From owner-freebsd-net Sun May 3 09:33:38 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id JAA17939 for freebsd-net-outgoing; Sun, 3 May 1998 09:33:38 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from yunus.cmpe.boun.edu.tr (agayev@[208.218.37.218]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id JAA17929 for ; Sun, 3 May 1998 09:33:26 -0700 (PDT) (envelope-from agayev@yunus.cmpe.boun.edu.tr) Received: (from agayev@localhost) by yunus.cmpe.boun.edu.tr (8.8.7/8.8.7) id TAA23885 for freebsd-net@freebsd.org; Sun, 3 May 1998 19:32:11 +0300 Date: Sun, 3 May 1998 19:32:11 +0300 From: Agayev Ebutalip Message-Id: <199805031632.TAA23885@yunus.cmpe.boun.edu.tr> To: freebsd-net@FreeBSD.ORG Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org unsubscribe freebsd-net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sun May 3 16:44:25 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA24865 for freebsd-net-outgoing; Sun, 3 May 1998 16:44:25 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from frmug.org (frmug-gw.frmug.org [193.56.58.252]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA24854 for ; Sun, 3 May 1998 16:44:09 -0700 (PDT) (envelope-from pb@fasterix.frmug.org) Received: (from uucp@localhost) by frmug.org (8.9.0.Beta7/frmug-2.3/nospam) with UUCP id BAA23707 for freebsd-net@freebsd.org; Mon, 4 May 1998 01:44:02 +0200 (CEST) (envelope-from pb@fasterix.frmug.org) Received: (from pb@localhost) by fasterix.frmug.org (8.8.8/8.8.5/pb-19970302) id BAA27510 for freebsd-net@freebsd.org; Mon, 4 May 1998 01:43:11 +0200 (CEST) Message-ID: <19980504014310.A22037@fasterix.frmug.fr.net> Date: Mon, 4 May 1998 01:43:10 +0200 From: Pierre Beyssac To: freebsd-net@FreeBSD.ORG Subject: patches for fast forwarding (with kernel option & sysctl) Mime-Version: 1.0 Content-Type: multipart/mixed; boundary=VS++wcV0S1rZb1Fb X-Mailer: Mutt 0.92.2 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org --VS++wcV0S1rZb1Fb Content-Type: text/plain; charset=us-ascii Here are new patches to use NetBSD's IPFLOW stuff on -current. To use, store ipflow.c in /sys/netinet, then add IPFLOW to your kernel config. After booting, activate with: sysctl -w net.inet.ip.fastforwarding=1 Two attachments follow: the patches and the additional file. -- Pierre Beyssac pb@fasterix.frmug.org pb@fasterix.freenix.org {Free,Net,Open}BSD, Linux : il y a moins bien, mais c'est plus cher Free domains: http://www.eu.org/ or mail dns-manager@EU.org --VS++wcV0S1rZb1Fb Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename=patch2 Index: conf/files =================================================================== RCS file: /home/ncvs/src/sys/conf/files,v retrieving revision 1.136 diff -u -r1.136 files --- files 1998/04/22 18:12:29 1.136 +++ files 1998/05/03 23:30:21 @@ -268,6 +268,7 @@ netinet/ip_auth.c optional ipfilter inet netinet/ip_divert.c optional ipdivert netinet/ip_fil.c optional ipfilter inet +netinet/ip_flow.c optional ipflow inet netinet/ip_frag.c optional ipfilter inet netinet/ip_fw.c optional ipfirewall netinet/ip_icmp.c optional inet Index: conf/options =================================================================== RCS file: /home/ncvs/src/sys/conf/options,v retrieving revision 1.72 diff -u -r1.72 options --- options 1998/04/20 04:30:41 1.72 +++ options 1998/05/03 23:30:24 @@ -172,6 +172,7 @@ IPFIREWALL_VERBOSE opt_ipfw.h IPFIREWALL_VERBOSE_LIMIT opt_ipfw.h IPFIREWALL_DEFAULT_TO_ACCEPT opt_ipfw.h +IPFLOW opt_ipflow.h IPX opt_ipx.h IPXIP opt_ipx.h IPTUNNEL opt_ipx.h Index: i386/conf/LINT =================================================================== RCS file: /home/ncvs/src/sys/i386/conf/LINT,v retrieving revision 1.429 diff -u -r1.429 LINT --- LINT 1998/04/29 17:09:41 1.429 +++ LINT 1998/05/03 23:30:53 @@ -419,6 +419,9 @@ # IPFILTER_LOG enables ipfilter's logging. # IPFILTER_LKM enables LKM support for an ipfilter module (untested). # +# IPFLOW enables Matt Thomas' fast IP forwarding code +# (activate using "sysctl -w net.inet.ip.fastforwarding=1") +# # TCPDEBUG is undocumented. # options "TCP_COMPAT_42" #emulate 4.2BSD TCP bugs @@ -432,6 +435,7 @@ options IPFILTER #kernel ipfilter support options IPFILTER_LOG #ipfilter logging #options IPFILTER_LKM #kernel support for ip_fil.o LKM +options IPFLOW options TCPDEBUG Index: net/if_ethersubr.c =================================================================== RCS file: /home/ncvs/src/sys/net/if_ethersubr.c,v retrieving revision 1.47 diff -u -r1.47 if_ethersubr.c --- if_ethersubr.c 1998/03/30 09:51:39 1.47 +++ if_ethersubr.c 1998/05/03 23:31:06 @@ -36,6 +36,7 @@ #include "opt_atalk.h" #include "opt_inet.h" +#include "opt_ipflow.h" #include "opt_ipx.h" #include @@ -501,6 +502,10 @@ switch (ether_type) { #ifdef INET case ETHERTYPE_IP: +#ifdef IPFLOW + if (ipflow_fastforward(m)) + return; +#endif schednetisr(NETISR_IP); inq = &ipintrq; break; Index: net/if_fddisubr.c =================================================================== RCS file: /home/ncvs/src/sys/net/if_fddisubr.c,v retrieving revision 1.27 diff -u -r1.27 if_fddisubr.c --- if_fddisubr.c 1998/03/30 09:51:44 1.27 +++ if_fddisubr.c 1998/05/03 23:31:18 @@ -38,6 +38,7 @@ #include "opt_atalk.h" #include "opt_inet.h" +#include "opt_ipflow.h" #include "opt_ipx.h" #include @@ -533,6 +534,10 @@ switch (type) { #ifdef INET case ETHERTYPE_IP: +#ifdef IPFLOW + if (ipflow_fastforward(m)) + return; +#endif schednetisr(NETISR_IP); inq = &ipintrq; break; Index: net/if_ppp.c =================================================================== RCS file: /home/ncvs/src/sys/net/if_ppp.c,v retrieving revision 1.56 diff -u -r1.56 if_ppp.c --- if_ppp.c 1998/04/06 11:43:10 1.56 +++ if_ppp.c 1998/05/03 23:31:50 @@ -77,6 +77,7 @@ #if NPPP > 0 #include "opt_inet.h" +#include "opt_ipflow.h" #include "opt_ipx.h" #include "opt_ppp.h" @@ -1488,6 +1489,12 @@ m->m_pkthdr.len -= PPP_HDRLEN; m->m_data += PPP_HDRLEN; m->m_len -= PPP_HDRLEN; +#ifdef IPFLOW + if (ipflow_fastforward(m)) { + sc->sc_last_recv = time_second; + return; + } +#endif schednetisr(NETISR_IP); inq = &ipintrq; sc->sc_last_recv = time_second; /* update time of last pkt rcvd */ Index: netinet/in.h =================================================================== RCS file: /home/ncvs/src/sys/netinet/in.h,v retrieving revision 1.31 diff -u -r1.31 in.h --- in.h 1998/04/19 17:22:27 1.31 +++ in.h 1998/05/03 23:32:00 @@ -398,7 +398,8 @@ #define IPCTL_INTRQDROPS 11 /* number of netisr q drops */ #define IPCTL_STATS 12 /* ipstat structure */ #define IPCTL_ACCEPTSOURCEROUTE 13 /* may accept source routed packets */ -#define IPCTL_MAXID 14 +#define IPCTL_FASTFORWARDING 14 /* use fast IP forwarding code */ +#define IPCTL_MAXID 15 #define IPCTL_NAMES { \ { 0, 0 }, \ @@ -415,6 +416,7 @@ { "intr-queue-drops", CTLTYPE_INT }, \ { "stats", CTLTYPE_STRUCT }, \ { "accept_sourceroute", CTLTYPE_INT }, \ + { "fastforwarding", CTLTYPE_INT }, \ } Index: netinet/in_var.h =================================================================== RCS file: /home/ncvs/src/sys/netinet/in_var.h,v retrieving revision 1.27 diff -u -r1.27 in_var.h --- in_var.h 1997/09/07 05:26:43 1.27 +++ in_var.h 1998/05/03 23:32:02 @@ -211,6 +211,7 @@ IN_NEXT_MULTI((step), (inm)); \ } while(0) +struct route; struct in_multi *in_addmulti __P((struct in_addr *, struct ifnet *)); void in_delmulti __P((struct in_multi *)); int in_control __P((struct socket *, int, caddr_t, struct ifnet *, @@ -219,6 +220,9 @@ void ip_input __P((struct mbuf *)); int in_ifadown __P((struct ifaddr *ifa)); void in_ifscrub __P((struct ifnet *, struct in_ifaddr *)); +int ipflow_fastforward __P((struct mbuf *)); +void ipflow_create __P((const struct route *, struct mbuf *)); +void ipflow_slowtimo __P((void)); #endif /* KERNEL */ Index: netinet/ip_fw.c =================================================================== RCS file: /home/ncvs/src/sys/netinet/ip_fw.c,v retrieving revision 1.82 diff -u -r1.82 ip_fw.c --- ip_fw.c 1998/04/21 18:54:53 1.82 +++ ip_fw.c 1998/05/03 23:32:13 @@ -36,6 +36,7 @@ #include #include #include +#include #include #include #include Index: netinet/ip_input.c =================================================================== RCS file: /home/ncvs/src/sys/netinet/ip_input.c,v retrieving revision 1.82 diff -u -r1.82 ip_input.c --- ip_input.c 1998/04/13 17:27:08 1.82 +++ ip_input.c 1998/05/03 23:32:27 @@ -41,6 +41,7 @@ #include "opt_ipfw.h" #include "opt_ipdivert.h" #include "opt_ipfilter.h" +#include "opt_ipflow.h" #include @@ -80,7 +81,7 @@ static int ip_rsvp_on; struct socket *ip_rsvpd; -static int ipforwarding = 0; +int ipforwarding = 0; SYSCTL_INT(_net_inet_ip, IPCTL_FORWARDING, forwarding, CTLFLAG_RW, &ipforwarding, 0, ""); @@ -878,6 +879,9 @@ } } } +#ifdef IPFLOW + ipflow_slowtimo(); +#endif splx(s); } @@ -1381,8 +1385,12 @@ if (type) ipstat.ips_redirectsent++; else { - if (mcopy) + if (mcopy) { +#ifdef IPFLOW + ipflow_create(&ipforward_rt, mcopy); +#endif m_freem(mcopy); + } return; } } Index: netinet/ip_var.h =================================================================== RCS file: /home/ncvs/src/sys/netinet/ip_var.h,v retrieving revision 1.34 diff -u -r1.34 ip_var.h --- ip_var.h 1997/09/07 05:26:46 1.34 +++ ip_var.h 1998/05/03 23:32:29 @@ -132,6 +132,7 @@ u_long ips_fragdropped; /* frags dropped (dups, out of space) */ u_long ips_fragtimeout; /* fragments timed out */ u_long ips_forward; /* packets forwarded */ + u_long ips_fastforward; /* packets fast forwarded */ u_long ips_cantforward; /* packets rcvd for unreachable dest */ u_long ips_redirectsent; /* packets forwarded on same net */ u_long ips_noproto; /* unknown or unsupported protocol */ @@ -150,6 +151,21 @@ u_long ips_notmember; /* multicasts for unregistered grps */ }; +#define IPFLOW_HASHBITS 6 /* should not be a multiple of 8 */ +struct ipflow { + LIST_ENTRY(ipflow) ipf_next; /* next ipflow in bucket */ + struct in_addr ipf_dst; /* destination address */ + struct in_addr ipf_src; /* source address */ + u_int8_t ipf_tos; /* type-of-service */ + struct route ipf_ro; /* associated route entry */ + u_long ipf_uses; /* number of uses in this period */ + u_long ipf_last_uses; /* number of uses in last period */ + u_long ipf_dropped; /* ENOBUFS returned by if_output */ + u_long ipf_errors; /* other errors returned by if_output */ + int ipf_timer; /* remaining lifetime of this entry */ + time_t ipf_start; /* creation time */ +}; + #ifdef KERNEL /* flags passed to ip_output as last parameter */ #define IP_FORWARDING 0x1 /* most of ip header exists */ @@ -163,6 +179,7 @@ extern struct ipstat ipstat; extern u_short ip_id; /* ip packet ctr, for ids */ extern int ip_defttl; /* default IP ttl */ +extern int ipforwarding; /* ip forwarding */ extern u_char ip_protox[]; extern struct socket *ip_rsvpd; /* reservation protocol daemon */ extern struct socket *ip_mrouter; /* multicast routing daemon */ --VS++wcV0S1rZb1Fb Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename="ip_flow.c" /* $NetBSD: ip_flow.c,v 1.1 1998/04/29 21:37:55 matt Exp $ */ /*- * Copyright (c) 1998 The NetBSD Foundation, Inc. * All rights reserved. * * This code is derived from software contributed to The NetBSD Foundation * by the 3am Software Foundry ("3am"). It was developed by Matt Thomas. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by the NetBSD * Foundation, Inc. and its contributors. * 4. Neither the name of The NetBSD Foundation nor the names of its * contributors may be used to endorse or promote products derived * from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define IPFLOW_TIMER (5 * PR_SLOWHZ) #define IPFLOW_HASHSIZE (1 << IPFLOW_HASHBITS) static LIST_HEAD(ipflowhead, ipflow) ipflows[IPFLOW_HASHSIZE]; static int ipflow_inuse; #define IPFLOW_MAX 256 static int ipflow_active = 0; SYSCTL_INT(_net_inet_ip, IPCTL_FASTFORWARDING, fastforwarding, CTLFLAG_RW, &ipflow_active, 0, ""); MALLOC_DEFINE(M_IPFLOW, "ip_flow", "IP flow"); static unsigned ipflow_hash( struct in_addr dst, struct in_addr src, unsigned tos) { unsigned hash = tos; int idx; for (idx = 0; idx < 32; idx += IPFLOW_HASHBITS) hash += (dst.s_addr >> (32 - idx)) + (src.s_addr >> idx); return hash & (IPFLOW_HASHSIZE-1); } static struct ipflow * ipflow_lookup( const struct ip *ip) { unsigned hash; struct ipflow *ipf; hash = ipflow_hash(ip->ip_dst, ip->ip_src, ip->ip_tos); ipf = LIST_FIRST(&ipflows[hash]); while (ipf != NULL) { if (ip->ip_dst.s_addr == ipf->ipf_dst.s_addr && ip->ip_src.s_addr == ipf->ipf_src.s_addr && ip->ip_tos == ipf->ipf_tos) break; ipf = LIST_NEXT(ipf, ipf_next); } return ipf; } int ipflow_fastforward( struct mbuf *m) { struct ip *ip; struct ipflow *ipf; struct rtentry *rt; u_int32_t sum; int error; /* * Are we forwarding packets? Big enough for an IP packet? */ if (!ipforwarding || !ipflow_active || m->m_len < sizeof(struct ip)) return 0; /* * IP header with no option and valid version and length */ ip = mtod(m, struct ip *); if (ip->ip_v != IPVERSION || ip->ip_hl != (sizeof(struct ip) >> 2) || ntohs(ip->ip_len) > m->m_pkthdr.len) return 0; /* * Find a flow. */ if ((ipf = ipflow_lookup(ip)) == NULL) return 0; /* * Route and interface still up? */ rt = ipf->ipf_ro.ro_rt; if ((rt->rt_flags & RTF_UP) == 0 || (rt->rt_ifp->if_flags & IFF_UP) == 0) return 0; /* * Packet size OK? TTL? */ if (m->m_pkthdr.len > rt->rt_ifp->if_mtu || ip->ip_ttl <= IPTTLDEC) return 0; /* * Everything checks out and so we can forward this packet. * Modify the TTL and incrementally change the checksum. * On little endian machine, the TTL is in LSB position * (so we can simply add) while on big-endian it's in the * MSB position (so we have to do two calculation; the first * is the add and second is to wrap the results into 17 bits, * 16 bits and a carry). */ ip->ip_ttl -= IPTTLDEC; #if BYTE_ORDER == LITTLE_ENDIAN sum = ip->ip_sum + IPTTLDEC; #endif #if BYTE_ORDER == BIG_ENDIAN sum = ip->ip_sum + (IPTTLDEC << 8); sum = (sum & 0xFFFF) + (sum >> 16); #endif if (sum > 0x10000) /* add in carry if needed */ sum++; ip->ip_sum = sum; /* bit 16 is dropped */ /* * Send the packet on its way. All we can get back is ENOBUFS */ ipf->ipf_uses++; ipf->ipf_timer = IPFLOW_TIMER; if ((error = (*rt->rt_ifp->if_output)(rt->rt_ifp, m, &ipf->ipf_ro.ro_dst, rt)) != 0) { if (error == ENOBUFS) ipf->ipf_dropped++; else ipf->ipf_errors++; } return 1; } static void ipflow_addstats( struct ipflow *ipf) { ipf->ipf_ro.ro_rt->rt_use += ipf->ipf_uses; ipstat.ips_cantforward += ipf->ipf_errors + ipf->ipf_dropped; ipstat.ips_forward += ipf->ipf_uses; ipstat.ips_fastforward += ipf->ipf_uses; } static void ipflow_free( struct ipflow *ipf) { int s; /* * Remove the flow from the hash table (at elevated IPL). * Once it's off the list, we can deal with it at normal * network IPL. */ s = splimp(); LIST_REMOVE(ipf, ipf_next); splx(s); ipflow_addstats(ipf); RTFREE(ipf->ipf_ro.ro_rt); ipflow_inuse--; FREE(ipf, M_IPFLOW); } static struct ipflow * ipflow_reap( void) { struct ipflow *ipf, *maybe_ipf = NULL; int idx; int s; for (idx = 0; idx < IPFLOW_HASHSIZE; idx++) { ipf = LIST_FIRST(&ipflows[idx]); while (ipf != NULL) { /* * If this no longer points to a valid route * reclaim it. */ if ((ipf->ipf_ro.ro_rt->rt_flags & RTF_UP) == 0) goto done; /* * choose the one that's been least recently used * or has had the least uses in the last 1.5 * intervals. */ if (ipf == NULL || ipf->ipf_timer < maybe_ipf->ipf_timer || (ipf->ipf_timer == maybe_ipf->ipf_timer && ipf->ipf_last_uses + ipf->ipf_uses < maybe_ipf->ipf_last_uses + maybe_ipf->ipf_uses)) maybe_ipf = ipf; ipf = LIST_NEXT(ipf, ipf_next); } } ipf = maybe_ipf; done: /* * Remove the entry from the flow table. */ s = splimp(); LIST_REMOVE(ipf, ipf_next); splx(s); ipflow_addstats(ipf); RTFREE(ipf->ipf_ro.ro_rt); return ipf; } void ipflow_slowtimo( void) { struct ipflow *ipf; int idx; for (idx = 0; idx < IPFLOW_HASHSIZE; idx++) { ipf = LIST_FIRST(&ipflows[idx]); while (ipf != NULL) { struct ipflow *next_ipf = LIST_NEXT(ipf, ipf_next); if (--ipf->ipf_timer == 0) { ipflow_free(ipf); } else { ipf->ipf_last_uses = ipf->ipf_uses; ipf->ipf_ro.ro_rt->rt_use += ipf->ipf_uses; ipstat.ips_forward += ipf->ipf_uses; ipstat.ips_fastforward += ipf->ipf_uses; ipf->ipf_uses = 0; } ipf = next_ipf; } } } void ipflow_create( const struct route *ro, struct mbuf *m) { const struct ip *const ip = mtod(m, struct ip *); struct ipflow *ipf; unsigned hash; int s; /* * Don't create cache entries for ICMP messages. */ if (!ipflow_active || ip->ip_p == IPPROTO_ICMP) return; /* * See if an existing flow struct exists. If so remove it from it's * list and free the old route. If not, try to malloc a new one * (if we aren't at our limit). */ ipf = ipflow_lookup(ip); if (ipf == NULL) { if (ipflow_inuse == IPFLOW_MAX) { ipf = ipflow_reap(); } else { ipf = (struct ipflow *) malloc(sizeof(*ipf), M_IPFLOW, M_NOWAIT); if (ipf == NULL) return; ipflow_inuse++; } bzero((caddr_t) ipf, sizeof(*ipf)); } else { s = splimp(); LIST_REMOVE(ipf, ipf_next); splx(s); ipflow_addstats(ipf); RTFREE(ipf->ipf_ro.ro_rt); ipf->ipf_uses = ipf->ipf_last_uses = 0; ipf->ipf_errors = ipf->ipf_dropped = 0; } /* * Fill in the updated information. */ ipf->ipf_ro = *ro; ro->ro_rt->rt_refcnt++; ipf->ipf_dst = ip->ip_dst; ipf->ipf_src = ip->ip_src; ipf->ipf_tos = ip->ip_tos; ipf->ipf_timer = IPFLOW_TIMER; ipf->ipf_start = time_second; /* * Insert into the approriate bucket of the flow table. */ hash = ipflow_hash(ip->ip_dst, ip->ip_src, ip->ip_tos); s = splimp(); LIST_INSERT_HEAD(&ipflows[hash], ipf, ipf_next); splx(s); } --VS++wcV0S1rZb1Fb-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sun May 3 17:09:23 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id RAA28886 for freebsd-net-outgoing; Sun, 3 May 1998 17:09:23 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from implode.root.com (implode.root.com [198.145.90.17]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id RAA28873 for ; Sun, 3 May 1998 17:09:13 -0700 (PDT) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.5/8.8.5) with ESMTP id RAA02612; Sun, 3 May 1998 17:07:30 -0700 (PDT) Message-Id: <199805040007.RAA02612@implode.root.com> To: Pierre Beyssac cc: freebsd-net@FreeBSD.ORG Subject: Re: patches for fast forwarding (with kernel option & sysctl) In-reply-to: Your message of "Mon, 04 May 1998 01:43:10 +0200." <19980504014310.A22037@fasterix.frmug.fr.net> From: David Greenman Reply-To: dg@root.com Date: Sun, 03 May 1998 17:07:30 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Here are new patches to use NetBSD's IPFLOW stuff on -current. To use, store ipflow.c in /sys/netinet, then add IPFLOW to your kernel config. After booting, activate with: These appear to contain the broken header checksumming code. That'll need to be fixed before these are integrated. There is a NetBSD #5380 open on this - author Dennis Ferguson . Text from PR#5380 for proposed fix: >Given an htons() macro which compiles to a constant with a constant >argument the following code fragment avoids generating a 0xffff >checksum, produces a correct checksum in all cases, avoids the #if's, >and does the operation with a single comparison and addition on >machines of either byte order. > > ip->ip_ttl -= IPTTLDEC; > if (ip->ip_sum >= htons(0xffff - (IPTTLDEC << 8))) { > ip->ip_sum += htons(IPTTLDEC << 8) + 1; > } else { > ip->ip_sum += htons(IPTTLDEC << 8); > } -DG David Greenman Co-founder/Principal Architect, The FreeBSD Project To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sun May 3 17:28:34 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id RAA02666 for freebsd-net-outgoing; Sun, 3 May 1998 17:28:34 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from mh2.cts.com (root@mh2.cts.com [205.163.24.68]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id RAA02653 for ; Sun, 3 May 1998 17:28:23 -0700 (PDT) (envelope-from danx@cts.com) Received: from king.cts.com (danx@king.cts.com [198.68.168.21]) by mh2.cts.com (8.8.7/8.8.5) with SMTP id RAA23378 for ; Sun, 3 May 1998 17:28:23 -0700 (PDT) Date: Sun, 3 May 1998 17:28:23 -0700 (PDT) From: Dan Anderson To: FreeBSD-net List Subject: Help w/IPv6 config for FreeBSD Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi There, I have the IPv6 patches by INRIA.fr for FreeBSD 2.2.5 and have it up and running locally. I'm trying to connect to the 6bone through a tunnel but cannot ping. (Gory details below for those interested). I've done it successfully for Linux. The INRIA instructions are very sketchy, so . . . if you tunneled successfully to IPv6 with sit tunnel devices, please e-mail me your /etc/rc.ipv6 startup file. Thanks. - Dan Anderson, , San Diego, California # cat /etc/rc.ipv6 #!/bin/sh # /etc/rc.ipv6 autoconf6 -v >/tmp/autoconf6 2>&1 ifconfig ed1 inet6 firstalias 3ffe:1d01:2:2::/64 eui64 #danx.att # Tunnel ifconfig sit0 inet6 ::204.94.80.94/96 ifconfig sit0 up ifconfig sit1 inet6 ::207.211.160.111/128 ifconfig sit1 up route -n add -inet6 3ffe::/16 ::207.211.160.111 sysctl -w net.inet6.ipv6.forwarding=1 sysctl -w net.inet6.ipv6.mforwarding=1 # netstat -rn Routing tables Internet: Destination Gateway Flags Refs Use Mtu Netif Expire default 192.188.72.30 UGSc 1 2 1500 ppp0 10.1.1/24 link#1 UC 0 0 1500 ed1 - 127.0.0.1 127.0.0.1 UHl 1 1 16384 lo0 192.188.72.30 204.94.80.94 UH 2 0 1500 ppp0 204.94.80.94 127.0.0.1 UGHS 0 1 16384 lo0 IPv6: Destination Gateway Flags Refs Use Mtu Netif Expire ::/96 0.0.0.0 UC 0 0 - sit0 - ::1 ::1 UHl 0 0 16384 lo0 ::207.211.160.111 207.211.160.111 UHLl 1 0 16384 sit1 3ffe::/16 ::207.211.160.111 UGS 0 3 1480 sit1 3ffe:1d01:2:2::/64 link#1 UC 0 0 1500 ed1 - 3ffe:1d01:2:2:240:5ff:fe19:2f85 0:40:5:19:2f:85 UHDLWl 0 2 16384 ed1 fe80::/64 link#1 UC 0 0 1500 ed1 - 12#ff01::/16 ::1 US 0 0 16384 lo0 1#ff02::/16 fe80::240:5ff:fe19:2f85 US 0 3 1500 ed1 12#ff11::/16 ::1 US 0 0 16384 lo0 1#ff12::/16 fe80::240:5ff:fe19:2f85 US 0 0 1500 ed1 # ifconfig -a ed1: flags=8a43 mtu 1500 inet 10.1.1.1 netmask 0xffffff00 broadcast 10.1.1.255 inet6 3ffe:1d01:2:2:240:5ff:fe19:2f85/64 inet6 fe80::240:5ff:fe19:2f85/64 ether 00:40:05:19:2f:85 lp0: flags=8810 mtu 1500 sit0: flags=1041 mtu 1480 inet6 ::204.94.80.94/96 lladdr 0a:01:01:01 sit1: flags=1001 mtu 1480 inet6 ::207.211.160.111/128 sit2: flags=1000 mtu 1480 cti0: flags=8010 mtu 1480 cti1: flags=8010 mtu 1480 ll0: flags=2 mtu 1500 tun0: flags=8010 mtu 1500 sl0: flags=c010 mtu 552 ppp0: flags=8051 mtu 1500 inet 204.94.80.94 --> 192.188.72.30 netmask 0xffffff00 lo0: flags=8049 mtu 16384 inet 127.0.0.1 netmask 0xff000000 inet6 ::1/128 # ping6 danx.ipv6 trying to get source for danx.ipv6 source should be 3ffe:1d01:2:2:240:5ff:fe19:2f85 PING danx.ipv6.yosemite.ca.us (3ffe:1d01:2:2:240:5ff:fe19:2f85): 56 data bytes 64 bytes from 3ffe:1d01:2:2:240:5ff:fe19:2f85: icmp6_seq=0 ttl=255 time=0.234 ms 64 bytes from 3ffe:1d01:2:2:240:5ff:fe19:2f85: icmp6_seq=1 ttl=255 time=0.189 ms ^C --- danx.ipv6.yosemite.ca.us ping statistics --- 2 packets transmitted, 2 packets received, 0% packet loss round-trip min/avg/max = 0.189/0.211/0.234 ms # ping6 dan.ipv6 trying to get source for dan.ipv6 source should be ::207.211.160.111 PING dan.ipv6.yosemite.ca.us (3ffe:1d01:2:1:200:c0ff:feee:f097): 56 data bytes ping: wrote dan.ipv6.yosemite.ca.us 64 chars, ret=-1 ping: wrote dan.ipv6.yosemite.ca.us 64 chars, ret=-1 --- dan.ipv6.yosemite.ca.us ping statistics --- 2 packets transmitted, 0 packets received, 100% packet loss ping: sendto: Network is down ping: sendto: Network is down ^C To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sun May 3 19:22:39 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id TAA21546 for freebsd-net-outgoing; Sun, 3 May 1998 19:22:39 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from apollo.iecs.fcu.edu.tw (apollo.iecs.fcu.edu.tw [140.134.24.110]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id TAA21444 for ; Sun, 3 May 1998 19:21:57 -0700 (PDT) (envelope-from jzlu@apollo.iecs.fcu.edu.tw) Received: from himalayas (himalayas.netlab.fcu.edu.tw) by apollo.iecs.fcu.edu.tw (4.1/SMI-4.1) id AA04040; Mon, 4 May 98 08:42:26 CST Message-Id: <000b01bd76f6$3d82da20$653a868c@himalayas.netlab.fcu.edu.tw> From: "=?big5?B?p2ar2KfT?=" To: Subject: subscribe Date: Mon, 4 May 1998 08:47:33 +0800 Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0006_01BD7739.47AF6480" X-Priority: 3 X-Msmail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.2106.4 X-Mimeole: Produced By Microsoft MimeOLE V4.72.2106.4 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org This is a multi-part message in MIME format. ------=_NextPart_000_0006_01BD7739.47AF6480 Content-Type: text/plain; charset="big5" Content-Transfer-Encoding: quoted-printable subscribe ------=_NextPart_000_0006_01BD7739.47AF6480 Content-Type: text/html; charset="big5" Content-Transfer-Encoding: quoted-printable
subscribe
