From owner-svn-src-all@FreeBSD.ORG Fri Feb 20 00:39:39 2009 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D9B8C10656DF; Fri, 20 Feb 2009 00:39:39 +0000 (UTC) (envelope-from luigi@FreeBSD.org) Received: from svn.freebsd.org (svn.freebsd.org [IPv6:2001:4f8:fff6::2c]) by mx1.freebsd.org (Postfix) with ESMTP id C40338FC23; Fri, 20 Feb 2009 00:39:39 +0000 (UTC) (envelope-from luigi@FreeBSD.org) Received: from svn.freebsd.org (localhost [127.0.0.1]) by svn.freebsd.org (8.14.3/8.14.3) with ESMTP id n1K0ddKA011951; Fri, 20 Feb 2009 00:39:39 GMT (envelope-from luigi@svn.freebsd.org) Received: (from luigi@localhost) by svn.freebsd.org (8.14.3/8.14.3/Submit) id n1K0ddVa011946; Fri, 20 Feb 2009 00:39:39 GMT (envelope-from luigi@svn.freebsd.org) Message-Id: <200902200039.n1K0ddVa011946@svn.freebsd.org> From: Luigi Rizzo Date: Fri, 20 Feb 2009 00:39:39 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-7@freebsd.org X-SVN-Group: stable-7 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Cc: Subject: svn commit: r188836 - stable/7/sbin/ipfw X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 20 Feb 2009 00:39:42 -0000 Author: luigi Date: Fri Feb 20 00:39:39 2009 New Revision: 188836 URL: http://svn.freebsd.org/changeset/base/188836 Log: MFC: sync the ipfw code with the version in HEAD. The only new feature is that now one can write "table all {flush | list}" to act on all tables. Just for the records, there is one difference which probably has no practical importance; two "tos" flags are represented differently now: @@ -182,8 +182,8 @@ static struct _s_x f_iptos[] = { { "throughput", IPTOS_THROUGHPUT}, { "reliability", IPTOS_RELIABILITY}, { "mincost", IPTOS_MINCOST}, - { "congestion", IPTOS_CE}, - { "ecntransport", IPTOS_ECT}, + { "congestion", IPTOS_ECN_CE}, + { "ecntransport", IPTOS_ECN_ECT0}, { "ip tos option", 0}, { NULL, 0 } }; IPTOS_ECT = IPTOS_ECN_ECT0 = 2 so 'ecntransport' is the same. IPTOS_CE = 1, IPTOS_ECN_CE = 3 so 'congestion' is represented by a different codepoint, but this also reflects a different specification (RFC3168 obsoletes RFC2481) so the change is just adopting the new spec. Added: stable/7/sbin/ipfw/altq.c (contents, props changed) stable/7/sbin/ipfw/dummynet.c (contents, props changed) stable/7/sbin/ipfw/ipfw2.h (contents, props changed) stable/7/sbin/ipfw/ipv6.c (contents, props changed) stable/7/sbin/ipfw/main.c (contents, props changed) stable/7/sbin/ipfw/nat.c (contents, props changed) Modified: stable/7/sbin/ipfw/Makefile stable/7/sbin/ipfw/ipfw.8 stable/7/sbin/ipfw/ipfw2.c Modified: stable/7/sbin/ipfw/Makefile ============================================================================== --- stable/7/sbin/ipfw/Makefile Fri Feb 20 00:05:33 2009 (r188835) +++ stable/7/sbin/ipfw/Makefile Fri Feb 20 00:39:39 2009 (r188836) @@ -1,8 +1,8 @@ # $FreeBSD$ PROG= ipfw -SRCS= ipfw2.c -WARNS?= 0 +SRCS= ipfw2.c dummynet.c ipv6.c main.c nat.c altq.c +WARNS?