.4.0/net/ipv4/netfilter/Makefile Wed Jan 24 12:07:54 2001 @@ -54,6 +54,7 @@ obj-$(CONFIG_IP_NF_MATCH_TOS) += ipt_tos.o obj-$(CONFIG_IP_NF_MATCH_STATE) += ipt_state.o obj-$(CONFIG_IP_NF_MATCH_UNCLEAN) += ipt_unclean.o +obj-$(CONFIG_IP_NF_MATCH_TCPMSS) += ipt_tcpmss.o # targets obj-$(CONFIG_IP_NF_TARGET_REJECT) += ipt_REJECT.o @@ -63,6 +64,7 @@ obj-$(CONFIG_IP_NF_TARGET_MASQUERADE) += ipt_MASQUERADE.o obj-$(CONFIG_IP_NF_TARGET_REDIRECT) += ipt_REDIRECT.o obj-$(CONFIG_IP_NF_TARGET_LOG) += ipt_LOG.o +obj-$(CONFIG_IP_NF_TARGET_TCPMSS) += ipt_TCPMSS.o # backwards compatibility obj-$(CONFIG_IP_NF_COMPAT_IPCHAINS) += ipchains.o diff -ruN linux-2.4.0.orig/net/ipv4/netfilter/ipt_TCPMSS.c linux-2.4.0/net/ipv4/netfilter/ipt_TCPMSS.c --- linux-2.4.0.orig/net/ipv4/netfilter/ipt_TCPMSS.c Wed Dec 31 19:00:00 1969 +++ linux-2.4.0/net/ipv4/netfilter/ipt_TCPMSS.c Wed Jan 24 12:07:54 2001 @@ -0,0 +1,245 @@ +/* + * This is a module which is used for setting the MSS option in TCP packets. + * + * Copyright (c) 2000 Marc Boucher + */ +#include +#include + +#include +#include + +#include +#include + +#if 0 +#define DEBUGP printk +#else +#define DEBUGP(format, args...) +#endif + +static u_int16_t +cheat_check(u_int32_t oldvalinv, u_int32_t newval, u_int16_t oldcheck) +{ + u_int32_t diffs[] = { oldvalinv, newval }; + return csum_fold(csum_partial((char *)diffs, sizeof(diffs), + oldcheck^0xFFFF)); +} + +static inline unsigned int +optlen(const u_int8_t *opt, unsigned int offset) +{ + /* Beware zero-length options: make finite progress */ + if (opt[offset] <= TCPOPT_NOP || opt[offset+1] == 0) return 1; + else return opt[offset+1]; +} + +static unsigned int +ipt_tcpmss_target(struct sk_buff **pskb, + unsigned int hooknum, + const struct net_device *in, + const struct net_device *out, + const void *targinfo, + void *userinfo) +{ + const struct ipt_tcpmss_info *tcpmssinfo = targinfo; + struct tcphdr *tcph; + struct iphdr *iph = (*pskb)->nh.iph; + u_int16_t tcplen, newtotlen, oldval, newmss; + unsigned int i; + u_int8_t *opt; + + tcplen = (*pskb)->len - iph->ihl*4; + + tcph = (void *)iph + iph->ihl*4; + + /* Since it passed flags test in tcp match, we know it is is + not a fragment, and has data >= tcp header length. SYN + packets should not contain data: if they did, then we risk + running over MTU, sending Frag Needed and breaking things + badly. --RR */ + if (tcplen != tcph->doff*4) { + if (net_ratelimit()) + printk(KERN_ERR + "ipt_tcpmss_target: bad length (%d bytes)\n", + (*pskb)->len); + return NF_DROP; + } + + if(tcpmssinfo->mss == IPT_TCPMSS_CLAMP_PMTU) { + if(!(*pskb)->dst) { + if (net_ratelimit()) + printk(KERN_ERR + "ipt_tcpmss_target: no dst?! can't determine path-MTU\n"); + return NF_DROP; /* or IPT_CONTINUE ?? */ + } + + if((*pskb)->dst->pmtu <= (sizeof(struct iphdr) + sizeof(struct tcphdr))) { + if (net_ratelimit()) + printk(KERN_ERR + "ipt_tcpmss_target: unknown or invalid path-MTU (%d)\n", (*pskb)->dst->pmtu); + return NF_DROP; /* or IPT_CONTINUE ?? */ + } + + newmss = (*pskb)->dst->pmtu - sizeof(struct iphdr) - sizeof(struct tcphdr); + } else + newmss = tcpmssinfo->mss; + + opt = (u_int8_t *)tcph; + for (i = sizeof(struct tcphdr); i < tcph->doff*4; i += optlen(opt, i)){ + if ((opt[i] == TCPOPT_MSS) && + ((tcph->doff*4 - i) >= TCPOLEN_MSS) && + (opt[i+1] == TCPOLEN_MSS)) { + u_int16_t oldmss; + + oldmss = (opt[i+2] << 8) | opt[i+3]; + + if((tcpmssinfo->mss == IPT_TCPMSS_CLAMP_PMTU) && + (oldmss <= newmss)) + return IPT_CONTINUE; + + opt[i+2] = (newmss & 0xff00) >> 8; + opt[i+3] = (newmss & 0x00ff); + + tcph->check = cheat_check(htons(oldmss)^0xFFFF, + htons(newmss), + tcph->check); + + DEBUGP(KERN_INFO "ipt_tcpmss_target: %u.