------=_NextPart_000_0006_01BD7739.47AF6480-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Mon May 4 02:56:47 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id CAA23476 for freebsd-net-outgoing; Mon, 4 May 1998 02:56:47 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from itesec.hsc.fr (root@itesec.hsc.fr [192.70.106.33]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id CAA23446 for ; Mon, 4 May 1998 02:56:39 -0700 (PDT) (envelope-from pb@hsc.fr) Received: from mars.hsc.fr (pb@mars.hsc.fr [192.70.106.44]) by itesec.hsc.fr (8.8.8/8.8.5/itesec-1.12-nospam) with ESMTP id LAA27986; Mon, 4 May 1998 11:56:35 +0200 (MET DST) Received: (from pb@localhost) by mars.hsc.fr (8.8.5/8.8.5/pb-19970301) id LAA03947; Mon, 4 May 1998 11:56:33 +0200 (MET DST) Message-ID: <19980504115633.YX23090@mars.hsc.fr> Date: Mon, 4 May 1998 11:56:33 +0200 From: pb@fasterix.frmug.org (Pierre Beyssac) To: danx@cts.com (Dan Anderson) Cc: freebsd-net@FreeBSD.ORG (FreeBSD-net List) Subject: Re: Help w/IPv6 config for FreeBSD References: X-Mailer: Mutt 0.59.1e Mime-Version: 1.0 In-Reply-To: ; from Dan Anderson on May 3, 1998 17:28:23 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org According to Dan Anderson: > The INRIA instructions are very sketchy, so . . . if you tunneled successfully > to IPv6 with sit tunnel devices, please e-mail me your /etc/rc.ipv6 startup The sit device is for automated host to host tunnels. For configured tunnels, what you want is the cti device. Try this (I don't have my config handy and I generally need to tweak it by hand because my outgoing tunnel interface is on a dialup link, and cticonfig doesn't like that at boot time). This should be placed after autoconf6. Here's the idea. You might need to tweak it because I'm not sure there's nothing missing (I have no way to test this right now). In the following, just replace REMOTE_IPV4_ADDRESS/LOCAL_IPV4_ADDRESS with your tunnel endpoints, and LOCAL_IPV6_ADDRESS with the IPv6 address of your machine (your global scope address, i.e. 3ffe:something). The cticonfig sets the IPv4 tunnel endpoint addresses. The ifconfig firstalias is needed to ensure that your machine puts a correct source address when it emits tunneled packets. Then you add a default route pointing to the IPv4-compatible address of the other endpoint. cticonfig -v -4 -s LOCAL_IPV4_ADDRESS cti0 REMOTE_IPV4_ADDRESS > /tmp/cticonfig 2>&1 ifconfig cti0 inet6 firstalias LOCAL_IPV6_ADDRESS/128 ::REMOTE_IPV4_ADDRESS route -n add -inet6 ::/0 ::REMOTE_IPV4_ADDRESS -- Pierre Beyssac pb@fasterix.frmug.org pb@fasterix.freenix.org {Free,Net,Open}BSD, Linux : il y a moins bien, mais c'est plus cher Free domains: http://www.eu.org/ or mail dns-manager@EU.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 04:02:22 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id EAA18219 for freebsd-net-outgoing; Tue, 5 May 1998 04:02:22 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from avrasya.ispro.net.tr (avrasya.ispro.net.tr [195.174.18.1]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id DAA17314 for ; Tue, 5 May 1998 03:58:25 -0700 (PDT) (envelope-from yurtesen@ispro.net.tr) Received: from localhost (yurtesen@localhost) by avrasya.ispro.net.tr (8.8.6/8.7.3) with SMTP id OAA14863 for ; Tue, 5 May 1998 14:53:23 +0300 Date: Tue, 5 May 1998 14:53:23 +0300 (EET DST) From: Evren Yurtesen To: freebsd-net@FreeBSD.ORG Subject: ethernet Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org when I first installed freebsd I used a PCI RL2000 ethernet adapter and it was able to recognize my ethernet card and install driver for it Jan 1 02:13:47 papua /kernel: ed2 rev 10 int a $ Jan 1 02:13:47 papua /kernel: ed2: address 00:80:48:eb:ac:2b, type NE2000 (16 $ then I recompiled my kernel and, Jan 1 07:44:23 papua /kernel: pci0:18: vendor=0x11f6, device=0x1401, class=network (ethernet) int a irq 11 [no driver assigned] it started to give this error message, and I compiled kernel again with support to a isa ethernet card... Jan 1 07:34:19 papua /kernel: ed2 rev 10 int a $ Jan 1 07:34:19 papua /kernel: ed2: address 00:80:48:eb:ac:2b, type NE2000 (16 $ Jan 1 07:34:19 papua /kernel: vga0 rev 68 int $ Jan 1 07:34:19 papua /kernel: Probing for devices on the ISA bus: Jan 1 07:34:19 papua /kernel: sc0 at 0x60-0x6f irq 1 on motherboard Jan 1 07:34:19 papua /kernel: sc0: VGA color <16 virtual consoles, flags=0x0> Jan 1 07:34:19 papua /kernel: ed0 not found at 0x300 well it found the pci ethernet card again...but when I take out the line device ed0 at isa? port 0x300 net irq 9 iomem 0xd8000 vector edintr from kernel configuration it does not assign a driver for the ethernet card... why is this happening? +--------------------------------------------------------+ | Name : Evren Yurtesen - yurtesen@ispro.net.tr | | S-mail: Mithatpasa Cad. No:1079/13 35290 Guzelyali | | Home:+90-232-2857604 Work:+90-232-2463992 Izmir/TURKEY | +--------------------------------------------------------+ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 04:34:33 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id EAA23155 for freebsd-net-outgoing; Tue, 5 May 1998 04:34:33 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from avrasya.ispro.net.tr (avrasya.ispro.net.tr [195.174.18.1]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id EAA23150 for ; Tue, 5 May 1998 04:34:21 -0700 (PDT) (envelope-from yurtesen@ispro.net.tr) Received: from localhost (yurtesen@localhost) by avrasya.ispro.net.tr (8.8.6/8.7.3) with SMTP id PAA15525 for ; Tue, 5 May 1998 15:29:24 +0300 Date: Tue, 5 May 1998 15:29:24 +0300 (EET DST) From: Evren Yurtesen To: freebsd-net@FreeBSD.ORG Subject: ethernet Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org hello when I first installed freebsd I used a PCI RL2000 ethernet adapter and it was able to recognize my ethernet card and install driver for it Jan 1 02:13:47 papua /kernel: ed2 rev 10 int a $ Jan 1 02:13:47 papua /kernel: ed2: address 00:80:48:eb:ac:2b, type NE2000 (16 $ then I recompiled my kernel and, Jan 1 07:44:23 papua /kernel: pci0:18: vendor=0x11f6, device=0x1401, class=network (ethernet) int a irq 11 [no driver assigned] it started to give this error message, and I compiled kernel again with support to a isa ethernet card... Jan 1 07:34:19 papua /kernel: ed2 rev 10 int a $ Jan 1 07:34:19 papua /kernel: ed2: address 00:80:48:eb:ac:2b, type NE2000 (16 $ Jan 1 07:34:19 papua /kernel: vga0 rev 68 int $ Jan 1 07:34:19 papua /kernel: Probing for devices on the ISA bus: Jan 1 07:34:19 papua /kernel: sc0 at 0x60-0x6f irq 1 on motherboard Jan 1 07:34:19 papua /kernel: sc0: VGA color <16 virtual consoles, flags=0x0> Jan 1 07:34:19 papua /kernel: ed0 not found at 0x300 well it found the pci ethernet card again...but when I take out the line device ed0 at isa? port 0x300 net irq 9 iomem 0xd8000 vector edintr from kernel configuration it does not assign a driver for the ethernet card... why is this happening? +--------------------------------------------------------+ | Name : Evren Yurtesen - yurtesen@ispro.net.tr | | S-mail: Mithatpasa Cad. No:1079/13 35290 Guzelyali | | Home:+90-232-2857604 Work:+90-232-2463992 Izmir/TURKEY | +--------------------------------------------------------+ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 07:35:19 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA24873 for freebsd-net-outgoing; Tue, 5 May 1998 07:35:19 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from jingoro.prevmed.sunysb.edu (jingoro.prevmed.sunysb.edu [129.49.123.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA24855 for ; Tue, 5 May 1998 07:35:04 -0700 (PDT) (envelope-from cfanning@jingoro.prevmed.sunysb.edu) Received: (from cfanning@localhost) by jingoro.prevmed.sunysb.edu (8.8.8/8.8.5) id KAA01507 for freebsd-net@freebsd.org; Tue, 5 May 1998 10:26:08 GMT From: Chris Fanning Message-Id: <199805051026.KAA01507@jingoro.prevmed.sunysb.edu> Subject: sudden MBUF problem To: freebsd-net@FreeBSD.ORG Date: Tue, 5 May 1998 10:26:08 +0000 (GMT) X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org All was fine and good with FreeBSD's networking (for years!) until this last weekend. I first noticed the problem dialed in - frequent disconnection, packets would get delayed for a crazy amount of time, etc. In fact, it's happening right now, but only a lot less often. It's not easy diagnosing a problem remotely when you get kicked out every minute! I found this in my syslog: May 5 09:54:04 jingoro /kernel: Out of mbuf clusters - increase maxusers! So I searched for "mbuf clusters" and found a kernel setting to increase them. All was fine with the standard settings until this weekend, so I was a bit puzzled. Nothing changed on the machine and it gets very light loading. The output of my "netstat -m" is: 4704 mbufs in use: 4434 mbufs allocated to data 4 mbufs allocated to packet headers 9 mbufs allocated to protocol control blocks 253 mbufs allocated to fragment reassembly queue headers 4 mbufs allocated to socket names and addresses 2248/2262 mbuf clusters in use 5112 Kbytes allocated to network (99% in use) 0 requests for memory denied 15 requests for memory delayed 15 calls to protocol drain routines And my uptime right now is less than 20 minutes. I'm guessing something's going on where I'm getting a crazy amount of mbufs being allocated to data for no good reason. I was running 2.2-STABLE and upgraded to 2.2.6-STABLE with the same results. Could I be the victim of some degenerative TCP/IP attack? I ran Ethload on my network and found no packets being specifically destined for my MAC... After a protocol drain, the mbuf count in use climbs a few hundred per second until it gets to the top where it stays for a while. Anyone have any ideas on this? Or what I can do to determine what's causing the problem? Thanks, Chris To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 07:48:59 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA26740 for freebsd-net-outgoing; Tue, 5 May 1998 07:48:59 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from softweyr.com ([204.68.178.33]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA26727 for ; Tue, 5 May 1998 07:48:50 -0700 (PDT) (envelope-from wes@softweyr.com) Received: from softweyr.com (localhost.softweyr.com [127.0.0.1]) by softweyr.com (8.8.8/8.8.7) with ESMTP id IAA27951; Tue, 5 May 1998 08:56:43 -0600 (MDT) (envelope-from wes@softweyr.com) Message-ID: <354F28AB.4E57CCB0@softweyr.com> Date: Tue, 05 May 1998 08:56:43 -0600 From: Wes Peters Organization: Softweyr llc X-Mailer: Mozilla 4.04 [en] (X11; I; FreeBSD 2.2.6-RELEASE i386) MIME-Version: 1.0 To: Evren Yurtesen , freebsd-net@FreeBSD.ORG Subject: Re: ethernet References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Evren Yurtesen wrote: > well it found the pci ethernet card again...but when I take out the line > > device ed0 at isa? port 0x300 net irq 9 iomem 0xd8000 vector edintr > > from kernel configuration it does not assign a driver for the ethernet > card... > why is this happening? The PCI NE2000 is handled as a special case of the NE2000 driver. If you don't include the ed0 line, the ed driver isn't compiled into the kernel, so it cannot possible work without it. -- "Where am I, and what am I doing in this handbasket?" Wes Peters Softweyr LLC http://www.softweyr.com/~softweyr wes@softweyr.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 08:30:48 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA04677 for freebsd-net-outgoing; Tue, 5 May 1998 08:30:48 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from implode.root.com (implode.root.com [198.145.90.17]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA04513 for ; Tue, 5 May 1998 08:30:01 -0700 (PDT) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.5/8.8.5) with ESMTP id IAA29909; Tue, 5 May 1998 08:28:45 -0700 (PDT) Message-Id: <199805051528.IAA29909@implode.root.com> To: Chris Fanning cc: freebsd-net@FreeBSD.ORG Subject: Re: sudden MBUF problem In-reply-to: Your message of "Tue, 05 May 1998 10:26:08 -0000." <199805051026.KAA01507@jingoro.prevmed.sunysb.edu> From: David Greenman Reply-To: dg@root.com Date: Tue, 05 May 1998 08:28:44 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >I'm guessing something's going on where I'm getting a crazy amount of mbufs >being allocated to data for no good reason. Probably to new connections. I'd recommend looking at "netstat -n", and "netstat -s". >Could I be the victim of some degenerative TCP/IP attack? Probably. Or some local application that is going crazy. -DG David Greenman Co-founder/Principal Architect, The FreeBSD Project To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 08:36:32 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA06396 for freebsd-net-outgoing; Tue, 5 May 1998 08:36:32 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from lestat.nas.nasa.gov (lestat.nas.nasa.gov [129.99.50.29]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA06384 for ; Tue, 5 May 1998 08:36:25 -0700 (PDT) (envelope-from thorpej@lestat.nas.nasa.gov) Received: from localhost (localhost [127.0.0.1]) by lestat.nas.nasa.gov (8.8.8/8.6.12) with SMTP id IAA22952; Tue, 5 May 1998 08:26:06 -0700 (PDT) Message-Id: <199805051526.IAA22952@lestat.nas.nasa.gov> X-Authentication-Warning: lestat.nas.nasa.gov: localhost [127.0.0.1] didn't use HELO protocol To: Evren Yurtesen Cc: freebsd-net@FreeBSD.ORG Subject: Re: ethernet Reply-To: Jason Thorpe From: Jason Thorpe Date: Tue, 05 May 1998 08:26:05 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 5 May 1998 14:53:23 +0300 (EET DST) Evren Yurtesen wrote: > device ed0 at isa? port 0x300 net irq 9 iomem 0xd8000 vector edintr > > from kernel configuration it does not assign a driver for the ethernet > card... > why is this happening? Because FreeBSD doesn't have a flexible autoconfiguration system, so they require hacks that glue drivers for PCI devices which are similar to ISA devices to the ISA drivers themselves. We addressed this in NetBSD ... quite some time ago. Jason R. Thorpe thorpej@nas.nasa.gov NASA Ames Research Center Home: +1 408 866 1912 NAS: M/S 258-5 Work: +1 650 604 0935 Moffett Field, CA 94035 Pager: +1 415 428 6939 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 11:43:06 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA12013 for freebsd-net-outgoing; Tue, 5 May 1998 11:43:06 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from colin.muc.de (root@colin.muc.de [193.174.4.1]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id LAA11859 for ; Tue, 5 May 1998 11:42:16 -0700 (PDT) (envelope-from lutz@muc.de) Received: from tavari.muc.de ([193.174.4.22]) by colin.muc.de with SMTP id <140562-1>; Tue, 5 May 1998 20:42:06 +0200 Received: from muc.de (abraxas [192.168.42.5]) by tavari.muc.de (8.8.7/8.8.7) with ESMTP id UAA12937; Tue, 5 May 1998 20:42:31 +0200 (CEST) Message-ID: <354F5D95.629EF2D@muc.de> Date: Tue, 5 May 1998 20:42:29 +0200 From: Lutz Albers X-Mailer: Mozilla 4.04 [en] (X11; U; FreeBSD 2.2.6-RELEASE i386) MIME-Version: 1.0 To: Evren Yurtesen CC: freebsd-net@FreeBSD.ORG Subject: Re: ethernet References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Evren Yurtesen wrote: > > when I first installed freebsd I used a PCI RL2000 ethernet adapter > and it was able to recognize my ethernet card and install driver for it > ... > well it found the pci ethernet card again...but when I take out the line > > device ed0 at isa? port 0x300 net irq 9 iomem 0xd8000 vector edintr > > from kernel configuration it does not assign a driver for the ethernet > card... just add a device ed0 line and it will find your card again (as ed1) ciao lutz -- Lutz Albers, lutz@muc.de Do not take life too seriously, you will never get out of it alive. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 11:56:18 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA14430 for freebsd-net-outgoing; Tue, 5 May 1998 11:56:18 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from spring.edu.tw (root@Spring.edu.tw [140.111.1.29]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA14168 for ; Tue, 5 May 1998 11:54:08 -0700 (PDT) (envelope-from u8534523@cc.nctu.edu.tw) Received: from cc.NCTU.edu.tw (root@ccserv2.cc.nctu.edu.tw [140.113.6.2]) by spring.edu.tw (8.8.7/8.8.7) with ESMTP id BAA27328 for ; Wed, 6 May 1998 01:56:57 +0800 (CST) Received: from cc.nctu.edu.tw (Hsieh.GradDorm2.NCTU.edu.tw [140.113.129.5]) by cc.NCTU.edu.tw (8.8.7/8.8.7) with ESMTP id BAA06417 for ; Wed, 6 May 1998 01:59:41 +0800 (CST) Message-ID: <354F5436.8F7FAD56@cc.nctu.edu.tw> Date: Wed, 06 May 1998 02:02:30 +0800 From: Administrator Organization: 交大資管所 X-Mailer: Mozilla 4.05 [en] (Win95; I) MIME-Version: 1.0 To: freebsd-net@FreeBSD.ORG Subject: Is there any program provide functions like the 'lanstat' or 'trafshow' ? Content-Type: text/plain; charset=big5 Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I use the lanstat and trafshow, but they just for 'interactive' interface. I want to make a log. In lanstat, it can't make a log. In the manual of lanstat, it says that there is a program called 'lanstatd' which can make log, but I can't find the lanstatd. In trafshow, it use the ANSI color, I can't redirect it to file... My problem is: Is there any program can do the same thing as the lanstat or trafshow ? Regards. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 13:02:10 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id NAA25518 for freebsd-net-outgoing; Tue, 5 May 1998 13:02:10 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from shrimp.dataplex.net (shrimp.dataplex.net [208.2.87.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id NAA25316 for ; Tue, 5 May 1998 13:01:59 -0700 (PDT) (envelope-from rkw@dataplex.net) Received: from [208.2.87.7] (user10.dataplex.net [208.2.87.10]) by shrimp.dataplex.net (8.8.8/8.8.5) with ESMTP id PAA09176 for ; Tue, 5 May 1998 15:01:25 -0500 (CDT) X-Sender: rkw@mail.dataplex.net Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Tue, 5 May 1998 15:00:21 -0500 To: freebsd-net@FreeBSD.ORG From: Richard Wackerbarth Subject: Have I left something out? Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I set up a test network and it did not behave as I had hoped/expected. I wanted to have redundant connections so that if one goes down, another can be used. Consider the simple case of a loop. For discussion, we will have 6 nets. 10.0.1.0/24, 10.0.2.0/24, ..., 10.0.6.0/24. We will join them with routers 10.0.1.1 == 10.0.2.1 10.0.2.2 == 10.0.3.2 ... 10.0.5.5 == 10.0.6.5 and to close the loop 10.0.6.6 == 10.0.1.6 Now we bring them all up and let them exchange routing info. All is well. A packet from net 1 ==> net 3 goes via net 2. A packet from net 1 ==> net 5 goes via net 6. Now we break the loop by declaring net 6 (or one of the router interfaces thereon) down. The routes reconfigure and the packet to net 5 now takes the longer (but still available) route via nets 2,3,4. However, and this is my problem, if I declared the 10.0.1.6 interface down, I cannot reach the machine. Host 10.0.1.1 INSISTS that the packet go out the net 1 interface and cannot reach 10.0.1.6. It never finds the "backdoor", 10.0.6.6. Any suggestions? Richard Wackerbarth To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 13:52:29 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id NAA05046 for freebsd-net-outgoing; Tue, 5 May 1998 13:52:29 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from dirty.research.bell-labs.com (dirty.research.bell-labs.com [204.178.16.6]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id NAA04924 for ; Tue, 5 May 1998 13:52:08 -0700 (PDT) (envelope-from tnaik@dnrc.bell-labs.com) Received: from chair.dnrc.bell-labs.com ([135.180.161.201]) by dirty; Tue May 5 16:50:22 EDT 1998 Received: from dnrc.bell-labs.com ([135.17.249.50]) by chair.dnrc.bell-labs.com (8.8.8/8.8.8) with ESMTP id QAA04027 for ; Tue, 5 May 1998 16:50:17 -0400 (EDT) Message-ID: <354F7B83.B0B4DC04@dnrc.bell-labs.com> Date: Tue, 05 May 1998 16:50:11 -0400 From: Tejas Naik X-Mailer: Mozilla 4.04 [en] (Win95; I) MIME-Version: 1.0 To: FreeBSDTCP/IP Subject: Broadcast on Multihomed Router Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org FreeBSD kernel does not broadcast the (IP)broadcast packet to all interfaces attached to multihomed machine. It broadcasts that packet only on default interface. What is the hack if I want network wide broadcast? Thanks Tejas To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 15:23:34 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA25337 for freebsd-net-outgoing; Tue, 5 May 1998 15:23:34 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from mail.vera.net (enya.vera.net [200.33.116.58]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA25229 for ; Tue, 5 May 1998 15:23:06 -0700 (PDT) (envelope-from mdragon@vera.net) Received: from cache.reco (mail.vera.net [200.33.116.55]) by mail.vera.net (8.8.8/8.8.7) with SMTP id RAA07189; Tue, 5 May 1998 17:22:38 -0500 (CDT) (envelope-from mdragon@vera.net) Date: Tue, 5 May 1998 17:22:38 -0500 (CDT) From: "Victor M. Mondragon A." X-Sender: mdragon@cache.reco To: Richard Wackerbarth cc: freebsd-net@FreeBSD.ORG Subject: Re: Have I left something out? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Tue, 5 May 1998, Richard Wackerbarth wrote: [...munch...] > I wanted to have redundant connections so that if one goes down, another > can be used. [...munch...] Try routed(8) or gated(1), these take care of dynamic routing. --------------------------------------------------------------------------- Victor Manuel Mondragon mdragon@vera.net --------------------------------------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 15:55:16 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA01475 for freebsd-net-outgoing; Tue, 5 May 1998 15:55:16 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from shrimp.dataplex.net (shrimp.dataplex.net [208.2.87.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA01460 for ; Tue, 5 May 1998 15:55:01 -0700 (PDT) (envelope-from rkw@dataplex.net) Received: from [208.2.87.7] (user10.dataplex.net [208.2.87.10]) by shrimp.dataplex.net (8.8.8/8.8.5) with ESMTP id RAA09912; Tue, 5 May 1998 17:54:36 -0500 (CDT) X-Sender: rkw@mail.dataplex.net Message-Id: In-Reply-To: References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Tue, 5 May 1998 17:54:46 -0500 To: "Victor M. Mondragon A." From: Richard Wackerbarth Subject: Re: Have I left something out? Cc: freebsd-net@FreeBSD.ORG Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 5:22 PM -0500 5/5/98, Victor M. Mondragon A. wrote: >On Tue, 5 May 1998, Richard Wackerbarth wrote: >[...munch...] > >> I wanted to have redundant connections so that if one goes down, another >> can be used. > >[...munch...] > >Try routed(8) or gated(1), these take care of dynamic routing. I am already doing that. That is what changes the tables. However, it does not seem to find the back door. Richard Wackerbarth To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 18:46:14 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id SAA29834 for freebsd-net-outgoing; Tue, 5 May 1998 18:46:14 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from socrates.i-pi.com (socrates.i-pi.com [198.49.217.5]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA29725 for ; Tue, 5 May 1998 18:46:02 -0700 (PDT) (envelope-from ingham@i-pi.com) Received: (from ingham@localhost) by socrates.i-pi.com (8.8.8/8.8.7) id TAA12573; Tue, 5 May 1998 19:44:03 -0600 (MDT) (envelope-from ingham) Message-ID: <19980505194403.22825@i-pi.com> Date: Tue, 5 May 1998 19:44:03 -0600 From: Kenneth Ingham To: Administrator , freebsd-net@FreeBSD.ORG Subject: Re: Is there any program provide functions like the 'lanstat' or 'trafshow' ? References: <354F5436.8F7FAD56@cc.nctu.edu.tw> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.89i In-Reply-To: <354F5436.8F7FAD56@cc.nctu.edu.tw>; from Administrator on Wed, May 06, 1998 at 02:02:30AM +0800 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > My problem is: Is there any program can do the same thing as > the lanstat or trafshow ? Check out tcpdump(1) Kenneth To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Tue May 5 22:41:01 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA11547 for freebsd-net-outgoing; Tue, 5 May 1998 22:41:01 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from gras-varg.worldgate.com (root@gras-varg.worldgate.com [198.161.84.12]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA11458 for ; Tue, 5 May 1998 22:40:34 -0700 (PDT) (envelope-from skafte@worldgate.com) Received: (from skafte@localhost) by gras-varg.worldgate.com (8.8.8/8.6.12) id XAA21382; Tue, 5 May 1998 23:26:52 -0600 (MDT) Message-ID: <19980505232651.54171@worldgate.com> Date: Tue, 5 May 1998 23:26:51 -0600 From: Greg Skafte To: Richard Wackerbarth Cc: "Victor M. Mondragon A." , freebsd-net@FreeBSD.ORG Subject: Re: Have I left something out? References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88 In-Reply-To: ; from Richard Wackerbarth on Tue, May 05, 1998 at 05:54:46PM -0500 Organization: WorldGate Inc. X-PGP-Fingerprint: 42 9C 2C A8 4D 2B C9 C4 7D B6 00 B0 50 47 20 97 X-URL: http://gras-varg.worldgate.com/~skafte Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org the default BSD kernels do not accept multiple routes to the same desitation... on hackers a few months back someone talked about a multiple paths kernel hack .... in turn you would then have to modify gated or routed to accept and process multipath routes .... as a side note what dynamic routing protocol are you using to attempt to do this .... Quoting Richard Wackerbarth (rkw@dataplex.net) On Subject: Re: Have I left something out? Date: Tue, May 05, 1998 at 05:54:46PM -0500 > At 5:22 PM -0500 5/5/98, Victor M. Mondragon A. wrote: > >On Tue, 5 May 1998, Richard Wackerbarth wrote: > >[...munch...] > > > >> I wanted to have redundant connections so that if one goes down, another > >> can be used. > > > >[...munch...] > > > >Try routed(8) or gated(1), these take care of dynamic routing. > > I am already doing that. That is what changes the tables. > However, it does not seem to find the back door. > > Richard Wackerbarth > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-net" in the body of the message -- Email: skafte@worldgate.com Voice: +403 413 1910 Fax: +403 421 4929 #575 Sun Life Place * 10123 99 Street * Edmonton, AB * Canada * T5J 3H1 -- -- When things can't get any worse, they simplify themselves by getting a whole lot worse then complicated. A complete and utter disaster is the simplest thing in the world; it's preventing one that's complex. (Janet Morris) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Wed May 6 04:28:42 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id EAA11326 for freebsd-net-outgoing; Wed, 6 May 1998 04:28:42 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from shrimp.dataplex.net (shrimp.dataplex.net [208.2.87.