= 2 MAN= ipfw.8 .include Added: stable/7/sbin/ipfw/altq.c ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ stable/7/sbin/ipfw/altq.c Fri Feb 20 00:39:39 2009 (r188836) @@ -0,0 +1,150 @@ +/* + * Copyright (c) 2002-2003 Luigi Rizzo + * Copyright (c) 1996 Alex Nash, Paul Traina, Poul-Henning Kamp + * Copyright (c) 1994 Ugen J.S.Antsilevich + * + * Idea and grammar partially left from: + * Copyright (c) 1993 Daniel Boulet + * + * Redistribution and use in source forms, with and without modification, + * are permitted provided that this entire comment appears intact. + * + * Redistribution in binary form may occur without any restrictions. + * Obviously, it would be nice if you gave credit where credit is due + * but requiring it would be too onerous. + * + * This software is provided ``AS IS'' without any warranties of any kind. + * + * NEW command line interface for IP firewall facility + * + * $FreeBSD$ + * + * altq interface + */ + +#include +#include +#include + +#include "ipfw2.h" + +#include +#include +#include +#include +#include +#include +#include +#include + +#include /* IFNAMSIZ */ +#include +#include + +/* + * Map between current altq queue id numbers and names. + */ +static TAILQ_HEAD(, pf_altq) altq_entries = + TAILQ_HEAD_INITIALIZER(altq_entries); + +void +altq_set_enabled(int enabled) +{ + int pffd; + + pffd = open("/dev/pf", O_RDWR); + if (pffd == -1) + err(EX_UNAVAILABLE, + "altq support opening pf(4) control device"); + if (enabled) { + if (ioctl(pffd, DIOCSTARTALTQ) != 0 && errno != EEXIST) + err(EX_UNAVAILABLE, "enabling altq"); + } else { + if (ioctl(pffd, DIOCSTOPALTQ) != 0 && errno != ENOENT) + err(EX_UNAVAILABLE, "disabling altq"); + } + close(pffd); +} + +static void +altq_fetch(void) +{ + struct pfioc_altq pfioc; + struct pf_altq *altq; + int pffd; + unsigned int mnr; + static int altq_fetched = 0; + + if (altq_fetched) + return; + altq_fetched = 1; + pffd = open("/dev/pf", O_RDONLY); + if (pffd == -1) { + warn("altq support opening pf(4) control device"); + return; + } + bzero(&pfioc, sizeof(pfioc)); + if (ioctl(pffd, DIOCGETALTQS, &pfioc) != 0) { + warn("altq support getting queue list"); + close(pffd); + return; + } + mnr = pfioc.nr; + for (pfioc.nr = 0; pfioc.nr < mnr; pfioc.nr++) { + if (ioctl(pffd, DIOCGETALTQ, &pfioc) != 0) { + if (errno == EBUSY) + break; + warn("altq support getting queue list"); + close(pffd); + return; + } + if (pfioc.altq.qid == 0) + continue; + altq = safe_calloc(1, sizeof(*altq)); + *altq = pfioc.altq; + TAILQ_INSERT_TAIL(&altq_entries, altq, entries); + } + close(pffd); +} + +u_int32_t +altq_name_to_qid(const char *name) +{ + struct pf_altq *altq; + + altq_fetch(); + TAILQ_FOREACH(altq, &altq_entries, entries) + if (strcmp(name, altq->qname) == 0) + break; + if (altq == NULL) + errx(EX_DATAERR, "altq has no queue named `%s'", name); + return altq->qid; +} + +const char * +altq_qid_to_name(u_int32_t qid) +{ + struct pf_altq *altq; + + altq_fetch(); + TAILQ_FOREACH(altq, &altq_entries, entries) + if (qid == altq->qid) + break; + if (altq == NULL) + return NULL; + return altq->qname; +} + +void +print_altq_cmd(ipfw_insn_altq *altqptr) +{ + if (altqptr) { + const char *qname; + + qname = altq_qid_to_name(altqptr->qid); + if (qname == NULL) + printf(" altq ?<%u>", altqptr->qid); + else + printf(" altq %s", qname); + } +} Added: stable/7/sbin/ipfw/dummynet.c ============================================================================== --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ stable/7/sbin/ipfw/dummynet.c Fri Feb 20 00:39:39 2009 (r188836) @@ -0,0 +1,719 @@ +/* + * Copyright (c) 2002-2003 Luigi Rizzo + * Copyright (c) 1996 Alex Nash, Paul Traina, Poul-Henning Kamp + * Copyright (c) 1994 Ugen J.S.Antsilevich + * + * Idea and grammar partially left from: + * Copyright (c) 