%u.%u.%u:%hu" + "->%u.%u.%u.%u:%hu changed TCP MSS option" + " (from %u to %u)\n", + NIPQUAD((*pskb)->nh.iph->saddr), + ntohs(tcph->source), + NIPQUAD((*pskb)->nh.iph->daddr), + ntohs(tcph->dest), + oldmss, newmss); + goto retmodified; + } + } + + /* + * MSS Option not found ?! add it.. + */ + if (skb_tailroom((*pskb)) < TCPOLEN_MSS) { + struct sk_buff *newskb; + + newskb = skb_copy_expand(*pskb, skb_headroom(*pskb), + TCPOLEN_MSS, GFP_ATOMIC); + if (!newskb) { + if (net_ratelimit()) + printk(KERN_ERR "ipt_tcpmss_target:" + " unable to allocate larger skb\n"); + return NF_DROP; + } + + kfree_skb(*pskb); + *pskb = newskb; + iph = (*pskb)->nh.iph; + tcph = (void *)iph + iph->ihl*4; + } + + skb_put((*pskb), TCPOLEN_MSS); + + opt = (u_int8_t *)tcph + sizeof(struct tcphdr); + memmove(opt + TCPOLEN_MSS, opt, tcplen - sizeof(struct tcphdr)); + + tcph->check = cheat_check(htons(tcplen) ^ 0xFFFF, + htons(tcplen + TCPOLEN_MSS), tcph->check); + tcplen += TCPOLEN_MSS; + + opt[0] = TCPOPT_MSS; + opt[1] = TCPOLEN_MSS; + opt[2] = (newmss & 0xff00) >> 8; + opt[3] = (newmss & 0x00ff); + + tcph->check = cheat_check(~0, *((u_int32_t *)opt), tcph->check); + + oldval = ((u_int16_t *)tcph)[6]; + tcph->doff += TCPOLEN_MSS/4; + tcph->check = cheat_check(oldval ^ 0xFFFF, + ((u_int16_t *)tcph)[6], tcph->check); + + newtotlen = htons(ntohs(iph->tot_len) + TCPOLEN_MSS); + iph->check = cheat_check(iph->tot_len ^ 0xFFFF, + newtotlen, iph->check); + iph->tot_len = newtotlen; + + DEBUGP(KERN_INFO "ipt_tcpmss_target: %u.%u.%u.%u:%hu" + "->%u.%u.%u.%u:%hu added TCP MSS option (%u)\n", + NIPQUAD((*pskb)->nh.iph->saddr), + ntohs(tcph->source), + NIPQUAD((*pskb)->nh.iph->daddr), + ntohs(tcph->dest), + newmss); + + retmodified: + /* If we had a hardware checksum before, it's now invalid */ + (*pskb)->ip_summed = CHECKSUM_NONE; + (*pskb)->nfcache |= NFC_UNKNOWN | NFC_ALTERED; + return IPT_CONTINUE; +} + +#define TH_SYN 0x02 + +static inline int find_syn_match(const struct ipt_entry_match *m) +{ + const struct ipt_tcp *tcpinfo = (const struct ipt_tcp *)m->data; + + if (strcmp(m->u.kernel.match->name, "tcp") == 0 + && (tcpinfo->flg_cmp & TH_SYN) + && !(tcpinfo->invflags & IPT_TCP_INV_FLAGS)) + return 1; + + return 0; +} + +/* Must specify -p tcp --syn/--tcp-flags SYN */ +static int +ipt_tcpmss_checkentry(const char *tablename, + const struct ipt_entry *e, + void *targinfo, + unsigned int targinfosize, + unsigned int hook_mask) +{ + const struct ipt_tcpmss_info *tcpmssinfo = targinfo; + + if (targinfosize != IPT_ALIGN(sizeof(struct ipt_tcpmss_info))) { + DEBUGP("ipt_tcpmss_checkentry: targinfosize %u != %u\n", + targinfosize, IPT_ALIGN(sizeof(struct ipt_tcpmss_info))); + return 0; + } + + + if((tcpmssinfo->mss == IPT_TCPMSS_CLAMP_PMTU) && + ((hook_mask & ~((1 << NF_IP_FORWARD) + | (1 << NF_IP_LOCAL_OUT) + | (1 << NF_IP_POST_ROUTING))) != 0)) { + printk("TCPMSS: path-MTU clamping only supported in FORWARD, OUTPUT and POSTROUTING hooks\n"); + return 0; + } + + if (e->ip.proto == IPPROTO_TCP + && !