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id EAA11321 for ; Wed, 6 May 1998 04:28:38 -0700 (PDT) (envelope-from rkw@dataplex.net) Received: from [208.2.87.10] (user10.dataplex.net [208.2.87.10]) by shrimp.dataplex.net (8.8.8/8.8.5) with ESMTP id GAA21182; Wed, 6 May 1998 06:28:01 -0500 (CDT) X-Sender: rkw@mail.dataplex.net Message-Id: In-Reply-To: <19980505232651.54171@worldgate.com> References: ; from Richard Wackerbarth on Tue, May 05, 1998 at 05:54:46PM -0500 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Wed, 6 May 1998 06:26:43 -0500 To: Greg Skafte From: Richard Wackerbarth Subject: Re: Have I left something out? Cc: "Victor M. Mondragon A." , freebsd-net@FreeBSD.ORG Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 12:26 AM -0500 5/6/98, Greg Skafte wrote: >the default BSD kernels do not accept multiple routes to the same >desitation... >on hackers a few months back someone talked about a multiple paths kernel >hack .... in turn you would then have to modify gated or routed to accept and >process multipath routes .... > > >as a side note what dynamic routing protocol are you using to attempt to do >this .... > >Quoting Richard Wackerbarth (rkw@dataplex.net) >On Subject: Re: Have I left something out? >Date: Tue, May 05, 1998 at 05:54:46PM -0500 > >> At 5:22 PM -0500 5/5/98, Victor M. Mondragon A. wrote: >> >On Tue, 5 May 1998, Richard Wackerbarth wrote: >> >[...munch...] >> > >> >> I wanted to have redundant connections so that if one goes down, another >> >> can be used. >> > >> >[...munch...] >> > >> >Try routed(8) or gated(1), these take care of dynamic routing. >> >> I am already doing that. That is what changes the tables. >> However, it does not seem to find the back door. Thanks for the reply. Since I am not really looking for multi-path routing, the simple route discovery SHOULD work OK. In fact, it does on all machines other than the one where the down network is attached. When the once removed machine quits getting the route from the attached machine, it quits broadcasting it. As a result, the route decays and gets replaced with the alternate route that is still active. However on the machine which has the point of failure, even though the interface is marked "down", it still insists that that is THE route to any machines which would be on that net. I think that I will try moving an "essential service" daemon onto an alias in the lo0 interface and see if a route to it gets spread properly. Richard Wackerbarth To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Wed May 6 11:11:38 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA13306 for freebsd-net-outgoing; Wed, 6 May 1998 11:11:38 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA13232 for ; Wed, 6 May 1998 11:11:27 -0700 (PDT) (envelope-from gibbs@narnia.plutotech.com) Received: (from gibbs@localhost) by narnia.plutotech.com (8.8.8/8.7.3) id MAA05103; Wed, 6 May 1998 12:07:28 -0600 (MDT) Date: Wed, 6 May 1998 12:07:28 -0600 (MDT) From: "Justin T. Gibbs" Message-Id: <199805061807.MAA05103@narnia.plutotech.com> To: Jason Thorpe cc: net@FreeBSD.ORG Subject: Re: ethernet Newsgroups: pluto.freebsd.net In-Reply-To: <199805051526.IAA22952@lestat.nas.nasa.gov> User-Agent: tin/pre-1.4-971204 (UNIX) (FreeBSD/3.0-CURRENT (i386)) Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > device ed0 at isa? port 0x300 net irq 9 iomem 0xd8000 vector edintr > > > > from kernel configuration it does not assign a driver for the ethernet > > card... > > why is this happening? > > Because FreeBSD doesn't have a flexible autoconfiguration system, so they > require hacks that glue drivers for PCI devices which are similar to ISA > devices to the ISA drivers themselves. > > We addressed this in NetBSD ... quite some time ago. Although I will agree that FreeBSD's configuration system sucks rocks, your analysis here is incorrect. You can, with the exception of some trickery with unit number handling, cleanly create PCI/EISA/ISA attach stubs for a "generic" driver. All of the CAM SCSI drivers do this, but there are certainly "older drivers" that could use cleanup in this area. I'm still waiting to see NetBSD cleanly deal with LKMs in it's autoconfiguration code. Config.new is not the answer to all configuration problems, although I will admit that it is much better than FreeBSD's current config nightmare. > Jason R. Thorpe thorpej@nas.nasa.gov > NASA Ames Research Center Home: +1 408 866 1912 > NAS: M/S 258-5 Work: +1 650 604 0935 > Moffett Field, CA 94035 Pager: +1 415 428 6939 -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Wed May 6 12:18:33 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id MAA26905 for freebsd-net-outgoing; Wed, 6 May 1998 12:18:33 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from lestat.nas.nasa.gov (lestat.nas.nasa.gov [129.99.50.29]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA26889 for ; Wed, 6 May 1998 12:18:24 -0700 (PDT) (envelope-from thorpej@lestat.nas.nasa.gov) Received: from localhost (localhost [127.0.0.1]) by lestat.nas.nasa.gov (8.8.8/8.6.12) with SMTP id MAA07625; Wed, 6 May 1998 12:08:20 -0700 (PDT) Message-Id: <199805061908.MAA07625@lestat.nas.nasa.gov> X-Authentication-Warning: lestat.nas.nasa.gov: localhost [127.0.0.1] didn't use HELO protocol To: "Justin T. Gibbs" Cc: freebsd-net@FreeBSD.ORG Subject: Re: ethernet Reply-To: Jason Thorpe From: Jason Thorpe Date: Wed, 06 May 1998 12:08:19 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Wed, 6 May 1998 12:07:28 -0600 (MDT) "Justin T. Gibbs" wrote: > I'm still waiting to see NetBSD cleanly deal with LKMs in it's > autoconfiguration code. Config.new is not the answer to all > configuration problems, although I will admit that it is much > better than FreeBSD's current config nightmare. Dealing with LKMs really only requires defining a way of specifying the locator arguments. Then, the autoconfiguration code locates the parent node, and attempts to match/probe the device given the new driver/locators. The reason THAT hasn't been done is because the LKM interface sucks rocks (and it's basically the same as FreeBSD's), but that's been known for a while. Jason R. Thorpe thorpej@nas.nasa.gov NASA Ames Research Center Home: +1 408 866 1912 NAS: M/S 258-5 Work: +1 650 604 0935 Moffett Field, CA 94035 Pager: +1 650 428 6939 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Wed May 6 12:32:32 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id MAA00353 for freebsd-net-outgoing; Wed, 6 May 1998 12:32:32 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from pluto.plutotech.com (mail.plutotech.com [206.168.67.137]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA00308 for ; Wed, 6 May 1998 12:32:16 -0700 (PDT) (envelope-from gibbs@plutotech.com) Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by pluto.plutotech.com (8.8.7/8.8.5) with ESMTP id NAA24276; Wed, 6 May 1998 13:32:15 -0600 (MDT) Message-Id: <199805061932.NAA24276@pluto.plutotech.com> X-Mailer: exmh version 2.0.1 12/23/97 To: Jason Thorpe cc: "Justin T. Gibbs" , freebsd-net@FreeBSD.ORG Subject: Re: ethernet In-reply-to: Your message of "Wed, 06 May 1998 12:08:19 PDT." <199805061908.MAA07625@lestat.nas.nasa.gov> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 06 May 1998 13:28:28 -0600 From: "Justin T. Gibbs" Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >Dealing with LKMs really only requires defining a way of specifying >the locator arguments. Then, the autoconfiguration code locates the >parent node, and attempts to match/probe the device given the new >driver/locators. No, you have to go beyond this. If I attempt to load an EISA driver, it may need to load the EISA bus driver, etc. This means that further dependency information and the ability to dynamically add and remove locators is required. You want to be able to support devices and their locators going away too. The transition from statically compiled locaters to a mixture of static and dynamically registered ones, along with the requisite reference counting and callbacks for when nodes disappear seems a somewhat non-trivial addition to how config.new works now. >The reason THAT hasn't been done is because the LKM interface sucks >rocks (and it's basically the same as FreeBSD's), but that's been known >for a while. LKMs don't just suck rocks in *BSD, they are rocks. 8-) >Jason R. Thorpe thorpej@nas.nasa.gov >NASA Ames Research Center Home: +1 408 866 1912 >NAS: M/S 258-5 Work: +1 650 604 0935 >Moffett Field, CA 94035 Pager: +1 650 428 6939 -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Wed May 6 17:40:20 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id RAA28331 for freebsd-net-outgoing; Wed, 6 May 1998 17:40:20 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from thuule.pair.com (thuule.pair.com [209.68.1.107]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id RAA28171; Wed, 6 May 1998 17:39:42 -0700 (PDT) (envelope-from asr@millburn.net) Received: from localhost (asr@localhost) by thuule.pair.com (8.8.8/8.6.12) with SMTP id UAA29041; Wed, 6 May 1998 20:39:39 -0400 (EDT) X-Envelope-To: freebsd-security@freebsd.org Date: Wed, 6 May 1998 20:39:39 -0400 (EDT) From: Adam Rothschild X-Sender: asr@thuule.pair.com To: freebsd-net@FreeBSD.ORG, freebsd-security@FreeBSD.ORG Subject: NAT woes! Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hey everyone... I'm trying to setup the following, and I'm failing miserably. Internal network -[ed0]- FreeBSD (2.2.5) box -[ex0]- public IP 192.168.0.0/24 1.2.3.4 - FreeBSD box blocks all incoming connections, except on TCP port 25 Here's what I have. Note that the NAT for internal machines accessing the 'net works nicely. However, I can't figure out how to allow tcp port 25 into the box on it's public interface [ex0]. o FreeBSD box is assigned 192.168.0.1 on the ed0 interface, and 1.2.3.4 on the ex0 interface. o The following script is executed to initiate things: natd -log -redirect_address 192.168.0.0 0.0.0.0 -n ex0 -u ipfw -f flush ipfw add divert natd all from any to any ipfw add pass all from any to any Any help would be GREATLY appreciated!!! Thanks, Adam To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Wed May 6 21:11:40 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id VAA03209 for freebsd-net-outgoing; Wed, 6 May 1998 21:11:40 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from roma.coe.ufrj.br (jonny@roma.coe.ufrj.br [146.164.53.65]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id VAA03150 for ; Wed, 6 May 1998 21:11:19 -0700 (PDT) (envelope-from jonny@coe.ufrj.br) Received: (from jonny@localhost) by roma.coe.ufrj.br (8.8.8/8.8.8) id BAA05587; Thu, 7 May 1998 01:10:27 -0300 (EST) (envelope-from jonny) From: Joao Carlos Mendes Luis Message-Id: <199805070410.BAA05587@roma.coe.ufrj.br> Subject: Re: Have I left something out? In-Reply-To: from Richard Wackerbarth at "May 6, 98 06:26:43 am" To: rkw@dataplex.net (Richard Wackerbarth) Date: Thu, 7 May 1998 01:10:27 -0300 (EST) Cc: skafte@worldgate.com, mdragon@vera.net, freebsd-net@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL40 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org #define quoting(Richard Wackerbarth) // However on the machine which has the point of failure, even though the // interface is marked "down", it still insists that that is THE route to // any machines which would be on that net. Do you mean in the same subnetwork ? Local routes always have preference over RIP routes. And marking an interface down will not remove the local route through it. For networks other than local, I think that there's no problem. // I think that I will try moving an "essential service" daemon onto an alias in // the lo0 interface and see if a route to it gets spread properly. This is how I do things here. My IP aliases are in the ethernet interface, but in the loopback interface. If they're in the same subnet, export with arp published. If they're in different subnets, export with gated. For your case, the second approach is probably better. In the end, since there's no local address, RIP will always be used, and destination router discovery will succeed. There's a penalty: You lose one full subnet with each machine with "essential services". Maybe using subnet variant routing protocols like OSPF could help, but I don't know them enough to talk about. Jonny -- Joao Carlos Mendes Luis jonny@gta.ufrj.br +55 21 290-4698 ( Job ) jonny@coppe.ufrj.br M.Sc. Student Electrical Engineering Universidade Federal do Rio de Janeiro To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 00:08:53 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id AAA29168 for freebsd-net-outgoing; Thu, 7 May 1998 00:08:53 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from alpo.whistle.com (alpo.whistle.com [207.76.204.38]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id AAA29057 for ; Thu, 7 May 1998 00:07:49 -0700 (PDT) (envelope-from julian@whistle.com) Received: (from daemon@localhost) by alpo.whistle.com (8.8.5/8.8.5) id AAA15837; Thu, 7 May 1998 00:02:36 -0700 (PDT) Received: from current1.whistle.com(207.76.205.22) via SMTP by alpo.whistle.com, id smtpd015835; Thu May 7 07:02:36 1998 Date: Thu, 7 May 1998 00:02:32 -0700 (PDT) From: Julian Elischer To: Joao Carlos Mendes Luis cc: Richard Wackerbarth , skafte@worldgate.com, mdragon@vera.net, freebsd-net@FreeBSD.ORG Subject: Re: Have I left something out? In-Reply-To: <199805070410.BAA05587@roma.coe.ufrj.br> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, 7 May 1998, Joao Carlos Mendes Luis wrote: > #define quoting(Richard Wackerbarth) > // However on the machine which has the point of failure, even though the > // interface is marked "down", it still insists that that is THE route to > // any machines which would be on that net. This is a known bug and there are fixes for it floating around but none are really 'correct' julian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 04:44:56 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id EAA02124 for freebsd-net-outgoing; Thu, 7 May 1998 04:44:56 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from ns1.sby.ywcn.or.id ([167.205.169.3]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id EAA02114 for ; Thu, 7 May 1998 04:44:47 -0700 (PDT) (envelope-from roy@its-sby.edu) Received: from its-sby.edu [167.205.169.82] by ns1.sby.ywcn.or.id with ESMTP (SMTPD32-4.02) id A2F435D00F8; Thu, 07 May 1998 18:42:12 GMT+7 Received: from ns2.its-sby.edu (ns2.its-sby.edu [167.205.169.82]) by its-sby.edu (8.8.4/8.8.4) with SMTP id SAA15789 for ; Thu, 7 May 1998 18:46:52 -0700 Date: Thu, 7 May 1998 18:46:52 -0700 (GMT+7) From: Royyana Muslim Ijtihadie To: freebsd-net@FreeBSD.ORG Subject: Re: [SYSOP-L] NAT