1993 Daniel Boulet + * + * Redistribution and use in source forms, with and without modification, + * are permitted provided that this entire comment appears intact. + * + * Redistribution in binary form may occur without any restrictions. + * Obviously, it would be nice if you gave credit where credit is due + * but requiring it would be too onerous. + * + * This software is provided ``AS IS'' without any warranties of any kind. + * + * NEW command line interface for IP firewall facility + * + * $FreeBSD$ + * + * dummynet support + */ + +#include +#include +#include +/* XXX there are several sysctl leftover here */ +#include + +#include "ipfw2.h" + +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include /* inet_ntoa */ + +static struct _s_x dummynet_params[] = { + { "plr", TOK_PLR }, + { "noerror", TOK_NOERROR }, + { "buckets", TOK_BUCKETS }, + { "dst-ip", TOK_DSTIP }, + { "src-ip", TOK_SRCIP }, + { "dst-port", TOK_DSTPORT }, + { "src-port", TOK_SRCPORT }, + { "proto", TOK_PROTO }, + { "weight", TOK_WEIGHT }, + { "all", TOK_ALL }, + { "mask", TOK_MASK }, + { "droptail", TOK_DROPTAIL }, + { "red", TOK_RED }, + { "gred", TOK_GRED }, + { "bw", TOK_BW }, + { "bandwidth", TOK_BW }, + { "delay", TOK_DELAY }, + { "pipe", TOK_PIPE }, + { "queue", TOK_QUEUE }, + { "flow-id", TOK_FLOWID}, + { "dst-ipv6", TOK_DSTIP6}, + { "dst-ip6", TOK_DSTIP6}, + { "src-ipv6", TOK_SRCIP6}, + { "src-ip6", TOK_SRCIP6}, + { "dummynet-params", TOK_NULL }, + { NULL, 0 } /* terminator */ +}; + +static int +sort_q(const void *pa, const void *pb) +{ + int rev = (co.do_sort < 0); + int field = rev ? -co.do_sort : co.do_sort; + long long res = 0; + const struct dn_flow_queue *a = pa; + const struct dn_flow_queue *b = pb; + + switch (field) { + case 1: /* pkts */ + res = a->len - b->len; + break; + case 2: /* bytes */ + res = a->len_bytes - b->len_bytes; + break; + + case 3: /* tot pkts */ + res = a->tot_pkts - b->tot_pkts; + break; + + case 4: /* tot bytes */ + res = a->tot_bytes - b->tot_bytes; + break; + } + if (res < 0) + res = -1; + if (res > 0) + res = 1; + return (int)(rev ? res : -res); +} + +static void +list_queues(struct dn_flow_set *fs, struct dn_flow_queue *q) +{ + int l; + int index_printed, indexes = 0; + char buff[255]; + struct protoent *pe; + + if (fs->rq_elements == 0) + return; + + if (co.do_sort != 0) + heapsort(q, fs->rq_elements, sizeof *q, sort_q); + + /* Print IPv4 flows */ + index_printed = 0; + for (l = 0; l < fs->rq_elements; l++) { + struct in_addr ina; + + /* XXX: Should check for IPv4 flows */ + if (IS_IP6_FLOW_ID(&(q[l].id))) + continue; + + if (!index_printed) { + index_printed = 1; + if (indexes > 0) /* currently a no-op */ + printf("\n"); + indexes++; + printf(" " + "mask: 0x%02x 0x%08x/0x%04x -> 0x%08x/0x%04x\n", + fs->flow_mask.proto, + fs->flow_mask.src_ip, fs->flow_mask.src_port, + fs->flow_mask.dst_ip, fs->flow_mask.dst_port); + + printf("BKT Prot ___Source IP/port____ " + "____Dest. IP/port____ " + "Tot_pkt/bytes Pkt/Byte Drp\n"); + } + + printf("%3d ", q[l].hash_slot); + pe = getprotobynumber(q[l].id.proto); + if (pe) + printf("%-4s ", pe->p_name); + else + printf("%4u ", q[l].id.proto); + ina.s_addr = htonl(q[l].id.src_ip); + printf("%15s/%-5d ", + inet_ntoa(ina), q[l].id.src_port); + ina.s_addr = htonl(q[l].id.dst_ip); + printf("%15s/%-5d ", + inet_ntoa(ina), q[l].id.dst_port); + printf("%4llu %8llu %2u %4u %3u\n", + align_uint64(&q[l].tot_pkts), + align_uint64(&q[l].tot_bytes), + q[l].len, q[l].len_bytes, q[l].drops); + if (co.verbose) + printf(" S %20llu F %20llu\n", + align_uint64(&q[l].S), align_uint64(&q[l].F)); + } + + /* Print IPv6 flows */ + index_printed = 0; + for (l = 0; l < fs->rq_elements; l++) { + if (!IS_IP6_FLOW_ID(&(q[l].id))) + continue; + + if (!index_printed) { + index_printed = 1; + if (indexes > 0) + printf("\n"); + indexes++; + printf("\n mask: proto: 0x%02x, flow_id: 0x%08x, ", + fs->flow_mask.proto, fs->flow_mask.flow_id6); + inet_ntop(AF_INET6, &(fs->flow_mask.src_ip6), + buff, sizeof(buff)); + printf("%s/0x%04x -> ", buff, fs->flow_mask.src_port); + inet_ntop( AF_INET6, &(fs->flow_mask.dst_ip6), + buff, sizeof(buff) ); + printf("%s/0x%04x\n", buff, fs->flow_mask.dst_port); + + printf("BKT ___Prot___ _flow-id_ " + "______________Source IPv6/port_______________ " + "_______________Dest. IPv6/port_______________ " + "Tot_pkt/bytes Pkt/Byte Drp\n"); + } + printf("%3d ", q[l].hash_slot); + pe = getprotobynumber(q[l].id.proto); + if (pe != NULL) + printf("%9s ", pe->p_name); + else + printf("%9u ", q[l].id.proto); + printf("%7d %39s/%-5d ", q[l].id.flow_id6, + inet_ntop(AF_INET6, &(q[l].id.src_ip6), buff, sizeof(buff)), + q[l].id.src_port); + printf(" %39s/%-5d ", + inet_ntop(AF_INET6, &(q[l].id.dst_ip6), buff, sizeof(buff)), + q[l].id.dst_port); + printf(" %4llu %8llu %2u %4u %3u\n", + align_uint64(&q[l].tot_pkts), + align_uint64(&q[l].tot_bytes), + q[l].len, q[l].len_bytes, q[l].drops); + if (co.verbose) + printf(" S %20llu F %20llu\n", + align_uint64(&q[l].S), + align_uint64(&q[l].F)); + } +} + +static void +print_flowset_parms(struct dn_flow_set *fs, char *prefix) +{ + int l; + char qs[30]; + char plr[30]; + char red[90]; /* Display RED parameters */ + + l = fs->qsize; + if (fs->flags_fs & DN_QSIZE_IS_BYTES) { + if (l >= 8192) + sprintf(qs, "%d KB", l / 1024); + else + sprintf(qs, "%d B", l); + } else + sprintf(qs, "%3d sl.", l); + if (fs->plr) + sprintf(plr, "plr %f", 1.0 * fs->plr / (double)(0x7fffffff)); + else + plr[0] = '\0'; + if (fs->flags_fs & DN_IS_RED) /* RED parameters */ + sprintf(red, + "\n\t %cRED w_q %f min_th %d max_th %d max_p %f", + (fs->flags_fs & DN_IS_GENTLE_RED) ? 'G' : ' ', + 1.0 * fs->w_q / (double)(1 << SCALE_RED), + SCALE_VAL(fs->min_th), + SCALE_VAL(fs->max_th), + 1.0 * fs->max_p / (double)(1 << SCALE_RED)); + else + sprintf(red, "droptail"); + + printf("%s %s%s %d queues (%d buckets) %s\n", + prefix, qs, plr, fs->rq_elements, fs->rq_size, red); +} + +void +ipfw_list_pipes(void *data, uint nbytes, int ac, char *av[]) +{ + int rulenum; + void *next = data; + struct dn_pipe *p = (struct dn_pipe *) data; + struct dn_flow_set *fs; + struct dn_flow_queue *q; + int l; + + if (ac > 0) + rulenum = strtoul(*av++, NULL, 10); + else + rulenum = 0; + for (; nbytes >= sizeof *p; p = (struct dn_pipe *)next) { + double b = p->bandwidth; + char buf[30]; + char prefix[80]; + + if (SLIST_NEXT(p, next) != (struct dn_pipe *)DN_IS_PIPE) + break; /* done with pipes, now queues */ + + /* + * compute length, as pipe have variable size + */ + l = sizeof(*p) + p->fs.rq_elements * sizeof(*q); + next = (char *)p + l; + nbytes -= l; + + if ((rulenum != 0 && rulenum != p->pipe_nr) || co.do_pipe == 2) + continue; + + /* + * Print rate (or clocking interface) + */ + if (p->if_name[0] != '\0') + sprintf(buf, "%s", p->if_name); + else if (b == 0) + sprintf(buf, "unlimited"); + else if (b >= 1000000) + sprintf(buf, "%7.3f Mbit/s", b/1000000); + else if (b >= 1000) + sprintf(buf, "%7.3f Kbit/s", b/1000); + else + sprintf(buf, "%7.3f bit/s ", b); + + sprintf(prefix, "%05d: %s %4d ms ", + p->pipe_nr, buf, p->delay); + print_flowset_parms(&(p->fs), prefix); + if (co.verbose) + printf(" V %20llu\n", align_uint64(&p->V) >> MY_M); + + q = (struct dn_flow_queue *)(p+1); + list_queues(&(p->fs), q); + } + for (fs = next; nbytes >= sizeof *fs; fs = next) { + char prefix[80]; + + if (SLIST_NEXT(fs, next) != (struct dn_flow_set *)DN_IS_QUEUE) + break; + l = sizeof(*fs) + fs->rq_elements * sizeof(*q); + next = (char *)fs + l; + nbytes -= l; + + if (rulenum != 0 && ((rulenum != fs->fs_nr && co.do_pipe == 2) || + (rulenum != fs->parent_nr && co.do_pipe == 1))) { + continue; + } + + q = (struct dn_flow_queue *)(fs+1); + sprintf(prefix, "q%05d: weight %d pipe %d ", + fs->fs_nr, fs->weight, fs->parent_nr); + print_flowset_parms(fs, prefix); + list_queues(fs, q); + } +} + +/* + * Delete pipe or queue i + */ +int +ipfw_delete_pipe(int pipe_or_queue, int i) +{ + struct dn_pipe p; + + memset(&p, 0, sizeof p); + if (pipe_or_queue == 1) + p.pipe_nr = i; /* pipe */ + else + p.fs.fs_nr = i; /* queue */ + i = do_cmd(IP_DUMMYNET_DEL, &p, sizeof p); + if (i) { + i = 1; + warn("rule %u: setsockopt(IP_DUMMYNET_DEL)", i); + } + return i; +} + +void +ipfw_config_pipe(int ac, char **av) +{ + struct dn_pipe p; + int i; + char *end; + void *par = NULL; + + memset(&p, 0, sizeof p); + + av++; ac--; + /* Pipe number */ + if (ac && isdigit(**av)) { + i = atoi(*av); av++; ac--; + if (co.do_pipe == 1) + p.pipe_nr = i; + else + p.fs.fs_nr = i; + } + while (ac > 0) { + double d; + int tok = match_token(dummynet_params, *av); + ac--; av++; + + switch(tok) { + case TOK_NOERROR: + p.fs.flags_fs |= DN_NOERROR; + break; + + case TOK_PLR: + NEED1("plr needs argument 0..1\n"); + d = strtod(av[0], NULL); + if (d > 1) + d = 1; + else if (d < 0) + d = 0; + p.fs.plr = (int)(d*0x7fffffff); + ac--; av++; + break; + + case TOK_QUEUE: + NEED1("queue needs queue size\n"); + end = NULL; + p.fs.qsize = strtoul(av[0], &end, 0); + if (*end == 'K' || *end == 'k') { + p.fs.flags_fs |= DN_QSIZE_IS_BYTES; + p.fs.qsize *= 1024; + } else if (*end == 'B' || + _substrcmp2(end, "by", "bytes") == 0) { + p.fs.flags_fs |= DN_QSIZE_IS_BYTES; + } + ac--; av++; + break; + + case TOK_BUCKETS: + NEED1("buckets needs argument\n"); + p.fs.rq_size = strtoul(av[0], NULL, 0); + ac--; av++; + break; + + case TOK_MASK: + NEED1("mask needs mask specifier\n"); + /* + * per-flow queue, mask is dst_ip, dst_port, + * src_ip, src_port, proto measured in bits + */ + par = NULL; + + bzero(&p.fs.flow_mask, sizeof(p.fs.flow_mask)); + end = NULL; + + while (ac >= 1) { + uint32_t *p32 = NULL; + uint16_t *p16 = NULL; + uint32_t *p20 = NULL; + struct in6_addr *pa6 = NULL; + uint32_t a; + + tok = match_token(dummynet_params, *av); + ac--; av++; + switch(tok) { + case TOK_ALL: + /* + * special case, all bits significant + */ + p.fs.flow_mask.dst_ip = ~0; + p.fs.flow_mask.src_ip = ~0; + p.fs.flow_mask.dst_port = ~0; + p.fs.flow_mask.src_port = ~0; + p.fs.flow_mask.proto = ~0; + n2mask(&(p.fs.flow_mask.dst_ip6), 128); + n2mask(&(p.fs.flow_mask.src_ip6), 128); + p.fs.flow_mask.flow_id6 = ~0; + p.fs.flags_fs |= DN_HAVE_FLOW_MASK; + goto end_mask; + + case TOK_DSTIP: + p32 = &p.fs.flow_mask.dst_ip; + break; + + case TOK_SRCIP: + p32 = &p.fs.flow_mask.src_ip; + break; + + case TOK_DSTIP6: + pa6 = &(p.fs.flow_mask.dst_ip6); + break; + + case TOK_SRCIP6: + pa6 = &(p.fs.flow_mask.src_ip6); + break; + + case TOK_FLOWID: + p20 = &p.fs.flow_mask.flow_id6; + break; + + case TOK_DSTPORT: + p16 = &p.fs.flow_mask.dst_port; + break; + + case TOK_SRCPORT: + p16 = &p.fs.flow_mask.src_port; + break; + + case TOK_PROTO: + break; + + default: + ac++; av--; /* backtrack */ + goto end_mask; + } + if (ac < 1) + errx(EX_USAGE, "mask: value missing"); + if (*av[0] == '/') { + a = strtoul(av[0]+1, &end, 0); + if (pa6 == NULL) + a = (a == 32) ? ~0 : (1 << a) - 1; + } else + a = strtoul(av[0], &end, 0); + if (p32 != NULL) + *p32 = a; + else if (p16 != NULL) { + if (a > 0xFFFF) + errx(EX_DATAERR, + "port mask must be 16 bit"); + *p16 = (uint16_t)a; + } else if (p20 != NULL) { + if (a > 0xfffff) + errx(EX_DATAERR, + "flow_id mask must be 20 bit"); + *p20 = (uint32_t)a; + } else if (pa6 != NULL) { + if (a > 128) + errx(EX_DATAERR, + "in6addr invalid mask len"); + else + n2mask(pa6, a); + } else { + if (a > 0xFF) + errx(EX_DATAERR, + "proto mask must be 8 bit"); + p.fs.flow_mask.proto = (uint8_t)a; + } + if (a != 0) + p.fs.flags_fs |= DN_HAVE_FLOW_MASK; + ac--; av++; + } /* end while, config masks */ +end_mask: + break; + + case TOK_RED: + case TOK_GRED: + NEED1("red/gred needs w_q/min_th/max_th/max_p\n"); + p.fs.flags_fs |= DN_IS_RED; + if (tok == TOK_GRED) + p.fs.flags_fs |= DN_IS_GENTLE_RED; + /* + * the format for parameters is w_q/min_th/max_th/max_p + */ + if ((end = strsep(&av[0], "/"))) { + double w_q = strtod(end, NULL); + if (w_q > 1 || w_q <= 0) + errx(EX_DATAERR, "0 < w_q <= 1"); + p.fs.w_q = (int) (w_q * (1 << SCALE_RED)); + } + if ((end = strsep(&av[0], "/"))) { + p.fs.min_th = strtoul(end, &end, 0); + if (*end == 'K' || *end == 'k') + p.fs.min_th *= 1024; + } + if ((end = strsep(&av[0], "/"))) { + p.fs.max_th = strtoul(end, &end, 0); + if (*end == 'K' || *end == 'k') + p.fs.max_th *= 1024; + } + if ((end = strsep(&av[0], "/"))) { + double max_p = strtod(end, NULL); + if (max_p > 1 || max_p <= 0) + errx(EX_DATAERR, "0 < max_p <= 1"); + p.fs.max_p = (int)(max_p * (1 << SCALE_RED)); + } + ac--; av++; + break; + + case TOK_DROPTAIL: + p.fs.flags_fs &= ~(DN_IS_RED|DN_IS_GENTLE_RED); + break; + + case TOK_BW: + NEED1("bw needs bandwidth or interface\n"); + if (co.do_pipe != 1) + errx(EX_DATAERR, "bandwidth only valid for pipes"); + /* + * set clocking interface or bandwidth value + */ + if (av[0][0] >= 'a' && av[0][0] <= 'z') { + int l = sizeof(p.if_name)-1; + /* interface name */ + strncpy(p.if_name, av[0], l); + p.if_name[l] = '\0'; + p.bandwidth = 0; + } else { + p.if_name[0] = '\0'; + p.bandwidth = strtoul(av[0], &end, 0); + if (*end == 'K' || *end == 'k') { + end++; + p.bandwidth *= 1000; + } else if (*end == 'M') { + end++; + p.bandwidth *= 1000000; + } + if ((*end == 'B' && + _substrcmp2(end, "Bi", "Bit/s") != 0) || + _substrcmp2(end, "by", "bytes") == 0) + p.bandwidth *= 8; + if (p.bandwidth < 0) + errx(EX_DATAERR, "bandwidth too large"); + } + ac--; av++; + break; + + case TOK_DELAY: + if (co.do_pipe != 1) + errx(EX_DATAERR, "delay only valid for pipes"); + NEED1("delay needs argument 