(e->ip.invflags & IPT_INV_PROTO) + && IPT_MATCH_ITERATE(e, find_syn_match)) + return 1; + + printk("TCPMSS: Only works on TCP SYN packets\n"); + return 0; +} + +static struct ipt_target ipt_tcpmss_reg += { { NULL, NULL }, "TCPMSS", + ipt_tcpmss_target, ipt_tcpmss_checkentry, NULL, THIS_MODULE }; + +static int __init init(void) +{ + return ipt_register_target(&ipt_tcpmss_reg); +} + +static void __exit fini(void) +{ + ipt_unregister_target(&ipt_tcpmss_reg); +} + +module_init(init); +module_exit(fini); diff -ruN linux-2.4.0.orig/net/ipv4/netfilter/ipt_tcpmss.c linux-2.4.0/net/ipv4/netfilter/ipt_tcpmss.c --- linux-2.4.0.orig/net/ipv4/netfilter/ipt_tcpmss.c Wed Dec 31 19:00:00 1969 +++ linux-2.4.0/net/ipv4/netfilter/ipt_tcpmss.c Wed Jan 24 12:07:54 2001 @@ -0,0 +1,108 @@ +/* Kernel module to match TCP MSS values. */ +#include +#include +#include + +#include +#include + +#define TH_SYN 0x02 + +/* Returns 1 if the mss option is set and matched by the range, 0 otherwise */ +static inline int +mssoption_match(u_int16_t min, u_int16_t max, + const struct tcphdr *tcp, + u_int16_t datalen, + int invert, + int *hotdrop) +{ + unsigned int i; + const u_int8_t *opt = (u_int8_t *)tcp; + + /* If we don't have the whole header, drop packet. */ + if (tcp->doff * 4 > datalen) { + *hotdrop = 1; + return 0; + } + + for (i = sizeof(struct tcphdr); i < tcp->doff * 4; ) { + if ((opt[i] == TCPOPT_MSS) + && ((tcp->doff * 4 - i) >= TCPOLEN_MSS) + && (opt[i+1] == TCPOLEN_MSS)) { + u_int16_t mssval; + + mssval = (opt[i+2] << 8) | opt[i+3]; + + return (mssval >= min && mssval <= max) ^ invert; + } + if (opt[i] < 2) i++; + else i += opt[i+1]?:1; + } + + return invert; +} + +static int +match(const struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + const void *matchinfo, + int offset, + const void *hdr, + u_int16_t datalen, + int *hotdrop) +{ + const struct ipt_tcpmss_match_info *info = matchinfo; + const struct tcphdr *tcph = (void *)skb->nh.iph + skb->nh.iph->ihl*4; + + return mssoption_match(info->mss_min, info->mss_max, tcph, + skb->len - skb->nh.iph->ihl*4, + info->invert, hotdrop); +} + +static inline int find_syn_match(const struct ipt_entry_match *m) +{ + const struct ipt_tcp *tcpinfo = (const struct ipt_tcp *)m->data; + + if (strcmp(m->u.kernel.match->name, "tcp") == 0 + && (tcpinfo->flg_cmp & TH_SYN) + && !(tcpinfo->invflags & IPT_TCP_INV_FLAGS)) + return 1; + + return 0; +} + +static int +checkentry(const char *tablename, + const struct ipt_ip *ip, + void *matchinfo, + unsigned int matchsize, + unsigned int hook_mask) +{ + if (matchsize != IPT_ALIGN(sizeof(struct ipt_tcpmss_match_info))) + return 0; + + /* Must specify -p tcp */ + if (ip->proto != IPPROTO_TCP || (ip->invflags & IPT_INV_PROTO)) { + printk("tcpmss: Only works on TCP packets\n"); + return 0; + } + + return 1; +} + +static struct ipt_match tcpmss_match += { { NULL, NULL }, "tcpmss", &match, &checkentry, NULL, THIS_MODULE }; + +static int __init init(void) +{ + return ipt_register_match(&tcpmss_match); +} + +static void __exit fini(void) +{ + ipt_unregister_match(&tcpmss_match); +} + +module_init(init); +module_exit(fini);