woes! (fwd) In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, 7 May 1998, Dali Ahmad wrote: > On Thu, 7 May 1998, Royyana Muslim Ijtihadie wrote: > > > > > > > ---------- Forwarded message ---------- > > Date: Wed, 6 May 1998 20:39:39 -0400 (EDT) > > From: Adam Rothschild > > To: freebsd-net@FreeBSD.ORG, freebsd-security@FreeBSD.ORG > > Subject: NAT woes! > > > > Hey everyone... > > > > I'm trying to setup the following, and I'm failing miserably. > > > > Internal network -[ed0]- FreeBSD (2.2.5) box -[ex0]- public IP > > 192.168.0.0/24 1.2.3.4 > > > > - FreeBSD box blocks all incoming connections, except on TCP port 25 > > > > Here's what I have. Note that the NAT for internal machines accessing the > > 'net works nicely. However, I can't figure out how to allow tcp port 25 > > into the box on it's public interface [ex0]. > > > > o FreeBSD box is assigned 192.168.0.1 on the ed0 interface, and 1.2.3.4 on > > the ex0 interface. > > o The following script is executed to initiate things: > > > > natd -log -redirect_address 192.168.0.0 0.0.0.0 -n ex0 -u > > ipfw -f flush > > ipfw add divert natd all from any to any > > ipfw add pass all from any to any > > how about using this setting : > > > natd -interface ex0 > /sbin/ipfw add 32768 divert natd all from any to any via ex0 > > it works just fine on my box. > > ------------------------------------------- > Dali Ahmad > GIS & Remote Sensing Project Coordinator > Volcanological Survey of Indonesia > Jalan Diponegoro 57 Bandung 40122 > Tel. +62 22 774 706 > Fax. +62 22 702 761 > E-mail dali@vsi.dpe.go.id > daliahmad@ibm.net > ------------------------------------------ > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 05:08:50 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id FAA04549 for freebsd-net-outgoing; Thu, 7 May 1998 05:08:50 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from osku.suutari.iki.fi (kn6-045.ktvlpr.inet.fi [194.197.169.45]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id FAA04390; Thu, 7 May 1998 05:07:15 -0700 (PDT) (envelope-from ari@osku.suutari.iki.fi) Received: from localhost (ari@localhost) by osku.suutari.iki.fi (8.8.7/8.8.5) with SMTP id PAA27980; Thu, 7 May 1998 15:06:18 +0300 (EET DST) Date: Thu, 7 May 1998 15:06:18 +0300 (EET DST) From: Ari Suutari To: Adam Rothschild cc: freebsd-net@FreeBSD.ORG, freebsd-security@FreeBSD.ORG Subject: Re: NAT woes! In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Wed, 6 May 1998, Adam Rothschild wrote: Hi, > natd -log -redirect_address 192.168.0.0 0.0.0.0 -n ex0 -u I would leave -redirect_address out since it is not required for accepting incoming connection and also because it doesn't map networks - it maps host addresses. > ipfw -f flush > ipfw add divert natd all from any to any add "via ex0" to end of this rule to pass only packets of ex0 to natd. Ari Lappeenranta, Finland To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 12:35:31 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id MAA16387 for freebsd-net-outgoing; Thu, 7 May 1998 12:35:31 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from oldyeller.comtest.com (comtest.hits.net [206.127.244.117]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA16328 for ; Thu, 7 May 1998 12:35:12 -0700 (PDT) (envelope-from randal@comtest.com) Received: from graphics.comtest.com (graphics.comtest.com [206.127.245.194]) by oldyeller.comtest.com (8.8.8/8.8.8) with SMTP id JAA03049; Thu, 7 May 1998 09:26:14 -1000 (HST) (envelope-from randal@comtest.com) Message-Id: <199805071926.JAA03049@oldyeller.comtest.com> Comments: Authenticated sender is From: "Randal S. Masutani" Organization: ComTest Technologies, Inc. To: freebsd-net@FreeBSD.ORG Date: Thu, 7 May 1998 09:42:58 -1000 Subject: Re: Have I left something out? Reply-to: randal@comtest.com CC: Richard Wackerbarth In-reply-to: X-mailer: Pegasus Mail for Windows (v2.54) Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 5 May 98 at 15:00, Richard Wackerbarth wrote: > I set up a test network and it did not behave as I had hoped/expected. > > I wanted to have redundant connections so that if one goes down, another > can be used. > > Consider the simple case of a loop. For discussion, we will have 6 nets. > 10.0.1.0/24, 10.0.2.0/24, ..., 10.0.6.0/24. > We will join them with routers > 10.0.1.1 == 10.0.2.1 > 10.0.2.2 == 10.0.3.2 > ... > 10.0.5.5 == 10.0.6.5 > and to close the loop > 10.0.6.6 == 10.0.1.6 > > Now we bring them all up and let them exchange routing info. > All is well. > A packet from net 1 ==> net 3 goes via net 2. > A packet from net 1 ==> net 5 goes via net 6. > > Now we break the loop by declaring net 6 (or one of the router interfaces > thereon) down. > The routes reconfigure and the packet to net 5 now takes the longer (but > still available) > route via nets 2,3,4. > > However, and this is my problem, if I declared the 10.0.1.6 interface down, > I cannot reach the machine. > > Host 10.0.1.1 INSISTS that the packet go out the net 1 interface and cannot > reach 10.0.1.6. > It never finds the "backdoor", 10.0.6.6. > > Any suggestions? > > Richard Wackerbarth > Well is supposed to work that way. If you down the interface 10.0.1.6 then by definition that host is down. Host 10.0.1.1 cannot get to 10.0.1.6 by any other route, other than by its own interface 10.0.1.1 for network 10.0.1.0. What you are asking is that Host 10.0.1.1 redirect its interface for net 10.0.1.0 to route thru 10.0.2.1. This is not possible. As far as Host 10.0.1.1 is concerned it sees Host 10.0.1.6 as down. However, Host 10.0.1.1 can access the Host 10.0.6.1 indirectly as Host 10.0.6.6. Randal Masutani ------------------------------------------------------------------------- ComTest Technologies, Inc. 3049 Ualena St., Suite 1005 Honolulu, Hawaii 96819 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 12:58:16 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id MAA20302 for freebsd-net-outgoing; Thu, 7 May 1998 12:58:16 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from jingoro.prevmed.sunysb.edu (jingoro.prevmed.sunysb.edu [129.49.123.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id MAA20165 for ; Thu, 7 May 1998 12:57:54 -0700 (PDT) (envelope-from cfanning@jingoro.prevmed.sunysb.edu) Received: (from cfanning@localhost) by jingoro.prevmed.sunysb.edu (8.8.8/8.8.5) id PAA05805 for freebsd-net@freebsd.org; Thu, 7 May 1998 15:49:03 GMT From: Chris Fanning Message-Id: <199805071549.PAA05805@jingoro.prevmed.sunysb.edu> Subject: MBUFs and IPFW revisited To: freebsd-net@FreeBSD.ORG Date: Thu, 7 May 1998 15:49:03 +0000 (GMT) X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Two questions, but first some history. A few days ago I wrote about running out of mbuf clusters. I think because of some attack. In any case, I configured my kernel for ipfw and set it up to deny fragmented packets and pings (fragmented pings pass through the "frag" rule??). This blocked the garbage from getting as far as my valuable mbuf clusters and I'm back to pause free operation... However, I now have very few mbuf clusters! I have: options "NMBCLUSTERS=1024" in my config file and have tried: options NMBCLUSTERS=1024 as well which comes from LINT. Before I configured ipfw, I had 1024 clusters and after, only 100something. I'm not about to compile another kernel w/o ipfw to see if this is the cause because at present I really need it. :) 1. So, the question becomes, does configuring ipfw into the kernel change the behavior of NMBCLUSTERS or reset it somewhere? As an aside, from tcpdump I'm getting LOTS of entries like: 15:42:14.989641 150.66.64.10 > 129.49.123.10: (frag 31221:1480@1480+) 15:42:14.990925 202.25.238.1 > 129.49.123.10: (frag 26973:1480@22200+) 15:42:15.010306 202.236.112.2 > 129.49.123.10: (frag 54785:1480@7400+) 15:42:15.083450 150.66.64.10 > 129.49.123.10: (frag 31221:1480@2960+) 15:42:15.084686 202.25.238.1 > 129.49.123.10: (frag 26973:1480@39960+) 15:42:15.090866 202.25.238.1 > 129.49.123.10: (frag 26973:1480@41440+) 15:42:15.110137 202.25.238.1 > 129.49.123.10: (frag 26973:1480@45880+) 15:42:15.118523 202.25.238.1 > 129.49.123.10: (frag 26973:1480@47360+) 15:42:15.129750 202.25.238.1 > 129.49.123.10: (frag 26973:1480@48840+) 15:43:14.373910 148.161.33.10 > 129.49.123.10: icmp: echo request (frag 51997:1480@0+) 15:43:14.427491 150.66.64.10 > 129.49.123.10: icmp: echo request (frag 31339:1480@0+) with: tcpdump -n -c 20 -p host not 129.49.123.9 2. This is an attack yes? The only problem I have with this is that I can't see this traffic with a sniffer on another machine (not plugged into the same hub). So unless someone slipped a switch under my nose I'm starting to believe Jingoro's becoming schiztophrenic. Chris To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 21:42:34 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id VAA17550 for freebsd-net-outgoing; Thu, 7 May 1998 21:42:34 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from implode.root.com (implode.root.com [198.145.90.17]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id VAA17543 for ; Thu, 7 May 1998 21:42:22 -0700 (PDT) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.5/8.8.5) with ESMTP id VAA07894; Thu, 7 May 1998 21:40:58 -0700 (PDT) Message-Id: <199805080440.VAA07894@implode.root.com> To: Chris Fanning cc: freebsd-net@FreeBSD.ORG Subject: Re: MBUFs and IPFW revisited In-reply-to: Your message of "Thu, 07 May 1998 15:49:03 -0000." <199805071549.PAA05805@jingoro.prevmed.sunysb.edu> From: David Greenman Reply-To: dg@root.com Date: Thu, 07 May 1998 21:40:58 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >in my config file and have tried: >options NMBCLUSTERS=1024 > >as well which comes from LINT. Before I configured ipfw, I had 1024 clusters >and after, only 100something. I'm not about to compile another kernel w/o >ipfw to see if this is the cause because at present I really need it. :) > >1. So, the question becomes, does configuring ipfw into the kernel change the >behavior of NMBCLUSTERS or reset it somewhere? netstat -m reports current/peak. The 'current' should shrink as expected, and the peak will indicate the most that have been in-use at one time since the system was booted. Does this answer your question? netstat in -current reports current/peak/max and is thus easier to figure out how many are actually configured (max). -DG David Greenman Co-founder/Principal Architect, The FreeBSD Project To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 22:12:41 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA21621 for freebsd-net-outgoing; Thu, 7 May 1998 22:12:41 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from jingoro.prevmed.sunysb.edu (jingoro.prevmed.sunysb.edu [129.49.123.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA21612 for ; Thu, 7 May 1998 22:12:27 -0700 (PDT) (envelope-from cfanning@jingoro.prevmed.sunysb.edu) Received: (from cfanning@localhost) by jingoro.prevmed.sunysb.edu (8.8.8/8.8.5) id BAA25377; Fri, 8 May 1998 01:03:33 GMT From: Chris Fanning Message-Id: <199805080103.BAA25377@jingoro.prevmed.sunysb.edu> Subject: Re: MBUFs and IPFW revisited In-Reply-To: <199805080440.VAA07894@implode.root.com> from David Greenman at "May 7, 98 09:40:58 pm" To: dg@root.com Date: Fri, 8 May 1998 01:03:33 +0000 (GMT) Cc: freebsd-net@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > >1. So, the question becomes, does configuring ipfw into the kernel change the > >behavior of NMBCLUSTERS or reset it somewhere? > > netstat -m reports current/peak. The 'current' should shrink as expected, > and the peak will indicate the most that have been in-use at one time since > the system was booted. Does this answer your question? netstat in -current > reports current/peak/max and is thus easier to figure out how many are > actually configured (max). Yes sir, it certainly does. I assumed it was current/max since the first time I ever looked at netstat -m was when peak == max. So when it says 60% in use, it's referring to 60% of the peak? That certainly didn't help my confusion either. It seems a little strange, but OK. Makes sense if the buffers aren't allocated until they're used - which is why "peak" would never drop because they're not getting freed. (If they were getting freed, we'd have to rename peak!) Next time maybe I'll look at the source code. FreeBSD does come with sources right? :) Thanks for your help on this whole episode, it's been greatly appreciated! Chris To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 22:32:10 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA24065 for freebsd-net-outgoing; Thu, 7 May 1998 22:32:10 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from implode.root.com (implode.root.com [198.145.90.17]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA23974 for ; Thu, 7 May 1998 22:31:54 -0700 (PDT) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.5/8.8.5) with ESMTP id WAA08726; Thu, 7 May 1998 22:30:30 -0700 (PDT) Message-Id: <199805080530.WAA08726@implode.root.com> To: Chris Fanning cc: freebsd-net@FreeBSD.ORG Subject: Re: MBUFs and IPFW revisited In-reply-to: Your message of "Fri, 08 May 1998 01:03:33 -0000." <199805080103.BAA25377@jingoro.prevmed.sunysb.edu> From: David Greenman Reply-To: dg@root.com Date: Thu, 07 May 1998 22:30:30 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> >1. So, the