0..10000ms\n"); + p.delay = strtoul(av[0], NULL, 0); + ac--; av++; + break; + + case TOK_WEIGHT: + if (co.do_pipe == 1) + errx(EX_DATAERR,"weight only valid for queues"); + NEED1("weight needs argument 0..100\n"); + p.fs.weight = strtoul(av[0], &end, 0); + ac--; av++; + break; + + case TOK_PIPE: + if (co.do_pipe == 1) + errx(EX_DATAERR,"pipe only valid for queues"); + NEED1("pipe needs pipe_number\n"); + p.fs.parent_nr = strtoul(av[0], &end, 0); + ac--; av++; + break; + + default: + errx(EX_DATAERR, "unrecognised option ``%s''", av[-1]); + } + } + if (co.do_pipe == 1) { + if (p.pipe_nr == 0) + errx(EX_DATAERR, "pipe_nr must be > 0"); + if (p.delay > 10000) + errx(EX_DATAERR, "delay must be < 10000"); + } else { /* co.do_pipe == 2, queue */ + if (p.fs.parent_nr == 0) + errx(EX_DATAERR, "pipe must be > 0"); + if (p.fs.weight >100) + errx(EX_DATAERR, "weight must be <= 100"); + } + if (p.fs.flags_fs & DN_QSIZE_IS_BYTES) { + size_t len; + long limit; + + len = sizeof(limit); + if (sysctlbyname("net.inet.ip.dummynet.pipe_byte_limit", + &limit, &len, NULL, 0) == -1) + limit = 1024*1024; + if (p.fs.qsize > limit) + errx(EX_DATAERR, "queue size must be < %ldB", limit); + } else { + size_t len; + long limit; + + len = sizeof(limit); + if (sysctlbyname("net.inet.ip.dummynet.pipe_slot_limit", + &limit, &len, NULL, 0) == -1) + limit = 100; + if (p.fs.qsize > limit) + errx(EX_DATAERR, "2 <= queue size <= %ld", limit); + } + if (p.fs.flags_fs & DN_IS_RED) { + size_t len; + int lookup_depth, avg_pkt_size; + double s, idle, weight, w_q; + struct clockinfo ck; + int t; + + if (p.fs.min_th >= p.fs.max_th) + errx(EX_DATAERR, "min_th %d must be < than max_th %d", + p.fs.min_th, p.fs.max_th); + if (p.fs.max_th == 0) + errx(EX_DATAERR, "max_th must be > 0"); + + len = sizeof(int); + if (sysctlbyname("net.inet.ip.dummynet.red_lookup_depth", + &lookup_depth, &len, NULL, 0) == -1) + errx(1, "sysctlbyname(\"%s\")", + "net.inet.ip.dummynet.red_lookup_depth"); + if (lookup_depth == 0) + errx(EX_DATAERR, "net.inet.ip.dummynet.red_lookup_depth" + " must be greater than zero"); + + len = sizeof(int); + if (sysctlbyname("net.inet.ip.dummynet.red_avg_pkt_size", + &avg_pkt_size, &len, NULL, 0) == -1) + + errx(1, "sysctlbyname(\"%s\")", + "net.inet.ip.dummynet.red_avg_pkt_size"); + if (avg_pkt_size == 0) + errx(EX_DATAERR, + "net.inet.ip.dummynet.red_avg_pkt_size must" + " be greater than zero"); + + len = sizeof(struct clockinfo); + if (sysctlbyname("kern.clockrate", &ck, &len, NULL, 0) == -1) + errx(1, "sysctlbyname(\"%s\")", "kern.clockrate"); + + /* + * Ticks needed for sending a medium-sized packet. + * Unfortunately, when we are configuring a WF2Q+ queue, we + * do not have bandwidth information, because that is stored + * in the parent pipe, and also we have multiple queues + * competing for it. So we set s=0, which is not very + * correct. But on the other hand, why do we want RED with + * WF2Q+ ? + */ + if (p.bandwidth==0) /* this is a WF2Q+ queue */ + s = 0; + else + s = (double)ck.hz * avg_pkt_size * 8 / p.bandwidth; + + /* + * max idle time (in ticks) before avg queue size becomes 0. + * NOTA: (3/w_q) is approx the value x so that + * (1-w_q)^x < 10^-3. + */ + w_q = ((double)p.fs.w_q) / (1 << SCALE_RED); + idle = s * 3. / w_q; + p.fs.lookup_step = (int)idle / lookup_depth; + if (!p.fs.lookup_step) + p.fs.lookup_step = 1; + weight = 1 - w_q; + for (t = p.fs.lookup_step; t > 1; --t) + weight *= 1 - w_q; + p.fs.lookup_weight = (int)(weight * (1 << SCALE_RED)); + } + i = do_cmd(IP_DUMMYNET_CONFIGURE, &p, sizeof p); + if (i) + err(1, "setsockopt(%s)", "IP_DUMMYNET_CONFIGURE"); +} Modified: stable/7/sbin/ipfw/ipfw.8 ============================================================================== --- stable/7/sbin/ipfw/ipfw.8 Fri Feb 20 00:05:33 2009 (r188835) +++ stable/7/sbin/ipfw/ipfw.8 Fri Feb 20 00:39:39 2009 (r188836) @@ -1,7 +1,7 @@ .