question becomes, does configuring ipfw into the kernel change the >> >behavior of NMBCLUSTERS or reset it somewhere? >> >> netstat -m reports current/peak. The 'current' should shrink as expected, >> and the peak will indicate the most that have been in-use at one time since >> the system was booted. Does this answer your question? netstat in -current >> reports current/peak/max and is thus easier to figure out how many are >> actually configured (max). > >Yes sir, it certainly does. > >I assumed it was current/max since the first time I ever looked at netstat -m >was when peak == max. So when it says 60% in use, it's referring to 60% of >the peak? That certainly didn't help my confusion either. It seems a little >strange, but OK. Yeah, it's certainly confusing. I added the "max" to netstat recently in an effort to reduce the confusion. >Makes sense if the buffers aren't allocated until they're used - which is >why "peak" would never drop because they're not getting freed. (If they >were getting freed, we'd have to rename peak!) Right, they don't get allocated until they are needed, after which time they are never completely freed (but of course can be reused as network buffers). It's (CPU) expensive to coalesce the freed buffers into whole pages in order to free them, so we don't. >Next time maybe I'll look at the source code. FreeBSD does come with sources >right? :) Also called "documentation". :-) -DG David Greenman Co-founder/Principal Architect, The FreeBSD Project To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Thu May 7 23:16:02 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id XAA29777 for freebsd-net-outgoing; Thu, 7 May 1998 23:16:02 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from oldyeller.comtest.com (comtest.hits.net [206.127.244.117]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id XAA29758 for ; Thu, 7 May 1998 23:15:53 -0700 (PDT) (envelope-from randal@comtest.com) Received: from graphics.comtest.com (graphics.comtest.com [206.127.245.194]) by oldyeller.comtest.com (8.8.8/8.8.8) with SMTP id UAA04245; Thu, 7 May 1998 20:06:48 -1000 (HST) (envelope-from randal@comtest.com) Message-Id: <199805080606.UAA04245@oldyeller.comtest.com> Comments: Authenticated sender is From: "Randal S. Masutani" Organization: ComTest Technologies, Inc. To: freebsd-net@FreeBSD.ORG Date: Thu, 7 May 1998 20:23:24 -1000 Subject: Re: Have I left something out? Reply-to: randal@comtest.com CC: Richard Wackerbarth In-reply-to: References: <199805071926.JAA03049@oldyeller.comtest.com> X-mailer: Pegasus Mail for Windows (v2.54) Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On 7 May 98 at 21:02, Richard Wackerbarth wrote: > At 2:42 PM -0500 5/7/98, Randal S. Masutani wrote: > >> Consider the simple case of a loop. For discussion, we will have 6 nets. > >> 10.0.1.0/24, 10.0.2.0/24, ..., 10.0.6.0/24. > >> We will join them with routers > >> 10.0.1.1 == 10.0.2.1 > >> 10.0.2.2 == 10.0.3.2 > >> ... > >> 10.0.5.5 == 10.0.6.5 > >> and to close the loop > >> 10.0.6.6 == 10.0.1.6 > > >> It never finds the "backdoor", 10.0.6.6. > > >Well is supposed to work that way. If you down the interface 10.0.1.6 then by > >definition that host is down. Host 10.0.1.1 cannot get to 10.0.1.6 by > >any other route, other than by its own interface 10.0.1.1 for network 10.0.1.0. > >What you are asking is that Host 10.0.1.1 redirect its interface for net > >10.0.1.0 to route thru 10.0.2.1. This is not possible. As far as Host > >10.0.1.1 is concerned it sees Host 10.0.1.6 as down. However, Host 10.0.1.1 can > >access the Host 10.0.6.1 indirectly as Host 10.0.6.6. > > I disagree. If that were the case, NO machine should be able to find it. > However, machine 5 thinks that it is one hop away; > Machine 4 thinks that it is 2 hops away; > Machine 3 finds that it is 3 hops away; > etc. > > Only the last machine fails to realize that there is another way to get there. That is true. All other machines can still route to it because they are not directly connected. What I said above only applies to that last machine, because it is directly connected to the 10.0.1.0 net. The Host 10.0.1.1 is setup to access subnet 10.0.1.0 only thru its directly connect interface. Since your subnet MASK is set to 255.255.255.0 it will try to access any host number from 10.0.1.2 to 10.0.1.254 thru that interface 10.0.1.1. So you cannot reroute 10.0.1.6 to another route because it falls within you subnet mask and according to routing rules it will only route it to the directly connected network 10.0.1.0. Only if the host fall outside of the subnet mask range then you will be able to reroute the address thru another interface. > >However, Host 10.0.1.1 can access the Host 10.0.6.1 indirectly as Host 10.0.6.6. > > Well, there is a logical inconsistency here. Am I addressing the host or > the interface? both, the host address is bound to the interface. you cannot separate the two. sorry, i made a typo above(10.0.6.1 is meant to be 10.0.1.6.) Let me explain it again. Since the host 10.0.1.6 has a second interface as 10.0.6.6, others computers can still access host 10.0.1.6 thru the 10.0.6.6 interface. But any computer that is on the 10.0.1.0 net cannot access host 10.0.1.6 because they are directly connected. And as I explained above their subnet mask will require them to ARP for the address 10.0.1.6, because it is directly connected. However, the hosts on net 10.0.1.0 can still access that compter(10.0.1.6) as host 10.0.6.6. > Anyone who cannot directly access the interface ends up getting the host. > When it is up, anyone addressing the interface, also gets the host. > > If a host has multiple interfaces, how do I address a SERVICE on that host is > such a way that I get access whenever possible? If I ask DNS for an address, > I get the address of an interface (which may be down). > > Richard Wackerbarth DNS as it currently is, is setup thru static files. What you are asking for is Dynamic DNS lookups. So if one of your interfaces goes down then a dynamic DNS server could change the entry to the active address(which is on the other interface.) I know that there is talk about Dynamic DNS, but I do not know when or how it will be implemented. Randal Masutani ------------------------------------------------------------------------- ComTest Technologies, Inc. 3049 Ualena St., Suite 1005 Honolulu, Hawaii 96819 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Fri May 8 02:48:09 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id CAA23048 for freebsd-net-outgoing; Fri, 8 May 1998 02:48:09 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from shrimp.dataplex.net (shrimp.dataplex.net [208.2.87.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id CAA23012 for ; Fri, 8 May 1998 02:48:04 -0700 (PDT) (envelope-from rkw@dataplex.net) Received: from [208.2.87.10] (user10.dataplex.net [208.2.87.10]) by shrimp.dataplex.net (8.8.8/8.8.5) with ESMTP id EAA25551; Fri, 8 May 1998 04:47:58 -0500 (CDT) X-Sender: rkw@mail.dataplex.net Message-Id: In-Reply-To: <199805080606.UAA04245@oldyeller.comtest.com> References: <199805071926.JAA03049@oldyeller.comtest.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Fri, 8 May 1998 04:47:21 -0500 To: randal@comtest.com From: Richard Wackerbarth Subject: Re: Have I left something out? Cc: freebsd-net@FreeBSD.ORG Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 1:23 AM -0500 5/8/98, Randal S. Masutani wrote: >On 7 May 98 at 21:02, Richard Wackerbarth wrote: >both, the host address is bound to the interface. you cannot separate the two. >> If a host has multiple interfaces, how do I address a SERVICE on that >>host is >> such a way that I get access whenever possible? If I ask DNS for an address, >> I get the address of an interface (which may be down). >DNS server could change the entry to the active address(which is on the other >interface.) I know that there is talk about Dynamic DNS, but I do not know >when >or how it will be implemented. I think that I actually have a scheme that might solve this situation for me. Since I want to address the HOST rather than the INTERFACE, and the host can be addressed by ANY of its interfaces, I simply need to chose an address that is NOT on a DIRECTLY connected network. That way, the packet, on its way to the interface in question will pass through the host and be trapped there (at its intended destination). Thus if host 1 (10.0.1.1, 10.0.6.1) addresses host 2 (10.0.1.2, 10.0.2.2) by its 10.0.2.2 address, it will always consult the routing tables and find a path via another host rather than the direct connection. If all the interfaces are up and working, the shortest path will be via the 10.0.1.2 address and packets will take the expected path. The only problem that I see is that the reply needs to be addressed to the far side of the originator (10.0.6.1) so that it does not suffer the same fate that keeps me from using 10.0.1.2. The other solution might be to assign the host an address on its loopback channel and not matching any of its network addresses. Richard Wackerbarth To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Fri May 8 03:54:10 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id DAA00438 for freebsd-net-outgoing; Fri, 8 May 1998 03:54:10 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from ifi.uio.no (0@ifi.uio.no [129.240.64.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id DAA00432 for ; Fri, 8 May 1998 03:54:06 -0700 (PDT) (envelope-from dag-erli@ifi.uio.no) Received: from skejdbrimir.ifi.uio.no (skejdbrimir.ifi.uio.no [129.240.65.2]) by ifi.uio.no (8.8.8/8.8.7/ifi0.2) with SMTP id MAA13248; Fri, 8 May 1998 12:54:06 +0200 (MET DST) Received: from localhost (dag-erli@localhost) by skejdbrimir.ifi.uio.no ; Fri, 8 May 1998 10:54:04 GMT Mime-Version: 1.0 To: Chris Fanning Cc: freebsd-net@FreeBSD.ORG Subject: Re: MBUFs and IPFW revisited References: <199805071549.PAA05805@jingoro.prevmed.sunysb.edu> Organization: Gutteklubben Terrasse / KRST / PUMS / YASMW X-url: http://www.stud.ifi.uio.no/~dag-erli/ X-Stop-Spam: http://www.cauce.org From: dag-erli@ifi.uio.no (Dag-Erling Coidan =?iso-8859-1?Q?Sm=F8rgrav?= ) Date: 08 May 1998 12:53:56 +0200 In-Reply-To: Chris Fanning's message of "Thu, 7 May 1998 15:49:03 +0000 (GMT)" Message-ID: Lines: 16 X-Mailer: Gnus v5.5/Emacs 19.34 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Chris Fanning writes: > options "NMBCLUSTERS=1024" As DG pointed out, the Fine Documentation (otherwise known as Source Code, or /usr/src/sys/conf/param.c) is a good place to look for information on mbufs. Carfeul reading of the aforementioned file uncovers the fact that the default value for NMBCLUSTERS is (512 + MAXUSERS * 16). This means that unless you had MAXUSERS set to something below 32, setting NMBCLUSTERS to 1024 may not have had the effect you intended, since 512 + 32 * 16 = 1024. I have plenty of RAM to blow on my home -stable box, so I have MAXUSERS set to 64 and NMBCLUSTERS set to 8192. -- Noone else has a .sig like this one. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Fri May 8 06:03:49 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA17157 for freebsd-net-outgoing; Fri, 8 May 1998 06:03:49 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from bells.cs.ucl.ac.uk (bells.cs.ucl.ac.uk [128.16.5.31]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id GAA17145 for ; Fri, 8 May 1998 06:03:40 -0700 (PDT) (envelope-from P.Gevros@cs.ucl.ac.uk) Received: from mickey.cs.ucl.ac.uk by bells.cs.ucl.ac.uk with local SMTP id ; Fri, 8 May 1998 13:12:06 +0100 X-Mailer: exmh version 1.6.6 3/24/96 To: dag-erli@ifi.uio.no (Dag-Erling Coidan =?iso-8859-1?Q?Sm=F8rgrav?= ) cc: Chris Fanning , freebsd-net@FreeBSD.ORG Subject: Re: MBUFs and IPFW revisited In-reply-to: Your message of "08 May 1998 12:53:56 +0200." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 08 May 1998 13:12:05 +0100 Message-ID: <29428.894629525@cs.ucl.ac.uk> From: Panos GEVROS Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org i 'll take the chance here, i've captured instances of netstat -m which reported 99% (even 100%) in use and occasionally the machine crashed (and then i couldnt tell of course) and i was suspecting that "running out of mbufs" could well be the reason. At least that's what i liked to think since we are talking about "sligthly" modified TCP code with several connections opened simultaneoulsy with large cwnds (around 17K each if i remember well). Am i right in assuming : - "100% in use" will cause a crash ? - the way to increase memory allocated to network is options "NMBCLUSTERS=XXXX" ? Cheers, Panos To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Fri May 8 06:58:27 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id GAA25167 for freebsd-net-outgoing; Fri, 8 May 1998 06:58:27 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from jingoro.prevmed.sunysb.edu (jingoro.prevmed.sunysb.edu [129.49.123.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id GAA25147 for ; Fri, 8 May 1998 06:58:10 -0700 (PDT) (envelope-from cfanning@jingoro.prevmed.sunysb.edu) Received: (from cfanning@localhost) by jingoro.prevmed.sunysb.edu (8.8.8/8.8.5) id JAA11923; Fri, 8 May 1998 09:49:16 GMT From: Chris Fanning Message-Id: <199805080949.JAA11923@jingoro.prevmed.sunysb.edu> Subject: Re: MBUFs and IPFW revisited In-Reply-To: from =?ISO-8859-1?Q?Dag=2DErling_Coidan_Sm=F8rgrav?