\" .\" $FreeBSD$ .\" -.Dd August 5, 2007 +.Dd September 27, 2008 .Dt IPFW 8 .Os .Sh NAME @@ -49,9 +49,13 @@ .Nm .Cm table Ar number Cm delete Ar addr Ns Op / Ns Ar masklen .Nm -.Cm table Ar number Cm flush +.Cm table +.Brq Ar number | all +.Cm flush .Nm -.Cm table Ar number Cm list +.Cm table +.Brq Ar number | all +.Cm list .Pp .Nm .Brq Cm pipe | queue @@ -537,7 +541,7 @@ A match is only declared with the specif This can be useful for a number of applications such as random packet drop or (in conjunction with -.Xr dummynet 4 ) +.Nm dummynet ) to simulate the effect of multiple paths leading to out-of-order packet delivery. .Pp @@ -553,7 +557,7 @@ with a .Dv LOG_SECURITY facility. The logging only occurs if the sysctl variable -.Em net.inet.ip.fw.verbose +.Va net.inet.ip.fw.verbose is set to 1 (which is the default when the kernel is compiled with .Dv IPFIREWALL_VERBOSE ) @@ -564,7 +568,7 @@ parameter. If no .Cm logamount is specified, the limit is taken from the sysctl variable -.Em net.inet.ip.fw.verbose_limit . +.Va net.inet.ip.fw.verbose_limit . In both cases, a value of 0 removes the logging limit. .Pp Once the limit is reached, logging can be re-enabled by @@ -666,7 +670,7 @@ and .Nm .Cm disable Ar altq . The usage of -.Em net.inet.ip.fw.one_pass +.Va net.inet.ip.fw.one_pass is irrelevant to ALTQ traffic shaping, as the actual rule action is followed always after adding an ALTQ tag. .El @@ -760,7 +764,7 @@ see the Section for further information. .It Cm pipe Ar pipe_nr Pass packet to a -.Xr dummynet 4 +.Nm dummynet .Dq pipe (for bandwidth limitation, delay, etc.). See the @@ -770,12 +774,12 @@ The search terminates; however, on exit the .Xr sysctl 8 variable -.Em net.inet.ip.fw.one_pass +.Va net.inet.ip.fw.one_pass is not set, the packet is passed again to the firewall code starting from the next rule. .It Cm queue Ar queue_nr Pass packet to a -.Xr dummynet 4 +.Nm dummynet .Dq queue (for bandwidth limitation using WF2Q+). .It Cm reject @@ -841,12 +845,12 @@ Divert packet into netgraph with given The search terminates. If packet is later returned from netgraph it is either accepted or continues with the next rule, depending on -.Em net.inet.ip.fw.one_pass +.Va net.inet.ip.fw.one_pass sysctl variable. .It Cm ngtee Ar cookie A copy of packet is diverted into netgraph, original packet is either accepted or continues with the next rule, depending on -.Em net.inet.ip.fw.one_pass +.Va net.inet.ip.fw.one_pass sysctl variable. See .Xr ng_ipfw 4 @@ -1193,7 +1197,7 @@ Matches ICMP packets whose ICMP type is .Ar types . The list may be specified as any combination of individual types (numeric) separated by commas. -.Em Ranges are not allowed. +.Em Ranges are not allowed . The supported ICMP types are: .Pp echo reply @@ -1231,7 +1235,7 @@ Matches ICMP6 packets whose ICMP6 type i *** DIFF OUTPUT TRUNCATED AT 1000 LINES ***