= at "May 8, 98 12:53:56 pm" To: dag-erli@ifi.uio.no (Dag-Erling Coidan =?iso-8859-1?Q?Sm=F8rgrav?=) Date: Fri, 8 May 1998 09:49:16 +0000 (GMT) Cc: freebsd-net@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Chris Fanning writes: > > options "NMBCLUSTERS=1024" > > As DG pointed out, the Fine Documentation (otherwise known as Source > Code, or /usr/src/sys/conf/param.c) is a good place to look for > information on mbufs. Carfeul reading of the aforementioned file > uncovers the fact that the default value for NMBCLUSTERS is (512 + > MAXUSERS * 16). This means that unless you had MAXUSERS set to > something below 32, setting NMBCLUSTERS to 1024 may not have had the > effect you intended, since 512 + 32 * 16 = 1024. > > I have plenty of RAM to blow on my home -stable box, so I have > MAXUSERS set to 64 and NMBCLUSTERS set to 8192. Good point. I originally had it set to 2048 (maxusers=20) but I changed things around seeing if I could get the max number to change. Of course, it wouldn't because it was peak, not max. :) Jingoro's got 128MB RAM which is probably enough for a largish NMBCLUSTERS setting. I have to admit, I don't look at the sources much, -stable has been very kind to me for the 2+ years I've been running it. Chris To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Fri May 8 07:09:32 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA26801 for freebsd-net-outgoing; Fri, 8 May 1998 07:09:32 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from jingoro.prevmed.sunysb.edu (jingoro.prevmed.sunysb.edu [129.49.123.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA26723 for ; Fri, 8 May 1998 07:09:14 -0700 (PDT) (envelope-from cfanning@jingoro.prevmed.sunysb.edu) Received: (from cfanning@localhost) by jingoro.prevmed.sunysb.edu (8.8.8/8.8.5) id JAA12235; Fri, 8 May 1998 09:59:19 GMT From: Chris Fanning Message-Id: <199805080959.JAA12235@jingoro.prevmed.sunysb.edu> Subject: Re: MBUFs and IPFW revisited In-Reply-To: <29428.894629525@cs.ucl.ac.uk> from Panos GEVROS at "May 8, 98 01:12:05 pm" To: P.Gevros@cs.ucl.ac.uk (Panos GEVROS) Date: Fri, 8 May 1998 09:59:19 +0000 (GMT) Cc: dag-erli@ifi.uio.no, freebsd-net@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I'll take a stab at this based on real info and observation... > i 'll take the chance here, > > i've captured instances of netstat -m which reported 99% (even 100%) in use > and occasionally the machine crashed (and then i couldnt tell of course) and i was suspecting that "running out of mbufs" could well be the reason. > > At least that's what i liked to think since we are talking about "sligthly" modified TCP code with several connections opened simultaneoulsy with > large cwnds (around 17K each if i remember well). > > Am i right in assuming : > - "100% in use" will cause a crash ? I believe this will generate a call to the protocol drain routines. This shouldn't crash your box. It may freeze/pause network activity for a few seconds (up to 7ish on mine) until it's done with its thing. (This is based solely on observation.) 100% in use is also assuming that we're talking about 100% of max in use - which is probably the only condition you're likely to see this number anyway. > - the way to increase memory allocated to network is > options "NMBCLUSTERS=XXXX" ? Yes, so long as this number is greater than: #define NMBCLUSTERS (512 + MAXUSERS * 16) Otherwise it could decrease the number of clusters as was pointed out. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sat May 9 09:02:34 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id JAA25465 for freebsd-net-outgoing; Sat, 9 May 1998 09:02:34 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from softweyr.com ([204.68.178.33]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id JAA25420 for ; Sat, 9 May 1998 09:02:20 -0700 (PDT) (envelope-from wes@softweyr.com) Received: from softweyr.com (localhost.softweyr.com [127.0.0.1]) by softweyr.com (8.8.8/8.8.7) with ESMTP id KAA07999; Sat, 9 May 1998 10:02:07 -0600 (MDT) (envelope-from wes@softweyr.com) Message-ID: <35547DFB.950E6F33@softweyr.com> Date: Sat, 09 May 1998 10:02:03 -0600 From: Wes Peters Organization: Softweyr llc X-Mailer: Mozilla 4.04 [en] (X11; I; FreeBSD 2.2.6-RELEASE i386) MIME-Version: 1.0 To: randal@comtest.com CC: freebsd-net@FreeBSD.ORG, Richard Wackerbarth Subject: Re: Have I left something out? References: <199805071926.JAA03049@oldyeller.comtest.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Randal S. Masutani wrote: > Well is supposed to work that way. If you down the interface 10.0.1.6 then by > definition that host is down. Host 10.0.1.1 cannot get to 10.0.1.6 by > any other route, other than by its own interface 10.0.1.1 for network 10.0.1.0. > What you are asking is that Host 10.0.1.1 redirect its interface for net > 10.0.1.0 to route thru 10.0.2.1. This is not possible. Not only is it possible, it's perfectly reasonable. TCP/IP was designed to route around network failures of this sort. It works marvelously on Xylan switches, if you have XOS 3.1.7.G or 3.2.3.>=10. ;^) -- "Where am I, and what am I doing in this handbasket?" Wes Peters Softweyr LLC http://www.softweyr.com/~softweyr wes@softweyr.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sat May 9 21:05:04 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id VAA09959 for freebsd-net-outgoing; Sat, 9 May 1998 21:05:04 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from wna1.wna-linknet.com (root@wna1.wna-linknet.com [208.128.88.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id VAA09926 for ; Sat, 9 May 1998 21:04:54 -0700 (PDT) (envelope-from x5gb7uk9@wna-linknet.com) Received: from RED_DWARF (cust142.wna-linknet.com [208.128.88.142]) by wna1.wna-linknet.com (8.8.8/8.8.8) with SMTP id AAA27995 for ; Sun, 10 May 1998 00:04:42 -0400 (EDT) Message-ID: <35552667.55DF@wna-linknet.com> Date: Sat, 09 May 1998 23:00:39 -0500 From: "Arnold J. Rimmer" Reply-To: x5gb7uk9@wna-linknet.com Organization: Jupiter Mining Corporation X-Mailer: Mozilla 3.01Gold (WinNT; U) MIME-Version: 1.0 To: freebsd-net@FreeBSD.ORG Subject: homenetwork Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org i am trying to setup a 2 box network in my home with the object of using the freebsd box as a gateway for a dialup internet connection. i have followed the instructions in the ppp tutorial. the second box is running windoze nt 4.0 and each box has an smc 8416 ethernet card on isa bus. each box seems to be setup properly, but they do not acknowledge each others existence. ping to localhost and hostname work on each end, but ping from freebsd box to windoze box results: host is down. ping from windoze box to freebsd box results: timed out. i have even checked the cable for breaks...none. any suggestions as to what i am doing wrong will be greatly appreciated. / If you get to it / and you cannot do it / there you jolly well are / aren't you? ...Lord Buckley To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sat May 9 22:13:35 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA15212 for freebsd-net-outgoing; Sat, 9 May 1998 22:13:35 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from softweyr.com ([204.68.178.33]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA15203 for ; Sat, 9 May 1998 22:13:28 -0700 (PDT) (envelope-from wes@softweyr.com) Received: from softweyr.com (localhost.softweyr.com [127.0.0.1]) by softweyr.com (8.8.8/8.8.7) with ESMTP id XAA09071; Sat, 9 May 1998 23:13:20 -0600 (MDT) (envelope-from wes@softweyr.com) Message-ID: <3555376F.7FBAA205@softweyr.com> Date: Sat, 09 May 1998 23:13:19 -0600 From: Wes Peters Organization: Softweyr llc X-Mailer: Mozilla 4.04 [en] (X11; I; FreeBSD 2.2.6-RELEASE i386) MIME-Version: 1.0 To: Julian Elischer CC: freebsd-net@FreeBSD.ORG Subject: Re: Have I left something out? References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Julian Elischer wrote: > > On Thu, 7 May 1998, Joao Carlos Mendes Luis wrote: > > > #define quoting(Richard Wackerbarth) > > // However on the machine which has the point of failure, even though the > > // interface is marked "down", it still insists that that is THE route to > > // any machines which would be on that net. > > This is a known bug and there are fixes for it floating around > but none are really 'correct' > "The Right Way" to fix it is with a function that walks the route table and deletes all the routes whose ifp points to the ifnet being detached. I haven't had time to poke into the FreeBSD sources yet. If nobody else gets this, I'll look into it as soon as I get my plate cleared at work a bit. -- "Where am I, and what am I doing in this handbasket?" Wes Peters Softweyr LLC http://www.softweyr.com/~softweyr wes@softweyr.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sat May 9 22:34:26 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA16925 for freebsd-net-outgoing; Sat, 9 May 1998 22:34:26 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from implode.root.com (implode.root.com [198.145.90.17]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA16919 for ; Sat, 9 May 1998 22:34:21 -0700 (PDT) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.5/8.8.5) with ESMTP id WAA16228; Sat, 9 May 1998 22:32:33 -0700 (PDT) Message-Id: <199805100532.WAA16228@implode.root.com> To: Wes Peters cc: Julian Elischer , freebsd-net@FreeBSD.ORG Subject: Re: Have I left something out? In-reply-to: Your message of "Sat, 09 May 1998 23:13:19 MDT." <3555376F.7FBAA205@softweyr.com> From: David Greenman Reply-To: dg@root.com Date: Sat, 09 May 1998 22:32:33 -0700 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> > // interface is marked "down", it still insists that that is THE route to >> > // any machines which would be on that net. >> >> This is a known bug and there are fixes for it floating around >> but none are really 'correct' >> > >"The Right Way" to fix it is with a function that walks the route table >and deletes all the routes whose ifp points to the ifnet being detached. >I haven't had time to poke into the FreeBSD sources yet. If nobody else >gets this, I'll look into it as soon as I get my plate cleared at work >a bit. This is normally done by the routing daemon. In the case of routed, it scans the list of interfaces periodically for changes. For gated, it has a socket open for up/down change events and takes care of business as soon as it sees the change. I'm inclined to think that this functionality should stay in the route daemon where it has always been since deleting the routes to the interface may not be the correct course of action. -DG David Greenman Co-founder/Principal Architect, The FreeBSD Project To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message From owner-freebsd-net Sat May 9 23:54:23 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id XAA23774 for freebsd-net-outgoing; Sat, 9 May 1998 23:54:23 -0700 (PDT) (envelope-from owner-freebsd-net@FreeBSD.ORG) Received: from alpo.whistle.com (alpo.whistle.com [207.76.204.38]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id XAA23757 for ; Sat, 9 May 1998 23:54:18 -0700 (PDT) (envelope-from julian@whistle.com) Received: (from daemon@localhost) by alpo.whistle.com (8.8.5/8.8.5) id XAA04939; Sat, 9 May 1998 23:42:24 -0700 (PDT) Received: from current1.whistle.com(207.76.205.22) via SMTP by alpo.whistle.com, id smtpd004937; Sun May 10 06:42:17 1998 Date: Sat, 9 May 1998 23:42:13 -0700 (PDT) From: Julian Elischer To: Wes Peters cc: freebsd-net@FreeBSD.ORG Subject: Re: Have I left something out? In-Reply-To: <3555376F.7FBAA205@softweyr.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I have such a function it's checked into the tree under the tag WHISTLE_NET1 orr something similar (on a branch) look a the tags and branches on route.c for the actual tag name.. unfortunatly the changes a re far-reaching and have lots of unexpected side effects.. (e.g just deleting rtentries can sometimes lead to odd errors if you don't do it correctly) I hope to get back to it one day (I was actually told by my boss this year that it was on my official list but other things became more important) julian On Sat, 9 May 1998, Wes Peters wrote: > Julian Elischer wrote: > > > > On Thu, 7 May 1998, Joao Carlos Mendes Luis wrote: > > > > > #define quoting(Richard Wackerbarth) > > > // However on the machine which has the point of failure, even though the > > > // interface is marked "down", it still insists that that is THE route to > > > // any machines which would be on that net. > > > > This is a known bug and there are fixes for it floating around > > but none are really 'correct' > > > > "The Right Way" to fix it is with a function that walks the route table > and deletes all the routes whose ifp points to the ifnet being detached. > I haven't had time to poke into the FreeBSD sources yet. If nobody else > gets this, I'll look into it as soon as I get my plate cleared at work > a bit. > > -- > "Where am I, and what am I doing in this handbasket?" > > Wes Peters Softweyr LLC > http://www.softweyr.com/~softweyr wes@softweyr.com > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message